=============================================================== =============================================================== Tag name: ctsm1.0.dev113 Originator(s): negins (Negin Sobhani,UCAR/TSS,303-497-1224) Date: Tue Sep 8 15:57:49 MDT 2020 One-line Summary: Some bit-for-bit changes needed for the Perturbed Parameter Ensemble work Purpose of changes ------------------ Some BFB fixes including bug fixes needed for PPE simulations. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/CTSM#1060 (Some code cleanup of luna bug fixes) - Resolves ESCOMP/CTSM#1062 (PPE namelist parameter defaults) - Resolves ESCOMP/CTSM#1053 (Non-standard namelist parameter) - Resolves ESCOMP/CTSM#1055 (Expand potential vcmax:jmax ratio) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): cmb_cmplt_fact was a parameter under namelist control pertaining to the fire model (combusion completion factor [unitless]), which previously included an array. This parameter is now splitted into two real values i.e. cmb_cmplt_fact_litter and cmb_cmplt_fact_cwd. If the users have cmb_cmplt_fact in a user_nl_clm file that worked before, it won't work anymore since it is now splitted into cmb_cmplt_fact_litter and cmb_cmplt_fact_cwd. Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): Previously, denitrif_nitrateconc_exponent and denitrif_nitrateconc_coefficient were on the params file but ALSO on the namelist. They are now removed from the namelist. Changes made to namelist defaults (e.g., changed parameter values): None Changes to the datasets (e.g., parameter, surface or initial files): None Substantial timing or memory changes: [For timing changes, can check PFS test(s) in the test suite] None Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): None Changes to tests or testing: None CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- PASS (except for NLCOMP) izumi ------- PASS (except for NLCOMP) If the tag used for baseline comparisons was NOT the previous tag, note that here: ctsm1.0.dev111 Answer changes -------------- Changes answers relative to baseline: None Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): None Pull Requests that document the changes (include PR ids): https://github.com/ESCOMP/CTSM/pull/1102 =============================================================== =============================================================== Tag name: ctsm1.0.dev112 Originator(s): sacks (Bill Sacks), wwieder (Will Wieder) Date: Fri Aug 28 17:10:57 MDT 2020 One-line Summary: Small changes to mksurfdata_map, singlept tool and run_sys_tests Purpose of changes ------------------ (1) Move mksurfdata_map rounding of special landunits to before wetland fill: I think this may be important to avoid errors and/or wrong behavior in some edge cases where we have < 0.5% cover of glacier and/or lake, and other cases. See https://github.com/ESCOMP/CTSM/issues/1118 for detailed thoughts. May change answers in rare cases, but no answer changes observed for present-day surface datasets at standard resolutions (f09, f19, f10, f45). (2) run_sys_tests: record command-line arguments in SRCROOT_GIT_STATUS file (3) (From Will Wieder) Updates to current surfdata & landuse files in tools/contrib/singlept Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/CTSM#1118 (mksurfdata_map rounding/truncation of special landunits should happen before wetland filling) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: self CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - PASS Note: ran tools testing on code based off of ctsm1.0.dev110 (before merging with ctsm1.0.dev111) PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - PASS on mac (bishorn) regular tests (aux_clm): cheyenne ---- not run izumi ------- not run If the tag used for baseline comparisons was NOT the previous tag, note that here: ctsm1.0.dev110 Answer changes -------------- Changes answers relative to baseline: NO In principle, the changes to mksurfdata_map could change answers for rare cases. No answer changes were observed for present-day surface datasets at resolutions f09, f19, f10 or f45. Also note that no surface datasets were actually updated in this tag, so if any answer changes appear from these changes, they would appear the next time we update surface datasets. Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): https://github.com/ESCOMP/CTSM/pull/1119 https://github.com/ESCOMP/CTSM/pull/1120 https://github.com/ESCOMP/CTSM/pull/1122 =============================================================== =============================================================== Tag name: ctsm1.0.dev111 Originator(s): erik (Erik Kluzek,UCAR/TSS,303-497-1326) Date: Fri Aug 28 14:08:25 MDT 2020 One-line Summary: Compsets don't use 2014 for GSWP3 forcing, LUNA prevyr changed back Purpose of changes ------------------ Update cime so that default end year for GSWP3 forcing is 2013 because 2014 data is bad. Also update the prevyr default for vcmax/jmax to what it was before ctsm1.0.dev102. Fix some issues for finding initial condition files for Clm45 compsets. Also bring in updated CDEPS/CMEPS as they were required with the cime update. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): Fixes #1107 -- Change initial values of LUNA back to previous global values Fixes #1067 -- Stop using PIO2 in the LILAC build Fixes #1121 -- FMOZ test fails CIME Issues fixed (include issue #): Fixes ESMCI/cime#3683 -- Avoid artificial limit on string lengths in shr_string_listMerge Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): Last time-step in Dec/2013 for GSWP3 data still has zero humidity Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): None Changes made to namelist defaults (e.g., changed parameter values): None Changes to the datasets (e.g., parameter, surface or initial files): None Substantial timing or memory changes: None Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): cime is pointing to a branch_tag we should update it to a trunk tag as soon as possible Changes to tests or testing: build-namelist unit testing improved to find problems like #1121 a #1125 Code reviewed by: self CTSM testing: regular [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - PASS (all identical to previous tag) tools-tests (test/tools): cheyenne - PASS python testing (see instructions in python/README.md; document testing done): cheyenne - PASS regular tests (aux_clm): cheyenne ---- OK izumi ------- PASS If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: Yes (similar climate) Summarize any changes to answers, i.e., - what code configurations: (2010 and SSP compsets) and Clm50 with LUNA on (Bgc or Sp) - what platforms/compilers: All - nature of change: similar climate Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): cime, cmeps, cdeps (fox removed as a separate checkout) cime to branch_tags/cime5.8.31_a01 Pull Requests that document the changes (include PR ids): #1114 (https://github.com/ESCOMP/ctsm/pull) #1114 -- Update cime so default end year for GSWP3 forcing is 2013 and change prevyr default for vcmax/jmax =============================================================== =============================================================== Tag name: ctsm1.0.dev110 Originator(s): jedwards (Jim Edwards) Date: Fri Aug 21 10:49:08 MDT 2020 One-line Summary: Fixes needed for PIO2 Purpose of changes ------------------ Fixes needed for CTSM to work with PIO2. There is an additional fix needed in PIO2 itself; this is available in a later version of cime that we will bring in soon. Note that this PR also includes a change in config_files.xml that is needed for a newer version of cime when running compsets with dlnd, slnd or xlnd; this may be incompatible with the current version of the cime external, but this does not impact compsets with CTSM. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): The changes in this tag, together with a cime update that will be coming in soon, resolve the following issues (I'm not sure off-hand which issues are resolved by the changes here alone and which ones require the cime update): - Resolves ESCOMP/CTSM#1029 (FATES tests fail with PIO2) - Resolves ESCOMP/CTSM#1030 (mpi-serial nag case fails with pio2 when defining h1 history restart file) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): - May not be able to run configurations with dlnd, slnd or xlnd from this tag Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: Not investigated, but almost definitely not Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: Bill Sacks CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- PASS izumi ------- PASS Note: Standard testing was done using PIO1 (as this is still the default), but separate testing was also done using PIO2 (using an updated cime) and tests were found to pass. If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: NO Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): https://github.com/ESCOMP/CTSM/pull/1095 =============================================================== =============================================================== Tag name: ctsm1.0.dev109 Originator(s): negins (Negin Sobhani,UCAR/TSS,303-497-1224) Date: Thu Aug 20 11:40:07 MDT 2020 One-line Summary: Allow for resorbtion in transition from live to dead wood N Purpose of changes ------------------ This PR allows for resorbtion in transition from live to dead wood N, which also move to NPOOL for free. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): Issue #443 Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [x] clm5_0 [x] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): None Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): None Changes made to namelist defaults (e.g., changed parameter values): None Changes to the datasets (e.g., parameter, surface or initial files): None Substantial timing or memory changes: [For timing changes, can check PFS test(s) in the test suite] Notes of particular relevance for developers: None --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): Changes to tests or testing: None Code reviewed by: Erik Kluzek CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- PASS izumi ------- PASS If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: YES (similar climate) Summarize any changes to answers, i.e., - what code configurations: Clm50 with Bgc or Cn - what platforms/compilers: all - nature of change (roundoff; larger than roundoff/same climate; new climate): larger than roundoff/same climate If this tag changes climate describe the run(s) done to evaluate the new climate (put details of the simulations in the experiment database) Will Wieder thought the changes were small enough that we didn't need to do a long simulation. Erik Kluzek thinks the changes were verified with single point simulations. Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): Pull Requests that document the changes (include PR ids): (https://github.com/ESCOMP/ctsm/pull) =============================================================== =============================================================== Tag name: ctsm1.0.dev108 Originator(s): erik (Erik Kluzek,UCAR/TSS,303-497-1326) Date: Wed Aug 19 17:23:47 MDT 2020 One-line Summary: Update default PE layouts for new SE/FV3 grids Purpose of changes ------------------ Change PE layouts for new high resolution SE/FV3 grids. Have a version for cheyenne and a default one as well. Update externals: cime, cism, CMEPS, CDEPS Change single point tests to use stub ROF Change tests to use gx1v7 rather than gx1v6 Update waccmx_offline to use START=1979 Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): Fixes #1108 Change single point tests to use stub ROF Fixes #1098 ARCTICGRIS PE layout is very slow Fixes #1105 Default new FV3/SE grid PE layouts are problematic Fixes #1103 Some changes to prevent missing file for WACCMX testing Fixes #1113 ARCTIC test is failing Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? No (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): None Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): None Changes made to namelist defaults (e.g., changed parameter values): None Changes to the datasets (e.g., parameter, surface or initial files): None Substantial timing or memory changes: None Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): Some tests were chagned to use stub ROF and one to stub GLC. @billsacks suggest we should also change other tests in this same way. Changes to tests or testing: Yes Code reviewed by: self CTSM testing: regular, tools [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - PASS tools-tests (test/tools): cheyenne - PASS PTCLM testing (tools/shared/PTCLM/test): cheyenne - OK python testing (see instructions in python/README.md; document testing done): cheyenne -- PASS regular tests (aux_clm): cheyenne ---- PASS izumi ------- PASS If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: No (except for waccmx_offline test) Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): cime, cism, CDEPS, CMEPS cism to cism2_1_69 cime to cime5.8.30 Update hash for CDEPS and CMEPS (so that nuopc test will run) Pull Requests that document the changes (include PR ids): #1111 (https://github.com/ESCOMP/ctsm/pull) #1111 -- Adjust FV3/SE PE layouts =============================================================== =============================================================== Tag name: ctsm1.0.dev107 Originator(s): erik (Erik Kluzek) Date: Mon Aug 10 02:21:12 MDT 2020 One-line Summary: Answer changes needed for CESM2.2.0 Purpose of changes ------------------ A list of small answer changes needed for CESM2.2.0. Nitrogen deposition file now properly ends in 2015 rather than 2005. Have 2010_control compset match the 2000 IOC file rather than 2010. Modify the waccmx_offline tests to more closely match the CAM tests. Interpolate IC files for an exact match. Answers are identical for an exact match, but different when used for other resolutions. Two small bit-for-bit changes remove levurb and string_len dimensions on restart files. Remove xsmrpool_loss when use_crop is NOT true. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): [If none, remove this line] Answer changing fixes: Fixes #1004 -- stream_year_last_ndep is set to 2005 instead of 2015 for historical with ctsm1.0.dev093 Fixes #1012 -- Help prevent problems with WACCM-X testing by modifying our waccmx_offline test Fixes #1090 -- Have 2010_control compset match 2000 simulation-year rather than 2010 Fixes #1089 -- Failure in FXHIST test when running with CAM Fixes #1037 -- Interpolate out-of-the-box initial conditions files for CESM2.2 release (only changes answers for resolutions other than IC file resolution) Bit-for-bit changes: Fixes #1101 -- levurb and string_len dimensions no longer needed on the restart files Fixes #1097 -- xsmrpool_loss should only be written to the restart file if use_crop is true Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? No (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): None Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): 2010_control doesn't match 2010 finidat file Changes made to namelist defaults (e.g., changed parameter values): None Changes to the datasets (e.g., parameter, surface or initial files): Interpolated IC files Add f19 2003 IC file, f09 and 19 1979 IC file Remove 2010 IC file Substantial timing or memory changes: None Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): None Changes to tests or testing: waccm_offline test changed to more closely match the FXHIST CAM test Code reviewed by: self CTSM testing: regular [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - PASS (115 tests are different than before) python testing (see instructions in python/README.md; document testing done): cheyenne - PASS regular tests (aux_clm): cheyenne ---- PASS izumi ------- PASS If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: Yes! (very small) Hist and SSP BGC compsets change answers because of the correction of the end year for ndep. You wouldn't think this would matter until you get to the end of the time-sequence -- but it does for the first time-step. The 2010 compset changes answers because of using the 2000 finidat file rather than 2010 Some finidat files interpolated to the resolution they are designed for. This causes an apparent change of answers when these datasets are used for other resolutions (even though it does NOT for the resolution it's interpolated to) waccm_offline tests are also changed in order to more closely match the FXHIST tests in CAM. Summarize any changes to answers, i.e., - what code configurations: Hist, SSP, 2010 compsets, Clm45 and some f10, 1x1 because of IC files - what platforms/compilers: All - nature of change: same climate Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): None Pull Requests that document the changes (include PR ids): #1100 (https://github.com/ESCOMP/ctsm/pull) #1100 -- Answer changes needed for CESM2.2.0 =============================================================== =============================================================== Tag name: ctsm1.0.dev106 Originator(s): erik (Erik Kluzek) Date: Thu Aug 6 15:35:47 MDT 2020 One-line Summary: Bit-for-bit updates for the CESM2.2.0 release Purpose of changes ------------------ Fix some bit-for-bit things needed for the CESM2.2.0 release Bring changes needed from release-clm5.0.31-34 to trunk. ndep change. Brings in some changes for PtVg datasets for no-anthro compset. Do some refactoring of Fire class so that it makes sense for FATES to use the base class. Update RTM and MOSART with NUOPC changes and an update of a NetCDF-4 file to NetCDF-5 needed for the cesm2.2.0 release. The 2010 finidat file is now set in the 2010_control use-case rather than relying on the finidat dataset matching. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): Fixes #1087 -- wrong finidat file Fixes #1096 -- failing tests Fixes #1036 -- ne30 case fails Fixes #946 --- better error message Fixes #983 --- Fire method renames Fixes #938 --- mscripgrid issue Known bugs found since the previous tag (include github issue ID): #1098 -- ARCTICGRIS PE-layout is very slow... #1097 -- xsmrpool_loss should only be written to the restart file if use_crop is true Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? No (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): if ndep file isn't found die with a warning rather than fatal error. The warning allows you to override with CLM_BLDNML_OPTS="-ignore_warnings" finidat file matching will now use the sim_year rather than start-year for a control case (transient cases used the start-year) Add 2010 f09 finidat file for clm5_0_cam6.0 Changes made to namelist defaults (e.g., changed parameter values): Changes to the datasets (e.g., parameter, surface or initial files): Substantial timing or memory changes: None Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): Changes to tests or testing: Add clm45cam6LndTuningMode testmod Some tests needed more wallclock Code reviewed by: self CTSM testing: regular [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - PASS (one test is different than ctsm1.0.dev105) python testing (see instructions in python/README.md; document testing done): cheyenne -= PASS regular tests (aux_clm): cheyenne ---- PASS izumi ------- PASS If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: No (bit-for-bit) Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): rtm, mosart RTM to rtm1_0_72 MOSART to mosart1_0_37 Pull Requests that document the changes (include PR ids): #1079 (https://github.com/ESCOMP/ctsm/pull) #1079 -- Bit-for-bit updates for the CESM2.2.0 release =============================================================== =============================================================== Tag name: ctsm1.0.dev105 Originator(s): erik (Erik Kluzek,UCAR/TSS,303-497-1326) / Chris Fischer Date: Thu Jul 23 08:59:37 MDT 2020 One-line Summary: Bring in some new SE grids for CAM as well as initial condition files for them Purpose of changes ------------------ Add new SE (Spectral Element) grids: ne30np4.pg2, ne30np4.pg3, ne30np4, ne60np4, ne120np4, ne120np4.pg2, ne120np4.pg3, ne0np4CONUS.ne30x8, ne0np4.ARCTIC.ne30x4, ne0np4.ARCTICGRIS.ne30x8 Add new datasets for them, some new tests, and add support for their creation. Also add some new initial condition (IC) files for 1979, 2000, and 2013 for use when coupled to CAM The process for picking initial condition files was improved so that there can be more than one year that matches and it will do the best to pick the best match. Also added the possibility for 1982 and 2013 simulation years for specific grid matches. Also update cime to a newer version that supports these new grids. And update CMEPS (and hence add in fox, CDEPS) to latest used on the nuopc_dev branch. Testing with nuopc wouldn't work without this update. The cime version updates the tables for dry-deposition and changes answers for dry-deposition when it's turned on. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): #992 #994 Fixes #992 -- New initial conditions for 1979, 2000 Fixes #994 -- New CAM SE grids Fixes #888 -- Add support for various "physics grids" needed by CAM CIME Issues fixed (include issue #): #3593, #3569, #3564, PR#3557 cime/#3593 -- Fix mapping files for new grids cime/#3569 -- Fix cime regression testing cime/#3564 -- Change alias for arctic grids cime/#3557 -- updates to dry deposition data (PR) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? No (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): Initial conditions (IC) for 1979 start date is technically 1982 New initial condition files are ONLY for when coupled to CAM (B and F compsets) Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): Allow for new grids User updates to cime: - Allow a user supplied config_grids.xml file. - Improved error message for incorrect setting of cime model. Changes made to namelist defaults (e.g., changed parameter values): Add IC files for 1982 and 2013 years Changes to the datasets (e.g., parameter, surface or initial files): New SE grid datasets Remove conus_30_x8 grid (previous name for new ne0np4CONUS.ne30x8 grid Add new grids: ne30np4.pg2, ne30np4.pg3, ne30np4, ne60np4, ne120np4, ne120np4.pg2, ne120np4.pg3, ne0np4CONUS.ne30x8, ne0np4.ARCTIC.ne30x4, ne0np4.ARCTICGRIS.ne30x8 Add new mask: tx0.1v3 New IC files: 1982 for ARCTIC and ARCTICGRIS, ne120np4 for 2000 and 2013 for CONUS all when coupled to CAM Substantial timing or memory changes: None Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): Initial condition picking will pick the file it finds with the closet simulation year to the request for the lnd_tuning_mode desired. It will pick based on year over an exact match in resolution. You can get around this by setting use_init_interp to be specific for a resolution Also realized that there is a dependence between the namelist_defaults_ctsm.xml file and namelist_defaults_overall.xml for the clm_start_type item. It must be coordinated with use_init_interp, init_interp_attributes and finidat to work properly. I added a comment about this. We should probably just move them into the same file. Changes to tests or testing: Added tests for new grids Code reviewed by: self, fischer-ncar, adamrher CTSM testing: regular [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - OK (130 tests fail comparison to previous verison due to new IC files) tools-tests (test/tools): izumi - OK PTCLM testing (tools/shared/PTCLM/test): izumi - OK python testing (see instructions in python/README.md; document testing done): izumi -- PASS regular tests (aux_clm): cheyenne ---- OK izumi ------- PASS If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: No bit-for-bit (unless running with CAM because of new finidat files) Running with nuopc changes answers because of updates to nuopc drivers Dry-deposition changes answers when turned on (but doesn't effect anything else) Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): cime, cmeps, cdeps, fox cime updated to cime5.8.28 (from branch_tags/cime5.8.24_a01) fox, and cdeps added in cmeps, cdeps and fox updated to hashes used in the nuopc_dev branch of CESM Pull Requests that document the changes (include PR ids): #1038 (https://github.com/ESCOMP/ctsm/pull) #1038 -- New fsufdat and landuse time series files for ne grids =============================================================== =============================================================== Tag name: ctsm1.0.dev104 Originator(s): sacks (Bill Sacks) Date: Mon Jul 6 09:58:15 MDT 2020 One-line Summary: Add LILAC Purpose of changes ------------------ Add LILAC: The Lightweight Infrastructure for Land-Atmosphere Coupling. This infrastructure consists of two major pieces: (1) A lightweight coupling infrastructure built on top of ESMF that makes it easier for atmosphere models to call CTSM directly, rather than using the hub-and-spoke architecture that is used by CESM. (2) A set of python-based tools for building CTSM and creating its runtime inputs when running in an atmosphere model via LILAC. Although these tools are built on top of cime, details of the create_newcase / case.setup / case.build process are hidden from the user, because many of the aspects of this workflow don't make sense in the LILAC context. So far we have used LILAC to couple CTSM to WRF. There are plans to use the same infrastructure to couple CTSM to other regional atmosphere models. Documentation of LILAC is provided in https://escomp.github.io/ctsm-docs/versions/master/html/lilac/index.html (though there are still some missing sections), as well as in various presentations on the wiki (https://github.com/ESCOMP/CTSM/wiki/Presentations). There have been many contributors besides myself to the development, testing and documentation of LILAC; chief among them being Mariana Vertenstein, Negin Sobhani, Joe Hamman, Sam Levis, Mike Barlage and Dave Lawrence. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - See issues in the Done column of https://github.com/ESCOMP/CTSM/projects/23 Known bugs introduced in this tag (include github issue ID): - Although LILAC is working to first order, there is still some work to do. See outstanding issues in https://github.com/ESCOMP/CTSM/projects/23 Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): - dtime no longer specified on the namelist: just obtained from driver Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: - Added LILACSMOKE test Code reviewed by: self CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - ok Baseline comparisons fail as expected. One test failed, but it also failed for me on master (https://github.com/ESCOMP/CTSM/issues/1074) tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - pass on my mac regular tests (aux_clm): cheyenne ---- pass izumi ------- pass If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: NO Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): https://github.com/ESCOMP/CTSM/pull/1068 =============================================================== =============================================================== Tag name: ctsm1.0.dev103 Originator(s): slevis (Samuel Levis, SLevis Consulting LLC,303-665-1310) Date: Mon Jun 29 17:16:29 MDT 2020 One-line Summary: Gridcell-level balance-check for methane (CH4) Purpose of changes ------------------ Bracket the model time-step loop to calculate balance checks at the gridcell level, as detailed in issue #315. The column-level check remains unchanged. Subroutine ch4_init_balance_check is replaced with ch4_init_column_balance_check. Subroutine ch4_init_gridcell_balance_check is added. The implementation is similar to the one for carbon and nitrogen (see tag ctsm1.0.dev096). Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): #315 Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): None Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): None Changes made to namelist defaults (e.g., changed parameter values): None Changes to the datasets (e.g., parameter, surface or initial files): None Substantial timing or memory changes: No Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): None Changes to tests or testing: None Code reviewed by: @billsacks CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - tools-tests (test/tools): cheyenne - PTCLM testing (tools/shared/PTCLM/test): cheyenne - python testing (see instructions in python/README.md; document testing done): (any machine) - regular tests (aux_clm): cheyenne ---- PASS izumi ------- PASS If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: No Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): None Pull Requests that document the changes (include PR ids): https://github.com/ESCOMP/CTSM/pull/1022 =============================================================== =============================================================== Tag name: ctsm1.0.dev102 Originator(s): erik (Erik Kluzek,UCAR/TSS,303-497-1326) Date: Fri Jun 26 01:32:04 MDT 2020 One-line Summary: Some important fixes for LUNA in clm5_0, and small urban issue in clm5_0 Purpose of changes ------------------ Fix some important issues that were found in LUNA for the arctic. These changes were large enough that we had to re-tune the clm5_0 parameters to give a similar simulation. Leah Birch noticed that the formulations in the code disagreed with the paper. Correcting this caused arctic plants to be less productive. In order to compensate parameters for leafcn, and slatop were adjusted for broadleaf_deciduous_boreal_shrub and c3_arctic_grass for clm5_0. The total vegetation carbon threshold on exit_spinup (from accelerated decomposition (AD) mode) was increased from 0.1 to 1.0. Without this change some PFT's can die out in exit spinup even when they have significant carbon stores from the AD spinup mode. There is an implicit assumption in the urban building energy model that building width equals street width. However, this assumption can/should be relaxed and building width can be derived from the morphology dataset. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): #738 #803 #953 #958 #1056 Fixes #1056 -- Remove allpfts dimension Fixes #953 -- Incorrect formula in LUNA uses day time rather than average of day/night Fixes #958 -- LUNA day length factor missing Fixes #738 -- Totvegc threshold increased from 0.1 to 1. Fixes #803 -- No longer assume building width Known bugs introduced in this tag (include github issue ID): #1060 #1060 -- Some code cleanup of luna bug fixes Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [x] clm5_0 [x] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): clm5_0 climate adjusted for arctic plants Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): Changes to the datasets (e.g., parameter, surface or initial files): New parameter files Substantial timing or memory changes: None Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): We chose some bad names for variables, so we'll clean this up in a seperate step (see #1060) Changes to tests or testing: None Code reviewed by: self, olyson, bill-sacks, wweider CTSM testing: regular [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - PASS (266 compare differently because of update in paramsfile and jmaxb1) python testing (see instructions in python/README.md; document testing done): cheyenne - PASS regular tests (aux_clm): cheyenne ---- PASS izumi ------- PASS If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: Yes for clm5_0 (important change required retuning) Summarize any changes to answers, i.e., - what code configurations: clm5_0 when Luna on as well as smaller changes for urban - what platforms/compilers: all - nature of change: climate changing, but tuned to respond similar to clm5.0 If this tag changes climate describe the run(s) done to evaluate the new climate (put details of the simulations in the experiment database) luna changes: oleson case clm50_ctsm10d089_2deg_GSWP3V1_luna3_jmaxb1-0.17_slatopA_leafcnA_hist URL for LMWG diagnostics output used to validate new climate: urban changes: http://webext.cgd.ucar.edu/I2000/clm50_ctsm10d098_1deg_GSWP3V1_CON_FIXBUILDENERGY_2000/lnd/clm50_ctsm10d098_1deg_GSWP3V1_CON_FIXBUILDENERGY_2000.1991_2010-clm50_ctsm10d098_1deg_GSWP3V1_CON_2000.1991_2010/setsIndex.html Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): None Pull Requests that document the changes (include PR ids): #1034 and #962 (https://github.com/ESCOMP/ctsm/pull) #1034 -- Building energy fix #962 --- Fixes for the LUNA dayl bugs =============================================================== =============================================================== Tag name: ctsm1.0.dev101 Originator(s): erik (Erik Kluzek,UCAR/TSS,303-497-1326), oleson (Keith Oleson) Date: Wed Jun 17 23:51:22 MDT 2020 One-line Summary: Changes from Keith to bring a list of variables to the parameter file Purpose of changes ------------------ Change some hardcoded parameters to go on the parameter files. This is needed in preparation of running the Perturbed Parameter Ensemble. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): #995 #138, #206 Fixes #995 -- Soil turnover times are on paramfile and hardcoded Fixes #138 -- use tau_ values on paramfile Fixes #206 -- add prh30 to paramfile Some work on #290 -- LUNA refactoring Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? No (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): None Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): None Changes made to namelist defaults (e.g., changed parameter values): None Changes to the datasets (e.g., parameter, surface or initial files): paramfile for both clm4_5 and clm5_0 Add 51 new parameters to the paramsfile: bsw_adjustfactor,cpha,csol_clay,csol_om,csol_sand,denitrif_nitrateconc_coefficient,denitrif_nitrateconc_exponent, denitrif_respiration_coefficient,denitrif_respiration_exponent,drift_gs,e_ice,enzyme_turnover_daily,eta0_anderson, eta0_vionnet,fnps,hksat_adjustfactor,ignition_efficiency,jmaxb0,jmaxha,jmaxhd,kcha,koha,kp25ratio,lmrha,lmrhd, lmrse,luna_theta_cj,minrelh,n_baseflow,pd,perched_baseflow_scalar,prh30,relhExp,ssi,sucsat_adjustfactor,theta_cj, theta_ip,theta_psii,tpuha,tpuhd,tkd_clay,tkd_om,tkd_sand,tkm_om,tpu25ratio,vcmaxha,vcmaxhd,watsat_adjustfactor, wc2wjb0,wimp,wind_min k_nitr_max removed and replaced with k_nitr_max_perday meta-data for the tau_* variables changed, and a roundoff difference in the clm4_5 paramfile Substantial timing or memory changes: None Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): prh30 not used for clm4_5 prh30 was added in but won't be used for the CNFireLi2014 formulation. It's only used for the clm5_0 and following formulations Changes to tests or testing: None Code reviewed by: self, oleson CTSM testing: regular [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - PASS (170 tests differ from baseline because of new paramfiles) python testing (see instructions in python/README.md; document testing done): cheyenne - PASS regular tests (aux_clm): cheyenne ---- OK izumi ------- OK If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: No bit-for-bit Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): None Pull Requests that document the changes (include PR ids): #1039 (https://github.com/ESCOMP/ctsm/pull) #1039 -- move hardcoded parameters to paramfile =============================================================== =============================================================== Tag name: ctsm1.0.dev100 Originator(s): sacks (Bill Sacks) Date: Tue Jun 9 06:15:15 MDT 2020 One-line Summary: Deallocate memory after running init_interp Purpose of changes ------------------ Many of the pointer arrays in init_interp were not being deallocated when init_interp finished. This was a particular problem because some of these arrays were allocated for the full global source grid on each processor, so the memory use was significant. This tag fixes this issue. This issue is documented in https://github.com/ESCOMP/CTSM/issues/1010 and https://github.com/ESCOMP/CTSM/issues/1032. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/CTSM#1032 (Memory isn't deallocated in some of the init_interp code) Known bugs found since the previous tag (include github issue ID): - This does NOT resolve ESCOMP/CTSM#1010 (WACCMX test fails with ctsm1.0.dev093 in cesm2_2_alpha05c), even though that was the original hope. Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none - The initial run of the PFS test showed a very long time taken in initialization. However, it appears that this was a system issue: rerunning twice showed times back to normal. Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: self CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - pass (on my mac) regular tests (aux_clm): cheyenne ---- pass izumi ------- pass Most testing was done on f332fafc. After that, I made a minor change that was needed for the unit tests to pass. I then reran (1) clm_short on cheyenne, and (2) this subset of the test list on izumi: SMS.f10_f10_musgs.I2000Clm50BgcCrop.izumi_gnu.clm-crop SMS_D.f10_f10_musgs.I2000Clm50BgcCrop.izumi_gnu.clm-crop PEM_Ld1.f10_f10_musgs.I2000Clm50BgcCropGs.izumi_intel.clm-crop SMS_D.f10_f10_musgs.I2000Clm50BgcCrop.izumi_intel.clm-crop SMS.f10_f10_musgs.I2000Clm50BgcCrop.izumi_pgi.clm-crop SMS_D.f10_f10_musgs.I2000Clm50BgcCrop.izumi_pgi.clm-crop SMS_Ld5_D_P48x1.f10_f10_musgs.IHistClm50Bgc.izumi_nag.clm-monthly ERP_Ld5_P48x1.f10_f10_musgs.I1850Clm50Bgc.izumi_nag.clm-default If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: NO Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): none =============================================================== =============================================================== Tag name: ctsm1.0.dev099 Originator(s): sacks (Bill Sacks) Date: Mon Jun 8 09:29:38 MDT 2020 One-line Summary: Update cime, needed for izumi machine updates Purpose of changes ------------------ Update cime. The most important change here is an update for izumi, which is needed for the recent izumi machine updates. Also: Change izumi PEM test to use SGLC (after the machine updates, the PEM test was failing; see also https://github.com/ESCOMP/CTSM/issues/840). Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: not investigated; none expected Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: - Changed izumi PEM test to use SGLC (after the machine updates, the PEM test was failing; see also https://github.com/ESCOMP/CTSM/issues/840). Code reviewed by: self CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- pass izumi ------- ok izumi: tests pass, some baselines fail as expected ERP_D_Ld3.f09_g17.I2000Clm50SpGs.cheyenne_intel.clm-prescribed showed differences from the stored baselines, but it appears that the stored baselines were incorrect: when I reran this test from master, the master version differed from the stored baselines and was identical to this branch. If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: YES, just on izumi due to machine updates Summarize any changes to answers, i.e., - what code configurations: some izumi tests - what platforms/compilers: izumi gnu & pgi - nature of change (roundoff; larger than roundoff/same climate; new climate): not investigated; expected to be roundoff Answer changes were observed in these tests: - SMS_D.f10_f10_musgs.I2000Clm50BgcCrop.izumi_gnu.clm-crop - SMS_D.f10_f10_musgs.I2000Clm50BgcCrop.izumi_pgi.clm-crop - SMS.f10_f10_musgs.I2000Clm50BgcCrop.izumi_pgi.clm-crop Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): - cime: branch_tags/cime5.8.23_a01 -> branch_tags/cime5.8.24_a01 Point to a cime branch tag that is equivalent to cime5.8.24 but uses pio1 rather than pio2, even for mpi-serial cases (the changes in branch_tags/cime5.8.24_a01 relative to cime5.8.24 are equivalent to the changes in branch_tags/cime5.8.23_a01 relative to cime5.8.23). Pull Requests that document the changes (include PR ids): none =============================================================== =============================================================== Tag name: ctsm1.0.dev098 Originator(s): John Truesdale,UCAR/AMP,303-497-1386 Date: Thu May 28 14:39:27 MDT 2020 One-line Summary: Bring in GFDL fv3 cubed sphere grids C24, C48, C96, C192, C384 Purpose of changes ------------------ Add support for new GFDL fv3 curbed sphere grids. Low resolution C24, and C48 Moderate resolution C96, and high resolution C192 and C384. Add in mapping files so all can be created. Make C96 a default resolution created with new surface datasets. For all of these only the crop versions of surface and landuse.timeseries are in place, as non-crop can run in CLM without needing different datasets. Also for historical landuse.timeseries we use the SSP5-8.5 timeseries files, so that it can be used for future, historical and beyond 2015 present day. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): #1025 Fixes #1025 -- By default only create crop version of surface/landuse.timeseries datasets Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): None Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): None Changes made to namelist defaults (e.g., changed parameter values): None Changes to the datasets (e.g., parameter, surface or initial files): New surface and landuse.timeseries files Substantial timing or memory changes: None Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): None Changes to tests or testing: Add C96 tests Code reviewed by: @ekluzek CTSM testing: regular, tools [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - PASS (15 new tests for new resolutions) tools-tests (test/tools): cheyenne - OK python testing (see instructions in python/README.md; document testing done): (any machine) - PASS (cheyenne) regular tests (aux_clm): cheyenne ---- PASS izumi ------- PASS If the tag used for baseline comparisons was NOT the previous tag, note that here: previous Answer changes -------------- Changes answers relative to baseline: No bit-for-bit Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): None Pull Requests that document the changes (include PR ids): #891 (https://github.com/ESCOMP/ctsm/pull) #891 -- add definitions/defaults for GFDL fv3 cubed sphere grids C24, C48, C96, C192, C384 =============================================================== =============================================================== Tag name: ctsm1.0.dev097 Originator(s): ycheng (Yanyan Cheng); final changes & testing by Bill Sacks Date: Mon May 25 15:14:56 MDT 2020 One-line Summary: Add two bioenergy crops (switchgrass and miscanthus) Purpose of changes ------------------ Main change is from Yanyan Cheng: adding parameters and code for two bioenergy crops, switchgrass and miscanthus. Along with this, there is a new potential flux from crop leaves and stems to the crop product pool at harvest, representing biofuel products; a new pft-specific parameter controls this flux (biofuel_harvfrac). Currently, the out-of-the-box surface datasets do not specify any area for these crops, but the new parameter file will allow them to be present if specified on the surface dataset or landuse_timeseries file. Note that this is only an option for CLM5.0, NOT for CLM4.5. (See https://github.com/ESCOMP/CTSM/pull/884 for details.) Also, some minor fixes from Bill Sacks: See notes under "bugs fixed" below. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/CTSM#203 - Fixes creation of harvest-related variables on surface datasets created with the all_veg option - smallville, PTCLM, etc. (documented in https://github.com/ESCOMP/CTSM/pull/1019) - Resolves ESCOMP/CTSM#930 - Makes gindex_ocn intent(inout) rather than intent(out) - Resolves ESCOMP/CTSM#1021 - Changes SSP test to only do symlink if needed Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): - For transient crop cases, now more crops in memory. This creates a need for interpolation of initial conditions for initial conditions generated from a transient case. Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): - Changed parameter values for switchgrass and miscanthus (only for clm5.0) - New pft-specific parameter, biofuel_harvfrac (for both clm4.5 and clm5.0) Changes to the datasets (e.g., parameter, surface or initial files): - New parameter files with the above changes Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: - Changed smallville ERS_Lm20 test to be cold start (see comment in commit 8368f222 for details) - Added a multi-year smallville test that includes these new crops Code reviewed by: Bill Sacks, Danica Lombardozzi CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - ok Pass except for expected baseline failures Ran on 633be0eb, with comparison against ctsm1.0.dev095; tests pass except these expected baseline failures: 030 blf84 TBLscript_tools.sh PTCLM PTCLMmkdata PTCLM_USUMB_clm4_5^buildtools ....................\c rc=7 FAIL 032 blfc4 TBLscript_tools.sh PTCLM PTCLMmkdata PTCLM_USUMB_Cycle_clm4_5^buildtools ..............\c rc=7 FAIL Diffs for these are just in the harvest-related fields on the surface dataset, as expected. PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- ok izumi ------- pass For cheyenne testing: a few tests had baseline failures. The differences just arise in the h1 (vector) files, due to differences in dimension sizes. These differences are expected. See below for details. Ran most testing on 8309c213. Reran two SSP tests on 4665c9ce If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: YES, but diagnostic only due to change in dimension sizes - does not change answers in any meaningful way Summarize any changes to answers, i.e., - what code configurations: Transient CLM50 cases with crops; just changes dimensions on vector (h1) history files - what platforms/compilers: All - nature of change (roundoff; larger than roundoff/same climate; new climate): diagnostic change only Differences were observed in the following tests: ERS_Ly3_Mmpi-serial.1x1_smallvilleIA.IHistClm50BgcCropQianGs.cheyenne_gnu.clm-cropMonthOutput SMS_Ld5.f10_f10_musgs.ISSP245Clm50BgcCrop.cheyenne_gnu.clm-ciso_dec2050Start SMS_Ld5.f10_f10_musgs.ISSP370Clm50BgcCrop.cheyenne_gnu.clm-ciso_dec2050Start ERP_D_Ld10_P36x2.f10_f10_musgs.IHistClm50BgcCrop.cheyenne_intel.clm-ciso_decStart ERP_D_Ld5.f10_f10_musgs.IHistClm50BgcCrop.cheyenne_intel.clm-allActive ERP_Ly3_P72x2.f10_f10_musgs.IHistClm50BgcCrop.cheyenne_intel.clm-cropMonthOutput ERS_D_Ld7_Mmpi-serial.1x1_smallvilleIA.IHistClm50BgcCropGs.cheyenne_intel.clm-decStart1851_noinitial ERS_Ly3_P72x2.f10_f10_musgs.IHistClm50BgcCropG.cheyenne_intel.clm-cropMonthOutput ERS_Ly5_P144x1.f10_f10_musgs.IHistClm50BgcCrop.cheyenne_intel.clm-cropMonthOutput ERS_Ly6_Mmpi-serial.1x1_smallvilleIA.IHistClm50BgcCropQianGs.cheyenne_intel.clm-cropMonthOutput LCISO_Lm13.f10_f10_musgs.IHistClm50BgcCrop.cheyenne_intel.clm-ciso_monthly SMS_Ld5.f10_f10_musgs.ISSP585Clm50BgcCrop.cheyenne_intel.clm-ciso_dec2050Start The differences just arise in the h1 files, due to differences in dimension sizes. I reran these tests with a paramfile where I reverted the mergeToPft variable; then they were all bfb with master. Other than this, there were FIELDLIST differences, as expected. Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): - https://github.com/ESCOMP/CTSM/pull/884 - https://github.com/ESCOMP/CTSM/pull/1019 =============================================================== =============================================================== Tag name: ctsm1.0.dev096 Originator(s): slevis (Samuel Levis,SLevis Consulting LLC,303-665-1310) Date: Thu May 21 19:18:24 MDT 2020 One-line Summary: Gridcell-level balance checks for carbon and nitrogen Purpose of changes ------------------ Bracket the model's time-step loop to calculate balance checks at the gridcell level because there are terms in the carbon and nitrogen cycles not accounted for at the column level. Balance checks at the column level remain unchanged for now. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): #314 Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): None Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): None Changes made to namelist defaults (e.g., changed parameter values): None Changes to the datasets (e.g., parameter, surface or initial files): None Substantial timing or memory changes: None Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): None Changes to tests or testing: None Code reviewed by: @billsacks CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - tools-tests (test/tools): cheyenne - PTCLM testing (tools/shared/PTCLM/test): cheyenne - python testing (see instructions in python/README.md; document testing done): (any machine) - regular tests (aux_clm): cheyenne ---- OK izumi ------- PASS If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: No Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): None Pull Requests that document the changes (include PR ids): https://github.com/ESCOMP/CTSM/pull/984 =============================================================== =============================================================== Tag name: ctsm1.0.dev095 Originator(s): sacks (Bill Sacks) Date: Thu May 21 10:25:16 MDT 2020 One-line Summary: Update cime and cmeps externals; rework initialization of CNFire object Purpose of changes ------------------ Main change is to update cime and cmeps externals. cime is now essentially at the version used in cesm2_2_beta05 (but with one change backed out); cmeps is at latest master. Also, reworks initialization of CNFire object to avoid occasional segmentation faults. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/CTSM#1018 (Occasional seg faults when initializing CNFire object) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none - Note: Because of a change in cime, datm now uses netcdf rather than pnetcdf, but it doesn't seem that that impacts performance significantly Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: self CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- ok izumi ------- pass cheyenne tests pass; some answer changes as expected (as noted below) If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: YES (small) Summarize any changes to answers, i.e., - what code configurations: - cheyenne_gnu - due to compiler version update - nuopc - presumably due to the major cmeps update - what platforms/compilers: see above - nature of change (roundoff; larger than roundoff/same climate; new climate): - Not investigated I have verified that gnu tests in 59d1dd4a are identical to those in dcaad375 - i.e., the rework of the initialization logic for fire didn't change answers Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): - cime: cime5.8.19 -> branch_tags/cime5.8.23_a01 (same as cime5.8.23, but reverts to using pio1 for mpi-serial cases, since some mpi-serial cases failed with pio2) - cmeps Pull Requests that document the changes (include PR ids): none =============================================================== =============================================================== Tag name: ctsm1.0.dev094 Originator(s): sacks (Bill Sacks) Date: Fri May 15 13:41:01 MDT 2020 One-line Summary: Minor bug fixes needed for latest cime Purpose of changes ------------------ (1) Refactor a conditional to avoid divide by 0 (ESCOMP/CTSM#1013) (2) Remove some FATES coordinate variables that were writing garbage (temporary workaround for ESCOMP/CTSM#1014) (3) Minor update to manage_externals (4) Document authorship in README.md and .zenodo.json Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/CTSM#1013 - Workaround for ESCOMP/CTSM#1014, but doesn't properly resolve that issue Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: not checked; none expected Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: self CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- pass izumi ------- pass If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: NO Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): - https://github.com/ESCOMP/ctsm/pull/1006 =============================================================== =============================================================== Tag name: ctsm1.0.dev093 Originator(s): erik (Erik Kluzek) Date: Fri May 1 22:27:16 MDT 2020 One-line Summary: Bring in changes from release-clm5.0 branch to master Purpose of changes ------------------ Move changes from release-clm5.0.12.13/14/16/17/18/20/21/22/23/24/25/26/27/28/29/30 to master (note excludes release-clm5.0.15, release-clm5.0.19) Answer changing! Updates FATES with PARTEH option. Remove CRUNCEP half degree, ne120np4, f05 surface datasets Landuse-timeseries files for just: f09, f19, f10, ne30np4, 1x1_numaIA, 1x1_smallvilleIA, 1x1_brazil Add datasets needed for SSP scenarios (only Tier I for ndep). Add ZWT_inversion dataset for methane. Add new ability to use prescribed soil moisture data (from streams files based on an earlier CTSM simulation). Add new MODISv2 mapping files for mksurfdata_map. Remove T5 and T21 mapping files. Add conus_30_x8 mapping files, and T62 mapping files Add contrib directory to tools with contributed tools and scripts. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): 31 issues.... #873 -- Release clm5.0.12 changes #871 -- don't check for valid resolution for CESM cases by default #836 -- add izumi as a place that can run tools tests #831 -- add namelist tint_algo settings for input streams #817 -- Have clm4_5 use the same ndep file as clm5_0 for 1850 #809 -- mkmapdata using an old version of intel #804 -- mksurfdata option #791 -- fix threading for prescribed LAI and soil-moisture #786 -- allow use_init_interp for a branch #746 -- Some CLM45 configurations use CLM50 initial conditions files #741 -- C13_HR, C13_NBP, FPI values result in numeric conversion not representable error (changes answers) #734 -- Isotope historical files off of NetCDF-4 #716 -- 2 degree CMIP6WACCMDECK needs Carbon isotopes off #707 -- Missing mapping files for hirespft #673 -- Ice shelf wetland fix in mksurfdata_map can lead to glacier+lake > 100% on surface datasets (changes surface dataset creation) #653 -- To save disk on fsurdat files use zwt0 on finindated file #629 -- Remove unused fields from surface datasets #624 -- QSNOEVAP history output incorrect #621 -- Fully remove rcp in place of only ssp_rcp #608 -- Odd (blocky) spatial patterns in generic crop leafc for BGC simulation when initialized from BGC-Crop (changes answers) #589 -- Existence of content in the lai_streams namelist makes it confusing to users #557 -- no-anthro compset and options #553 -- More robust mksurfdata_map logic for determining where to put wetlands #547 -- Add conus_30_x8 grid as valid option for CTSM and mksurfdata_map #545 -- Antarctica ice shelves are being treated as wetlands rather than glaciers #544 -- Transient cases are matching wrong IC file #533 -- Add -no-anthro option to mksurfdata #276 -- Tests for pre_alpha/pre_beta #262 -- hirespft option to mksurfdata.pl #175 -- Set stream_year_first_urbantv/last for CLM4.5 #150 -- mksurfdata code cleanup Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): New surface datasets do mean initial conditions will need to be interpolated Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): Add 2010 compset option New namelist items: ndep_tintalgo, lai_tintalgo, lightng_tintalgo, urbantv_tintalgo, dribble_crophrv_xsmrpool_2atm New soil moisture namelist items: use_soil_moisture_streams, stream_year_first_soilm, stream_year_last_soilm, model_year_align_soilm, model_year_align_soilm, soilm_tintalgo, soilm_offset, soilm_ignore_data_if_missing New namelist items for FATES: fates_parteh_mode New 2010, no-anthro, SSP use cases New compsets: I2010Clm50Sp, I2000Clm50BgcCropQianGs, I1850Clm50BgcNoAnthro, I1850Clm50SpNoAnthro, IHistClm50SpCru, IHistClm50SpCru, IHistClm45SpGs, IHistClm45SpGs New SSP compsets: ISSP585Clm50BgcCrop, ISSP126Clm50BgcCrop, ISSP126Clm50BgcCrop, ISSP245Clm50BgcCrop, ISSP370Clm50BgcCrop, ISSP434Clm50BgcCrop, ISSP460Clm50BgcCrop, ISSP460Clm50BgcCrop LAI streams options don't show up in the namelist unless you ask for them finundation comes from streams files, no longer from the surface dataset FATES adds a list of new history variables and dimension types Changes made to namelist defaults (e.g., changed parameter values): New fsurdat files, SSP C13/C14 files Add PtVg surface dataset most finidat files now need to be interpolated, a few added in to get testing to work New ndep datasets Changes to the datasets (e.g., parameter, surface or initial files): New fsurdat files, finidat files need to be interpolated from now Substantial timing or memory changes: None Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): Threshold to reseed dead vegetation increased from zero to 1.0 as a hard coded parameter. Reseeding also happens for generic crop types and not just natural vegetation now. The ability to read in 3D streams was added (input files have to be at the same resolution as the model) Changes to tests or testing: Add no-anthro, 2010, f19 waccm and SSP tests Make sure prealpha/prebeta tests are also run by aux_clm Add prescribed soil moisture test Lengthen waccm_offline test Remove hcru_hcru and edison tests Add some aux_cime_baseline tests Code reviewed by: self CTSM testing: regular, tools [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - PASS (345 are different than baseline) tools-tests (test/tools): cheyenne - OK PTCLM testing (tools/shared/PTCLM/test): cheyenne - OK python testing (see instructions in python/README.md; document testing done): cheyennne -- PASS (when ncar_pylib run) regular tests (aux_clm): cheyenne ---- OK izumi ------- OK If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: Yes (small) Updates in all fsurdat files leads to changes in answers Also there are updated finidat files as they had to be interpolated to work with the new fsurdat files Reseeding change can effect answers on startup if activated Extra call to SoilBiogeochemPrecisionControl can cause a small answer change Summarize any changes to answers, i.e., - what code configurations: all global - what platforms/compilers: all - nature of change: Main change is around Antarctica CMIP6 simulations and release branch simulations were run with these surface datasets Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): fates, PTCLM fates to sci.1.30.0_api.8.0.0 PTCLM to PTCLM2_20200121 FATES now points to the NGEET repository rather than the NCAR fates-release Pull Requests that document the changes (include PR ids): #998, #883, #849, #567, #564, #561 (https://github.com/ESCOMP/ctsm/pull) #998 -- Remove Copyright file #883 -- Bring changes on release-clm5.0 branch to master PR status: ready priority: high type: enhancement #849 -- Fix link to new forums #567 -- Ignore patterns for vim #564 -- Update mksurfdata_map to include glaciers outside of pft landmask #561 -- Add list of checks that should be done when making new surface datasets =============================================================== =============================================================== Tag name: ctsm1.0.dev092 Originator(s): erik (Erik Kluzek,UCAR/TSS,303-497-1326) Date: Mon Apr 27 14:51:21 MDT 2020 One-line Summary: Update cime to version in cesm2_2_beta04 Purpose of changes ------------------ Update cime to latest in cesm2.2 beta tag (cesm2_2_beta04) Fix two small issues that had gone on the release-clm5.0 branch. Update manage_externals (to manic-v1.1.8) NOTE: Part of the reason for updating cime, was because of some problems Mike Mills was having on cheyenne with inputdata files being deleted out from under him before setting up his cases. We think system issues were interacting with issues with cime putting in place zero length missing files for the version of cime he was using. Those issues went away with a cime update, so we hope updating to a newer cime version here will help prevent those issues for CTSM users. cime updates: some cleanup from the last maint-5.6 update Add ne5np4.pg4 grid. case.setup: Add --keep option Add omip as option to DATM_CO2_TSERIES Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): #974 #966 Fixes #974 -- python 3 problem in ssp test Fixes #966 -- a few history fields need to be made CF compliant CIME Issues fixed (include issue #): 3459 3466 3459 -- case.setup refuses to fix corrupted env_mach_specfic.xml 3466 -- fix for aquap Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): None Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): ./case.setup now has a --keep option Changes made to namelist defaults (e.g., changed parameter values): None Changes to the datasets (e.g., parameter, surface or initial files): None Substantial timing or memory changes: None Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): None Changes to tests or testing: None Code reviewed by: self CTSM testing: regular [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - PASS tools-tests (test/tools): cheyenne - OK PTCLM testing (tools/shared/PTCLM/test): cheyenne - OK python testing (see instructions in python/README.md; document testing done): (any machine) - regular tests (aux_clm): cheyenne ---- OK izumi ------- OK If the tag used for baseline comparisons was NOT the previous tag, note that here: previous Answer changes -------------- Changes answers relative to baseline: None bit-for-bit Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): cime cime to cime5.8.19 (NOTE: CMEPS is pointing to a newer hash than used in the beta tag) Pull Requests that document the changes (include PR ids): (https://github.com/ESCOMP/ctsm/pull) #996 -- Update cime to latest cesm2.2 beta tag =============================================================== =============================================================== Tag name: ctsm1.0.dev091 Originator(s): sacks (Bill Sacks) Date: Fri Apr 24 07:43:46 MDT 2020 One-line Summary: Increase tolerance on near-zero truncation for a snow state update Purpose of changes ------------------ In UpdateState_TopLayerFluxes, the tolerance of 1.e-13 was occasionally exceeded. Although I haven't done a careful analysis, it seems okay to me to increase this tolerance slightly. Here I increase it to 1.e-12. Previously, if top-layer h2osoi_ice or h2osoi_liq were reduced to less than 1e-13 times the original value (in an absolute value sense), these masses were set to 0; now we set these masses to 0 if they are reduced to less than 1e-12 times the original. So we can now occasionally set a value to exactly 0 when before it was left at slightly different from zero. If the previous code led to a small positive value, between 1e-13 and 1e-12 times the original, this tag will change answers slightly. If the previous code led to a small negative number, it would cause the model to abort, leading to the issue reported in ESCOMP/CTSM#988; this change should fix those occasional aborts. This tag also introduces the general ability to set tolerances to a custom value in calls to truncate_small_values. See ESCOMP/CTSM#988 for details. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/CTSM#988 Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none expected (not checked) Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: Erik Kluzek CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- pass izumi ------- pass If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: YES, in theory (but none in test suite) Summarize any changes to answers, i.e., - what code configurations: potentially all, but very rarely - what platforms/compilers: all - nature of change (roundoff; larger than roundoff/same climate; new climate): Roundoff Potentially changes answers by roundoff (see notes above), but no changes were observed in any test in the test suite Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): https://github.com/ESCOMP/CTSM/pull/989 =============================================================== =============================================================== Tag name: ctsm1.0.dev090 Originator(s): negins (Negin Sobhani,UCAR/CSEG,303-497-1224) Date: Thu Apr 23 09:21:18 MDT 2020 One-line Summary: Refactoring banded diagonal matrix and solve system in SoilTemperature module Purpose of changes ------------------ Previously, the SoilTemperature module has a high level of nested subroutine calls, which hurt its performance and readability. There were many redundancies, logical issues, and nested conditionals in the modular code as discussed in issue #323. In this PR, SoilTemperature module is optimized and rewritten for better readability and performance while maintaining bfb compatibility. This PR addresses issue ESCOMP/ctsm#323 Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): -ESCOMP/ctsm#323 Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: The revised code improved the performance of the previous code by ~1.5-2x. Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: Bill Sacks CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- OK izumi ------- PASS Externals used for the testing: [cism] tag = cism2_1_68 [rtm] tag = rtm1_0_70 [mosart] tag = mosart1_0_35 [cime] tag = branch_tags/cime5.8.15_a01 If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: NO Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): (https://github.com/ESCOMP/CTSM/pull/979) =============================================================== =============================================================== Tag name: ctsm1.0.dev089 Originator(s): sacks (Bill Sacks) Date: Tue Apr 7 12:01:59 MDT 2020 One-line Summary: Bring documentation source to master Purpose of changes ------------------ 1. Bring documentation source to master: Pulls in the source from https://github.com/escomp/ctsm-docs. This is important so that documentation can remain in sync with changes in the model code. Images are stored here using git-lfs (Git Large File Storage). I also made some minor fixes to get the pdf build of the tech note working. 2. Use a different documentation theme that supports a version dropdown menu, and add the code needed to support this versioning on the documentation web pages. At a high level, the way the versioned documentation works is to have separate subdirectories in the gh-pages branch of the ctsm-docs repository for each version of the documentation we want to support. There is then a bit of JavaScript code which uses a json file in the gh-pages branch to determine which versions exist and how these should be named in the dropdown menu. Most of these changes were borrowed from ESMCI/cime#3439, which in turn borrowed from ESCOMP/CISM-wrapper#23, which in turn was a slight modification of an implementation provided by @mnlevy1981 for the MARBL documentation, which in turn borrowed from an implementation put together by Unidata (credit where credit is due). I am not aware of out-of-the-box support for a version pull-down in out-of-the-box sphinx themes (though the last time I looked was in Fall, 2018, so there may be something available now). However, support for a version dropdown exists in an open PR in the sphinx readthedocs theme repository: readthedocs/sphinx_rtd_theme#438. I have pushed this branch to a new repository in ESMCI (https://github.com/ESMCI/sphinx_rtd_theme) to facilitate long-term maintenance of this branch in case it disappears from the official sphinx_rtd_theme repository. I have also cherry-picked a commit onto that branch, which is needed to fix search functionality in sphinx1.8 (from readthedocs/sphinx_rtd_theme#672) (which is another reason for maintaining our own copy of this branch). The branch in this repository is now named version-dropdown-with-fixes (branching off of the version-dropdown branch in the sphinx_rtd_theme repository). In the long-term, I am a little concerned about using this theme that isn't showing any signs of being merged to the main branch of the readthedocs theme, but this has been working for us in other projects for the last 2 years, so I feel this is a reasonable approach in the short-medium term. The new process for building the documentation is given here: https://github.com/ESCOMP/CTSM/wiki/Directions-for-editing-CLM-documentation-on-github-and-sphinx Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/ctsm#239 (Move documentation source to master) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: Erik Kluzek CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- not run! izumi ------- not run! If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: NO Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): https://github.com/ESCOMP/CTSM/pull/954 =============================================================== =============================================================== Tag name: ctsm1.0.dev088 Originator(s): sacks (Bill Sacks) Date: Tue Apr 7 10:43:03 MDT 2020 One-line Summary: Fix bug in snow aerosol numerics (causes occasional HUGE aerosol values) Purpose of changes ------------------ Multiplications / divisions by dtime were missing in a few places, in some code that ensures that aerosol mass fluxes between the snow layers don't exceed the total mass. This bug was introduced in ctsm1.0.dev065, when a term was changed from a mass to a flux (per unit time), but some of the code wasn't changed correspondingly. In one case this bug caused a crash, but in other cases it just causes incorrect answers. The issue can be seen by examining the history variables SNOBCMCL, SNOOCMCL and SNODSTMCL. In the case that led to the crash, SNOBCMCL and SNOOCMCL became huge in one grid cell (~ 10^12) for a long period before the crash. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/ctsm#965 (Bug in snow aerosol numerics can cause aerosol masses to blow up, starting with ctsm1.0.dev065) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none (not checked, but none expected) Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: Erik Kluzek, Keith Oleson CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- ok izumi ------- pass ok means tests pass, some answer changes as expected If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: YES (Small) Summarize any changes to answers, i.e., - what code configurations: all - what platforms/compilers: all - nature of change (roundoff; larger than roundoff/same climate; new climate): larger than roundoff/same climate Answer changes arise due to a bug fix for a condition that only appears infrequently. When it does appear, answer changes can potentially be enormous (e.g., snow aerosol concentrations 20 orders of magnitude too large), though answer changes observed in the test suite all appear relatively small. Answer changes just appeared for the following tests: FAIL ERP_D_P36x2_Ld3.f10_f10_musgs.I2000Clm45BgcCrop.cheyenne_gnu.clm-no_subgrid_fluxes BASELINE ctsm1.0.dev087: DIFF FAIL SMS_Ld1.nldas2_rnldas2_mnldas2.I2000Ctsm50NwpSpNldasGs.cheyenne_gnu.clm-default BASELINE ctsm1.0.dev087: DIFF FAIL SMS_Ld1.nldas2_rnldas2_mnldas2.I2000Ctsm50NwpSpNldasRsGs.cheyenne_gnu.clm-default BASELINE ctsm1.0.dev087: DIFF FAIL ERI_D_Ld9.ne30_g16.I2000Clm50BgcCruGs.cheyenne_intel.clm-vrtlay BASELINE ctsm1.0.dev087: DIFF FAIL ERI_D_Ld9.T31_g37.I2000Clm50Sp.cheyenne_intel.clm-SNICARFRC BASELINE ctsm1.0.dev087: DIFF FAIL ERI_Ld9.f45_g37.I2000Clm50BgcCruGs.cheyenne_intel.clm-nofire BASELINE ctsm1.0.dev087: DIFF FAIL ERI_N2_Ld9.f19_g17.I2000Clm50BgcCrop.cheyenne_intel.clm-default BASELINE ctsm1.0.dev087: DIFF FAIL SMS_D_Lm6.f45_f45_mg37.I2000Clm50Fates.cheyenne_intel.clm-FatesColdDef BASELINE ctsm1.0.dev087: DIFF FAIL SMS_D_Lm6_P144x1.f45_f45_mg37.I2000Clm50Fates.cheyenne_intel.clm-FatesColdDef BASELINE ctsm1.0.dev087: DIFF FAIL SMS_Lm13.f19_g17.I2000Clm50BgcCrop.cheyenne_intel.clm-cropMonthOutput BASELINE ctsm1.0.dev087: DIFF Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): https://github.com/ESCOMP/CTSM/pull/971 =============================================================== =============================================================== Tag name: ctsm1.0.dev087 Originator(s): sacks (Bill Sacks) Date: Mon Apr 6 11:23:51 MDT 2020 One-line Summary: Change hard-coded soil layers in phenology to use a target depth Purpose of changes ------------------ Adds infrastructure for finding the soil layer containing a given depth (covered by unit tests). In phenology code, replaces hard-coded level 3 with a target soil depth. This value is read from the parameter file. The default value is set in a way that is bit-for-bit for standard CLM45 and CLM50 configurations, but this will change answers for non-SP configurations using the NWP soil layer structure or non-standard soil layer structures. This is important for the sake of supporting flexible soil layers, as in the NWP configuration. Also, some cleanup of the CNPhenology parameter reading code. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/ctsm#952 (Change hard coded soil layers in phenology to use a target depth) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [X] ctsm5_0-nwp -- only for non-SP cases [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): - New parameter added to NetCDF parameter file: phenology_soil_depth Substantial timing or memory changes: none expected (not checked) Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: Erik Kluzek CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- ok izumi ------- pass ok means tests pass, a few answer changes as expected If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: Yes (only Non-SP cases with NWP) Summarize any changes to answers, i.e., - what code configurations: Non-SP cases with NWP or non-standard soil layer structures - what platforms/compilers: all - nature of change (roundoff; larger than roundoff/same climate; new climate): larger than roundoff; same vs. new climate not investigated Differences in non-SP cases with soil layer structure different from that of standard clm45 or clm50. Differences observed in tests ERS_D_Ld3.f10_f10_musgs.I2000Clm50BgcCruGs.cheyenne_intel.clm-deepsoil_bedrock and SOILSTRUCTUD_Ld5.f10_f10_musgs.I2000Clm50BgcCropGs.cheyenne_intel.clm-default. ERP_P36x2_D_Ld5.f10_f10_musgs.I2000Ctsm50NwpBgcCropGswpGs.cheyenne_intel.clm-default had no differences; I'm guessing this is just because there didn't happen to be the appropriate PFTs present to show differences in this short, coarse-resolution test. Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): https://github.com/ESCOMP/CTSM/pull/963 =============================================================== =============================================================== Tag name: ctsm1.0.dev086 Originator(s): mvertens (Mariana Vertenstein) Date: Thu Mar 19 16:03:26 MDT 2020 One-line Summary: Updates to NUOPC cap and minor fixes to ncdio_pio Purpose of changes ------------------ This tag has updates to the NUOPC cap that are needed for the latest version of CMEPS. Now each component NUOPC cap needs to compute possibly time varying orbital parameters. Before, in cpl7, this was done in the driver and passed to the components via seq_infodata. However, now the driver is really just a simple ascii file - so those calculations needed to migrate to the component caps. In addition, this tag contains changes to ncdio_pio.F90 that improve performance in some situations. This also updates MOSART and RTM tags to include fixes for the NUOPC cap for MOSART and implementation of an initial NUOPC cap for RTM. Also, I1Pt compsets now use SROF rather than using an active runoff model in 'null' mode. This leads to differences in namelists, but not in results. Mariana Vertenstein implemented most of the changes in this tag; Jim Edwards implemented the ncdio_pio changes; Bill Sacks did final testing and integration. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/ctsm#943 (NUOPC tests will fail in ctsm1.0.dev085) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: - I1Pt tests and cases now use SROF rather than an active runoff model with a null grid (only impacts namelists, not results) - Changed nuopc test to use MOSART rather than SROF - Changed finidat file used by LII test to solve a problem that was uncovered with the ncdio_pio changes in this tag Code reviewed by: Bill Sacks CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- pass izumi ------- pass If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: YES (just for NUOPC configurations) Summarize any changes to answers, i.e., - what code configurations: Just some NUOPC configurations - what platforms/compilers: - nature of change (roundoff; larger than roundoff/same climate; new climate): Not investigated carefully, but likely to be larger than roundoff / same climate (due to introduction of time-varying orbital parameters) Differences are expected in NUOPC configurations using time-varying orbital parameters. Note that no differences showed up in the test suite, however, because the only nuopc test in the test suite is new in this tag (so did not have baselines). Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): - mosart: mosart1_0_35 -> mosart1_0_36 - rtm: rtm1_0_70 -> rtm1_0_71 - cmeps Pull Requests that document the changes (include PR ids): https://github.com/ESCOMP/CTSM/issues/943 =============================================================== =============================================================== Tag name: ctsm1.0.dev085 Originator(s): sacks (Bill Sacks) Date: Mon Mar 16 09:19:57 MDT 2020 One-line Summary: Update cime Purpose of changes ------------------ Update to latest version of cime master Bugs fixed or introduced ------------------------ Known bugs introduced in this tag (include github issue ID): - https://github.com/ESCOMP/CTSM/issues/943 - NUOPC tests will fail in ctsm1.0.dev085 Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): - NUOPC cases will fail from a standalone checkout Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: self CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- ok izumi ------- pass Two test failures: PEND SMS_D_Vnuopc.f10_f10_musgs.I2000Clm50SpRsGs.cheyenne_intel.clm-default SHAREDLIB_BUILD I have added this to the expected fails list for now FAIL SMS_P720x1_Ln6.hcru_hcru.I2000Clm50BgcCruGs.cheyenne_intel.clm-coldStart BASELINE ctsm1.0.dev084: ERROR BFAIL baseline directory '/glade/p/cgd/tss/ctsm_baselines/ctsm1.0.dev084/SMS_P720x1_Ln6.hcru_hcru.I2000Clm50BgcCruGs.cheyenne_intel.clm-coldStart' does not exist This test was added in the previous tag, but apparently wasn't run from there Note: I did testing using cime hash 4f1880756a66ff12d304937d3c166d451ea50cf9, then updated to the tag (cime5.8.17). The two are identical except for the cime ChangeLog. If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: NO Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): cime: branch_tags/cime5.8.15_a01 -> cime5.8.17 Pull Requests that document the changes (include PR ids): none =============================================================== =============================================================== Tag name: ctsm1.0.dev084 Originator(s): negins (Negin Sobhani,UCAR/CSEG,303-497-1224) Date: Wed Feb 19 15:05:03 MST 2020 One-line Summary: Bounds assertion for C isotopes when threading is on Purpose of changes ------------------ Previously, the bounds of c13 and c14 in subroutines TruncateCStates and TruncateCandNStates were not checked when threading was on. This PR allows us to handle c13 and c14 arrays like everything else in CTSM including (bounds checking). This PR resolves issue ESCOMP/ctsm#811 Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): -ESCOMP/ctsm#811 Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: - There was an issue (wrong wallclock time ) for one of the tests in the test definitions which got fixed: (SMS_D_Vnuopc.f10_f10_musgs.I2000Clm50SpRsGs..) - Also, I tried making the wallclocktime for all tests in cime_config/testdefs/testlist_clm.xml consistent in the format of hh:mm:ss. Code reviewed by: Bill Sacks CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- OK izumi ------- PASS If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: NO Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): (https://github.com/ESCOMP/CTSM/pull/923) =============================================================== =============================================================== Tag name: ctsm1.0.dev083 Originator(s): negins (Negin Sobhani,UCAR/CSEG,303-497-1224) Date: Thu Feb 6 12:12:26 MST 2020 One-line Summary: Some BFB Fixes: Resolve issues #683, # 874, #878, # 885, # 745, #838 Purpose of changes ------------------ Resolved issues: ESCOMP/ctsm#683 : Writing out FATES parameters only on masterproc. ESCOMP/ctsm#874 : Changing the number of timesteps to run for SMS_P720x1_Ln3.hcru_hcru.I2000Clm50BgcCruGs.cheyenne_intel.clm-coldStar to prevent failing with CMEPS. ESCOMP/ctsm#878 : Removing the unused atm2lnd field. ESCOMP/ctsm#885 : Changing the defaults for br_root in namelist_defaults to enable FUN with CLM4.5. ESCOMP/ctsm#745 : Removing ReadNL private subroutine from initVerticalMod.F90. ESCOMP/ctsm#838 : Clarifying ZISNO in the variable long name. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): Resolves : - ESCOMP/ctsm#683 - ESCOMP/ctsm#874 - ESCOMP/ctsm#878 - ESCOMP/ctsm#885 - ESCOMP/ctsm#745 - ESCOMP/ctsm#838 Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): Not any changes in parameter values. changing the default for br_root in namelist_defaults to enable FUN with CLM4.5. Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: Changed the number of timesteps from Ln3 to Ln6 for SMS_P720x1_Ln3.hcru_hcru.I2000Clm50BgcCruGs.cheyenne_intel.clm-coldStar Code reviewed by: Bill Sacks CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- OK izumi ------- PASS If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: NO Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): https://github.com/ESCOMP/CTSM/pull/897 =============================================================== =============================================================== Tag name: ctsm1.0.dev082 Originator(s): oleson (Keith Oleson) Date: Sat Feb 1 09:28:41 MST 2020 One-line Summary: Rename variables to avoid confusion; fix QSNOEVAP diagnostic Purpose of changes ------------------ Renamed variables as discussed in ESCOMP/ctsm#118 throughout the code. Also made a couple of minor changes to fix a couple of potential problems with these variables as described in the branch commit logs. Tested for bfb before changing the history field variables themselves. These changes are all bfb (with the exception of QSNOEVAP) for a 1 month global 2deg simulation, but may not be bfb under all conditions. QSNOEVAP is not bfb because I've added in qflx_ev_snow for lakes. Also, update cime slightly with a fix for izumi machine updates Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/ctsm#118 Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none expected (not checked) Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: Bill Sacks CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- ok izumi ------- ok ok: tests pass, answer changes in QSNOEVAP but nothing else (as expected) If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: YES (only one diagnostic field) Summarize any changes to answers, i.e., - what code configurations: all - what platforms/compilers: all - nature of change (roundoff; larger than roundoff/same climate; new climate): Larger than roundoff, but only in the diagnostic field QSNOEVAP Possibility of other answer changes in rare circumstances, not observed in testing. Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): - cime: cime5.8.15 -> branch_tags/cime5.8.15_a01 Minor update for izumi machine changes Pull Requests that document the changes (include PR ids): https://github.com/ESCOMP/CTSM/pull/893 =============================================================== =============================================================== Tag name: ctsm1.0.dev081 Originator(s): slevis (Samuel Levis,SLevis Consulting LLC,303-665-1310) Date: Mon Jan 13 15:37:07 MST 2020 One-line Summary: Speed up restart writes Purpose of changes ------------------ E3SM#3163 made some changes to speed up restart writes. I use E3SM#3163 function GetGlobalIndexArray in place of GetGlobalIndex so as to get global indices by array rather than looping over every element. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): #801 Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: Limited testing did not show substantial timing changes Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: @billsacks @bishtgautam CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - tools-tests (test/tools): cheyenne - PTCLM testing (tools/shared/PTCLM/test): cheyenne - python testing (see instructions in python/README.md; document testing done): (any machine) - regular tests (aux_clm): cheyenne ---- OK izumi ------- PASS If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: No @billsacks recommended running a global test in the branch vs. in master (i.e. with vs. without the code mods) to compare restart files because the test suite does not compare restart files and an error in the subgrid indices on the restart file might go undetected. I ran this test: ERS_D_Ld6.f10_f10_musgs.I1850Clm45BgcCrop.cheyenne_intel.clm-clm50CMIP6frc and got identical values for all variables in the two restart files. Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): https://github.com/ESCOMP/ctsm/pull/880 =============================================================== =============================================================== Tag name: ctsm1.0.dev080 Originator(s): sacks (Bill Sacks) Date: Mon Dec 2 12:23:22 MST 2019 One-line Summary: Update externals, minor fixes to work with latest cime, get nuopc cap working Purpose of changes ------------------ (1) Update to latest version of externals. This includes a significant cime update. (2) Minor fixes to get things working with the latest version of cime. (3) Fix from Mariana Vertenstein to get NUOPC cap working, fixes to externals cfg files to allow running NUOPC cases out of the box, and add a NUOPC test. Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): CIME update brings in some changes, e.g., in mapping files; otherwise, none. Substantial timing or memory changes: not investigated Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: - Added NUOPC test: SMS_D_Vnuopc.f10_f10_musgs.I2000Clm50SpRsGs.cheyenne_intel.clm-default Code reviewed by: self CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- ok izumi ------- ok ok means tests passed, some baseline comparisons failed as expected Note: Most testing was done on ad1ad752; then, from the latest version, I ran a 2-test test list of: - SMS_D_Vnuopc.f10_f10_musgs.I2000Clm50SpRsGs.cheyenne_intel.clm-default - SMS_D.f10_f10_musgs.I2000Clm50SpRsGs.cheyenne_intel.clm-default In addition to verifying that the above tests passed, I also inspected the sharedlib build area to verify that separate builds of the clm library were done for mct and nuopc. If the tag used for baseline comparisons was NOT the previous tag, note that here: - cheyenne tests compared against ctsm1.0.dev078, since cheyenne testing was not run on ctsm1.0.dev079. Answer changes -------------- Changes answers relative to baseline: YES (only f10, same climate) Summarize any changes to answers, i.e., - what code configurations: Cases at f10 resolution - what platforms/compilers: all - nature of change (roundoff; larger than roundoff/same climate; new climate): Larger than roundoff, but probably same climate Answer changes are due to https://github.com/esmci/cime/pull/3186 Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): - cime: branch_tags/cime5.8.3_chint17-05 -> cime5.8.15 - rtm: rtm1_0_68 -> rtm1_0_70 - mosart: mosart1_0_34 -> mosart1_0_35 - cmeps: 471ac52 -> 181ff1ed9dfb279e619e8a2173f43baf7bf1dce3 Also, note that cmeps is now checked out automatically Pull Requests that document the changes (include PR ids): https://github.com/ESCOMP/CTSM/pull/845 =============================================================== =============================================================== Tag name: ctsm1.0.dev079 Originator(s): sacks (Bill Sacks) Date: Mon Nov 4 14:41:41 MST 2019 One-line Summary: Change a few uses of shr_kind Purpose of changes ------------------ Change a few uses of shr_kind to work with latest cime. These changes are from Mariana Vertenstein. Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: Bill Sacks CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- NOT RUN! izumi ------- pass If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: NO Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): none =============================================================== =============================================================== Tag name: ctsm1.0.dev078 Originator(s): oleson (Keith Oleson) Date: Thu Oct 31 12:30:00 MDT 2019 One-line Summary: Fix rootr calculation with use_hydrstress true Purpose of changes ------------------ In cases with use_hydrstress=.true., both rootr_patch and rootr_col had problems. This affected the diagnostic fields ROOTR and ROOTR_COLUMN, and (in BGC cases) fed back on other parts of the system via affecting the methane code. See https://github.com/ESCOMP/CTSM/issues/812 for more details. This tag fixes this issue via these specific changes: 1. Calculate btran in SoilMoistStressMod.F90 regardless of use_hydrstress 2. Comment on the applicability of this calculation and that of rootr in SoilMoistStressMod.F90 3. Add the following comment to uses of rootr and rootr_col throughout the code: "SMS method only" 4. Allow ROOTR_COL history field only if use_hydrstress = .false. 5. Remove the calculation of rootr_col in subroutine Compute_EffecRootFrac_And_VertTranSink_HydStress within module SoilWaterPlantSinkMod.F90 since that calculation is invalid for PHS. 6. Normalize root resistances for pervious road over nlevgrnd, not nlevsoi Note that the rootr calculation is done using the non-PHS (i.e., SMS) method. So rootr is not really consistent with PHS; this is something that should probably be revisited later. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Fixes ESCOMP/ctsm#812 (Problems with rootr when use_hydrstress=.true.) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: not checked; none expected Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: Bill Sacks CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- ok izumi ------- ok ok means tests pass, answers change as expected If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: YES (PHS and methane) (same climate) Summarize any changes to answers, i.e., - what code configurations: Clm50Bgc (cases with both use_hydrstress=.true. and methane active) - what platforms/compilers: all - nature of change (roundoff; larger than roundoff/same climate; new climate): larger than roundoff / same climate For details on differences, see https://github.com/ESCOMP/CTSM/issues/812#issuecomment-547924289 If this tag changes climate describe the run(s) done to evaluate the new climate (put details of the simulations in the experiment database) - casename: clm50_ctsm10d074_2deg_GSWP3V1_rootr_hist URL for LMWG diagnostics output used to validate new climate: http://webext.cgd.ucar.edu/I20TR/clm50_ctsm10d074_2deg_GSWP3V1_rootr_hist/lnd/clm50_ctsm10d074_2deg_GSWP3V1_rootr_hist.1995_2014-clm50_ctsm10d074_2deg_GSWP3V1_hist.1995_2014/setsIndex.html Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): none =============================================================== =============================================================== Tag name: ctsm1.0.dev077 Originator(s): sacks (Bill Sacks) Date: Sun Oct 27 16:23:39 MDT 2019 One-line Summary: Consistently use frac_sno_eff rather than frac_sno in qg calculations Purpose of changes ------------------ Consistently use frac_sno_eff rather than frac_sno in qg calculations The change here only changes answers for (1) urban pervious road, and (2) more broadly with use_subgrid_fluxes = .false. Previously, code in CalculateSurfaceHumidity was inconsistent regarding the use of frac_sno vs. frac_sno_eff. I wanted to make this consistent to facilitate upcoming cleanup of CalculateSurfaceHumidity. I went back and forth as to which one we should use here. (See also https://github.com/ESCOMP/CTSM/issues/822). Argument for frac_sno_eff: This is consistent with the use of frac_sno_eff elsewhere in the calculation of surface fluxes. Argument for frac_sno: Even in CLM4.0 (prior to the introduction of frac_sno_eff and frac_h2osfc), frac_sno was used to calculate surface relative humidity. That makes me think that the parameterizations were improved by using frac_sno here, even though frac_sno at that point was mainly just used for albedo / radiation calculations. So sticking with frac_sno would be more consistent with what has been done for a long time, and would presumably be a bit better scientifically. However, when I tried changing this to use frac_sno rather than frac_sno_eff: qg(c) = frac_sno_eff(c)*qg_snow(c) + (1._r8 - frac_sno_eff(c) - frac_h2osfc(c))*qg_soil(c) & + frac_h2osfc(c) * qg_h2osfc(c) I got a crash with use_subgrid_fluxes false (test SMS_D_Ld1_P4x1.f10_f10_musgs.I2000Clm45BgcCropQianRsGs.bishorn_gnu.clm-no_subgrid_fluxes), due to a check I put in relatively recently to ensure that the top-layer soil liquid never goes significantly negative: WJS 218: frac_sno, frac_sno_eff, frac_h2osfc, snl = 6.6447358408778726E-005 1.0000000000000000 0.0000000000000000 -1 qg_snow, qg_soil, qg_h2osfc, qg = 3.8414454029812475E-003 3.6634834595404343E-003 3.6635176705822736E-003 3.6634952846414733E-003 ERROR: In UpdateState_TopLayerFluxes, h2osoi_liq has gone significantly negative Bulk/tracer name = bulk c, lev_top(c) = 218 0 h2osoi_liq_top_orig = 6.6448689668604415E-004 h2osoi_liq = -2.6754040213948407E-004 frac_sno_eff = 1.0000000000000000 qflx_liq_grnd*dtime = 0.0000000000000000 qflx_dew_grnd*dtime = 0.0000000000000000 qflx_evap_grnd*dtime = 9.3202729882552822E-004 ENDRUN: ERROR: In UpdateState_TopLayerFluxes, h2osoi_liq has gone significantly negative What seems to be going on is: With a little bit of snow present, when I use frac_sno rather than frac_sno_eff, we end up with a qg value very close to that of soil. But I think the later evaporation happens from the snow, at which point I guess the dryer air above leads to there being too large evaporation from the snow. So in the end, I decided to consistently use frac_sno_eff, even though this might be a small step backwards in terms of the accuracy of surface humidity for urban pervious road and with use_subgrid_fluxes = .false. Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: self (but approved conceptually by Sean Swenson and Keith Oleson) CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- ok izumi ------- ok ok means tests pass, answers change as expected If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: YES (Small) Summarize any changes to answers, i.e., - what code configurations: all - what platforms/compilers: all - nature of change (roundoff; larger than roundoff/same climate; new climate): Expected to be larger than roundoff/same climate, but not investigated carefully For the default configuration with use_subgrid_fluxes = .true., differences are restricted to urban pervious road. I spot-checked a few tests to ensure that we only saw differences in FSH for urban landunits; this was generally true, with the exception of ERP_Ly3_P72x2.f10_f10_musgs.IHistClm50BgcCrop.cheyenne_intel.clm-cropMonthOutput, which also had differences over crop and natural veg landunits. This spillover effect seems to be related to the non-local interactions that arise from methane's dependence on TWS, though I don't understand why I only saw this in a transient case, and not in similar non-transient cases. See https://github.com/ESCOMP/CTSM/issues/658#issuecomment-546740771 for more details. For the non-default use_subgrid_fluxes = .false., there differences in the natural vegetation and crop landunits, too. Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): none =============================================================== =============================================================== Tag name: ctsm1.0.dev076 Originator(s): sacks (Bill Sacks) Date: Fri Oct 25 16:04:20 MDT 2019 One-line Summary: Set frac_sno_eff=0 if frac_sno is 0; avoid unnecessary calls to QSat Purpose of changes ------------------ Two changes that are bit-for-bit except for changes in the FSNO_EFF diagnostic field, both with an eye towards cleaning up SurfaceHumidityMod in preparation for adding water tracers to that module: (1) Set frac_sno_eff to 0 if frac_sno is 0. Previously, for cases where frac_sno_eff was a binary 0/1 rather than being set equal to frac_sno, frac_sno_eff was always being set to 1, even if there was no snow. Sean Swenson thinks that didn't matter, but I found it confusing, and it could be more problematic with some upcoming changes I plan to make (where I'm going to use frac_sno_eff as a weighting factor in calculating qg terms, including for landunits that use a binary 0/1 frac_sno_eff.) (2) Avoid unnecessary calls to QSat, since this function is relatively expensive. Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: No substantial changes; from a single run of PFS_Ld20.f09_g17.I2000Clm50BgcCrop.cheyenne_intel, max timing of the main changed section (bgp1) decreased by about 10%; this may be within machine variability, though. Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: self CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- ok izumi ------- ok ok means tests pass, answer changes as expected just for FSNO_EFF If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: YES, but just for the diagnostic field FSNO_EFF Summarize any changes to answers, i.e., - what code configurations: all - what platforms/compilers: all - nature of change (roundoff; larger than roundoff/same climate; new climate): diagnostic only: just changes FSNO_EFF Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): none =============================================================== =============================================================== Tag name: ctsm1.0.dev075 Originator(s): sacks (Bill Sacks) Date: Fri Oct 25 15:44:52 MDT 2019 One-line Summary: Lakes: Adjust frac_sno after updating 0-layer snow pack for dew & sublimation Purpose of changes ------------------ Main change is: for lakes, possibly adjust frac_sno after updating 0-layer snow pack for dew & sub, ensuring that zero vs. non-zero values of frac_sno agree with whether the snow pack has zero or non-zero amounts of snow. The main motivation for this is to ensure that, if we have any snow, then frac_sno should be non-zero. Along with this, I'm also ensuring that, if we go down to 0 snow, then we also go down to 0 frac_sno. A related change was to move the code ensuring that h2osno_no_layers doesn't go negative: I have moved this to before h2osno_no_layers is referenced, to assure we don't get a slightly negative snow depth. (I haven't noticed any problems due to the old placement, but it seems safer to put the limiting before the reference to h2osno_no_layers.) Also, an unrelated change in run_sys_tests: Add option for rerunning existing failed tests. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/ctsm#827 (Lake frac_sno can remain 0 even when there is some snow addition from frost) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: None expected Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: self CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - pass (on my mac: 'make test' with python2 & python3 and 'make lint') regular tests (aux_clm): cheyenne ---- ok izumi ------- ok ok means tests pass, answers change as expected If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: YES (Small) Summarize any changes to answers, i.e., - what code configurations: all - what platforms/compilers: all - nature of change (roundoff; larger than roundoff/same climate; new climate): Expected to be larger than roundoff/same climate Changes are introduced only over lake columns, and these changes should be small, since they only relate to the setting of frac_sno when there is dew or sublimation on a very thin snow pack. Although the direct changes should only affect lakes, I did observe changes in crop columns in a transient test that I spot-checked (ERP_Ly3_P72x2.f10_f10_musgs.IHistClm50BgcCrop.cheyenne_intel.clm-cropMonthOutput) - maybe due to some of the non-local interactions in https://github.com/ESCOMP/CTSM/wiki/Parameterizations-that-can-cause-nonlocal-interactions, or maybe due to some other non-local interaction, either known or buggy. (In non-transient tests that I spot checked (ERP_P72x2_Lm36.f10_f10_musgs.I2000Clm50BgcCrop.cheyenne_intel.clm-clm50cropIrrigMonth_interp and SMS_Lm13.f19_g17.I2000Clm50BgcCrop.cheyenne_intel.clm-cropMonthOutput): diffs in FSH in h1 file are just over lake landunits, as expected.) Update: it seems that these changes in crop columns arise from the nonlocal interactions caused by the dependence of methane on TWS, though I don't understand why I only saw this in a transient case, and not in similar non-transient cases. See https://github.com/ESCOMP/CTSM/issues/658#issuecomment-546740771 for more details. Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): none =============================================================== =============================================================== Tag name: ctsm1.0.dev074 Originator(s): sacks (Bill Sacks) Date: Wed Oct 23 19:25:21 MDT 2019 One-line Summary: For lakes: when reading finidat, set frac_sno=1 if h2osno_total > 0 Purpose of changes ------------------ Due to ESCOMP/ctsm#783, frac_sno used to be 0 for all lake points. That was mostly fixed, but the issue is still present on initial conditions files that were generated prior to that fix (which includes most or all of our out-of-the-box restart files). This causes frac_sno to be 0 for lake points at the start of the simulation, even if there is a snow pack present. Currently this doesn't cause significant problems, but I'd like to change the calculation of frac_sno_eff so that it is 0 if frac_sno is 0 - and then this frac_sno problem becomes an issue (causing divide by 0 in at least one place). However, further problems were introduced if I tried to always apply this correction: It appears that sometimes the restart file legitimately has frac_sno == 0 for lake columns with non-zero snow cover. To avoid changing answers for newer restart files, I am writing metadata to the restart file documenting whether the fix has already been applied on that restart file, and if so, we avoid reapplying the fix. Getting this restart metadata correct was tricky, especially when running init_interp. I have introduced a new module to encapsulate this complexity. This can be used in general for writing / reading metadata on which issues have been fixed on a given restart file. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): [If none, remove this line] - Resolves ESCOMP/ctsm#783 Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): - The setting of initial lake frac_sno to 1 isn't exactly correct, but it's better than the old value of 0. Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none - I didn't check the PFS test, but the only change that could have likely led to performance changes was the addition of a metadata copy in init_interp. I checked the timing of this piece, and found it to be a negligible portion of the total init_interp time (https://github.com/ESCOMP/CTSM/pull/825#issuecomment-545219601). Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: Erik Kluzek gave this an initial look and gave general approval for the restart file metadata mechanism I have added. He may give it a more careful look, possibly accompanied by a further cleanup tag addressing any issues he finds. CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- ok izumi ------- ok ok means tests passed, baselines fail as expected. Note that the FUNIT test failed initially; I got it passing on my Mac, but didn't rerun it on cheyenne. If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: YES (same climate, only at startup) Summarize any changes to answers, i.e., - what code configurations: all - what platforms/compilers: all - nature of change (roundoff; larger than roundoff/same climate; new climate): Larger than roundoff/same climate The only change is in initial lake frac_sno. This should only affect the start of the simulation, and only over lake points. I verified that changes are only over lake points via: For both SMS_D_Ld1.f09_g17.I1850Clm50Sp.cheyenne_intel.clm-default (which doesn't use init_interp) and SMS_D_Ld3.f10_f10_musgs.I2000Clm50BgcCruGs.cheyenne_gnu.clm-default (which uses init_interp): I checked which points have differences in FSH in the h1 (vector) history file. Diffs were only over lake points. Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): https://github.com/ESCOMP/CTSM/pull/825 =============================================================== =============================================================== Tag name: ctsm1.0.dev073 Originator(s): sacks (Bill Sacks) Date: Tue Oct 22 06:14:24 MDT 2019 One-line Summary: Fix bug in calculation of dqgdT Purpose of changes ------------------ The calculation of dqgdT, which gives the derivative, d(qg)/dT, was incorrect for soil/crop landunits when there is no snow. This tag fixes that bug. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/CTSM#824 (dqgdT incorrect when there is no snow) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: not checked, but none expected Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: Sean Swenson reviewed and approved my plan, though not the final code CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- ok izumi ------- ok ok means tests pass, answers change as expected If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: YES (small) Summarize any changes to answers, i.e., - what code configurations: all - what platforms/compilers: all - nature of change (roundoff; larger than roundoff/same climate; new climate): Larger than roundoff/same climate I checked the impact from one-day averages in test SMS_D_Ld1_P4x1.f10_f10_musgs.I2000Clm50BgcCropQianRsGs.bishorn_gnu.clm-default, by looking at differences in EFLX_LH_TOT. The impact is small: daily-average differences in that variable are typically 0.01 W m-2 or less; most grid cells have significantly less difference than that; only one or a few f10 grid cells have significantly greater differences, with the biggest difference being about 0.1 W m-2. Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): none =============================================================== =============================================================== Tag name: ctsm1.0.dev072 Originator(s): mvertens (Mariana Vertenstein) Date: Tue Oct 15 09:41:50 MDT 2019 One-line Summary: Add NUOPC cap Purpose of changes ------------------ Add NUOPC cap in order to run with the new NUOPC driver/mediator. There was some code movement / refactoring of existing code to support this, with the biggest change being the introduction of a new subdirectory under src/cpl. Most changes are from Mariana Vertenstein and Jim Edwards. Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none expected (not checked) Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): - NUOPC cap currently does not have any tests in the aux_clm test suite - There are currently no tests ensuring that the nuopc and mct caps give bit-for-bit answers, though the nuopc_cap_bfb testmod could be used to support such a test Changes to tests or testing: none Code reviewed by: Bill Sacks CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- pass izumi ------- pass If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: NO Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): - mosart: mosart1_0_33 -> mosart1_0_34 (supports nuopc cap) Pull Requests that document the changes (include PR ids): none =============================================================== =============================================================== Tag name: ctsm1.0.dev071 Originator(s): sacks (Bill Sacks) Date: Fri Oct 11 07:03:40 MDT 2019 One-line Summary: Split CanopyTemperature into separate pieces Purpose of changes ------------------ CanopyTemperature was doing a few unrelated things, none of which actually had anything to do with calculating canopy temperature. At the driver level, I have split CanopyTemperature into BiogeophysPreFluxCalcs and CalculateSurfaceHumidity. BiogeophysPreFluxCalcs calls out to a number of separate routines, some in that same module and some elsewhere, including a new routine in FrictionVelocityMod. Also some minor related cleanup, including - Made FrictionVelocityMod object-oriented - A bit of rework of how qflx_evap_veg and qflx_tran_veg are set. This changes answers for the (default inactive) QFLX_EVAP_VEG history field (see also ESCOMP/ctsm#816). - Removed qsat conditionals that should never be invoked Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Fixes ESCOMP/ctsm#816 (QFLX_EVAP_VEG should be 0 over lakes) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: self CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- pass izumi ------- pass If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: YES, but only for the (default-inactive) field QFLX_EVAP_VEG Summarize any changes to answers, i.e., - what code configurations: all - what platforms/compilers: all - nature of change (roundoff; larger than roundoff/same climate; new climate): larger than roundoff, but only for the default-inactive field QFLX_EVAP_VEG Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): none =============================================================== =============================================================== Tag name: ctsm1.0.dev070 Originator(s): sacks (Bill Sacks) Date: Wed Oct 9 06:16:39 MDT 2019 One-line Summary: Fix for writing 0-d variables with PIO2 Purpose of changes ------------------ Previously, there were problems when writing 0-d (scalar) variables with PIO2 in DEBUG mode. This change fixes that issue. Jim Edwards put in place an initial fix; I put in place the final fix based on changes Mariana Vertenstein made in MOSART. I have not tested this with PIO2, but am hopeful that it fixes the PIO2 problem since the similar MOSART changes fixed that problem. Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: self; earlier version reviewed by Erik Kluzek and Mariana Vertenstein CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- pass izumi ------- pass If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: NO Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): https://github.com/ESCOMP/CTSM/pull/810 =============================================================== =============================================================== Tag name: ctsm1.0.dev069 Originator(s): sacks (Bill Sacks) Date: Mon Oct 7 10:22:22 MDT 2019 One-line Summary: Misc. code cleanup and minor bug fixes Purpose of changes ------------------ Resolve a variety of "simple bfb" issues. Note that some of these result in answer changes for select diagnostic fields. See specific notes below for details, under "Issues fixed" and answer changes. The fix for ESCOMP/ctsm#27 changes answers for some diagnostic fields, as noted below. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/ctsm#27 (BTRAN should be spval over all special landunits) - Also fixed some related fields, as noted below, under answer changes - Resolves ESCOMP/ctsm#31 (some CLM history fields don't restart properly) - This was mostly resolved previously; here I just did some final cleanup and fixing of NFIRE - Resolves ESCOMP/ctsm#48 (code for initializing some isotope carbon state variables looks wrong) - Removed the incorrect code, which was never being exercised - Resolves ESCOMP/ctsm#58 (Change SHR_ASSERT calls to avoid using errMsg) - Purpose was to speed up DEBUG runs on some machines/compilers - Resolves ESCOMP/ctsm#111 (Precision of constants in SoilTemperature.F90 is not double) - No answer changes, because 1 (single prec) = 1 (double prec) - Resolves ESCOMP/ctsm#334 (Introduce a new activelayer_type) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: potentially speeds up debug runs on some machines/compilers Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: self CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- ok izumi ------- ok ok means tests pass, some baseline comparisons fail as expected If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: YES, but just for select diagnostic fields Summarize any changes to answers, i.e., - what code configurations: ALL - what platforms/compilers: ALL - nature of change (roundoff; larger than roundoff/same climate; new climate): Greater than roundoff, but only in select diagnostic field - no impact on climate There were changes in the following diagnostic fields, which are now averaged over only vegetated landunits (see ESCOMP/ctsm#27): - BTRAN - BTRANMN - RSSHA - RSSUN - ROOTR (inactive by default) - ROOTR_COLUMN (inactive by default) - RRESIS (inactive by default) Changes are also possible in rare circumstances for the NFIRE diagnostic field, in transient cases, although no differences were observed for this field in the test suite (see ESCOMP/ctsm#31). Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): none =============================================================== =============================================================== Tag name: ctsm1.0.dev068 Originator(s): sacks (Bill Sacks) Date: Mon Sep 30 09:57:04 MDT 2019 One-line Summary: Add water tracers to CombineSnowLayers, DivideSnowLayers, ZeroEmptySnowLayers Purpose of changes ------------------ Add water tracers to CombineSnowLayers, DivideSnowLayers and ZeroEmptySnowLayers. This completes the tracerization of the core snow code: most or all of the code in SnowHydrologyMod is now tracerized. In contrast to most routines that I have tracerized, the tracerization of these routines is done by embedding loops over tracers inside existing code structures, rather than splitting out separate routines that operate on bulk or tracer quantities. This is because (1) it didn't make sense to introduce explicit fluxes for the state updates done here, so we didn't have the extra need of having separate routines for flux calculations vs. state updates; (2) the state updates were embedded in loops over levels and conditionals in a way that would have been awkward and messy to try to separately update bulk-and-tracer quantities (h2osoi_liq, h2osoi_ice, etc.) from bulk-only quantities (dz, etc.). Also, a bit of reordering of LakeHydrologyMod in order to keep the snow code more together. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/ctsm#778 (Implement water tracers for combining and dividing snow layers) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: I don't think so Based on the changes introduced, I expect *very* small increases in runtime of the modified routines (which would show up in the hydro_without_drainage timer), even for cases without water tracers, but I wouldn't expect this to have a significant impact. From PFS test, and specifically looking at the timing of hydro_without_drainage in that test, I see a possible small increase (a few percent) in the timing of that section, though the differences are within typical machine variability, so don't seem necessarily significant. Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: self CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- pass izumi ------- pass If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: NO Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): none =============================================================== =============================================================== Tag name: ctsm1.0.dev067 Originator(s): erik (Erik Kluzek,UCAR/TSS,303-497-1326) Date: Thu Sep 26 15:02:39 MDT 2019 One-line Summary: Add config_archive.xml to CTSM cime_config -- no testing Purpose of changes ------------------ Add config_archive.xml to cime_config for CTSM. Update the CODE_OF_CONDUCT to the August/2019 UCAR version. No testing done. Bugs fixed or introduced ------------------------ Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): None Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): None Changes made to namelist defaults (e.g., changed parameter values): None Changes to the datasets (e.g., parameter, surface or initial files): None Substantial timing or memory changes: None Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): None Changes to tests or testing: None Code reviewed by: self CTSM testing: doc (no standard testing done) [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - PASS src unit-tests: cheyenne - PASS regular tests (aux_clm): None If the tag used for baseline comparisons was NOT the previous tag, note that here: previous Answer changes -------------- Changes answers relative to baseline: No! Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): None Pull Requests that document the changes (include PR ids): None (https://github.com/ESCOMP/ctsm/pull) =============================================================== =============================================================== Tag name: ctsm1.0.dev066 Originator(s): sacks (Bill Sacks) Date: Fri Sep 20 09:07:23 MDT 2019 One-line Summary: Add water tracers to SnowCapping Purpose of changes ------------------ Update water tracer terms in SnowCapping. As I have done elsewhere, this has involved splitting each piece of SnowCapping into its own subroutine, which operates either on bulk, tracers, or bulk + tracers. This is supported by calls to ResetCheckedTracers prior to the call to SnowCapping, since some intervening code is not yet tracerized. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/ctsm#777 (Implement water tracers for SnowCapping) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Small increase in total runtime of PFS test, but the relevant timing section, hydro_without_drainage, took slightly *less* time compared with the last tag. Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: self CTSM testing: build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- pass izumi ------- pass If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: NO Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): none =============================================================== =============================================================== Tag name: ctsm1.0.dev065 Originator(s): sacks (Bill Sacks) Date: Sat Sep 14 12:52:42 MDT 2019 One-line Summary: Add water tracers to SnowWater Purpose of changes ------------------ Update water tracer terms in SnowWater. As I have done elsewhere, this has involved splitting each piece of SnowWater into its own subroutine, which operates either on bulk, tracers, or bulk + tracers. This is supported by calls to ResetCheckedTracers prior to the call to SnowWater, since some intervening code is not yet tracerized. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/ctsm#776 (Implement water tracers for SnowWater) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: self CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- ok izumi ------- ok ok means tests pass, baselines fail as expected If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: YES (roundoff) Summarize any changes to answers, i.e., - what code configurations: all - what platforms/compilers: all - nature of change (roundoff; larger than roundoff/same climate; new climate): Roundoff level: answer changes just arise because I have made qflx_snow_percolation_col a true flux, with units of per-unit-time; this involved dividing by dtime, and later multiplying by dtime when it is applied to states. Previously, the corresponding variable (qout / qin) was not per-unit-time. If bitwise differences were observed, how did you show they were no worse than roundoff? I ran comparisons against a version off of master that had the minimal changes to exhibit these roundoff-level changes. Diffs for this are below. My branch is bit-for-bit with this version. diff --git a/src/biogeophys/SnowHydrologyMod.F90 b/src/biogeophys/SnowHydrologyMod.F90 index cad47e28..0abcc49f 100644 --- a/src/biogeophys/SnowHydrologyMod.F90 +++ b/src/biogeophys/SnowHydrologyMod.F90 @@ -1140,17 +1140,17 @@ subroutine SnowWater(bounds, & c = filter_snowc(fc) if (j >= snl(c)+1) then - h2osoi_liq(c,j) = h2osoi_liq(c,j) + qin(c) + h2osoi_liq(c,j) = h2osoi_liq(c,j) + qin(c)*dtime - mss_bcphi(c,j) = mss_bcphi(c,j) + qin_bc_phi(c) - mss_bcpho(c,j) = mss_bcpho(c,j) + qin_bc_pho(c) - mss_ocphi(c,j) = mss_ocphi(c,j) + qin_oc_phi(c) - mss_ocpho(c,j) = mss_ocpho(c,j) + qin_oc_pho(c) + mss_bcphi(c,j) = mss_bcphi(c,j) + qin_bc_phi(c)*dtime + mss_bcpho(c,j) = mss_bcpho(c,j) + qin_bc_pho(c)*dtime + mss_ocphi(c,j) = mss_ocphi(c,j) + qin_oc_phi(c)*dtime + mss_ocpho(c,j) = mss_ocpho(c,j) + qin_oc_pho(c)*dtime - mss_dst1(c,j) = mss_dst1(c,j) + qin_dst1(c) - mss_dst2(c,j) = mss_dst2(c,j) + qin_dst2(c) - mss_dst3(c,j) = mss_dst3(c,j) + qin_dst3(c) - mss_dst4(c,j) = mss_dst4(c,j) + qin_dst4(c) + mss_dst1(c,j) = mss_dst1(c,j) + qin_dst1(c)*dtime + mss_dst2(c,j) = mss_dst2(c,j) + qin_dst2(c)*dtime + mss_dst3(c,j) = mss_dst3(c,j) + qin_dst3(c)*dtime + mss_dst4(c,j) = mss_dst4(c,j) + qin_dst4(c)*dtime if (j <= -1) then ! No runoff over snow surface, just ponding on surface @@ -1167,8 +1167,8 @@ subroutine SnowWater(bounds, & qout(c) = max(0._r8,(vol_liq(c,j) & - ssi*eff_porosity(c,j))*dz(c,j)*frac_sno_eff(c)) end if - qout(c) = qout(c)*1000._r8 - h2osoi_liq(c,j) = h2osoi_liq(c,j) - qout(c) + qout(c) = (qout(c)*1000._r8)/dtime + h2osoi_liq(c,j) = h2osoi_liq(c,j) - qout(c)*dtime qin(c) = qout(c) ! mass of ice+water: in extremely rare circumstances, this can @@ -1187,7 +1187,7 @@ subroutine SnowWater(bounds, & if (qout_bc_phi(c) > mss_bcphi(c,j)) then qout_bc_phi(c) = mss_bcphi(c,j) endif - mss_bcphi(c,j) = mss_bcphi(c,j) - qout_bc_phi(c) + mss_bcphi(c,j) = mss_bcphi(c,j) - qout_bc_phi(c)*dtime qin_bc_phi(c) = qout_bc_phi(c) ! BCPHO: @@ -1196,7 +1196,7 @@ subroutine SnowWater(bounds, & if (qout_bc_pho(c) > mss_bcpho(c,j)) then qout_bc_pho(c) = mss_bcpho(c,j) endif - mss_bcpho(c,j) = mss_bcpho(c,j) - qout_bc_pho(c) + mss_bcpho(c,j) = mss_bcpho(c,j) - qout_bc_pho(c)*dtime qin_bc_pho(c) = qout_bc_pho(c) ! OCPHI: @@ -1205,7 +1205,7 @@ subroutine SnowWater(bounds, & if (qout_oc_phi(c) > mss_ocphi(c,j)) then qout_oc_phi(c) = mss_ocphi(c,j) endif - mss_ocphi(c,j) = mss_ocphi(c,j) - qout_oc_phi(c) + mss_ocphi(c,j) = mss_ocphi(c,j) - qout_oc_phi(c)*dtime qin_oc_phi(c) = qout_oc_phi(c) ! OCPHO: @@ -1214,7 +1214,7 @@ subroutine SnowWater(bounds, & if (qout_oc_pho(c) > mss_ocpho(c,j)) then qout_oc_pho(c) = mss_ocpho(c,j) endif - mss_ocpho(c,j) = mss_ocpho(c,j) - qout_oc_pho(c) + mss_ocpho(c,j) = mss_ocpho(c,j) - qout_oc_pho(c)*dtime qin_oc_pho(c) = qout_oc_pho(c) ! DUST 1: @@ -1223,7 +1223,7 @@ subroutine SnowWater(bounds, & if (qout_dst1(c) > mss_dst1(c,j)) then qout_dst1(c) = mss_dst1(c,j) endif - mss_dst1(c,j) = mss_dst1(c,j) - qout_dst1(c) + mss_dst1(c,j) = mss_dst1(c,j) - qout_dst1(c)*dtime qin_dst1(c) = qout_dst1(c) ! DUST 2: @@ -1232,7 +1232,7 @@ subroutine SnowWater(bounds, & if (qout_dst2(c) > mss_dst2(c,j)) then qout_dst2(c) = mss_dst2(c,j) endif - mss_dst2(c,j) = mss_dst2(c,j) - qout_dst2(c) + mss_dst2(c,j) = mss_dst2(c,j) - qout_dst2(c)*dtime qin_dst2(c) = qout_dst2(c) ! DUST 3: @@ -1241,7 +1241,7 @@ subroutine SnowWater(bounds, & if (qout_dst3(c) > mss_dst3(c,j)) then qout_dst3(c) = mss_dst3(c,j) endif - mss_dst3(c,j) = mss_dst3(c,j) - qout_dst3(c) + mss_dst3(c,j) = mss_dst3(c,j) - qout_dst3(c)*dtime qin_dst3(c) = qout_dst3(c) ! DUST 4: @@ -1250,7 +1250,7 @@ subroutine SnowWater(bounds, & if (qout_dst4(c) > mss_dst4(c,j)) then qout_dst4(c) = mss_dst4(c,j) endif - mss_dst4(c,j) = mss_dst4(c,j) - qout_dst4(c) + mss_dst4(c,j) = mss_dst4(c,j) - qout_dst4(c)*dtime qin_dst4(c) = qout_dst4(c) end if @@ -1280,9 +1280,9 @@ subroutine SnowWater(bounds, & do fc = 1, num_snowc c = filter_snowc(fc) ! Qout from snow bottom - qflx_snow_drain(c) = qflx_snow_drain(c) + (qout(c) / dtime) + qflx_snow_drain(c) = qflx_snow_drain(c) + qout(c) - qflx_rain_plus_snomelt(c) = (qout(c) / dtime) & + qflx_rain_plus_snomelt(c) = qout(c) & + (1.0_r8 - frac_sno_eff(c)) * qflx_liq_grnd(c) int_snow(c) = int_snow(c) + frac_sno_eff(c) & * (qflx_dew_snow(c) + qflx_dew_grnd(c) + qflx_liq_grnd(c)) * dtime Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): none =============================================================== =============================================================== Tag name: ctsm1.0.dev064 Originator(s): slevis (Samuel Levis,SLevis Consulting LLC,303-665-1310) Date: Mon Sep 9 13:14:51 MDT 2019 One-line Summary: User defined top-two snow layers Purpose of changes ------------------ Instead of hardcoding, we now set: dzmin(1) = snow_dzmin_1 dzmax_l(1) = snow_dzmax_l_1 dzmax_u(1) = snow_dzmax_u_1 dzmin(2) = snow_dzmin_2 dzmax_l(2) = snow_dzmax_l_2 dzmax_u(2) = snow_dzmax_u_2 where the right-hand-side variables get namelist-defined values and the model calculates dzmin and dzmax_* for the remaining snow layers using recursive formulas. The calculation is intended to match the previously hardcoded values, except for the bottom snow layer when nlevsno < 12. The bottom layer now always gets dzmax_* = huge(1._r8). The formulas appear in a document titled "A firn model for CLM" that is linked to #729. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): #729 Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): New namelist parameters for specifying the top-two snow layers: snow_dzmin_1, snow_dzmax_l_1, snow_dzmax_u_1 snow_dzmin_2, snow_dzmax_l_2, snow_dzmax_u_2 Changes made to namelist defaults (e.g., changed parameter values): Introduced default values for the new parameters so as to maintain same answers to within roundoff Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: @billsacks CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - tools-tests (test/tools): cheyenne - PTCLM testing (tools/shared/PTCLM/test): cheyenne - python testing (see instructions in python/README.md; document testing done): (any machine) - regular tests (aux_clm): cheyenne ---- OK izumi ------- OK If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: YES (roundoff) Summarize any changes to answers, i.e., - what code configurations: all - what platforms/compilers: all - nature of change: roundoff If bitwise differences were observed, how did you show they were no worse than roundoff? Introduced temporary code to overwrite dzmin(j), dzmax_l(j), and dzmax_u(j) with their original values when different by less than 1e-13 from their original values for all but the bottom snow layer. The bottom layer was left as dzmax_l = dzmax_u(j) = huge(1._8), i.e. the new value. The test resulted in bit-for-bit same answers. Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): https://github.com/ESCOMP/ctsm/pull/792 =============================================================== =============================================================== Tag name: ctsm1.0.dev063 Originator(s): erik (Erik Kluzek) Date: Thu Sep 5 21:28:02 MDT 2019 One-line Summary: Two answer changing fixes (fire, DWT_SLASH) and fix for urban streams for Clm45 Purpose of changes ------------------ Fire bug fix and DWT_SLASH change and fix urban stream years for clm4_5. The latitude used in an expression for Bgc-Fire model Li2016 (used in Clm50) had latitude in degrees rather than in radians as it should have been. The term is used for ignition and the cosine was being taken in degrees rather than radians, hence the spatial pattern for ignition with latitude was incorrect. The history field DWT_SLASH_CFLUX was being calculated on the column, but really should have been a patch level variable. In the same way other fields were handled we made DWT_SLASH_CFLUX a grid-cell variable, and added DWT_SLASH_CFLUX_PATCH on the patch level that could be added to output. The same is true of the C13_ and C14_ versions of it. Clm45 was setting urbantv to year 2000, rather than 1850 or historical as it should have. The urbantv file for Clm45 doesn't actually change, so this doesn't actually make a difference. But, it does make the namelist look wrong. Note, that SP cases, Vic cases and Fates cases do NOT change answers. Or if fire is off, or certain single-point cases. Some cold-start cases don't seem to be sensitive to it either. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): #175, #787, #780 fixes #175 urban stream years for clm4_5 fixes #787 DWT_SLASH field fixes #780 Bug in CN Fire Li 2016 which used latitude in degrees rather than radians Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [X] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): None Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): None Changes made to namelist defaults (e.g., changed parameter values): Defaults for urbantv for clm4_5 Changes to the datasets (e.g., parameter, surface or initial files): None Substantial timing or memory changes: None Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): None Changes to tests or testing: None Code reviewed by: self, @olywon CTSM testing: regular [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - PASS (62 compare fail due to namelist changes) regular tests (aux_clm): cheyenne ---- OK izumi ------- OK If the tag used for baseline comparisons was NOT the previous tag, note that here: previous Answer changes -------------- Changes answers relative to baseline: Yes! (important clm5_0-BGC) Summarize any changes to answers, i.e., - what code configurations: Clm50Bgc and some Clm45Bgc cases - what platforms/compilers: All - nature of change: similar climate (but fire ignition pattern changes by latitude) DWT_SLASH_* history fields change because moved to gridcell quantity (so change is relatively minor) If this tag changes climate describe the run(s) done to evaluate the new climate (put details of the simulations in the experiment database) - casename: oleson/clm50_cesm20R_2deg_GSWP3V1_issue780_hist (for the fire change) URL for LMWG diagnostics output used to validate new climate: http://webext.cgd.ucar.edu/I20TR/clm50_cesm20R_2deg_GSWP3V1_issue780_hist/lnd/clm50_cesm20R_2deg_GSWP3V1_issue780_hist.1995_2014-clm50_cesm20R_2deg_GSWP3V1_hist.1995_2014/setsIndex.html Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): None Pull Requests that document the changes (include PR ids): #802 (https://github.com/ESCOMP/ctsm/pull) #802 -- Three answer changing fixes (fire, DWT_SLASH, urban streams for clm4_5) =============================================================== =============================================================== Tag name: ctsm1.0.dev062 Originator(s): sacks (Bill Sacks) Date: Tue Sep 3 16:04:28 MDT 2019 One-line Summary: Move hobart tests to izumi Purpose of changes ------------------ Move all hobart testing to izumi. Bugs fixed or introduced ------------------------ Known bugs introduced in this tag (include github issue ID): - Recurrence of ESCOMP/ctsm#174 (ne30 case fails on hobart_nag due to floating overflow) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): - Testing should now be run on izumi, not hobart Changes to tests or testing: - All hobart tests moved to izumi Code reviewed by: self CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- not run izumi ------- pass Also ran prealpha and prebeta tests that I moved from hobart to izumi. These passed except: - ERP_D_Ld5.ne30_g16.I1850Clm50BgcCrop.izumi_nag.clm-default (#174) - ERP_Ld5.f19_g17.I2000Clm50SpRtmFl.izumi_pgi.clm-default - SMS_D_Ld1.f19_g17.I1850Clm45Cn.izumi_pgi.clm-default The last two failed with errors like: i011.unified.ucar.edu.44619hfi_userinit: mmap of status page (dabbad0008040000) failed: Operation not permitted I wonder if the issue is that multi-node pgi cases don't work on izumi? If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: NO Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): - cime: cime5.8.3_chint17-04 -> cime5.8.3_chint17-05 Minor change on izumi to allow using run_sys_tests Pull Requests that document the changes (include PR ids): none =============================================================== =============================================================== Tag name: ctsm1.0.dev061 Originator(s): erik (Erik Kluzek) Date: Sun Sep 1 22:37:07 MDT 2019 One-line Summary: Simple b4b fixes: new params file, remove override_nsrest/anoxia_wtsat, DV deprecated Purpose of changes ------------------ New clm5 paramsfile that has the bad date of 631 changed to 701. Along with some simple bit for bit changes. Removing some problematic options. Making Dynamic Vegetation a deprecated option that requires you to use -ignore_warnings with it. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): #463 #79 #97 #98 #104 #122 #173 #794 #795 #796 Fixes #463 bad date on params file that didn't allow using ESMF library with Crop Fixes #79 correct documentation of qflx_evap_tot Fixes #97 remove override_nsrest Fixes #98 add warning when using single instance of a startup file for multi-instance cases Fixes #104 use get_step_size_real when putting into a real variable Fixes #122 decomp pool arrays should not start at 0 but 1 Fixes #173 remove psi_soil_ref as not used Fixes #794 remove anoxia_wtsat Fixes #795 Correct documentation of -light_res Fixes #796 Deprecate Dynamic Vegetation Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): None Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): Remove namelist items Remove namelist items: anoxia_wtsat, override_nsrest Turning dynamic Vegetation on, now displays a warning, and requires using -ignore_warnings in CLM_BLDNML_OPTS Changes made to namelist defaults (e.g., changed parameter values): New paramdata files Changes to the datasets (e.g., parameter, surface or initial files): clm5 paramdata file has correct date for crop of 701 (rather than incorrect 631) clm5 and clm4_5 paramdata file removed an unused variable Substantial timing or memory changes: None Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): None Changes to tests or testing: DV tests removed or changed Code reviewed by: self CTSM testing: regular [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - PASS (compare shows changes becausse of paramdata files) regular tests (aux_clm): cheyenne ---- OK hobart ------ OK An additional fail is ERP_P36x2_D_Ld5.f10_f10_musgs.I2000Clm50Cn.cheyenne_gnu.clm-default, because of #798. Because it's an issue with cime, I'm not marking it as an expected error. If the tag used for baseline comparisons was NOT the previous tag, note that here: previous Answer changes -------------- Changes answers relative to baseline: No bit-for-bit Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): None Pull Requests that document the changes (include PR ids): #797 (https://github.com/ESCOMP/ctsm/pull) #797 -- Simple b4b fixes: new params file, remove override_nsrest/anoxia_wtsat, DV deprecated =============================================================== =============================================================== Tag name: ctsm1.0.dev060 Originator(s): sacks (Bill Sacks) Date: Thu Aug 29 11:18:20 MDT 2019 One-line Summary: In SnowWater, truncate small h2osoi residuals Purpose of changes ------------------ Previously, in the handling of sublimation and evaporation from the snow pack, if these fluxes led to negative h2osoi_ice or h2osoi_liq, these negative values would be passed down the snow pack until they found a layer with a large enough positive state to absorb them. This logic was going to be a challenge for adding water tracers. I have done some tests to verify that these negative values never exceed roundoff-level in either an absolute or relative sense (with tolerance of 1.e-13). Specifically, I ran the full aux_clm test suite with the diffs below and verified that the endrun calls were never triggered. So here, I simplify this logic: After truncating any near-zero states to exactly zero, I check to confirm that the states are never negative. I then removed the code that passed these negative values down the snow pack: if they are only roundoff-level negative, then it seems safe to just throw them away rather than trying to handle them. Here were the diffs used to verify that the negative values are never greater than roundoff-level; these endrun calls were never triggered in a run of the full aux_clm test suite: diff --git a/src/biogeophys/SnowHydrologyMod.F90 b/src/biogeophys/SnowHydrologyMod.F90 index 3dd555cc..07776123 100644 --- a/src/biogeophys/SnowHydrologyMod.F90 +++ b/src/biogeophys/SnowHydrologyMod.F90 @@ -970,6 +970,8 @@ subroutine SnowWater(bounds, & real(r8) :: vol_ice(bounds%begc:bounds%endc,-nlevsno+1:0) ! partial volume of ice lens in layer real(r8) :: eff_porosity(bounds%begc:bounds%endc,-nlevsno+1:0) ! effective porosity = porosity - vol_ice real(r8) :: mss_liqice(bounds%begc:bounds%endc,-nlevsno+1:0) ! mass of liquid+ice in a layer + real(r8) :: h2osoi_ice_orig + real(r8) :: h2osoi_liq_orig !----------------------------------------------------------------------- associate( & @@ -1017,10 +1019,23 @@ subroutine SnowWater(bounds, & c = filter_snowc(fc) l=col%landunit(c) + h2osoi_ice_orig = h2osoi_ice(c,snl(c)+1) + h2osoi_liq_orig = h2osoi_liq(c,snl(c)+1) + wgdif = h2osoi_ice(c,snl(c)+1) & + frac_sno_eff(c) * (qflx_dew_snow(c) - qflx_sub_snow(c)) * dtime h2osoi_ice(c,snl(c)+1) = wgdif if (wgdif < 0._r8) then + if (wgdif < -1.e-13_r8) then + write(iulog,*) 'WJS: big abs wgdif ice: ', c, wgdif, h2osoi_ice_orig + call endrun('WJS: big abs wgdif ice') + end if + if (h2osoi_ice_orig > 0._r8) then + if (abs(wgdif / h2osoi_ice_orig) > 1.e-13_r8) then + write(iulog,*) 'WJS: big rel wgdif ice: ', c, wgdif, h2osoi_ice_orig + call endrun('WJS: big rel wgdif ice') + end if + end if h2osoi_ice(c,snl(c)+1) = 0._r8 h2osoi_liq(c,snl(c)+1) = h2osoi_liq(c,snl(c)+1) + wgdif end if @@ -1030,6 +1045,16 @@ subroutine SnowWater(bounds, & ! if negative, reduce deeper layer's liquid water content sequentially if(h2osoi_liq(c,snl(c)+1) < 0._r8) then + if (h2osoi_liq(c,snl(c)+1) < -1.e-13_r8) then + write(iulog,*) 'WJS: big abs wgdif liq: ', c, h2osoi_liq(c,snl(c)+1), h2osoi_liq_orig + call endrun('WJS: big abs wgdif liq') + end if + if (h2osoi_liq_orig > 0._r8) then + if (abs(h2osoi_liq(c,snl(c)+1) / h2osoi_liq_orig) > 1.e-13_r8) then + write(iulog,*) 'WJS: big rel wgdif liq: ', c, h2osoi_liq(c,snl(c)+1), h2osoi_liq_orig + call endrun('WJS: big rel wgdif liq') + end if + end if do j = snl(c)+1, 1 wgdif=h2osoi_liq(c,j) if (wgdif >= 0._r8) exit Bugs fixed or introduced ------------------------ None Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): - It's possible that the extra error checks I have added (to ensure we don't have greater-than-roundoff-level negative residuals) will be triggered in rare circumstances in a production run, even though they were never triggered in the test suite. Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Although there was a very small increase in total lnd run time, the timing of hydro_without_drainage actually decreased slightly in the PFS test relative to the last tag. Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): - It's possible that the extra error checks I have added (to ensure we don't have greater-than-roundoff-level negative residuals) will be triggered in rare circumstances in a production run, even though they were never triggered in the test suite. Changes to tests or testing: none Code reviewed by: self CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- ok hobart ------ ok ok means tests passed, some answer changes as expected If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: YES (roundoff) Summarize any changes to answers, i.e., - what code configurations: potentially all, though only show up in a few tests - what platforms/compilers: potentially all, though only show up in a few tests - nature of change (roundoff; larger than roundoff/same climate; new climate): roundoff If bitwise differences were observed, how did you show they were no worse than roundoff? First, I did some preliminary analyses as documented above, to ensure that any negative residuals were no greater than roundoff. Second, note that the use of truncate_small_values can only introduce roundoff-level changes (by truncating values that are roundoff-level different from zero to exactly zero). Third, only a few tests showed answer changes, and these were only in limited fields, and were fairly small RMS differences: ERP_Ly3_P72x2.f10_f10_musgs.IHistClm50BgcCrop.cheyenne_intel.clm-cropMonthOutput ERS_Ly3.f10_f10_musgs.I1850Clm50BgcCropCmip6.cheyenne_intel.clm-basic ERS_Ly3_P72x2.f10_f10_musgs.IHistClm50BgcCropG.cheyenne_intel.clm-cropMonthOutput ERS_Ly5_P144x1.f10_f10_musgs.IHistClm50BgcCrop.cheyenne_intel.clm-cropMonthOutput I actually expected more tests to show differences, but most tests were bit-for-bit. I'm guessing this is because the code added a roundoff-level term to a larger term, and this ended up not changing the larger term. Finally, as an extra precaution, I ran the test suite twice, on two different implementations of the truncation: the final one (which uses truncate_small_values), and one in which I did the truncation using custom, inline code (commit 0d21ccb5). These two were bit-for-bit with each other, giving me more confidence that I didn't make a mistake with either implementation. Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): none =============================================================== =============================================================== Tag name: ctsm1.0.dev059 Originator(s): sacks (Bill Sacks) Date: Sat Aug 24 15:14:06 MDT 2019 One-line Summary: Continue adding water tracers to LakeHydrology Purpose of changes ------------------ Finish implementing water tracers for initial snow-related code in LakeHydrology. This involved refactoring code in order to reuse existing code in SnowHydrology rather than having near-duplicates of that code in LakeHydrology. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/ctsm#775 (Implement water tracers for initial snow-related code in LakeHydrology) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Many tests took longer, including the PFS test. However, by looking at more detailed timing numbers in the PFS test, the main increases in time were in initialization and atm run: lnd run only showed a very small increase. This suggests that glade slowness was probably primarily responsible for the increased time. Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: Increased allowed wallclock time for a few tests These tests took longer, likely due to machine variability. Increasing their wallclock time to prevent them from timing out. Code reviewed by: self CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- pass hobart ------ pass If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: NO Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): none =============================================================== =============================================================== Tag name: ctsm1.0.dev058 Originator(s): slevis (Samuel Levis,SLevis Consulting LLC,303-665-1310) Date: Thu Aug 22 10:02:49 MDT 2019 One-line Summary: Soil texture interpolation bug-fix Purpose of changes ------------------ When assigning soil texture values from dataset soil levels to model soil levels, we should be comparing zsoi (NOT zisoi) in the model with zisoi on the file, i.e. we should be asking whether the node center (NOT interface) of a given model level falls between two interface depths on the file. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): #772 Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [X] clm5_0 [X] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: Unexpected failure of test SMS_D_Ld10.f10_f10_musgs.I2000Clm50BgcCropGs.hobart_nag.clm-tracer_consistency Bill Sacks did some follow-up testing and concluded that this may be a compiler-specific issue, so he changed the test from "nag" to "intel" and generated the new test's corresponding baseline Code reviewed by: @billsacks CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - tools-tests (test/tools): cheyenne - PTCLM testing (tools/shared/PTCLM/test): cheyenne - python testing (see instructions in python/README.md; document testing done): (any machine) - regular tests (aux_clm): cheyenne ---- OK hobart ------ OK If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: Yes (clm5) small Summarize any changes to answers, i.e., - what code configurations: not clm45 (clm5_0 and ctsm5_0-nwp) - what platforms/compilers: all - nature of change: larger than roundoff/same climate based on the assumption that shifting soil texture assignments by one level in model soil levels below the top-most level should not generate climate-changing differences. Detailed list of changes ------------------------ 1. Changed zisoi to zsoi as described in the Purpose section above. This caused the larger than roundoff changes. Before the fix some model soil levels were getting assigned soil textures from a deeper soil level in the dataset than they should. 2. In the same section of code I completed the if-block to address all zsoi vs. zisoifl comparisons. This ensures explicit assignment of soil textures for all model soil levels and avoids leaving some model soil levels with the texture assignment of the previous model soil level. Pull Requests that document the changes (include PR ids): (https://github.com/ESCOMP/ctsm/pull/788) =============================================================== =============================================================== Tag name: ctsm1.0.dev057 Originator(s): sacks (Bill Sacks) Date: Tue Aug 20 13:17:39 MDT 2019 One-line Summary: Fix frac_sno bugs Purpose of changes ------------------ Two bug fixes related to frac_sno, and a third change involving replacing a frac_sno calculation with a simpler equation that is algebraically equivalent: (1) Fix lake frac_sno always being 0 (https://github.com/ESCOMP/ctsm/issues/783). This fixes the albedo calculation (and possibly others) over snow-covered lake surfaces. (2) Fix threshold for explicit snow pack initiation to use frac_sno_eff, not frac_sno (https://github.com/ESCOMP/ctsm/issues/785). For standard runs (which have use_subgrid_fluxes = .true.), this just changes answers for urban columns. This also changes answers more widely for runs with use_subgrid_fluxes = .false. (3) Rewrite Swenson & Lawrence 2012 frac_sno equation to be more straightforward and less sensitive to roundoff errors (https://github.com/ESCOMP/ctsm/issues/784) Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/ctsm#783 (frac_sno is always 0 for lake points) - Resolves ESCOMP/ctsm#785 (Threshold for explicit snow pack initiation should use frac_sno_eff, not frac_sno) - Resolves ESCOMP/ctsm#784 (Suggested algebraic rework of frac_sno calculation) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [X] clm5_0 [X] ctsm5_0-nwp [X] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: self CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- ok hobart ------ ok ok means tests pass, answers change as expected Ran most testing on 95dad328 (before 8f1263a7, which removes a temporary endrun error check). Just reran SMS_Ld5_D_P48x1.f10_f10_musgs.IHistClm50Bgc.hobart_nag.clm-decStart on the final version. There was a memleak for this test: FAIL SMS_D_Ld10.f10_f10_musgs.I2000Clm50BgcCropGs.hobart_nag.clm-tracer_consistency MEMLEAK memleak detected, memory went from 1250.830000 to 1518.500000 in 8 days I have had memleak issues for this test before (https://github.com/ESCOMP/ctsm/issues/763), which I have chalked up to a compiler-specific issue. I have simply increased the tolerance for memleaks for this test moving forward. If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: YES (small) Summarize any changes to answers, i.e., - what code configurations: all - what platforms/compilers: all - nature of change (roundoff; larger than roundoff/same climate; new climate): Larger than roundoff; expected to be same climate in general, because the only impacts are on urban and lake points, which represent a small fraction of most grid cells. However, this needs further investigation. See above (under "Purpose of changes") for a description of the different answer changes in this tag. If bitwise differences were observed, how did you show they were no worse than roundoff? I verified that change (3) is no greater than roundoff by temporarily putting in an endrun if the new frac_sno differs from the old by more than 1e-13; this was not triggered for any test in the aux_clm test suite. Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): none =============================================================== =============================================================== Tag name: ctsm1.0.dev056 Originator(s): sacks (Bill Sacks) Date: Fri Aug 16 11:44:43 MDT 2019 One-line Summary: Start adding water tracers to LakeHydrology, and related refactoring Purpose of changes ------------------ Start adding water tracers to LakeHydrology, beginning with some initial things done for snow. This also includes some significant refactoring to allow LakeHydrology to reuse some of the same snow code used for non-lake columns. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Partially addresses ESCOMP/ctsm#775 (Implement water tracers for initial snow-related code in LakeHydrology) Known bugs found since the previous tag (include github issue ID): The following will be fixed in an upcoming tag: - ESCOMP/ctsm#783 (frac_sno is always 0 for lake points) - ESCOMP/ctsm#784 (Suggested algebraic rework of frac_sno calculation) - ESCOMP/ctsm#785 (Threshold for explicit snow pack initiation should use frac_sno_eff, not frac_sno) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: self CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- ok hobart ------ ok ok means tests pass, answers change as expected Note: Did most testing on 054dc95b (before the small change in 7e4e52a9); just ran LWISO_Ld10.f10_f10_musgs.I2000Clm50BgcCropGs.cheyenne_gnu.clm-coldStart on the final commit If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: YES (small) Summarize any changes to answers, i.e., - what code configurations: all - what platforms/compilers: all - nature of change (roundoff; larger than roundoff/same climate; new climate): roundoff There are two answer-changes here: (1) Roundoff-level changes due to changing some order of operations for the updating of snow_depth and dz for lakes: The previous code effectively used (qflx_snow_grnd(c)/bifall(c))*dtime, whereas the new code uses (qflx_snow_grnd(c)*dtime)/bifall(c); similarly, the dz update differs. (2) A change in FSNO_EFF for lakes: previously, this was set to 0 when there was no snow. Now this is set consistently with other landunits, for which it is 1 even if there is no snow. Note that this only affects the FSNO_EFF diagnostic field, and nothing else. If bitwise differences were observed, how did you show they were no worse than roundoff? I confirmed that the order-of-operations change is the only answer-changing part of this tag by comparing against a one-off from master with the following diffs; this comparison was bit-for-bit except for changes in the FSNO_EFF diagnostic field: diff --git a/src/biogeophys/LakeHydrologyMod.F90 b/src/biogeophys/LakeHydrologyMod.F90 index 91ddfc79..9711dad7 100644 --- a/src/biogeophys/LakeHydrologyMod.F90 +++ b/src/biogeophys/LakeHydrologyMod.F90 @@ -126,6 +126,8 @@ subroutine LakeHydrology(bounds, & real(r8) :: heatsum(bounds%begc:bounds%endc) ! used in case above [J/m^2] real(r8) :: snowmass ! liquid+ice snow mass in a layer [kg/m2] real(r8) :: snowcap_scl_fct ! temporary factor used to correct for snow capping + real(r8) :: temp_snow_depth + real(r8) :: newsnow real(r8), parameter :: snow_bd = 250._r8 ! assumed snow bulk density (for lakes w/out resolved snow layers) [kg/m^3] ! Should only be used for frost below. !----------------------------------------------------------------------- @@ -246,8 +248,10 @@ subroutine LakeHydrology(bounds, & ! U.S.Department of Agriculture Forest Service, Project F, ! Progress Rep. 1, Alta Avalanche Study Center:Snow Layer Densification. - dz_snowf = qflx_snow_grnd(c)/bifall(c) - snow_depth(c) = snow_depth(c) + dz_snowf*dtime + temp_snow_depth = snow_depth(c) + newsnow = qflx_snow_grnd(c)*dtime + snow_depth(c) = snow_depth(c) + newsnow/bifall(c) + dz_snowf = (snow_depth(c) - temp_snow_depth)/dtime if (snl(c) == 0) then h2osno_no_layers(c) = h2osno_no_layers(c) + qflx_snow_grnd(c)*dtime ! snow water equivalent (mm) else Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): none =============================================================== =============================================================== Tag name: ctsm1.0.dev055 Originator(s): sacks (Bill Sacks) Date: Tue Aug 6 14:06:50 MDT 2019 One-line Summary: Modularize snow cover fraction method Purpose of changes ------------------ This tag moves the calculation of frac_sno - and the related updates of snow_depth - into a new set of classes, with one class for each parameterization (Niu & Yang 2007 and Swenson & Lawrence 2012). Previously, the code always calculated frac_sno the new way, but then possibly overwrote it if using the older Niu & Yang method. The new code cleans this up, only doing the calculations that are needed for each method. In addition, other code that is specific to one of the two methods is now moved to a home that makes this dependence on method explicit. This includes the addition of newsnow to int_snow: previously, int_snow was always updated using an equation specific to the newer CLM5 parameterization of frac_sno, which was not appropriate if using the Niu & Yang parameterization; this doesn't make a difference currently, since int_snow is only referenced if using the Swenson & Lawrence parameterization, but this clears up some confusion. Also, time-constant parameters read from namelist or the netCDF parameter file now reside in the appropriate class rather than being more global. This tag also renames two namelist options to increase clarity: - subgridflag is renamed to use_subgrid_fluxes, and is now a logical - oldfflag is renamed to snow_cover_fraction_method, and is now a string Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/ctsm#502 (oldfflag can NOT be used with subgridflag==1, and rename subgridflag) - Resolves ESCOMP/ctsm#503 (Clean up CanopyHydrology) - This tag does the last piece of cleanup called for in that issue - Resolves ESCOMP/ctsm#571 (Add a system test: turning on water isotopes shouldn't change answers) - This is unrelated to the other changes in this tag Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): This tag renames two namelist options to increase clarity: - subgridflag is renamed to use_subgrid_fluxes, and is now a logical - oldfflag is renamed to snow_cover_fraction_method, and is now a string Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: - Added a LWISO test, which confirms that turning on water isotopes doesn't change answers for bulk - Added a test with the newer snow cover fraction method but use_subgrid_fluxes false (testmod no_subgrid_fluxes) Code reviewed by: Sean Swenson, Erik Kluzek CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - ok Tests pass; baseline comparisons not done (baseline comparisons expected to differ) tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- pass hobart ------ pass Additional tests: Compared both of these against ctsm1.0.dev052 (which should be bit-for-bit with ctsm1.0.dev054): Both were bit-for-bit: ERP_D_P36x2_Ld3.f10_f10_musgs.I2000Clm45BgcCrop.cheyenne_gnu.clm-no_subgrid_fluxes This is a new test, so did not have baselines; I created baselines from ctsm1.0.dev052 and compared against those SMS_Ly3.f10_f10_musgs.I2000Clm45BgcCrop.cheyenne_intel.clm-oldhyd_monthly This was a temporary test for the sake of running a longer test with the oldhyd testmod. This testmod was created as a one-off by changing the current oldhyd testmod to inherit from monthly rather than default. I ran baselines like this from ctsm1.0.dev052 then ran this test from this branch and compared against those baselines. If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: NO Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): https://github.com/ESCOMP/ctsm/pull/769 =============================================================== =============================================================== Tag name: ctsm1.0.dev054 Originator(s): sacks (Bill Sacks) Date: Fri Aug 2 07:57:51 MDT 2019 One-line Summary: Fix interpolation of surfdat soil layers so we can use interpolation for 10SL case Purpose of changes ------------------ Previously, on master, there was special-purpose code for the 10SL case that avoided doing interpolation from the surface dataset to the soil variables in the model. It would be cleaner - especially now that we allow user-defined soil layer structures - if we could use the general interpolation code always, rather than sometimes having special-purpose code that avoids doing the interpolation. This tag accomplishes that generality. In order to preserve answers for clm45 cases, I needed to make three changes: 1. The constant 0.025 needed an _r8 appended to it; otherwise, zisoifl could differ by roundoff from zisoi. 2. I changed which level is used when zisoi(lev) is exactly equal to zisoifl(j) for some j: I changed the conditional in the following: if (zisoi(lev) >= zisoifl(j) .AND. zisoi(lev) < zisoifl(j+1)) then clay = clay3d(g,j+1) sand = sand3d(g,j+1) om_frac = organic3d(g,j+1)/organic_max endif to: if (zisoi(lev) > zisoifl(j) .AND. zisoi(lev) <= zisoifl(j+1)) then Previously, when the zisoi values in the model exactly lined up with the zisoi values in the file, we would set clay in model level j equal to the value from level j+1 from the surface dataset (and similarly for sand and om_frac); in the new code, we use level j from the surface dataset for model level j in this case. 3. I changed the way zisoifl is calculated for the lowest layer, so that it matches zisoi(nlevsoi) when running with 10SL_3.5m. Previously, we had: zisoi(10) = 0.38018819123227207690E+01 zisoifl(10) = 0.34330930154359391437E+01 This tag changes zisoifl(10) to match zisoi(10). Bugs fixed or introduced ------------------------ Known bugs found since the previous tag (include github issue ID): - #772: Interpolation of clay, sand and om_frac can use the wrong layer (points out more fundamental issues with this interpolation of layers from the surface dataset) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: Sam Levis CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- pass hobart ------ pass If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: NO (but see note below) In principle, the changes in this tag could change answers for some soil layer structures, but there were no answer changes for any of the cases covered by the test suite. Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): https://github.com/ESCOMP/ctsm/pull/771 =============================================================== =============================================================== Tag name: ctsm1.0.dev053 Originator(s): slevis (Samuel Levis,SLevis Consulting LLC,303-665-1310) Date: Thu Aug 1 16:56:09 MDT 2019 One-line Summary: Soil layer definition clean-up and user-defined option Purpose of changes ------------------ Code clean-up clarifes that there are two types of soil layer definition: the node-based and the thickness-based. User-defined option allows user to specify a soil layer profile in the form of a dzsoi vector (values in meters) in the thickness-based approach. Default nlevsoi for NWP configurations had to change from 5 to 4 for consistency with the new error check described in known bugs below. Other code clean-up removes a couple of sections of repeating code. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): #279 #728 Known bugs found since the previous tag (include github issue ID): #759 (this PR) bug causes model to abort when nlevsoi = nlevgrnd; bug has been corrected with an error check Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [X] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): if neither soil_layerstruct_predefined nor soil_layerstruct_userdefined get specified in the namelist, then the model sets soil_layerstruct_predefined to the old default setting for soil_layerstruct (clm5: 20SL_8.5m, clm4.5: 10SL_3.5m) Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): renamed soil_layerstruct to soil_layerstruct_predefined and added soil_layerstruct_userdefined Changes made to namelist defaults (e.g., changed parameter values): Default nlevsoi for NWP configurations had to change from 5 to 4 for consistency with the new error check described in known bugs below Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers -------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: 1) New test... ERP_P36x2_D_Ld5.f10_f10_musgs.I2000Ctsm50NwpBgcCropGswpGs.cheyenne_intel.clm-default replaces existing test ERS_D_Ld10.f10_f10_musgs.I2000Clm50BgcCropGs.cheyenne_intel.clm-rm_indiv_lunits_and_collapse_to_dom to check the correction described in known bugs above; the new test together with existing unit tests cover what the old test was testing 2) New test and new test type... SOILSTRUCTUD_Ld5.f10_f10_musgs.I2000Clm50BgcCropGs.cheyenne_intel.clm-default ensures that soil_layerstruct_userdefined gives bfb same answers as soil_layerstruct_predefined = '4SL_2m' when set with the same dzsoi values. The new test type was put together by: - listing the test in (1) testlist_clm.xml (as all tests) and (2) config_tests.xml - creating the file .../cime_config/SystemTests/soilstructud.py named after the test in lower case Code reviewed by: @billsacks CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - tools-tests (test/tools): cheyenne - PTCLM testing (tools/shared/PTCLM/test): cheyenne - python testing (see instructions in python/README.md; document testing done): (any machine) - regular tests (aux_clm): cheyenne ---- OK hobart ------ OK If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: YES (just for NWP configurations) Summarize any changes to answers, i.e., - what code configurations: nwp - what platforms/compilers: all - nature of change: larger than roundoff/same climate This is due to the change of nlevsoi from 5 to 4 (more info above). I confirmed that nwp does return bfb same answers when I revert this change. Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): https://github.com/ESCOMP/ctsm/pull/759 =============================================================== =============================================================== Tag name: ctsm1.0.dev052 Originator(s): sacks (Bill Sacks) Date: Mon Jul 22 14:02:43 MDT 2019 One-line Summary: Fix rare soil color bug in mksurfdata_map Purpose of changes ------------------ Under rare conditions, mksurfdata_map could put the default soil color in an output cell where there is actually more information. This tag fixes that issue. None of the out-of-the-box surface datasets are impacted by this bug, so I have not recreated any surface datasets. (I checked all out-of-the-box surface datasets except for surfdata_0.125x0.125_mp24_simyr2000_c150114.nc, because it doesn't get remade cleanly out of the box.) Also: - Add some unit tests for the creation of soil color in mksurfdata_map - Point to correct (existing) surface dataset for year-1850 at f05 resolution Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/ctsm#4 (Minor bug in creation of soil color in mksurfdata_map: points can be given the default soil color, when they should have a real color) - Resolves ESCOMP/ctsm#765 (Year-1850 f05 surface dataset missing from inputdata repository) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): - Year-1850 f05 surface dataset now points to a file that actually exists Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): Changes to tests or testing: none Code reviewed by: self CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - ok Most tests pass, including baseline comparisons. The following tests fail, but also failed for me on master: 019 smiS4 TSMscript_tools.sh ncl_scripts getregional_datasets.pl getregional ....................\c rc=6 FAIL 020 bliS4 TBLscript_tools.sh ncl_scripts getregional_datasets.pl getregional ....................\c rc=4 FAIL 027 smf84 TSMscript_tools.sh PTCLM PTCLMmkdata PTCLM_USUMB_clm4_5^buildtools ....................\c rc=6 FAIL 028 blf84 TBLscript_tools.sh PTCLM PTCLMmkdata PTCLM_USUMB_clm4_5^buildtools ....................\c rc=4 FAIL 029 smfc4 TSMscript_tools.sh PTCLM PTCLMmkdata PTCLM_USUMB_Cycle_clm4_5^buildtools ..............\c rc=6 FAIL 030 blfc4 TBLscript_tools.sh PTCLM PTCLMmkdata PTCLM_USUMB_Cycle_clm4_5^buildtools ..............\c rc=4 FAIL 031 smfg4 TSMscript_tools.sh PTCLM PTCLMmkdata PTCLM_USUMB_Global_clm4_5^buildtools .............\c rc=6 FAIL 032 blfg4 TBLscript_tools.sh PTCLM PTCLMmkdata PTCLM_USUMB_Global_clm4_5^buildtools .............\c rc=4 FAIL PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- not run hobart ------ not run If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: NO Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): none =============================================================== =============================================================== Tag name: ctsm1.0.dev051 Originator(s): sacks (Bill Sacks) Date: Fri Jul 19 13:26:25 MDT 2019 One-line Summary: Update water tracers for remainder of first stage of hydrology Purpose of changes ------------------ Main change is to do tracer updates for final pieces of first stage of hydrology: HandleNewSnow and UpdateFracH2oSfc. Along with this, I have introduced greater modularity into these routines (breaking the operations down into more granular steps). Other changes included here (somewhat related to the main changes) are: - Added a routine that resets checked tracers to bulk*ratio to allow running the tracer consistency check for more than one time step; we now run it for 10 days - Change logic for rain-snow conversion terms: Use rain diff in rain to snow, snow diff in snow to rain. This is less sensitive to roundoff errors, and was needed in order to get the 10-day tracer consistency test to pass. - Setting of qflx_snow_h2osfc moved to CanopyInterceptionAndThroughfall - Add col%lun_itype. col%lun_itype(ci) is the same as lun%itype(col%landunit(ci)), but is often a more convenient way to access this type. I got tired of having to get the landunit index just for the sake of getting the landunit's type of a given column. I have NOT gone through the code and replaced lun%itype(l) with col%lun_itype(c) where applicable, but this will be useful for the future. - Removed unnecessary code for backwards compatibility of FH2OSFC from restart files - Point to slightly modified initial conditions files for two CLm45 initial conditions files. These two had FH2OSFC > 0 for some glc_mec points. I'm not sure how that came to be, but Sean Swenson and I think that shouldn't be the case. This caused soil balance errors in these tests: ERS_Ly5_P72x1.f10_f10_musgs.IHistClm45BgcCrop.cheyenne_intel.clm-cropMonthOutput SMS_D_Ld1.f09_g17.I1850Clm45BgcCruGs.cheyenne_intel.clm-default I'm fixing the problem by pointing to updated files with FH2OSFC set to 0 for special landunits, using the following python: ctypel = dat.variables['cols1d_ityplun'][:] dat.variables['FH2OSFC'][ctypel > 2] = 0 - Small cime update to fix the SHAREDLIBBUILD phase of ERS_D_Ln9_P480x3.f19_g16.I2000Clm50SpGs.cheyenne_intel.clm-waccmx_offline Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/ctsm#718 (Implement water tracers for HandleNewSnow and FracH2oSfc) - Resolves ESCOMP/ctsm#498 (Do tracer consistency checks every time step, not just first) CIME Issues fixed (include issue #): - ESCMI/cime#3717 (Cheyenne ESMF missing for WACCM X) Known bugs introduced in this tag (include github issue ID): - ESCOMP/ctsm#762 (Water tracers: Make sure calls to ResetCheckedTracers are removed) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): - If you have an initial conditions file with non-zero FH2OSFC for glacier (this was the case for some out-of-the-box clm45 files, which are now fixed), may run into soil balance errors. To fix this, set FH2OSFC to 0 in your restart file for all special landunits. Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): - Point to slightly modified initial conditions files for two CLm45 initial conditions files: see above for details. Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: - Changed tracer consistency test to 10 days rather than a single time step Code reviewed by: self CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- ok hobart ------ ok ok means tests pass, some roundoff-level diffs in baseline comparisons, as noted below Also, in order to have a more powerful test of the changes in SurfaceWaterMod (since the relevant code isn't triggered very often), I introduced the following diffs: diff --git a/src/biogeophys/SurfaceWaterMod.F90 b/src/biogeophys/SurfaceWaterMod.F90 index 1bc03cae..be656a15 100644 --- a/src/biogeophys/SurfaceWaterMod.F90 +++ b/src/biogeophys/SurfaceWaterMod.F90 @@ -12,6 +12,7 @@ module SurfaceWaterMod use shr_spfn_mod , only : erf => shr_spfn_erf use clm_varcon , only : denh2o, denice, roverg, wimp, tfrz, pc, mu, rpi use clm_varpar , only : nlevsno, nlevgrnd + use clm_varctl , only : iulog use clm_time_manager , only : get_step_size use column_varcon , only : icol_roof, icol_road_imperv, icol_sunwall, icol_shadewall, icol_road_perv use decompMod , only : bounds_type @@ -201,7 +202,7 @@ subroutine BulkDiag_FracH2oSfc(bounds, num_soilc, filter_soilc, & SHR_ASSERT_FL((ubound(qflx_too_small_h2osfc_to_soil, 1) == bounds%endc), sourcefile, __LINE__) ! arbitrary lower limit on h2osfc for safer numerics... - min_h2osfc=1.e-8_r8 + min_h2osfc=1.e0_r8 do f = 1, num_soilc c = filter_soilc(f) @@ -230,6 +231,9 @@ subroutine BulkDiag_FracH2oSfc(bounds, num_soilc, filter_soilc, & else frac_h2osfc(c) = 0._r8 + if (h2osfc(c) > 0._r8) then + write(iulog,*) 'WJS: in else' + end if qflx_too_small_h2osfc_to_soil(c) = h2osfc(c) / dtime ! The update of h2osfc is deferred to later, keeping with our standard ! separation of flux calculations from state updates, and because the state With those diffs, I ran the tracer consistency test; it passed. In addition, I ran SMS_D_Ld1_P4x1.f10_f10_musgs.I2000Clm50BgcCropQianRsGs.bishorn_gnu.clm-default with comparison against master, with the following diffs on master (the first change changes behavior to force the relevant code to be triggered, and is the same on master and on the branch; the second change on master divides then multiplies by dtime to avoid roundoff-level diffs): diff --git a/src/biogeophys/SurfaceWaterMod.F90 b/src/biogeophys/SurfaceWaterMod.F90 index 959d539b..4a4a4824 100644 --- a/src/biogeophys/SurfaceWaterMod.F90 +++ b/src/biogeophys/SurfaceWaterMod.F90 @@ -77,7 +77,7 @@ subroutine FracH2oSfc(bounds, num_nolakec, filter_nolakec, & ) ! arbitrary lower limit on h2osfc for safer numerics... - min_h2osfc=1.e-8_r8 + min_h2osfc=1.e0_r8 call waterstatebulk_inst%CalculateTotalH2osno(bounds, num_nolakec, filter_nolakec, & caller = 'FracH2oSfc', & @@ -112,7 +112,7 @@ subroutine FracH2oSfc(bounds, num_nolakec, filter_nolakec, & else frac_h2osfc(c) = 0._r8 - h2osoi_liq(c,1) = h2osoi_liq(c,1) + h2osfc(c) + h2osoi_liq(c,1) = h2osoi_liq(c,1) + (h2osfc(c) / get_step_size()) * get_step_size() h2osfc(c)=0._r8 endif This test passed and was bit-for-bit with master with the above changes. If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: Yes (roundoff level) Summarize any changes to answers, i.e., - what code configurations: Many/all - what platforms/compilers: All - nature of change (roundoff; larger than roundoff/same climate; new climate): Roundoff: just roundoff-level changes in FSH_PRECIP_CONVERSION, FSH_TO_COUPLER, l2x_Fall_sen due to refactored calculation of FSH_PRECIP_CONVERSION If bitwise differences were observed, how did you show they were no worse than roundoff? via summarize_cprnc_diffs Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): - cime: branch_tags/cime5.8.3_chint17-03 -> branch_tags/cime5.8.3_chint17-04 Fixes ESMCI/cime#3171 Pull Requests that document the changes (include PR ids): none =============================================================== =============================================================== Tag name: ctsm1.0.dev050 Originator(s): slevis (Samuel Levis,SLevisConsulting LLC,303-665-1310) Date: Mon Jul 15 11:40:09 MDT 2019 One-line Summary: dz --> dz_lake bug-fix in LakeTemperatureMod.F90 line 960 Purpose of changes ------------------ Bug-fix to prevent the model from aborting when running with fewer soil layers than lake layers; not to imply that this was not a bug when the model wasn't aborting. It was. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): #760 Known bugs found since the previous tag (include github issue ID): #759 bug causing model to abort when nlevsoi = nlevgrnd Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers -------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: @dlawrenncar @billsacks CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - tools-tests (test/tools): cheyenne - PTCLM testing (tools/shared/PTCLM/test): cheyenne - python testing (see instructions in python/README.md; document testing done): (any machine) - regular tests (aux_clm): cheyenne ---- OK hobart ------ OK If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: Yes (when methane on) (diagnostic only) Summarize any changes to answers, i.e., - what code configurations: use_lch4 = .true. - what platforms/compilers: all - nature of change: diagnostic variable WTGQ only Confirmed this on cheynne and hobart by running ./summarize_cprnc_diffs -baseline .../tests_0712... -testid '*' and inspecting the file cprnc.summary.*.by_varname Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): https://github.com/ESCOMP/ctsm/pull/761 =============================================================== =============================================================== Tag name: ctsm1.0.dev049 Originator(s): erik (Erik Kluzek) Date: Sun Jun 23 20:56:55 MDT 2019 One-line Summary: Update mosart and intel to intel-19 on cheyenne Purpose of changes ------------------ Update mosart to trunk tag (that has a roundoff change to history file output). And update intel compiler on cheyenne to intel-19. Bugs fixed or introduced None ------------------------ Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): None Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): None Changes made to namelist defaults (e.g., changed parameter values): None Changes to the datasets (e.g., parameter, surface or initial files): None Substantial timing or memory changes: [For timing changes, can check PFS test(s) in the test suite] None Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): None Changes to tests or testing: None Code reviewed by: self CTSM testing: regular [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - PASS regular tests (aux_clm): cheyenne ---- OK hobart ------ OK If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: Yes (roundoff level change when MOSART on) Summarize any changes to answers, i.e., - what code configurations: some compsets with mosart - what platforms/compilers: cheyenne/intel - nature of change: roundoff Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): cime, and mosart mosart to mosart1_0_33 cime to branch_tags/cime5.8.3_chint17-03 Pull Requests that document the changes (include PR ids): #753 (https://github.com/ESCOMP/ctsm/pull) #753 -- Update cheyenne to intel/19 and update mosart =============================================================== =============================================================== Tag name: ctsm1.0.dev048 Originator(s): erik (Erik Kluzek,UCAR/TSS,303-497-1326) Date: Sun Jun 23 15:16:01 MDT 2019 One-line Summary: Updates for buildlib changes and cime and externals updates Purpose of changes ------------------ Update of most externals: cime, mosart, rtm, cism. The cime update also includes a required CESM-wide change in buildlib. Uses a function to get the Macros filename "get_standard_makefile_args". The cime update brought in some answer chagnes, with an update of the intel compiler (but we backed those out by using a cime branch). There is also a roundoff change by going to the trunk version of RTM. Bugs fixed or introduced None ------------------------ Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): None Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): None Changes made to namelist defaults (e.g., changed parameter values): None Changes to the datasets (e.g., parameter, surface or initial files): None Substantial timing or memory changes: None Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): None Changes to tests or testing: None Code reviewed by: @jedwards, self CTSM testing: regular [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - PASS tools-tests (test/tools): cheyenne - PASS PTCLM testing (tools/shared/PTCLM/test): cheyenne - PASS python testing (see instructions in python/README.md; document testing done): cheyenne - PASS (for python3) regular tests (aux_clm): cheyenne ---- OK hobart ------ OK If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: Yes, but only for a few cases (compsets with RTM) Summarize any changes to answers, i.e., - what code configurations: With RTM - what platforms/compilers: All - nature of change: roundoff fieldlist is different for CISM in some cases Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): cime, mosart, rtm, cism cime to branch_tags/cime5.8.3_chint17-02 cism to cism2_0_68 rtm to rtm1_0_68 mosart to nldas-grid.n02_mosart1_0_31 Pull Requests that document the changes (include PR ids): #752 (https://github.com/ESCOMP/ctsm/pull) #752 -- buildlib, cime and externals update =============================================================== =============================================================== Tag name: ctsm1.0.dev047 Originator(s): sacks (Bill Sacks) Date: Sun Jun 16 13:04:38 MDT 2019 One-line Summary: Fix negative snow compaction during snow melt Purpose of changes ------------------ Fix an issue reported by @kjetilaas: In the presence of surface water, the old and new snow cover fractions are inconsistently calculated in subroutine "SnowCompaction". This can result in significant negative compaction (snow depth increase) during snow melt. More details from @kjetilaas (copied from https://github.com/escomp/ctsm/issues/573): There is an inconsistency in the calculation of snow cover fraction (FSNO) in the SnowCompaction subroutine when there is surface water present. The actual snow cover is limited to 0.99 when surface water is present (and the surface water fraction is set to 0.01). However, SnowCompaction recalculates the old FSNO without accounting for this limitation. This results in an apparent decrease in FSNO, and corresponding negative snow compaction from snow melt (ddz3). This can result in substantial, artificial increase in snow depth during summer (can be more than 0.5 m during a single season). The problem might be masked in monthly output, but could be significant whenever snow cover is close to 1 and surface water present. Snow mass is indirectly affected by this as radiation and temperature fluxes in the snow pack will be affected by artificially thick snow layers. This tag fixes this issue. The fix was implemented and tested by @swensosc (Sean Swenson). Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/ctsm#573 (Snow depth increase due to inconsistent snow cover fraction calculation) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [X] clm5_0 [X] ctsm5_0-nwp [X] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: Sean Swenson, @kjetilaas, Bill Sacks CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- ok hobart ------ ok ok means tests pass, answers change as expected If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: YES (similar climate) Summarize any changes to answers, i.e., - what code configurations: all - what platforms/compilers: all - nature of change (roundoff; larger than roundoff/same climate; new climate): Larger than roundoff/same climate Sean Swenson ran the LMWG diagnostics package and looked at the results with Keith Oleson. Their main findings were: - Snow depth is a bit smaller over Greenland and polar regions on average, as expected (up to 5-10% decrease in average snow depth averaged over the Canadian Arctic) - There is a small increase in h2osno over Greenland: about 30 mm averaged over Greenland; it looks like the main changes are around the coasts (which makes sense given that this bug only affects vegetated landunits) - There are small changes in runoff over Greenland - Changes in snow cover fraction are tiny - Overall, this doesn't look climate changing, but could potentially have a significant impact on select grid cells, and on some variables in polar regions Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): none =============================================================== =============================================================== Tag name: ctsm1.0.dev046 Originator(s): sacks (Bill Sacks) Date: Sat Jun 15 15:54:44 MDT 2019 One-line Summary: Separate the two uses of h2osno Purpose of changes ------------------ Until now, h2osno has been used in two ways: 1. When there are explicit snow layers, h2osno is the sum of snow (liquid + ice) in all layers 2. When there is a little bit of snow, but not enough to create a snow layer, h2osno gives the amount of this very thin snow pack (assumed to be all ice) This is confusing and complicates the addition of water tracers to the code. This tag separates h2osno into two different variables: 1. h2osno_no_layers is a fundamental state variable that gives the amount of snow in the case where we don't have explicit snow layers 2. h2osno_total is purely a diagnostic variable that is the sum of snow in all explicit layers, plus h2osno_unresolved. This exists in waterdiagnosticbulk_type, but the version there is just set once at the end of the time step, so that version is only meant for diagnostic (history) output. Code that needs the current sum of h2osno at a given point now computes it locally, via a new method on waterstate_type: CalculateTotalH2osno. If running in debug mode, that routine also performs some consistency checks. Also, an unrelated change to prevent user from trying to interpolate a file onto itself. Also, add a compset for testing, which is good for use on my mac. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/ctsm#733 (Make h2osno purely a diagnostic variable) - Resolves ESCOMP/ctsm#749 (h2osno inconsistent when interpolating from initial conditions with different snow layers) - Resolves ESCOMP/ctsm#750 (Prevent user from trying to use the same filename for init_interp source and destination files) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none (possibly a small improvement in timing based on PFS run time) Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: self CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- ok hobart ------ ok ok means tests pass; baseline comparisons fail as expected If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: Yes (mostly roundoff) Summarize any changes to answers, i.e., - what code configurations: all - what platforms/compilers: all - nature of change (roundoff; larger than roundoff/same climate; new climate): Mostly roundoff; greater than roundoff/same climate for cases for which number of snow layers is less in the case than in its initial conditions file, due to fixing #749 If bitwise differences were observed, how did you show they were no worse than roundoff? There was one point where h2osno_total differed from the old h2osno by roundoff, as verified by a run of the test suite (where I verified that (a) it only differed by 1e-11 or less, and (b) if I reset it, there were just roundoff-level changes in a few diagnostic fields). That one place with roundoff-level diffs can lead to propagation of diffs, causing differences in many fields in the final run of the test suite. In particular: I ran the test suite on 69823de4 (which had this temporary check and reset of h2osno_total vs. old h2osno). For nearly all tests, there were only roundoff-level changes in a few diagnostic fields. I then removed the temporary check and reset and ran the test suite on the result (4353cc90). Finally, I ran the test suite on the final, cleaned-up code, with comparison against 4353cc90 and verified bfb. In my testing on 69823de4: There were a few tests with greater than roundoff-level changes due just to differences in the initial value of h2osno_total (which was incorrect in the old due to #749). I verified that these only had roundoff-level changes in a few diagnostic fields if I used the following diffs in both the baseline and my branch, indicating that the diffs just arose due to the fix of #749: diff --git a/src/main/clm_initializeMod.F90 b/src/main/clm_initializeMod.F90 index 793b36ac..f10ee5e1 100644 --- a/src/main/clm_initializeMod.F90 +++ b/src/main/clm_initializeMod.F90 @@ -762,6 +762,18 @@ subroutine initialize2( ) !$OMP END PARALLEL DO end if + do c = bounds_proc%begc, bounds_proc%endc + if (col%snl(c) < 0) then + water_inst%waterstatebulk_inst%h2osno_col(c) = 0._r8 + do j = col%snl(c)+1, 0 + water_inst%waterstatebulk_inst%h2osno_col(c) = & + water_inst%waterstatebulk_inst%h2osno_col(c) + & + water_inst%waterstatebulk_inst%h2osoi_ice_col(c,j) + & + water_inst%waterstatebulk_inst%h2osoi_liq_col(c,j) + end do + end if + end do + end subroutine initialize2 end module clm_initializeMod Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): none =============================================================== =============================================================== Tag name: ctsm1.0.dev045 Originator(s): sacks (Bill Sacks) Date: Thu Jun 6 12:26:57 MDT 2019 One-line Summary: Recalculate h2osno for the sake of SnowCapping Purpose of changes ------------------ SnowCapping decides whether to remove excess snow based on h2osno. However, at the point where this is called,h2osno is out of sync with (h2osoi_liq + h2osoi_ice): Those variables can be changed slightly in SnowWater (which is called before SnowCapping), yet h2osno isn't recalculated based on those changes. I don't think this causes a big problem in practice, but it seems like it would improve code understandability if we had the general rule that any code that references h2osno is using a version of h2osno that is in sync with (h2osoi_liq + h2osoi_ice). This is also needed to prevent answer changes in an upcoming major refactor. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/ctsm#736 Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: Sean Swenson signed off on the change, but didn't review the final code CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - not run regular tests (aux_clm): cheyenne ---- ok hobart ------ ok ok means tests pass, answers change as expected If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: YES (should be small) Summarize any changes to answers, i.e., - what code configurations: all - what platforms/compilers: all - nature of change (roundoff; larger than roundoff/same climate; new climate): Larger than roundoff/same climate Answer changes are expected to be small, since this should just change the timing of snow capping a bit, but I haven't looked closely at the magnitude of the answer changes. Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): none =============================================================== =============================================================== Tag name: ctsm1.0.dev044 Originator(s): sacks (Bill Sacks) Date: Thu Jun 6 12:14:39 MDT 2019 One-line Summary: Make wetland snow resetting behavior more explicit Purpose of changes ------------------ There is some code that zeroes out h2osno over thawed wetlands. It seems like the intended behavior here is to zero out the entire snow pack. Currently, however, this only zeros out a very thin (zero-layer) snow pack. For now, I'm adding an snl==0 conditional to make this behavior explicit; long-term, we'd like to change this to actually zero out the whole snow pack. (At that time, this code block should probably be moved to a more appropriate home, as noted in comments in issue ESCOMP/ctsm#735.) Also, unrelated change in run_sys_tests: support machine-specific baseline directory, in order to support new default location for CTSM baselines on cheyenne. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Partially addresses, but does NOT fix, ESCOMP/ctsm#735 Known bugs introduced in this tag (include github issue ID): - Partially addresses, but does NOT fix, ESCOMP/ctsm#735 Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: Sean Swenson and Keith Oleson approved this change, though didn't actually review the final code. CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run python testing (see instructions in python/README.md; document testing done): (any machine) - pass on cgdm-bishorn make python=python2 test make pyhton=python3 test make lint regular tests (aux_clm): cheyenne ---- pass hobart ------ pass Also ran SMS_Ly1.f09_g17.I2000Clm50SpGs.cheyenne_intel.clm-monthly with comparison against baselines: passed and bit-for-bit. If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: NO (but see notes below) It seems theoretically possible for this to change answers, but no answer changes were observed in the aux_clm test suite or SMS_Ly1.f09_g17.I2000Clm50SpGs.cheyenne_intel.clm-monthly. Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): none =============================================================== =============================================================== Tag name: ctsm1.0.dev043 Originator(s): erik (Erik Kluzek) Date: Wed Jun 5 10:17:06 MDT 2019 One-line Summary: Fix FUN bug (frac_ideal_C_use was backwards in regard to delta_CN), and replace Ball-Berry mbbopt with the CLM4.5 version Purpose of changes ------------------ Carbon allocation to uptake responds to CN(uptake-cost) and CN(actual). The intended implementation is: For C:N less than the target C:N, C allocation is reduced with cost. For C:N greater than the target C:N, C allocation is increased with high C:N. However, the actual implementation is reversed. This change fixes that problem. Also the Ball-Berry mbbopt values had been changed in the creation of CLM5.0, as part of the tuning, but the new values don't work well (as our final tuning uses Medlyn photosynshesis rather than Ball-Berry). This brings the values for Ball-Berry for CLM5.0 back to the CLM4.5 values. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): #704 #705 Fixes #705 -- Ball-Berry parameters in CLM5.0 should go back to the previous values in CLM4.5 Fixes #704 -- FUN code logic to reduce or increase carbon allocation used for uptake is reversed Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [X] clm5_0 [X] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): None Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): None Changes made to namelist defaults (e.g., changed parameter values): New clm5.0 paramdata file Changes to the datasets (e.g., parameter, surface or initial files): mbbopt now use the clm4_5 values rather than the pretuned clm5_0 values (before Medlyn was being used) Substantial timing or memory changes: None Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): None Changes to tests or testing: None Code reviewed by: self, olyson CTSM testing: regular [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - PASS (134 tests are different from baseline as expected) regular tests (aux_clm): cheyenne ---- OK hobart ------ OK If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: Yes! (Important when FUN model on) Summarize any changes to answers, i.e., - what code configurations: Anything with use_FUN=T (so Clm50Bgc compsets) Also with Clm5.0 and Ball-Berry (stomatalcond_method='Ball-Berry1987') (so Clm50Nwp compsets) - what platforms/compilers: All - nature of change: climate is similar "Our assessment is that the impact is relatively small, but would affect climate; however it doesn’t appear to strongly affect transient C response." If this tag changes climate describe the run(s) done to evaluate the new climate (put details of the simulations in the experiment database) Experiments were done on the release branch - casename: oleson/clm50_release-clm5.0.15_delta_CN_FUNbug_2deg_GSWP3V1_1850 oleson/clm50_release-clm5.0.15_delta_CN_FUNbug_2deg_GSWP3V1_hist URL for LMWG diagnostics output used to validate new climate: http://webext.cgd.ucar.edu/I1850/clm50_release-clm5.0.15_delta_CN_FUNbug_2deg_GSWP3V1_1850/lnd/clm50_release-clm5.0.15_delta_CN_FUNbug_2deg_GSWP3V1_1850.21_40-clm50_release-clm5.0.15_2deg_GSWP3V1_1850.21_40/setsIndex.html We've also completed a historical run. Diagnostics are here: http://webext.cgd.ucar.edu/I20TR/clm50_release-clm5.0.15_delta_CN_FUNbug_2deg_GSWP3V1_hist/lnd/clm50_release-clm5.0.15_delta_CN_FUNbug_2deg_GSWP3V1_hist.1995_2014-clm50_release-clm5.0.15_2deg_GSWP3V1_hist.1995_2014/setsIndex.html Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): None Pull Requests that document the changes (include PR ids): #727 (https://github.com/ESCOMP/ctsm/pull) #727 -- Fix the FUN bug, where an if statement was backwards =============================================================== =============================================================== Tag name: ctsm1.0.dev042 Originator(s): slevis (Samuel Levis,SLevis Consulting LLC,303-665-1310) Date: Tue May 21 23:29:31 MDT 2019 One-line Summary: Rename, correct, and simplify parameters dewmx and sno_stor_max Purpose of changes ------------------ To rename dewmx to liq_canopy_storage_scalar sno_stor_max to sno_canopy_storage_scalar in the code and in the clm45 and clm50 param files. To change snocanmx, liqcanmx, and fcansno to... snocanmx = params_inst%snow_canopy_storage_scalar(p) * (elai(p) + esai(p)) liqcanmx = params_inst%liq_canopy_storage_scalar(p) * (elai(p) + esai(p)) fcansno(p) = (snocan(p) / (vegt * params_inst%sno_stor_max))**0.15_r8 Rewrote fwet(p) to follow the same format as fcansno(p). Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): #710 Fixes #710 -- rename dewmx Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): None Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): None Changes made to namelist defaults (e.g., changed parameter values): New default parameter file Changes to the datasets (e.g., parameter, surface or initial files): Pointing to new clm45 and clm50 param files containing the two renamed parameters listed above. Substantial timing or memory changes: None Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): None Changes to tests or testing: None Code reviewed by: @billsacks CTSM testing: regular [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - PASS (196 different because of new param file) regular tests (aux_clm): cheyenne ---- OK hobart ------ OK Roundoff changes were expected. If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: Yes! (roundoff) Summarize any changes to answers, i.e., - what code configurations: ALL - what platforms/compilers: ALL - nature of change: roundoff If bitwise differences were observed, how did you show they were no worse than roundoff? The code modifications causing the change were of the order-of-calculations type in two lines of code. Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): None Pull Requests that document the changes (include PR ids): https://github.com/ESCOMP/ctsm/pull/719 =============================================================== =============================================================== Tag name: ctsm1.0.dev041 Originator(s): sacks (Bill Sacks) Date: Fri May 17 06:02:38 MDT 2019 One-line Summary: Add water tracers to CanopyHydrologyMod Purpose of changes ------------------ Major overhaul of CanopyHydrologyMod in order to support water tracers. The part of CanopyHydrology that was actually dealing with canopy hydrology has now been broken out into a number of small routines in order to separate flux calculations from state updates. This is needed in order to calculate tracer versions of the relevant variables, which is now done here. The parts of CanopyHydrology that were NOT dealing with canopy hydrology have been moved to more appropriate homes. They have not yet been tracerized, but will be soon. This also makes some minor changes to diagnostic fields related to canopy hydrology. Of particular note: QDRIP now has a different meaning: it now corresponds to the excess liquid water that runs off from the canopy (and is inactive by default). The design of this broken-out code was developed in consultation with Mike Barlage, Erik Kluzek, Negin Sobhani, Sean Swenson and Mariana Vertenstein. We plan to apply this design moving forward to other parts of the code that need tracerization. See also . Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/ctsm#360 - Implement water isotopes for CanopyHydrology - Resolves ESCOMP/ctsm#709 - QDRIP diagnostic field misleadingly named - Partially addresses ESCOMP/ctsm#503 - Clean up CanopyHydrology (We still need to clean up / tracerize the parts that have now been moved elsewhere.) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): - Changed meaning of QDRIP diagnostic field: it now corresponds to qflx_liqcanfall rather than qflx_prec_grnd_patch, and is default inactive. (There is no longer a diagnostic field corresponding to qflx_prec_grnd_patch, but there are separate diagnostic fields for the rain and snow portions of this, both default inactive.) Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none - checked timing of hydro1 timer vs. previous canhydro timer from my dev039 testing: differences are negligible Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: - Design reviewed by Mike Barlage, Erik Kluzek, Negin Sobhani, Sean Swenson and Mariana Vertenstein - Final code reviewed only by self CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run regular tests (aux_clm): cheyenne ---- ok hobart ------ ok ok means tests pass, baseline comparisons fail as expected If the tag used for baseline comparisons was NOT the previous tag, note that here: Answer changes -------------- Changes answers relative to baseline: YES (roundoff) Summarize any changes to answers, i.e., - what code configurations: all - what platforms/compilers: all - nature of change (roundoff; larger than roundoff/same climate; new climate): roundoff If bitwise differences were observed, how did you show they were no worse than roundoff? Ensured no-greater-than-roundoff-level by: First, generated baselines using code with these diffs (note extra parentheses in multiplication): diff --git a/src/biogeophys/CanopyHydrologyMod.F90 b/src/biogeophys/CanopyHydrologyMod.F90 index f848f520..31c8c550 100644 --- a/src/biogeophys/CanopyHydrologyMod.F90 +++ b/src/biogeophys/CanopyHydrologyMod.F90 @@ -338,8 +338,8 @@ subroutine CanopyHydrology(bounds, & ! Intercepted precipitation [mm/s] qflx_prec_intr(p) = forc_snow(c)*fpisnow + qflx_liq_above_canopy(p)*fpi ! storage of intercepted snowfall, rain, and dew - snocan(p) = max(0._r8, snocan(p) + dtime*forc_snow(c)*fpisnow) - liqcan(p) = max(0._r8, liqcan(p) + dtime*qflx_liq_above_canopy(p)*fpi) + snocan(p) = max(0._r8, snocan(p) + dtime*(forc_snow(c)*fpisnow)) + liqcan(p) = max(0._r8, liqcan(p) + dtime*(qflx_liq_above_canopy(p)*fpi)) ! Initialize rate of canopy runoff and snow falling off canopy qflx_snocanfall(p) = 0._r8 Then ran test suite on 1cc9d38a. This is very close to the final branch, but had a few small tweaks in place to get bit-for-bit: 1. Under some conditions, the final branch code resets slightly negative snocan and liqcan to 0, where master did not. This commit puts in place temporary code to maintain the slightly negative values under these conditions, along with a check to ensure that the values are really always only slightly negative in this situation. 2. When snocan > snocanmx, master reset snocan to exactly snocanmx, whereas the final branch code can result in snocan being roundoff-level different from snocanmx, and similar for liqcan. This commit puts in place temporary code to set snocan to snocanmx under appropriate circumstances, and liqcan to liqcanmx, along with a check to ensure that this is only leading to roundoff-level changes in these state variables. 3. Multiplies then divides a flux by dtime in one place 4. Has parentheses around a partial sum to ensure same order of operations as before This was bfb with the baselines generated with the diff noted above, except for roundoff-level changes in SNOCAN in one grid cell in just one test (SMS_D_Ld1.f19_g17.I1850Clm45Cn.hobart_gnu.clm-default; old: -8.15e-322, new: -2e-323). Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): None. But see https://github.com/billsacks/prototypes-ctsm-canopyhydrology_tracers, which was the source of our design discussions and contains notes from these discussions. =============================================================== =============================================================== Tag name: ctsm1.0.dev040 Originator(s): slevis (Samuel Levis,SLevis Consulting LLC,303-665-1310) Date: Fri May 03 10:34:46 MDT 2019 One-line Summary: Move some hard-coded parameters from code to params.nc file Purpose of changes ------------------ Reduce the number of parameters that get hardwired directly in the code by placing in the params.nc file where they become visible and easier to modify. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): #680 Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users: None Changes to CTSM's user interface: Parameters that were hardwired in the code can now be seen and changed in the params.nc file. Changes made to namelist defaults (e.g., changed parameter values): None Changes to the datasets (e.g., parameter, surface or initial files): Various parameters have moved out of the code and into the params.nc file. Substantial timing or memory changes: [For timing changes, can check PFS test(s) in the test suite] Notes of particular relevance for developers: None --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): None Changes to tests or testing: None Code reviewed by: @ekluzek @billsacks CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - tools-tests (test/tools): cheyenne - PTCLM testing (tools/shared/PTCLM/test): cheyenne - regular tests (aux_clm): cheyenne ---- OK hobart ------ OK Answers change as expected. CTSM tag used for the baseline comparisons: ctsm1.0.dev039 Answer changes -------------- Changes answers relative to baseline: YES (single precision roundoff) Summarize any changes to answers, i.e., - what code configurations: ALL - what platforms/compilers: ALL - nature of change: single-precision roundoff Changes are due to four parameters: accum_factor ko25_coef kc25_coef frac_sat_soil_dsl_init that were not declared "_r8" when hardwired in the code, while now they are declared "double" in the params.nc files. I showed that the change was no greater than single-precision roundoff by writing the new and old values to the lnd and cesm log files and confirming differences smaller than 1.e-6. Detailed list of changes ------------------------ - Updated the clm5 and clm4.5 params.nc files to contain the previously hard-coded parameters that Katie Dagon recommended moving out of the code (listed here: https://docs.google.com/spreadsheets/d/1cRmcB0xz7CjVP6ljUoCehJx49Usxsore5ma4mWcvQwM/edit#gid=0) - Now read the parameters in the modules that use them Pull Requests that document the changes (include PR ids): https://github.com/ESCOMP/ctsm/pull/684 =============================================================== =============================================================== Tag name: ctsm1.0.dev039 Originator(s): sacks (Bill Sacks) Date: Wed May 1 11:12:34 MDT 2019 One-line Summary: Remove excess canopy liquid/snow regardless of temperature Purpose of changes ------------------ Previously: if it was above freezing, snow was allowed to accumulate on the canopy beyond the specified capacity, and if it was below freezing, liquid was allowed to accumulate beyond the specified capacity. Sean Swenson, Keith Oleson and Dave Lawrence agreed that this seems wrong. This tag changes the behavior so that both liquid and snow canopy capacities are enforced regardless of temperature. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/ctsm#699 (Canopy liquid and snow capacities not enforced consistently?) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [X] clm5_0 [X] ctsm5_0-nwp [X] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: Keith Oleson CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run regular tests (aux_clm): cheyenne ---- ok hobart ------ ok ok means tests pass, answers change as expected CTSM tag used for the baseline comparisons: ctsm1.0.dev038 Answer changes -------------- Changes answers relative to baseline: YES (small) Summarize any changes to answers, i.e., - what code configurations: all - what platforms/compilers: all - nature of change (roundoff; larger than roundoff/same climate; new climate): larger than roundoff; likely same climate, but this still requires investigation Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): https://github.com/ESCOMP/ctsm/pull/703 =============================================================== =============================================================== Tag name: ctsm1.0.dev038 Originator(s): sacks (Bill Sacks) Date: Wed May 1 10:53:02 MDT 2019 One-line Summary: Support for NWP configuration, NLDAS grid and NLDAS datm forcing Purpose of changes ------------------ This tag includes a set of changes that, together, provide support for running an initial Numerical Weather Prediction version of CTSM over the Continental U.S. (1) NWP configuration: The changes in this configuration relative to the climate (CLM5) configuration mainly target decreasing runtime by removing features that are expensive and not as important for NWP time scales (days to months rather than many years) and space scales (high resolution): - Single dominant landunit; if vegetated, single dominant PFT - Only 5 soil layers, down to 3 m - Only 5 snow layers - Plant hydraulic stress off - Ball-Berry rather than Medlyn stomatal conductance method (this is a side-effect of turning plant hydraulic stress off, and is not itself a critical aspect of the NWP configuration) - Maximum of 3 iterations to compute canopy fluxes - MEGAN is off Note that the NWP compset triggers changes in two new xml variables: CLM_CONFIGURATION ('clm' vs. 'nwp'; this controls things of a more scientific nature, like whether plant hydraulic stress is on or off), and CLM_STRUCTURE ('standard' vs. 'fast'; this controls structural things like the maximum number of landunits per grid cell and the soil layer structure). Thus, you can create a case that is a hybrid between the CLM and NWP configurations by changing these two xml variables independently. (2) NLDAS2 regional grid: 0.125 deg grid over the Continental U.S. (3) Updated version of MOSART with support for the same NLDAS2 regional grid (4) NLDAS2 data atmosphere forcing: over the same regional grid; forcing data are available from 1980 - 2018. (We have excluded 1979 because a small amount of data were missing for that year, and we didn't have a good way to handle those missing data.) These changes can be used independently or all together. However, note that you will get garbage if you try to use the new NLDAS2 datm forcing for a run that extends beyond the boundaries of the NLDAS2 domain. There are three new NWP compsets: - I2000Ctsm50NwpSpGswpGs: GSWP3 datm forcing; meant for global runs or regional runs outside the NLDAS domain - I2000Ctsm50NwpSpNldasGs: NLDAS2 datm forcing; meant for regional runs over the NLDAS domain or some portion of it - I2000Ctsm50NwpSpNldasRsGs: Same as above, but with a stub runoff model in place of MOSART, for runs that aren't interested in having a runoff model and for which you want improved throughput The alias for the new grid is: nldas2_rnldas2_mnldas2 (indicating that we're using the nldas2 grid for land/atmosphere, runoff ('r') and the ocean mask ('m'). Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/ctsm#451 (Add NWP physics option) - Resolves ESCOMP/ctsm#452 (Add an NWP compset) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] ctsm5_0-nwp [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): - The NWP configuration is still a work in progress - The NLDAS surface data set is still a work in progress (will soon update this to use high-res PFTs) Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): - New xml variables: - CLM_CONFIGURATION (will eventually be renamed to CTSM_CONFIGURATION) - CLM_STRUCTURE (will eventually be renamed to CTSM_STRUCTURE) - New namelist variable: - itmax_canopy_fluxes Changes made to namelist defaults (e.g., changed parameter values): No changes for existing CLM compsets; new defaults for NWP compsets Changes to the datasets (e.g., parameter, surface or initial files): No changes for existing grids; new surface dataset for nldas2 grid Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: - New tests covering NWP compsets, nldas grid and nldas datm forcing Code reviewed by: Mike Barlage, Erik Kluzek CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - ok tests pass; namelists change as expected tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run regular tests (aux_clm): cheyenne ---- pass hobart ------ pass CTSM tag used for the baseline comparisons: ctsm1.0.dev037 Answer changes -------------- Changes answers relative to baseline: NO Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): - MOSART: release-cesm2.0.00 -> nldas-grid.n01_mosart1_0_31 - CIME: ctsm/ctsm1.0/cime5.7.9/n04 -> ctsm/ctsm1.0/cime5.7.9/n05 Pull Requests that document the changes (include PR ids): - https://github.com/ESCOMP/ctsm/pull/682 - https://github.com/ESCOMP/ctsm/pull/685 =============================================================== =============================================================== Tag name: ctsm1.0.dev037 Originator(s): sacks (Bill Sacks) Date: Mon Apr 29 16:11:47 MDT 2019 One-line Summary: Change year alignment for present-day I compsets Purpose of changes ------------------ Change year alignment for present-day I compsets (typically, these compsets have aliases beginning with "I2000"), in order to work better for cases that are meant to replicate particular years. (1) Start the model in year 2000 rather than year 1 (2) Change datm's year alignment so that model year 2000 uses observations from year 2000, model year 2001 uses observations from year 2001, etc. (This change was requested by Sean Swenson and supported by Dave Lawrence.) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): - Year-2000 compsets now start in model year 2000 rather than 1 Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: cime changes reviewed by Erik Kluzek and others CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run regular tests (aux_clm): cheyenne ---- ok hobart ------ ok ok means tests pass, year-2000 cases change answers as expected CTSM tag used for the baseline comparisons: ctsm1.0.dev036 Answer changes -------------- Changes answers relative to baseline: YES (small) Summarize any changes to answers, i.e., - what code configurations: Present-day compsets (I2000, I1Pt; also I2010 once those compsets are added) - what platforms/compilers: all - nature of change (roundoff; larger than roundoff/same climate; new climate): larger than roundoff/same climate This changes the year alignment of datm. Climatological averages should be the same as before, but the alignment of model year to datm year changes. Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): - cime: ctsm/ctsm1.0/cime5.7.9/n03 -> ctsm/ctsm1.0/cime5.7.9/n04 Pull Requests that document the changes (include PR ids): - cime changes documented in https://github.com/ESMCI/cime/pull/3093 =============================================================== =============================================================== Tag name: ctsm1.0.dev036 Originator(s): erik (Erik Kluzek,UCAR/TSS,303-497-1326) Date: Fri Apr 26 01:20:57 MDT 2019 One-line Summary: Fix carbon isotope bug that caused wrong answers for isotopes under transient land-use change Purpose of changes ------------------ We currently have a bug so that for transient land-use change cases answers are different starting in soil carbon when Carbon isotopes (use_c13, or use_c14) are on versus off. Answers are identical if there is no land-use change. The bug does cause bulk Carbon (C12) to be slightly different, but qualitatively the same. There is a significant impact to the Carbon isotope simulation however. The bug was causing changes in isotopic pools to be directed to the bulk Carbon. Because, isotopes are so much smaller than bulk Carbon, this doesn't cause a marked difference in the bulk Carbon simulation. But, the lack of the change in the isotopic pool does have a meaningful impact on the simulation of the Carbon isotopic fields. Also added a general test for this and specific one to catch this case. And did some updates on documentation files, and a small bug fix for run_sys_tests. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): #692 #675 #686 #675 -- Answers differ when carbon isotopes are on under transient cases #686 -- Add a system test to make sure turning on carbon isotopes doesn't change answers #692 -- run_sys_test issue Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [X] clm5_0 [X] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): None Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): None Changes made to namelist defaults (e.g., changed parameter values): None Changes to the datasets (e.g., parameter, surface or initial files): None Substantial timing or memory changes: None Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- Caveats for developers (e.g., code that is duplicated that requires double maintenance): None Changes to tests or testing: Added LCISO test type and a 13 month Hist test to make sure carbon isotopes don't cause a change to answers Code reviewed by: dlawren, billsacks, klindsey, bishtgautum CTSM testing: regular [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - PASS tools-tests (test/tools): cheyenne - PASS PTCLM testing (tools/shared/PTCLM/test): cheyenne - PASS (compare tests fail, because of namelist updates) regular tests (aux_clm): cheyenne ---- OK hobart ------ OK CTSM tag used for the baseline comparisons: ctsm1.0.dev035 Answer changes -------------- Changes answers relative to baseline: Only for Carbon isotopes on under transient land-use change! Summarize any changes to answers, i.e., - what code configurations: Bgc with use_c13 or use_c14 T and with a transient land-use change after each year boundary - what platforms/compilers: All - nature of change: same climate for bulk Carbon, isotopic Carbon is quite different for regions with land-use change Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): cime cime from -- ctsm/ctsm1.0/cime5.7.9/n01 to ctsm/ctsm1.0/cime5.7.9/n03 cime update allows the new test to work and starts adding NLDAS grid in Pull Requests that document the changes (include PR ids): #696 #694 (https://github.com/ESCOMP/ctsm/pull) #696 -- run_sys_test small fix #694 -- Carbon isotope fix =============================================================== =============================================================== Tag name: ctsm1.0.dev035 Originator(s): sacks (Bill Sacks) Date: Mon Apr 22 20:17:40 MDT 2019 One-line Summary: Change h2ocan to a purely diagnostic variable Purpose of changes ------------------ Previously, h2ocan_patch was maintained through the driver loop as the sum of liqcan_patch and snocan_patch. However, there were places where it was updated separately (though consistently) - i.e., it was not simply a matter of 'h2ocan = liqcan + snocan'. This made the code more complex, especially for the upcoming addition of water tracers. This tag changes h2ocan_patch to be a purely diagnostic variable, set once at the end of the driver loop. Prognostic equations that used to use h2ocan now use liqcan and snocan, either directly or indirectly via a temporary variable set equal to the current value of liqcan+snocan. This tag also completely removes support for snowveg_flag = 'OFF' in order to clean up the code. In addition, this tag renames the namelist variable snowveg_flag to snowveg_affects_radiation (now a logical). Finally, this tag includes a minor change to always compute fcansno, even if snowveg_affects_radiation is false, in order to simplify the code. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/ctsm#199 Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): - snowveg_flag renamed to snowveg_affects_radiation, and changed to a logical flag Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: - Answer-changing portion of these changes reviewed by Keith Oleson and (at least to some extent) Sean Swenson CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run regular tests (aux_clm): cheyenne ---- ok hobart ------ ok ok means tests pass, answers change as expected CTSM tag used for the baseline comparisons: ctsm1.0.dev034 Answer changes -------------- Changes answers relative to baseline: YES (roundoff) Summarize any changes to answers, i.e., - what code configurations: all - what platforms/compilers: all - nature of change (roundoff; larger than roundoff/same climate; new climate): roundoff If bitwise differences were observed, how did you show they were no worse than roundoff? I split these changes into two pieces: (1) an answer-changing piece where I made minimal changes to replace uses of h2ocan_patch with liqcan+snocan (via temporary h2ocan variables holding this sum), and (2) non-answer-changing code cleanup. (1) consisted of the first two commits on this branch (461cafea and f195c105). With the changes in (1), I added checks near each change to ensure that the new expression differed by no more than roundoff from the old (using an absolute cutoff of 1e-13). I ran the full test suite with these checks in place, and generated new baselines. I then compared the changes in (2) with the baselines from (1); this cleanup step was bit-for-bit except for roundoff-level changes in the now-diagnostic-only field H2OCAN. As noted in ESCOMP/ctsm#689, the changes in (1) were larger than I expected, presumably because initially roundoff-level differences can lead to divergent code behavior. The most obvious example of this is in the check of h2ocan > 0 in subroutine FracWet. I have found that, due to roundoff errors, it is possible for h2ocan_patch > 0 by roundoff, while (liqcan_patch + snocan_patch) <= 0 by roundoff. (Presumably the reverse could also be true, though I haven't seen that – but I haven't spent much time looking for it.) Soon, we plan to do a more rigorous test to double check that this change is not climate-changing. Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): - First part of changes in https://github.com/ESCOMP/ctsm/pull/689 =============================================================== =============================================================== Tag name: ctsm1.0.dev034 Originator(s): sacks (Bill Sacks) Date: Sat Apr 20 06:11:23 MDT 2019 One-line Summary: Change clm4_5: Use Justin Perket snow on vegetation Purpose of changes ------------------ This tag changes the behavior of clm4_5, making it now use Justin Perket's snow on vegetation parameterization. This parameterization was originally introduced in clm4_5_1_r112. At the time it was only used for clm5_0 cases. This tag uses that parameterization for clm4_5 cases, too, so that (like clm5_0) they use snowveg_flag = 'ON_RAD' (i.e., vegetation snow canopy on with albedo influence). This changes climate for clm4_5, mainly through albedo changes, particularly in the shoulder seasons. The reason for this change is to allow an upcoming cleanup of the relationship between h2ocan, snocan and liqcan. (See https://github.com/escomp/ctsm/issues/199 for details.) This tag also no longer allows snowveg_flag = 'OFF'. However, code cleanup is deferred to a follow-on tag, in order to separate the answer changes in this tag from the no-greater-than-roundoff-level changes in the follow-on tag. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - First step towards addressing ESCOMP/ctsm#199 (Clean up relationship between h2ocan, snocan and liqcan) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [X] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): - Changes the scientific configuration of clm4_5 cases, as noted above Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): - For clm4_5: snowveg_flag = 'ON_RAD' rather than 'OFF' Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): - Need to clean up code involving snowveg_flag, liqcan, snocan and h2ocan. I will do this in a follow-on tag. Changes to tests or testing: - Added a test with snowveg_flag = 'ON'. (At least some of the code paths triggered by this setting were previously covered by clm45 cases. Now that both clm45 and clm50 use snowveg_flag = 'ON_RAD', there was some uncovered code. This new test covers this code.) - Changed long single-point tests to use Qian forcing: With the default GSWP3 forcing, the vast majority of the runtime was being spent in datm. In addition, I was getting repeated failures of ERS_Lm54_Mmpi-serial.1x1_numaIA.I2000Clm50BgcCropGs.cheyenne_intel.clm-cropMonthOutput, I think due to memory issues; I'm hopeful that changing the datm forcing will help with that. Code reviewed by: self CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run regular tests (aux_clm): cheyenne ---- ok hobart ------ ok ok means tests pass; NLCOMP and BASELINE failures for Clm45 cases, as expected CTSM tag used for the baseline comparisons: ctsm1.0.dev033 Answer changes -------------- Changes answers relative to baseline: YES (clm4_5) Important Summarize any changes to answers, i.e., - what code configurations: Clm45 compsets - what platforms/compilers: all - nature of change (roundoff; larger than roundoff/same climate; new climate): New climate (as in clm4_5_1_r112 for clm5_0); the main impact is through albedo changes, particularly in the shoulder seasons. Changes in clm4_5 cases are due to: (1) Using snowveg_flag = 'ON_RAD' rather than 'OFF' (2) Changed initial conditions for canopy water: Rather than using H2OCAN, which is the meaningful state variable on previous clm4_5 initial conditions files, now we use LIQCAN and SNOCAN. These may be 0, although at least some clm4_5 initial conditions files (including the two non-Fates out-of-the-box clm4_5 initial conditions files) have non-zero (though still reasonable) LIQCAN for reasons I can't understand. (Also changes answers for ERS_D_Ln9_P480x3.f19_g16.I2000Clm50SpGs.cheyenne_intel.clm-waccmx_offline, which uses a clm45 initial file.) Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): none =============================================================== =============================================================== Tag name: ctsm1.0.dev033 Originator(s): slevis (Samuel Levis, SLevis Consulting LLC, 303-665-1310) Date: Thu Apr 11 10:51:29 MDT 2019 One-line Summary: Limit landunit presence with thresholds set in the namelist Purpose of changes ------------------ Faster simulations. Motivated by NWP applications: - Introduce user-defined namelist parameters to use as thresholds above which landunit areas are kept and below which landunit areas are removed. - Introduce user-defined logical namelist parameter to collapse the up to three urban landunits to the dominant urban landunit. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): #581 Fixes #581 -- Allow zeroing out special landunits if their area is below some threshold Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? No Notes of particular relevance for users --------------------------------------- Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): - New namelist variables with valid real values from 0._r8 to 100._r8, where 0 means "do nothing" and other values determine the threshold above which to keep the landunit in a gridcell. The new variables are: toosmall_soil toosmall_crop toosmall_glacier toosmall_lake toosmall_wetland toosmall_urban - New logical namelist variable collapse_urban. If .true. then the up to three urban landunits collapse to the dominant urban landunit. --------------------------------------------- Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- Caveats for developers (e.g., code that is duplicated that requires double maintenance): There's a typo in the name for n_dom_landunits (missing the s) in bld/CLMBuildNamelist.pm build-namelist doesn't check for toosmall_* below 0 or above 100, but the code does Changes to tests or testing: Yes New unit tests subroutine test_collapse_individual_lunits and test_collapse_all_lunit_opts_active in test_surfrdUtils.pf Modified standard suite test ERS_D_Ld10.f10_f10_musgs.IHistClm50BgcCropGs.cheyenne_intel.clm-collapse_to_dominant_decStart to ERS_D_Ld10.f10_f10_musgs.I2000Clm50BgcCropGs.cheyenne_intel.clm-rm_indiv_lunits_and_collapse_to_dom and introduced fatal error message when namelist parameters toosmall_* > 0 or n_dom_* > 0 in transient simulations Code reviewed by: Erik Kluzek, Bill Sacks CTSM testing: regular [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - PASS (384 differences because of new namelist items) regular tests (aux_clm): cheyenne ---- ok hobart ------ ok CTSM tag used for the baseline comparisons: cheyenne ctsm1.0.dev031 hobart ctsm1.0.dev032 Answer changes -------------- Changes answers relative to baseline: No Note: Comparison to baseline gives NLCOMP failures because the new namelist parameters always appear in the namelist now, even when they have not been changed from their default value by the user. Detailed list of changes ------------------------ - Introduced call to new subroutine collapse_individual_lunits that uses the user-defined namelist variables listed above to determine whether a landunit gets kept or removed from a gridcell. - Introduced new call to collapse_to_dominant for when collapse_urban = .true. - Introduced error checks that do not allow coexistence of transient simulations and the toosmall_* variables OR the n_dom_pfts/n_dom_landunits variables OR the collapse_urban parameter. Pull Requests that document the changes (include PR ids): #641 https://github.com/ESCOMP/ctsm/pull/641 #641 -- Ability to remove individual landunits for faster runs =============================================================== =============================================================== Tag name: ctsm1.0.dev032 Originator(s): negins (Negin Sobhani,UCAR/CSEG,303-497-1224) Date: Mon Apr 8 08:26:57 MDT 2019 One-line Summary: Fixing the balance check to check for possible errors over all columns/patches Purpose of changes ------------------ Up until now, the BalanceCheckMod code in biogeophysics did not check all the possible errors over all columns or patches with the error threshold for aborting clm and it merely compared the last column/patch found (where warning threshold is met) with the error threshold. Therefore, there was always a possibility that the abort is not triggered when it should have been. This issue was first mentioned in ctsm issue #55 for soil energy balance check, but a similar issue is also present for other balance checks in BalanceCheckMod.F90 . Below is the list of the balance checks with a similar issue: 1. Water balance check 2. Snow balance check 3. Solar radiation energy balance check 4. Longwave radiation energy balance check 5. Surface energy balance check 6. Soil energy balance check (Issue #55) For remediating this issue, we used MAXVAL and MAXLOC Fortran intrinsic functions to compare the largest error with the warning and error thresholds instead of checking the last column or looping over all columns. In addition, this PR also makes slight changes to ./run_sys_tests outputi/case directory structure. Previously for each case in a test suite, there were two set of directories created by run_sys_tests as follows: * one directory which included run/ and bld/ and named as $SCRATCH/[case-name].... * The case directory which included logs , timings, CaseDocs/, and misc. scripts for ./case.build, ./case.submit... under $SCRATCH/tests-[testID]/[case-name].... This made debugging confusing and populated many folders in user's scratch. Now, for every case in a test, all directories related to that case including bld/ and run/ directories are nested under the case directory under $SCRATCH/tests-[testID]/[case-name].... This is now similar to ./create_test output folder structure. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/ctsm issue#55 Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): None Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): None Changes made to namelist defaults (e.g., changed parameter values): None Changes to the datasets (e.g., parameter, surface or initial files): None Substantial timing or memory changes: None Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): Changes to tests or testing: - ./run_sys_test is changed so that it nests bld/ and run/ directories under the case directory as $SCRATCH/tests-[testID]/[case-name]... The output folder structure now is similar to ./create_test for each case in a test. Code reviewed by: Bill Sacks CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run regular tests (aux_clm): cheyenne ---- ok hobart ------ ok CTSM tag used for the baseline comparisons: ctsm1.0.dev031 Answer changes -------------- Changes answers relative to baseline: No Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): None Pull Requests that document the changes (include PR ids): https://github.com/ESCOMP/ctsm/pull/670 =============================================================== =============================================================== Tag name: ctsm1.0.dev031 Originator(s): sacks (Bill Sacks) Date: Wed Mar 13 15:03:42 MDT 2019 One-line Summary: Subtract virtual states to reduce dynbal fluxes for transient glaciers Purpose of changes ------------------ Up until now, when computing the dynbal correction (conservation) fluxes for transient glacier columns, we have been (1) counting the mass and energy in the ice column as if that is a real state, but on the other hand (2) NOT accounting for the fact that glacier columns don't represent the soil under glacier. These two issues work in opposite directions, but (1) dominates, because there is much more ice (and energy, I think) in the roughly 50 m of glacial ice than there is in a typical 50 m soil column. In discussing this issue with Bill Lipscomb, we came up with the idea of subtracting some baseline value from each glacier column. I think that, as long as we subtract the same baseline value throughout an entire simulation for a given column, we will still conserve mass and energy through dynamic landunit transitions. So, here we subtract baseline values from glacier columns, accounting for the two issues mentioned above: (1) we subtract the water and energy in the glacier ice, because this is a virtual state in CTSM, and (2) we add the water and energy from the vegetated column(s), to account for the fact that we don't have an explicit representation of soil-under-glacier (this carries the assumption that the soil-under-glacier has the same state as the initial vegetated state in that grid cell). We set these baselines in initialization, so they begin equal to the cold start state. Water and ice in the glacial ice stay fixed over the course of a simulation, so the cold start values should be the same as the current values at any point in time. The heat content of the glacial ice does change over time, but by subtracting this baseline value, we can potentially reduce the dynbal sensible heat fluxes (however, note that it's also possible that these sensible heat fluxes could increase when subtracting the cold start value, if ice temperatures are closer to 0 deg C than to the cold start value, currently 250 K). In addition, this introduces a new namelist flag, reset_dynbal_baselines, which allows the user to reset these baselines at some desired point in the simulation. I think that, in general, this resetting would break conservation. But as long as it is done before the onset of transient glaciers, I think this should be okay. In this way, the user can minimize dynbal fluxes even further, by resetting the baselines after the system has spun up. If the states haven't changed much from this point to the point when glacier dynamics occur, then the dynbal fluxes should now be very small. (See the documentation of this flag in namelist_definition_ctsm.xml for more details.) Without this change, dynbal fluxes for a grid cell that undergoes 100% glaciation / deglaciation in a single year are around 50 m ice, 1 m liquid water, and a few tens of W m-2 (where those quantities of ice and water are dribbled throughout the year, so we end up with an ice flux of about 1.5e-6 mm/s throughout the year; and the energy flux is applied evenly throughout the year - i.e., tens of W m-2 every time step for a year). With this change, and with reset_dynbal_baselines set at an appropriate time, these fluxes can be reduced to close to zero. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Partially addresses ESCOMP/ctsm#274 (Dynamic landunits: improve water and energy conservation) CIME Issues fixed (include issue #): - Fixes three issues related to inputdata checksum: - ESCMI/cime#3033 - ESCMI/cime#3034 - ESCMI/cime#3036 Known bugs introduced in this tag (include github issue ID): - ESCOMP/ctsm#659 (Subtract dynbal baselines from begwb and endwb) - This can't be fixed until we're okay having this answer change on master Known bugs found since the previous tag (include github issue ID): - ESCOMP/ctsm#658 (Methane should not depend on gridcell-level TWS) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [X] clm5_0 [X] clm4_5 Significant answer changes for runs with transient glaciers. Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): - For runs with transient glaciers, careful thought should be given to when to set reset_dynbal_baselines - Inconsistency between begwb/endwb and liq1/liq2/ice1/ice2 (see ESCOMP/ctsm#659) Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): - New namelist variable: reset_dynbal_baselines Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): - Inconsistency between begwb/endwb and liq1/liq2/ice1/ice2 (see ESCOMP/ctsm#659) Changes to tests or testing: - Tweaked a test with transient glacier areas, including adding this new flag to the test Code reviewed by: Bill Lipscomb reviewed the conceptual ideas, but nobody has reviewed the actual code CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run regular tests (aux_clm): cheyenne ---- ok hobart ------ ok ok means tests pass, answers change as expected CTSM tag used for the baseline comparisons: ctsm1.0.dev030 Answer changes -------------- Changes answers relative to baseline: YES (Important) Summarize any changes to answers, i.e., - what code configurations: ALL, but different configurations show different levels of changes, as noted below - what platforms/compilers: all - nature of change (roundoff; larger than roundoff/same climate; new climate): - new climate (at least regionally) for some configurations; roundoff for others: see below There are three levels of changes here: (1) For runs with transient glaciers, there are large answer changes in regions with changing glacier area. These answer changes are in the various dynbal fluxes - for liquid water and ice runoff, and for the dynbal sensible heat flux. (2) For runs with transient vegetation (Hist, Dv and Fates), there are roundoff-level changes in the dynbal fluxes. For liquid water and ice, these changes appear to only occur in grid cells where there is some glacier area (because now both the before and after gridcell water contents have changed by the same fixed amount, which changes the difference at the roundoff level). For heat, there are roundoff-level changes everywhere (because of changes in the order of operations). (3) All runs have large changes in the diagnostic variables, ICE_CONTENT1, LIQUID_CONTENT1 and HEAT_CONTENT1, over grid cells containing some glacier area. Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): - cime: cime5.7.9 -> ctsm/ctsm1.0/cime5.7.9/n01 - Fixes issues related to inputdata checksum Pull Requests that document the changes (include PR ids): - https://github.com/ESCOMP/ctsm/pull/650 =============================================================== =============================================================== Tag name: ctsm1.0.dev030 Originator(s): sacks (Bill Sacks) Date: Fri Mar 8 15:08:34 MST 2019 One-line Summary: Update CIME; hookup expected test fails Purpose of changes ------------------ Two main changes: (1) Update to latest cime (from ctsm/ctsm1.0/cime5.7.5/n01 to cime5.7.9). Some of the important changes are: - Fix for recent cheyenne system changes - New domain file for f19_g17 (changes answers at the roundoff level) - New domain file for T31_g37 (changes answers by greater than roundoff) (2) In system tests: Annotate cs.status output with expected failures Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Fixes ESCOMP/ctsm#654 CIME Issues fixed (include issue #): (Many) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none? The PFS test showed a significant speedup (170s in the previous tag, 124s in this tag). However, some other tests showed either a slowdown or not much change. So I'm not sure there is a consistent pattern of speedup in this tag. Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: Erik Kluzek CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run regular tests (aux_clm): cheyenne ---- ok hobart ------ ok ok means tests pass, some baselines failed as noted below Most testing done with cime at eb2eb704f - excluded PR #3030, which updated support for CAM-SE grids (but the changes in this PR were very limited). Ran SSP tests with the final version of cime (so I'd have a few tests covering the latest cime). CTSM tag used for the baseline comparisons: ctsm1.0.dev029 Answer changes -------------- Changes answers relative to baseline: YES (f19_g17, T31_g37 small) Summarize any changes to answers, i.e., - what code configurations: f19_g17, T31_g37 - what platforms/compilers: all - nature of change (roundoff; larger than roundoff/same climate; new climate): - f19_g17: roundoff-level, presumably due to new domain file (some diffs grew to greater than roundoff-level: some methane variables in a long test - SMS_Lm13.f19_g17.I2000Clm50BgcCrop.cheyenne_intel.clm-cropMonthOutput; and many variables in test with flooding - ERP_P180x2_D.f19_g17.I2000Clm50SpRtmFl.cheyenne_intel.clm-default) - T31_g37: greater than roundoff-level, presumably due to new domain file from cime commit 4fcf592ee Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): - cime: ctsm/ctsm1.0/cime5.7.5/n01 -> cime5.7.9 Pull Requests that document the changes (include PR ids): - https://github.com/ESCOMP/ctsm/pull/602 =============================================================== =============================================================== Tag name: ctsm1.0.dev029 Originator(s): slevis (Samuel Levis, SLevis Consulting LLC, 303-665-1310) Date: Tue Feb 26 23:42:39 MST 2019 One-line Summary: Collapse landunits to the N most dominant Purpose of changes ------------------ Allows the collapse to fewer landunits using the concept of N most dominant for the purpose of faster runs, eg in the context of NWP (Numerical Weather Prediction). Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): #581 Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? No [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): None Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): New namelist variable n_dom_landunits with valid integer values from 0 to max_lunit currently equal to 9, where 0 means “do nothing” and other values determine the number of active landunits when running the model. Changes made to namelist defaults (e.g., changed parameter values): None Changes to the datasets (e.g., parameter, surface or initial files): None Substantial timing or memory changes: None Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- Caveats for developers (e.g., code that is duplicated that requires double maintenance): There is a question left in on whether the collapse call needs to also be moved to inside dynamic landunits reading. Changes to tests or testing: Yes ERS_D_Ld10.f10_f10_musgs.IHistClm50BgcCropGs.cheyenne_intel.clm-collapse_to_dominant_decStart now replaces (by combining with) this test introduced in ctsm1.0.dev026 ERS_D_Ld10.f10_f10_musgs.IHistClm50BgcCropGs.cheyenne_intel.clm-collapse_nat_pfts_decStart I added subroutine test_collapse_to_dom_landunits to test_surfrdUtils.pf to include new Unit Tests. In same file I renamed subroutine test_collapse_nat_pfts to test_collapse_to_dom_pfts to accurately describe the contents of the subroutine. Code reviewed by: Erik Kluzek CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - PASS regular tests (aux_clm): cheyenne ---- OK hobart ------ OK CTSM tag used for the baseline comparisons: ctsm1.0.dev028 Answer changes -------------- Changes answers relative to baseline: No Detailed list of changes ------------------------ Replaced existing subroutine collapse_nat_pfts with subroutine collapse_to_dominant which can collapse to the N dominant landunits or N dominant pfts by performing the same actions on the data. List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): None Pull Requests that document the changes (include PR ids): #639 https://github.com/ESCOMP/ctsm/pull/639 =============================================================== =============================================================== Tag name: ctsm1.0.dev028 Originator(s): sacks (Bill Sacks) Date: Tue Feb 26 14:35:01 MST 2019 One-line Summary: Interpolate out-of-the-box initial conditions and remove expensive tests Purpose of changes ------------------ Two main changes (plus some small additional changes): 1. Removed / reworked some expensive tests 2. Interpolated all out-of-the-box initial conditions, so that the out-of-the-box version is now compatible with our current configuration. The changes from before were (a) our standard configuration now uses the gx1v7 rather than gx1v6 land mask; (b) many inactive points are now absent in memory. See https://github.com/ESCOMP/ctsm/pull/622 for details. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/ctsm#312 - Partially addresses ESCOMP/ctsm#275 (just a bit) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): - New finidat files Substantial timing or memory changes: - Faster initialization times for cases that no longer need to interpolate initial conditions Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in README.CHECKLIST.master_tags as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: - Removed or reworked expensive tests: see details in https://github.com/ESCOMP/ctsm/pull/622 Code reviewed by: Erik Kluzek CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run regular tests (aux_clm): cheyenne ---- ok hobart ------ ok ok: tests pass, baselines fail as expected CTSM tag used for the baseline comparisons: ctsm1.0.dev027 Answer changes -------------- Changes answers relative to baseline: YES (just for initial conditions) Summarize any changes to answers, i.e., - what code configurations: all configurations that still need use_init_interp = .true. - what platforms/compilers: all - nature of change (roundoff; larger than roundoff/same climate; new climate): Not investigated carefully, but expected to be larger than roundoff/same climate Answers change for cases that still need to do interpolation from these out-of-the-box finidat files. I first confirmed that the new finidat files are good via: - Compared `ncdump -h` of old vs. new initial conditions files - Ran the following tests, with comparison against master; verified that these all pass and are bit-for-bit, that they use the new initial conditions files, and that they now have `use_init_interp` unset (whereas they used to have `use_init_interp` set to `.true.`); note that these cover 6 of the 7 new initial conditions files: ``` SMS_Ld1.f09_g17.I1850Clm45BgcGs.cheyenne_intel.clm-default SMS_Ld1.f09_g17.I1850Clm45BgcCruGs.cheyenne_intel.clm-default SMS_Ld1.f09_g17.I1850Clm50Sp.cheyenne_intel.clm-default SMS_Ld1.f09_g17.I1850Clm50BgcCrop.cheyenne_intel.clm-ciso SMS_Ld1.f09_g17.I1850Clm50BgcCropCru.cheyenne_intel.clm-ciso SMS_Ld1.f09_g17.I1850Clm50SpCru.cheyenne_intel.clm-default ``` - To cover the last new initial conditions file, ran the following test (with the same checks as above: passes and bit-for-bit, checked finidat, and checked `use_init_interp`), with `./xmlchange LND_TUNING_MODE=clm5_0_cam6.0` (for the baseline for this test, I set `init_interp_method = 'use_finidat_areas'`, which I also used in interpolating this initial conditions file): ``` SMS_Ld1.f09_g17.I1850Clm50BgcCrop.cheyenne_intel ``` Details on answer changes: As noted above, these changes are bit-for-bit for configurations that can now use the out-of-the-box initial conditions without interpolation. However, answers change for configurations that still require interpolation of the out-of-the-box initial conditions (e.g., because they are at a different resolution than the finidat file). I expected these answer changes around the Caspian Sea, due to the fact that we are now effectively interpolating twice from the original file: once to produce the new out-of-the-box file (which includes an interpolation from `gx1v6` to `gx1v7`, and so interpolates onto the Caspian Sea region) and once to go from `f09_g17` to the target resolution (rather than interpolating directly from `f09_g16` to the target resolution). However, when I ran a test with comparison to baselines - `ERP_Ld5.f19_g17.I1850Clm50Bgc.cheyenne_intel.clm-default` - I saw changes in regions further afield than simply around the Caspian Sea. See images in https://github.com/ESCOMP/ctsm/pull/622 for details. It's possible that the other differences are also due to the Caspian Sea change, but it's also possible that there is some other cause here that I don't understand. But Erik Kluzek and I agreed that it isn't worth trying to understand these other (isolated) differences. Also note: While most f09_g17 tests are bit-for-bit with the earlier tag (because now use_init_interp is .false. for those tests), `SMS_Ld2_D.f09_g17.I1850Clm50BgcCropCmip6.cheyenne_intel.clm-basic_interp` differs in some glc forcing fields for a few grid cells in Antarctica. This is perhaps not surprising, since this test needs to use interpolation to get fields for the Antarctica virtual columns; so, while I don't completely understand the cause of these answer changes, I suspect the root cause is similar to the isolated answer changes I saw sprinkled around the world for `ERP_Ld5.f19_g17.I1850Clm50Bgc.cheyenne_intel.clm-default`. Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): - https://github.com/ESCOMP/ctsm/pull/622 =============================================================== =============================================================== Tag name: ctsm1.0.dev027 Originator(s): negins (Negin Sobhani,UCAR/CSEG,303-497-1224) Date: Tue Feb 19 12:57:12 MST 2019 One-line Summary: Non-constant time initialization for soil hydrology types are moved to SoilHydrologyType.F90. Purpose of changes ------------------ Initalization of time-varying `zwt`, `zwt_perched`, and `frost_table` are moved to `InitCold` in `SoilHydrologyType.F90`. Previously, `zwt`, `zwt_perched`, and `frost_table` were initialized in `SoilHydrologyInitTimeConstMod.F90`, which caused confusions. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/ctsm#575 ("Some variables initialized in SoilHydrologyInitTimeConstMod are not time-constant") Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: Updated a test. The test added in previous tag included decStart with transient glaciers, which would not work. Hence, this test is modified so the compset does not include G. Test IHistClm50SpG is updated to IHistClm50Sp. Code reviewed by: Bill Sacks Erik Kluzek CTSM testing: build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run regular tests (aux_clm): cheyenne ---- ok hobart ------ ok ok means tests pass, except for the expected failures. CTSM tag used for the baseline comparisons: ctsm1.0.dev026 Answer changes ------------- Changes answers relative to baseline: no Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): Pull Requests that document the changes (include PR ids): PR #631 Soil hydrology initialization of time-constant variables moved to SoilHydrologyType https://github.com/ESCOMP/ctsm/pull/631 =============================================================== =============================================================== Tag name: ctsm1.0.dev026 Originator(s): slevis (Samuel Levis, SLevis Consulting LLC, 303-665-1310) Date: Tue Feb 5 17:08:48 MST 2019 One-line Summary: Collapse unmanaged PFTs to the N most dominant Purpose of changes ------------------ To run simulations fast for applications such as NWP (Numerical Weather Prediction), allow collapsing natural PFT's to the number the user wants. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): 457 Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): New namelist parameter n_dom_pfts with valid integer values from 0 to 14, where 0 means “do nothing” and other values determine the number of active pfts when running the model Changes made to namelist defaults (e.g., changed parameter values): None Changes to the datasets (e.g., parameter, surface or initial files): None Substantial timing or memory changes: None (unless new option invoked) Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in ../CTSMMasterChecklist as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: updated tests, added new test Updated standard testing test to include decStart (December start). Confirmed that answers were bit-for-bit same with baseline before introducing the change. Added standard testing test to test n_dom_pfts = 2 in a transient vegetation case with December start. Added subroutine test_collapse_nat_pfts to test_surfrdUtils.pf to include new Unit Tests. Code reviewed by: Erik Kluzek Bill Sacks CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - PASS regular tests (aux_clm): cheyenne ---- OK hobart ------ OK CTSM tag used for the baseline comparisons: ctsm1.0.dev025 Answer changes -------------- Changes answers relative to baseline: No (bit-for-bit) Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): None Pull Requests that document the changes (include PR ids): https://github.com/ESCOMP/ctsm/pull/588 -- NWP collapse pfts to the N most dominant https://github.com/ESCOMP/ctsm/pull/583 -- Add routine to find indices of max k values in data Added new subroutines collapse_nat_pfts and collapse_crop_var and called them from surfrdMod.F90. The second of these subroutines is not needed when collapsing unmanaged PFTs; however, I created it early in the PR and decided to keep it for use soon with crop-related variables like irrigation water and fertilizer. Bill Sacks made changes to include subroutine find_k_max_indices for the benefit of this PR (#583) =============================================================== =============================================================== Tag name: ctsm1.0.dev025 Originator(s): Keith Oleson, Bill Sacks Date: Wed Jan 23 10:48:01 MST 2019 One-line Summary: History fields for vertically-resolved sums of soil C and N, and minor fixes Purpose of changes ------------------ Main change is from Keith Oleson: Add history fields for vertically-resolved sums of SOIL1C, SOIL2C, and SOIL3C for C12, C13, C14, and similarly for N. New fields are SOILC_vr, C13_SOILC_vr, and C14_SOILC_vr, and SOILN_vr. For runs that use the output_bgc usermods, including cmip6 runs, we no longer output 'SOIL1C_vr', 'SOIL1N_vr', 'SOIL2C_vr', 'SOIL2N_vr', 'SOIL3C_vr', 'SOIL3N_vr'; instead we output 'SOILC_vr', 'SOILN_vr', and similarly for C isotopes. Also minor fixes: - Output cpl hist files in SSP test (resolves ESCOMP/ctsm#61) - Remove FATES-related commented-out code in OzoneMod (this has been moved to https://github.com/ESCOMP/ctsm/issues/618) - Minor tweak to run_sys_tests Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/ctsm#61 Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): - For runs that use the output_bgc usermods, including cmip6 runs, we no longer output 'SOIL1C_vr', 'SOIL1N_vr', 'SOIL2C_vr', 'SOIL2N_vr', 'SOIL3C_vr', 'SOIL3N_vr'; instead we output 'SOILC_vr', 'SOILN_vr', and similarly for C isotopes. Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in ../CTSMMasterChecklist as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: Bill Sacks; Erik Kluzek reviewed the changes to the SSP test CTSM testing: build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run regular tests (aux_clm): cheyenne ---- pass hobart ------ pass CTSM tag used for the baseline comparisons: ctsm1.0.dev024 Answer changes -------------- Changes answers relative to baseline: NO Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): https://github.com/ESCOMP/ctsm/pull/551 =============================================================== =============================================================== Tag name: ctsm1.0.dev024 Originator(s): slevis (Samuel Levis,SLevis Consulting, LLC, 303-665-1310) Date: Mon Jan 14 11:07:04 MST 2019 One-line Summary: Remove unnecessary restart variables Purpose of changes ------------------ Remove unnecessary restart variables to conserve on computing resources and disk space Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): 285 Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): Changes made to namelist defaults (e.g., changed parameter values): Changes to the datasets (e.g., parameter, surface or initial files): Substantial timing or memory changes: Notes of particular relevance for developers: Code reviewed by Bill Sacks. Standard testing (cheyenne and hobart) OK Unit testing PASS --------------------------------------------- Caveats for developers (e.g., code that is duplicated that requires double maintenance): Changes to tests or testing: Code reviewed by: Bill Sacks CTSM testing: Regular testing (cheyenne and hobart) OK Unit testing PASS CTSM tag used for the baseline comparisons: ctsm1.0.dev023 Answer changes -------------- Changes answers relative to baseline: NO (bfb) Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): Pull Requests that document the changes (include PR ids): https://github.com/ESCOMP/ctsm/pull/598 Variables removed from restart: plant_nalloc pot_f_nit_vr* f_nit_vr* root_depth qflx_floodg snounload qflx_snofrz_lyr sminn_to_plant_fun_patch =============================================================== =============================================================== Tag name: ctsm1.0.dev023 Originator(s): sacks (Bill Sacks) Date: Thu Jan 10 13:00:42 MST 2019 One-line Summary: Remove CLM4.0 Purpose of changes ------------------ Remove CLM4.0. This consists of removing code and other files that are specific to CLM4.0, and doing some resulting cleanup of code that was shared between CLM4.0 and later versions (primarily in build-namelist). Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves #455 (Remove CLM4.0) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] clm4_5 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): - Can no longer run CLM4.0 cases Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): - CLM4.0 compsets no longer supported Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in ../CTSMMasterChecklist as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: Removed CLM4.0 tests Code reviewed by: Erik Kluzek CTSM testing: build-namelist tests: cheyenne - pass tools-tests (test/tools): cheyenne - pass PTCLM testing (tools/shared/PTCLM/test): cheyenne - ok Total number of tests = 52 Number of tests that PASS = 51 Number of tests that Fail = 1 Number of compare tests that PASS = 37 Number of compare tests that Fail = 4 Number of tests without compare = 11 The one failing test is: FAIL fail-no_towerdata.-d+/glade/p/cesmdata/cseg/inputdata++-s+RF-Bra+--debug+--pftgrid+--soilgrid. These same failures occurred in ctsm1.0.dev022 (including the compare tests that Fail). Erik Kluzek says these failures are okay. regular tests (aux_clm): cheyenne ---- pass hobart ------ pass Notes on testing: - Most testing done on b5380c7e8. Since then, updated to a version of cime that fixes cime's configure tool (but this isn't invoked in system tests, so I didn't rerun the system tests) and pointed cime to a tag rather than a hash. - For tools tests: In baseline, cherry-picked commit 148bc2158 into cime CTSM tag used for the baseline comparisons: ctsm1.0.dev022 Answer changes -------------- Changes answers relative to baseline: NO Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): - cime: points to branch tag off of cime5.7.5 (ctsm/ctsm1.0/cime5.7.5/n01), with changes for: (1) Removal of support for CLM4.0 (equivalent to the changes in https://github.com/ESMCI/cime/pull/2968) (2) Fix for cime configure (cherry-picked 148bc2158) Pull Requests that document the changes (include PR ids): - https://github.com/ESCOMP/ctsm/pull/609 (Remove CLM4.0) - https://github.com/ESCOMP/ctsm/pull/601 (Print git status and related info when running run_sys_tests) =============================================================== =============================================================== Tag name: ctsm1.0.dev022 Originator(s): sacks (Bill Sacks) Date: Tue Jan 8 11:01:38 MST 2019 One-line Summary: Set tracer version of irrigation fluxes Purpose of changes ------------------ Set tracer version of irrigation fluxes. This required a substantial rewrite of ApplyIrrigation (now renamed to CalcIrrigationFluxes). The problem was that the code was written to compute the total irrigation withdrawal from groundwater, then use this to compute the application fluxes (drip/sprinkler), then later divide this total withdrawal by layer. But for tracer fluxes to be computed correctly, we needed to reorder this, so that the per-layer withdrawals are determined before determining the application fluxes. This is because the application flux needs to know the tracer concentrations of the source, which requires knowing which layers the source is drawing from. This was made more challenging because of the mix of patch-level and column-level variables at play: irrigation demand is patch-level, groundwater availability and extraction is column-level, but application is back to patch-level. In a somewhat-related change, I also reworked the passing of information between soil hydrology (groundwater availability) and irrigation: Previously, there was some near-duplicate code in CalcAvailableUnconfinedAquifer (which was called prior to ApplyIrrigation) and WithdrawGroundwaterIrrigation (which was called after ApplyIrrigation). I have reworked the code to remove this duplication, calling the new CalcIrrigWithdrawals in the midst of CalcIrrigationFluxes. In doing so, I reconciled an accidental discrepancy between the two original routines (see https://github.com/escomp/ctsm/issues/595). Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/ctsm#521 (Set tracer versions of fluxes set by ApplyIrrigation) - Resolves ESCOMP/ctsm#593 (Generalize groundwater irrigation availability to handle multiple patches per column) - Resolves ESCOMP/ctsm#595 (Inconsistency between CalcAvailableUnconfinedAquifer and WithdrawGroundwaterIrrigation) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] clm4_5 [ ] clm4_0 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: - Sean Swenson reviewed the initial parts of the rework of ApplyIrrigation and the passing of information between soil hydrology and irrigation CTSM testing: build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run regular tests (aux_clm): cheyenne ---- ok hobart ------ ok ok means tests pass, expected baseline failures as noted below Additional manual testing: - Out of the box, the one-timestep tracer consistency test (SMS_D_Ln1.f10_f10_musgs.I2000Clm50BgcCropGs.hobart_nag.clm-tracer_consistency) passed when the new check was inserted after the irrigation call, even without the changes in this tag. Presumably this is because no irrigation was being done in the first time step of this test. To confirm that the changes here were both necessary and effective, I hacked the code to force irrigation of all types (surface and groundwater) on the first time step (see https://github.com/billsacks/ctsm/commit/d28a03145). I confirmed that this test, with those code hacks, failed before the changes in this tag, and passed with these changes in place. - I used a multi-step process to confirm that these changes are only roundoff-level different for groundwater irrigation (because feedbacks in the system result in SMS_D_Ld5.f10_f10_musgs.I2000Clm50BgcCrop.hobart_nag.clm-irrig_alternate looking greater-than-roundoff-level different from the baseline): (1) Confirmed that e507fe603 is only roundoff-level different from baseline using cprnc output from this test. (2) Confirmed that a4a1a626c is only roundoff-level different from e507fe603. This was the tricky part. I did this by introducing some temporary code that (a) computed some fluxes in both the old and new ways, (b) compared the old and new methods for each point and time step, confirming that they were no more than roundoff-level different, then (c) set the fluxes to the old method. I confirmed that this was bit-for-bit with e507fe603, and the checks for greater-than-roundoff-level differences between the old and new methods were never triggered. (See https://github.com/billsacks/ctsm/commit/7d23e9e and the earlier commits on that branch.) (3) Confirmed that remaining changes an the branch are bit-for-bit with a4a1a626c. CTSM tag used for the baseline comparisons: ctsm1.0.dev021 Answer changes -------------- Changes answers relative to baseline: YES (small) Summarize any changes to answers, i.e., - what code configurations: - Small answer changes when groundwater irrigation is enabled. In principle, could change answers by greater than roundoff due to fix of #595, but I didn't see any answer changes in my 7-month test due to that change. Other than that, just roundoff-level changes. - Answer changes when tracers are enabled. - Roundoff-level changes in the diagnostic field, QIRRIG_FROM_SURFACE, for many tests - what platforms/compilers: all - nature of change (roundoff; larger than roundoff/same climate; new climate): See notes under "what code configurations" for details. Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): https://github.com/ESCOMP/ctsm/pull/600 =============================================================== =============================================================== Tag name: ctsm1.0.dev021 Originator(s): mvr (Mathew Rothstein,UCAR/CSEG,303-497-1304) Date: Wed Dec 26 16:29:06 MST 2018 One-line Summary: Added tracer ratio capability and included it in consistency checks Purpose of changes ------------------ These changes are needed to support the implementation of water isotopes by ensuring that isotope-related variables are being updated consistently throughout the model. The consistency checks include the testing of a bulk tracer that should always be equal to the bulk, and also test tracers that maintain a fixed ratio (other than one) to the bulk. This commit includes fixes to the cold start initialization for the water tracers with ratios other than one. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): [If none, remove this line] #541 Only do water tracer consistency checks on tracers that are supposed to maintain a fixed ratio #459 Fix cold start initialization of water tracers #508 Remove unused variables in new Water types #357 Put in place Water isotope consistency checks Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] clm4_5 [ ] clm4_0 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in ../CTSMMasterChecklist as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: added unit tests Code reviewed by: bill sacks CTSM testing: build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run regular tests (aux_clm): cheyenne ---- PASS hobart ------ PASS CTSM tag used for the baseline comparisons: ctsm1.0.dev020 Answer changes -------------- Changes answers relative to baseline: no, b4b Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): #599 Adding isotope tracer ratios (https://github.com/ESCOMP/ctsm/pull) =============================================================== =============================================================== Tag name: ctsm1.0.dev020 Originator(s): Sean Swenson, Bill Sacks Date: Mon Dec 3 11:51:24 MST 2018 One-line Summary: New options for irrigation and crop fsat Purpose of changes ------------------ Introduce three new options: (1) Ability to withdraw irrigation water from groundwater if not enough water is available from rivers. This is controlled via new namelist flag, use_groundwater_irrigation. Water can be withdrawn from both the unconfined (from the soil column) and confined (from wa) aquifers. (2) Irrigation method: sprinkler (above canopy) vs. drip (below canopy). This can be set on a per-crop and per-gridcell basis on the surface dataset, but out-of-the-box support for creating the necessary surface dataset field is not yet in place (see ESCOMP/ctsm#565). For now, it can be controlled globally via a new namelist flag, irrig_method_default. The default is drip, which was what we were previously using implicitly. (3) Set crop fsat to zero. This is controlled by a new namelist option, crop_fsat_equals_zero. Default behavior is the same as before for all three options. Also: - If use_aquifer_layer is false (which is the default for CLM50), no longer reset wa_col every time step - Adds indices to vector history files giving column, landunit and gridcell indices for each patch, etc. (Resolves ESCOMP/ctsm#81: "Restart files are different for CLM when run over different number of tasks" (issue name is a misnomer of the remaining to-dos in that issue.) - Writes 3d time-constant fields on first history file for all tapes, not just the first - Small performance improvements to irrigation Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Resolves ESCOMP/ctsm#81 ("Restart files are different for CLM when run over different number of tasks", but that name was a misnomer for the remaining work needed on the issue) Known bugs introduced in this tag (include github issue ID): - ESCOMP/ctsm#565: Surface dataset enhancements needed for irrigation method Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] clm4_5 [ ] clm4_0 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): - New namelist variables: - use_groundwater_irrigation - irrig_method_default - crop_fsat_equals_zero - Partial support for new surface dataset field: irrigation_method Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in ../CTSMMasterChecklist as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): Changes to tests or testing: - Added two tests of the new irrigation options - Added an _includes directory to hold things that should be included by testmods (not used directly) Code reviewed by: Bill Sacks and Sean Swenson reviewed each other's code CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - ok Tests pass, namelists differ as expected tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run regular tests (aux_clm): cheyenne ---- ok hobart ------ pass ok: tests pass, some answers change as noted below CTSM tag used for the baseline comparisons: ctsm1.0.dev019 Answer changes -------------- Changes answers relative to baseline: YES (same climate) Summarize any changes to answers, i.e., - what code configurations: Some CLM50 cases that are either cold start or (unusually) use a CLM45 initial conditions file; it's possible that non-cold-start configurations would see differences rarely. - what platforms/compilers: all - nature of change (roundoff; larger than roundoff/same climate; new climate): Larger than roundoff, same climate Differences arise due to no longer resetting wa_col to aquifer_water_baseline every time step in BeginWaterBalanceSingle if use_aquifer_layer is false (which is the default for CLM50). (This change is needed to conserve water properly with the new groundwater-based irrigation.) For the most part, wa_col wasn't being changed when use_aquifer_layer is false (so no longer resetting it has no effect if groundwater irrigation isn't being used), but the one exception is the "work around of the negative liquid water" added by Jinyun Tang in subroutine SoilWater (dated Jan 14, 2015), which can be exercised when use_flexibleCN is true. Sean Swenson did some experimentation, and found that the code that updates wa_col is only exercised in the first time step of a cold start run, and that finding is borne out in the test suite results. However, it's possible that there are rare cases when this is exercised later in a run. I have confirmed that this branch is bit-for-bit with the baseline if I revert the addition of the use_aquifer_layer conditional around the wa_col resetting in BeginWaterBalanceSingle. Alternatively, nearly all tests are bit-for-bit with the baseline if I remove the one line updating wa_col in subroutine SoilWater (and also remove the endrun for water balance errors from BalanceCheck); the one exception is the waccmx_offline test, which uses a CLM45 initial conditions file in a CLM50 configuration. The following tests had baseline failures: Cold start tests: ERP_D_Ld5.f10_f10_musgs.IHistClm50BgcCrop.cheyenne_intel.clm-allActive ERP_D_P36x2_Ld3.f10_f10_musgs.I2000Clm50BgcCrop.cheyenne_intel.clm-cropColdStart ERP_P180x2_D_Ld5.f19_g17.I2000Clm50BgcDvCrop.cheyenne_intel.clm-crop ERP_P36x2_Lm25.f10_f10_musgs.I2000Clm50BgcDvCrop.cheyenne_intel.clm-monthly ERP_P72x2_Lm25.f10_f10_musgs.I2000Clm50BgcDvCrop.cheyenne_intel.clm-monthly Test that uses a CLM45 initial conditions file in a CLM50 configuration: ERS_D_Ln9_P480x3.f19_g16.I2000Clm50SpGs.cheyenne_intel.clm-waccmx_offline Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): https://github.com/ESCOMP/ctsm/pull/523 =============================================================== =============================================================== Tag name: ctsm1.0.dev019 Originator(s): sacks (Bill Sacks) Date: Fri Nov 30 13:36:57 MST 2018 One-line Summary: Rework cold start initialization of wa and zwt Purpose of changes ------------------ Rework cold start initialization of wa and zwt when use_aquifer_layer is false to reduce answer changes in upcoming groundwater_irrigation branch. In the groundwater_irrigation branch (https://github.com/ESCOMP/ctsm/pull/523), Sean Swenson has stopped resetting wa_col each time step if use_aquifer_layer is false. However, this leads to having a substantially different value of wa_col when use_aquifer_layer is false: previously, it was reset to aquifer_water_baseline each time step, but with Sean's changes, it stays close to its initial values, which have been 4000 in most places. This tag changes the initial values to match the value it was being reset to, so it simply starts at aquifer_water_baseline - so the every-time-step-resetting to aquifer_water_baseline can be removed without massively changing the value of wa_col. In addition to changing the cold start initialization of wa_col, we are also changing the cold start initialization of zwt in this case where use_aquifer_layer is false: The old initialization of zwt wouldn't work as intended now that we have changed the initial value of wa_col; Sean Swenson suggested this new initialization method instead. This initialization to zi(c,nbedrock(c)) is correct if there are no saturated layers, and close enough for a decent cold start even if there are. Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] clm4_5 [ ] clm4_0 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in ../CTSMMasterChecklist as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: Sean Swenson CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run regular tests (aux_clm): cheyenne ---- ok hobart ------ ok ok means tests pass, answers change as expected for some cases CTSM tag used for the baseline comparisons: ctsm1.0.dev018 Answer changes -------------- Changes answers relative to baseline: YES (small) Summarize any changes to answers, i.e., - what code configurations: CLM50 cold start or transient (Hist) cases - what platforms/compilers: all - nature of change (roundoff; larger than roundoff/same climate; new climate): Larger than roundoff, same climate (not investigated completely rigorously through long simulations, but Sean Swenson and Bill Sacks are both pretty confident that the resulting changes will be small, partly based on difference statistics from the test suite, and partly based on the fact that the only change in this tag is in cold start initialization of some values). Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): https://github.com/ESCOMP/ctsm/pull/577 =============================================================== =============================================================== Tag name: ctsm1.0.dev018 Originator(s): sacks (Bill Sacks) Date: Thu Nov 29 16:03:50 MST 2018 One-line Summary: Water tracer updates for initial things in driver loop Purpose of changes ------------------ Update water tracers for initial stuff done in driver loop. This includes atm2lnd forcings (non-downscaled and downscaled), balance check initialization, and dyn subgrid updates. Broadly speaking, the changes here are: (1) Reworked WaterType to make it easier / more robust for other code to loop over tracers or bulk+tracers (2) The most interesting changes are probably the code to update the atm2lnd water tracers (in Wateratm2lndType.F90 and WaterTracerUtils.F90) (3) In various other places, do some infrastructurey stuff (initializing water balance, doing dyn subgrid stuff) for tracers as well as bulk (4) Supporting unit tests and unit test infrastructure Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): Resolves ESCOMP/ctsm#487 Resolves ESCOMP/ctsm#488 Resolves ESCOMP/ctsm#489 Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] clm4_5 [ ] clm4_0 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in ../CTSMMasterChecklist as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): - We have chosen not to set all water tracers as soon as possible, but instead to do these tracer settings later in the driver loop. This keeps the driver loop cleaner, but means that you cannot arbitrarily sprinkle calls to TracerConsistencyCheck throughout the driver. Specifically for this tag: the non-downscaled, gridcell-level atm2lnd water tracers are not updated until after the call to downscale_forcings, so tracer consistency checks before that point would fail. Changes to tests or testing: - Added a PFS test Code reviewed by: Portions of the design (and possibly code) have been reviewed by Mat Rothstein, David Noone and Mariana Vertenstein CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run regular tests (aux_clm): cheyenne ---- ok hobart ------ ok ok means tests pass, some answers change as expected CTSM tag used for the baseline comparisons: ctsm1.0.dev017 Answer changes -------------- Changes answers relative to baseline: YES (roundoff) Summarize any changes to answers, i.e., - what code configurations: many - what platforms/compilers: all - nature of change (roundoff; larger than roundoff/same climate; new climate): roundoff-level changes in sensible heat flux from precip conversion due to refactoring this calculation; everything else bit-for-bit If bitwise differences were observed, how did you show they were no worse than roundoff? via summarize_cprnc_diffs to see differences in the test suite Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): https://github.com/ESCOMP/ctsm/pull/572 =============================================================== =============================================================== Tag name: ctsm1.0.dev017 Originator(s): slevis (Samuel Levis, Slevis Consulting LLC,303-665-1310) Date: Wed Nov 28 14:27:50 MST 2018 One-line Summary: Merge the collapse2gencrop branch Purpose of changes ------------------ These changes allow the model to not need to read 16-pft datasets and rather read 78-pft datasets. The 78-pft datasets were developed for use with prognostic crops originally. The current changes allow the model to use the 78-pft datasets in 16-pft runs by collapsing the crop pfts (cfts) from specific types to the model's generic crop types. The changes are generic so that the model may still read 16-pft datasets for 16-pft runs. Ultimately these changes will permit users to run with any number of pfts, while the ctsm group maintains a single set of input pft data. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): Fixes #509 (partial) #509 -- irrigate in 1850 is off for runs with use_crop but on for those without Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [ ] clm5_0 [ ] clm4_5 [ ] clm4_0 Notes of particular relevance for users --------------------------------------- Changes made to namelist defaults (e.g., changed parameter values): - maxpatch_pft made obsolete - nnegcrit was increased because of failure in the TRENDY simulations. This was done in the source, rather than the namelist but does have an effect on user behavior. Substantial timing or memory changes: No Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in ../CTSMMasterChecklist as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): The "irrigate" attribute should now be removed from all fsurdat files for namelist generation. Thus irrigate on or off is now an option for non-crop cases. Renamed parameters: maxpatch_pft is obsolete; remove from namelist-related scripts in future PR numpft + 1 ---> maxsoil_patches maxpatch_pft ---> maxsoil_patches in CNDV numpft ---> maxveg numpft_ed ---> maxveg_fates Several modules were made default private. The list of variables in them that are public are explicitly listed now: LakeCon.F90 clm_initializeMod.F90 clm_instMod.F90 clm_varpar.F90 pftconMod.F90 subgridAveMod.F90 Changes to tests or testing: New test New unit tests in test_surfrdUtils.pf, all containing the prefix test_collapse_crop_types_* Code reviewed by: Erik Kluzek and Bill Sacks CTSM testing: Regular [PASS means all tests PASS and OK means tests PASS other than expected fails.] regular tests (aux_clm): cheyenne_intel ---- OK cheyenne_gnu ------ OK hobart_nag -------- PASS hobart_pgi -------- OK hobart_intel ------ PASS CTSM tag used for the baseline comparisons: ctsm1.0.dev016 Answer changes -------------- Changes answers relative to baseline: NO Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): None Pull Requests that document the changes (include PR ids): #483 Sam Levis cloned Erik Kluzek's collapse2gencrop branch and then tested, corrected errors, resolved conflicts, added unit tests, while updating to newer tags as needed. =============================================================== =============================================================== Tag name: ctsm1.0.dev016 Originator(s): sacks (Bill Sacks) Date: Thu Nov 1 07:27:48 MDT 2018 One-line Summary: Update cime, fix FATES DEBUG token, add script to easily run system tests Purpose of changes ------------------ (1) Update to latest version of cime master (2) Update FATES with a minor change: change DEBUG to debug, to allow working with a preprocessor-defined DEBUG token (3) Add a script (run_sys_tests) that allows easily running all system tests (see https://github.com/ESCOMP/ctsm/wiki/System-Testing-Guide#running-test-suites-with-the-run-sys-tests-wrapper for details) (4) As part of (3), start work on a CTSM python library and associated test infrastructure. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Fixes ESCOMP/ctsm#535 (Run Fortran unit tests as part of create_test) CIME Issues fixed (include issue #): - Various - see CIME ChangeLog for details Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [X] clm5_0 [X] clm4_5 [X] clm4_0 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in ../CTSMMasterChecklist as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: - New run_sys_tests wrapper - Fortran unit tests now run as part of aux_clm Code reviewed by: Erik Kluzek CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run tools-tests (test/tools): cheyenne - not run PTCLM testing (tools/shared/PTCLM/test): cheyenne - not run regular tests (aux_clm): cheyenne ---- ok hobart ------ ok ok means tests pass, baselines fail as expected CTSM tag used for the baseline comparisons: ctsm1.0.dev015 Answer changes -------------- Changes answers relative to baseline: YES (new presaesro and CO2 datasets) Summarize any changes to answers, i.e., - what code configurations: All I compsets (except spinup test that uses cplhist datm forcing) - what platforms/compilers: All - nature of change (roundoff; larger than roundoff/same climate; new climate): Potentially new climate Changes answers for I compsets due to datm update as part of the cime update: new presaero and CO2 datasets (https://github.com/esmci/cime/pull/2828) Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): - cime: cime5.7.3 -> cime5.7.5 - fates: fates_s1.8.1_a3.0.0 -> fates_s1.8.1_a3.0.0_rev2 Pull Requests that document the changes (include PR ids): - https://github.com/escomp/ctsm/pull/493 (Wrapper to system tests, and start of a ctsm python library) =============================================================== =============================================================== Tag name: ctsm1.0.dev015 Originator(s): sacks (Bill Sacks) Date: Sun Oct 28 14:29:30 MDT 2018 One-line Summary: CMIP6 compset modifiers, usermods for typical output, and other output enhancements Purpose of changes ------------------ (1) Support %BGC-CROP-CMIP6DECK and %BGC-CROP-CMIP6WACCMDECK compset modifiers, so that we can turn on the necessary options (output-related and others) via new CMIP6-specific compsets. (2) Turn on carbon isotopes in CMIP6 runs (from Erik Kluzek) (3) Remove setting of CCSM_BGC=CO2A in the cmip6 usermods (4) Add usermods directories for getting typical extra output that's wanted in many cases: output_crop, output_crop_highfreq, output_bgc, output_bgc_highfreq, output_sp, and output_sp_highfreq. These can be enabled by adding something like '--user-mods-dir output_crop' on the create_newcase line (that short-hand works for an I compset; for F or B compsets, you need to provide the full path to the usermod directory). (4) Allow holes in the number of history tapes. Holes are cases where, for example, we have h0, h1 and h3 tapes, but no h2 tape (because there are no fields on the h2 tape). (This is needed for (3).) (5) Fix reading and writing of 1-d logical global arrays. This fixes ESCOMP/ctsm#24 for real (rather than just preventing an attempt to read/write 1-d logical arrays, as was done in the previous 'fix'). (6) Add C13_NBP and C14_NBP diagnostic fields (from Keith Oleson) (7) Make a bunch of carbon isotope diagnostic fields inactive by default (8) Don't allow interpolation (use_init_interp) from a case without carbon isotopes to a case with carbon isotopes: Due to https://github.com/ESCOMP/ctsm/issues/67, interpolation from a case without carbon isotopes to a case with carbon isotopes yields incorrect initialization values for the carbon isotopes. Now that we're turning carbon isotopes on via some semi-out-of-the-box usermods (for cmip6), it is becoming more important to check to make sure someone doesn't shoot themselves in the foot this way. (9) Add tests of the new output usermods as well as of the CMIP6 compset modifiers Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Fixes #24 for real (ncd_io_1d_log_glob is broken) - Fixes #529 (Organize usermods_dirs to facilitate running cases with the right output) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] clm4_5 [ ] clm4_0 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): - New namelist variable, just for testing purposes: for_testing_allow_interp_non_ciso_to_ciso. This bypasses an error check, and should only be used in tests. Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in ../CTSMMasterChecklist as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: - Reworked tests of cmip6_deck usermods to use the new %BGC-CROP-CMIP6DECK compset modifier, and changed f09_g16 to f09_g17 - Added a test of the %BGC-CROP-CMIP6WACCMDECK compset modifier - Added tests of output_crop_highfreq, output_bgc_highfreq and output_sp_highfreq usermods directories - Note that there are NO tests covering the cmip6_evolving_icesheet usermods: this usermod directory adds very little beyond cmip6_deck, so it didn't seem worth its own test Code reviewed by: Erik Kluzek CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run unit-tests (components/clm/src): cheyenne - pass tools-tests (components/clm/test/tools): cheyenne - not run PTCLM testing (components/clm/tools/shared/PTCLM/test): cheyenne - not run regular tests (aux_clm): cheyenne_intel ---- pass cheyenne_gnu ------ pass hobart_nag -------- pass hobart_pgi -------- pass hobart_intel ------ pass ERI_D_Ld9.T31_g37.I2000Clm50Sp.cheyenne_intel.clm-SNICARFRC initially failed COMPARE_base_hybrid, COMPARE_base_rest and BASELINE comparisons. I reran it and it passed. Manually compared all history files from ERS_Ly3.f10_f10_musgs.I1850Clm50BgcCropCmip6.cheyenne_intel.clm-basic with baseline ERS_Ly3.f10_f10_musgs.I1850Clm50BgcCrop.cheyenne_intel.clm-cmip6. They were all identical. Also compared all history files from SMS_Ld5_D.f09_g17.I1850Clm50BgcCropCmip6.cheyenne_intel.clm-basic with baseline SMS_Ld5_D.f09_g17.I1850Clm50BgcCrop.cheyenne_intel.clm-cmip6. They were all identical (but note that this basically just compared the cpl.hi file: the test was too short to produce monthly files.) CTSM tag used for the baseline comparisons: ctsm1.0.dev014 Answer changes -------------- Changes answers relative to baseline: NO Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): - https://github.com/ESCOMP/ctsm/pull/536 - Allow holes in the number of history tapes and reorganize cmip6 usermods (main PR containing all of these changes) - https://github.com/ESCOMP/ctsm/pull/527 - Add carbon_isotope user-mod directory to turn on c13 and c14 - https://github.com/ESCOMP/ctsm/pull/539 - Support a %CMIP6DECK compset modifier =============================================================== =============================================================== Tag name: ctsm1.0.dev014 Originator(s): sacks (Bill Sacks) Date: Fri Oct 26 06:20:34 MDT 2018 One-line Summary: Miscellaneous minor, bit-for-bit bug fixes Purpose of changes ------------------ Four miscellaneous minor, bit-for-bit bug fixes: (1) Py3 pylint check and address cime issue ESMCI/cime#2822 (from Jim Edwards: ESCOMP/ctsm#526) (2) Change uppercase DEBUG variables to lowercase debug (requested by Jim Edwards to avoid conflicting with the DEBUG CPP token) (Fixes ESCOMP/ctsm#534) (3) Remove unnecessary line of code in LunaMod.F90 that was causing problems with some compilers due to an uninitialized variable (Fixes ESCOMP/ctsm#322) (4) Add r8 to 0 constant to fix build issue with XLF compiler (from Jim Edwards: ESCOMP/ctsm#531) Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Fixes ESCOMP/ctsm#322 (ERS_D_Ld5.f19_g16.I2000Clm50BgcCruGs run FAIL (intel)) - Fixes ESCOMP/ctsm#534 (Rename DEBUG to debug in a few places) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] clm4_5 [ ] clm4_0 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in ../CTSMMasterChecklist as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): Changes to tests or testing: none Code reviewed by: different pieces reviewed by different people CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run unit-tests (components/clm/src): cheyenne - pass tools-tests (components/clm/test/tools): cheyenne - not run PTCLM testing (components/clm/tools/shared/PTCLM/test): cheyenne - not run regular tests (aux_clm): cheyenne_intel ---- ok cheyenne_gnu ------ ok hobart_nag -------- ok hobart_pgi -------- ok hobart_intel ------ ok ok means tests and baseline comparisons pass. There were unexpected NLCOMP diffs. From spot-checking a few tests, I see the following: (1) On both cheyenne and hobart: Diffs in logfile. This looks like a problem in cime: it says: BASE: logfile = 'rof.log.136342.hobart.cgd.ucar.edu COMP: logfile = 'rof.log (2) On hobart: Diffs in pio_typename: netcdf vs. pnetcdf (says that new uses pnetcdf): but when I compare files by hand, they look the same in this respect (both baseline and new use netcdf), so maybe this is due to a problem in the timing of when nlcomp is run? Since these both look like problems in the comparison script rather than in the tag, I'm letting these go, but will open a cime issue if we see this again. CTSM tag used for the baseline comparisons: ctsm1.0.dev013 Answer changes -------------- Changes answers relative to baseline: NO Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): - https://github.com/ESCOMP/ctsm/pull/531 (fix build issue with xlf compiler) - https://github.com/ESCOMP/ctsm/pull/526 (fix cime issue 2822 and pylint chk) =============================================================== =============================================================== Tag name: ctsm1.0.dev013 Originator(s): erik (Erik Kluzek)/slevisconsulting Date: Thu Oct 25 18:09:47 MDT 2018 One-line Summary: Fix the fact that transient Bgc and SP cases had constant crop area in time Purpose of changes ------------------ Fix transient non-crop cases that had constant crop area so that crop area will change in time (and hence natural veg area will also change corresponding to it). Also bring in changes to update mksurfdata_map to handle SSP-RCP future scenarios. Right now it can handle SSP5-8.5 out of the box. Also add a new test for that. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): #538 #538 -- In transient pft simulations with use_crop=.false., %crop does not advance from 1850 values Known bugs found since the previous tag (include github issue ID): #545 #545 -- Antarctica ice shelves are being treated as wetlands rather than glaciers Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [X] clm5_0 [X] clm4_5 [ ] clm4_0 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): A few more namelist checks Added a check to make sure do_transient_pfts was the same as do_transient_crops Don't allow use_fertilizer for non-crop case namelist checks that create_crop_landunit=T except for FATES Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): New -ssp_rcp option to mksurfdata.pl Changes made to namelist defaults (e.g., changed parameter values): None Changes to the datasets (e.g., parameter, surface or initial files): New mksrfpft datasets in XML for SSP5-8.5 Substantial timing or memory changes: None Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in ../CTSMMasterChecklist as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): Now, do_transient_pfts and do_transient_crops only work in unison, and hence should be simplified to one logical called do_transient_veg Changes to tests or testing: Add new SSP5-8.5 mksurf test Code reviewed by: self, slevisconsulting, billsacks CTSM testing: regular, tools [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - PASS (18 of the transient test namelists are different to baseline as expected) unit-tests (components/clm/src): cheyenne - PASS tools-tests (components/clm/test/tools): cheyenne - OK (PTCLM tests fail) regular tests (aux_clm): cheyenne_intel ---- OK cheyenne_gnu ------ OK hobart_nag -------- OK hobart_pgi -------- OK hobart_intel ------ OK CTSM tag used for the baseline comparisons: ctsm1.0.dev012 Answer changes -------------- Changes answers relative to baseline: Yes! (Important) Summarize any changes to answers, i.e., - what code configurations: clm4_5/clm5_0 transient simulations without Crop (either Bgc or SP) - what platforms/compilers: All - nature of change: crop areas increase in time and impact simulations If this tag changes climate describe the run(s) done to evaluate the new climate (put details of the simulations in the experiment database) - casename: dll/clm50_r267_1deg_GSWP3V1_iso_hist_nocrop_transientfix Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): None Pull Requests that document the changes (include PR ids): #546, #540 (https://github.com/ESCOMP/ctsm/pull) #546 -- Add in ability for mksurfdata_map to handle ssp_rcp scenarios and specifically SSP5-8.5 $540 -- Transient PFT issue =============================================================== =============================================================== Tag name: ctsm1.0.dev012 Originator(s): erik (Erik Kluzek) Date: Sat Sep 29 11:49:35 MDT 2018 One-line Summary: Add snow-free fields for snowmip, fix several issues Purpose of changes ------------------ Bring in new diagnostic fields added by Justin Perket, Sean Swenson and Mark Flanner for Snow-MIP. Most of those are fields that represent "Snow Free" data. Also bring in fixes for a list of issues. Add handling of the new CO2 which includes both latitude-band and global average versions. Add some changes to make it easier for input data processing including NOT doing the slow 1km map file creation. Have the number of steps that are skipped at startup dependent on the time-step size. Add a test for some requirements of WACCMX (time-step and use of ESMF). Calculations of local time are now done in a global subroutine, that can handle negative longitudes. Fix how FFIX_TO_SMINN is handled for history output. The namelist logical "calc_human_stress_indices" changed from logical to a character string of three values: FAST, NONE, ALL. FAST is the default so the less expensive ones are output, NONE turns them all off, and ] ALL does all of them including the expensive ones. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): #428, #474, #475, #476, #450, #482, #481, #491 Fix #428 -- Update getco2_historical.ncl to handle latitude varying CO2 Fix #474 -- Add ability to send GRIDFILE to regridbatch.sh script Fix #475 -- Have number of steps to skip balance-check based on time Fix #476 -- Add a test for WACCMX standalone Fix #450 -- Add option to use global average of terrain standard deviation on surfdata files (partial fix with simplest option) Fix #482 -- Add extra field on CO2 streams file for global/time-averaged data Fix #481 -- FFIX_TO_SMINN needs to be output when FUN is on Fix #491 -- Calculations of local noon assume that longitude is 0 to 360 rather than -180 to 180 Known bugs introduced in this tag (include github issue ID): cime#2801 cime#2801 -- Problem building with ESMF_LIB Known bugs found since the previous tag (include github issue ID): #507, #505 #507 -- Albedo's are bad at night with negative longitudes #505 -- CTSM input data-set tools can only work on 0-360 grids, and require monotone increasing longitude Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] clm4_5 [ ] clm4_0 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): None Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): New namelist: use_ssre -- Turn on show free fields needed for SnowMIP Changed namelist: calc_human_stress_indices changed from logical to character with options: ALL, FAST, NONE New history fields: Mostly added Snow Free (SF) fields ALBDSF ALBISF FSRSF FSRSFND FSRSFNDLN FSRSFNI FSRSFVD FSRSFVDLN FSRSFVI SSRE_FSR SSRE_FSRND SSRE_FSRNDLN SSRE_FSRNI SSRE_FSRVD SSRE_FSRVDLN SSRE_FSRVI Changes made to namelist defaults (e.g., changed parameter values): calc_human_stress_indices = 'FAST' is now the default Changes to the datasets (e.g., parameter, surface or initial files): mkghg_bndtvghg -- Update with new CO2 files, both monthly, lat-bands and yearly, global Substantial timing or memory changes: None Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in .CTSMTrunkChecklist as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): I was able to reduce the duplication in SurfaceAlbedoMod where the original implementation added a cut and paste copy of code. But, there is still a lot of duplication in this file that could be improved, by making smaller functions/subroutines to do sections of code that are essentially repeated many times. There's a bit of an increase in complexity to reduce the duplication, but reducing the duplication was worth it. Changes to tests or testing: Add a new waccmx_offline test mods and test with it New test expected fail because of cime issue: ERS_D_Ln9_P480x3.f19_g16.I2000Clm50SpGs.cheyenne_intel.clm-waccmx_offline Turn use_ssre on for most tests, off for reducedOutput And set calc_human_stress_indices=NONE for reducedOutput, FAST for most, and ALL for KitchenSink and KSMOut tests Code reviewed by: self, @olyson, @billsacks CTSM testing: regular [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - PASS unit-tests (components/clm/src): cheyenne - PASS regular tests (aux_clm): cheyenne_intel ---- OK cheyenne_gnu ------ OK hobart_nag -------- OK hobart_pgi -------- OK hobart_intel ------ OK CTSM tag used for the baseline comparisons: ctsm1.0.dev011 Answer changes -------------- Changes answers relative to baseline: no bit-for-bit Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): No Pull Requests that document the changes (include PR ids): #462 #449 (https://github.com/ESCOMP/ctsm/pull) #462 -- Add namelist item to calculate all heat stress indices only if requested; to speed up model #449 -- Bring in snowmip diagnostic fields =============================================================== =============================================================== Tag name: ctsm1.0.dev011 Originator(s): sacks (Bill Sacks), mvr (Mathew Rothstein) Date: Wed Sep 12 10:50:31 MDT 2018 One-line Summary: Add water tracer consistency checks, and other water tracer work Purpose of changes ------------------ 1. Add water tracer consistency checks 2. Add infrastructure for looping over all water tracers - currently just used for the tracer consistency checks 3. Breakout of atm2lnd and lnd2atm water variables, needed for water tracers 4. Add some namelist control over the addition of water tracers 5. Add a system test that exercises the water tracer consistency checks 6. Add a 'ratio' variable for each water tracer 7. Add some unit tests of the new water tracer infrastructure Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Partially addresses #357 - Resolves #479 - Resolves #492 Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] clm4_5 [ ] clm4_0 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): - Don't be fooled by the new namelist variable, enable_water_isotopes: This is just a place-holder for now, not implying that water isotopes are actually working. Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): - New namelist variables: enable_water_tracer_consistency_checks and enable_water_isotopes. The latter is just a place-holder for now. Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in .CTSMTrunkChecklist as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: New test that runs the water tracer consistency check I ran this test on cheyenne_gnu and cheyenne_intel along with the out-of-the-box hobart_nag version Code reviewed by: Mat Rothstein and I have worked together on many of these changes, but not all code has been reviewed by the other. CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - ok (tests pass, clm4_5 and clm5_0 namelists differ from baseline as expected) unit-tests (components/clm/src): cheyenne - pass tools-tests (components/clm/test/tools): cheyenne - not run PTCLM testing (components/clm/tools/shared/PTCLM/test): cheyenne - not run regular tests (aux_clm): cheyenne_intel ---- pass cheyenne_gnu ------ pass hobart_nag -------- pass hobart_pgi -------- pass hobart_intel ------ pass CTSM tag used for the baseline comparisons: ctsm1.0.dev010 Answer changes -------------- Changes answers relative to baseline: NO Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): - https://github.com/ESCOMP/ctsm/pull/497 =============================================================== =============================================================== Tag name: ctsm1.0.dev010 Originator(s): sacks (Bill Sacks) Date: Thu Aug 30 17:14:28 MDT 2018 One-line Summary: Update cime to cime5.7.3 Purpose of changes ------------------ Update cime from cime5.6.10 to cime5.7.3. To support this change, there are also minor code changes related to the pause-resume implementation (from Erik Kluzek). Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - ESCOMP/ctsm#384 (VIC test is failing at f09 resolution with signal) (I'm not sure what fixed this, but it's passing now) CIME Issues fixed (include issue #): many Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] clm4_5 [ ] clm4_0 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in .CTSMTrunkChecklist as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: self CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run unit-tests (components/clm/src): cheyenne - pass tools-tests (components/clm/test/tools): cheyenne - not run PTCLM testing (components/clm/tools/shared/PTCLM/test): cheyenne - not run regular tests (aux_clm): cheyenne_intel ---- pass cheyenne_gnu ------ pass hobart_nag -------- pass hobart_pgi -------- pass hobart_intel ------ pass CTSM tag used for the baseline comparisons: ctsm1.0.dev009 Answer changes -------------- Changes answers relative to baseline: NO - bit-for-bit Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): cime: cime5.6.10 -> cime5.7.3 Pull Requests that document the changes (include PR ids): none =============================================================== =============================================================== Tag name: ctsm1.0.dev009 Originator(s): sacks (Bill Sacks) Date: Wed Aug 22 20:32:36 MDT 2018 One-line Summary: Fix initialization of AnnET in InitAccVars Purpose of changes ------------------ InitAccVars was mistakenly setting qflx_evap_tot_col rather than AnnET. This fix allows us to remove now-redundant cold start and restart code for AnnET. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Fixes #480 - Partially addresses #285 Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] clm4_5 [ ] clm4_0 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in .CTSMTrunkChecklist as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: basic proposed changes reviewed by Erik Kluzek CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run unit-tests (components/clm/src): cheyenne - pass tools-tests (components/clm/test/tools): cheyenne - not run PTCLM testing (components/clm/tools/shared/PTCLM/test): cheyenne - not run regular tests (aux_clm): cheyenne_intel ---- pass cheyenne_gnu ------ pass hobart_nag -------- pass hobart_pgi -------- pass hobart_intel ------ pass CTSM tag used for the baseline comparisons: ctsm1.0.dev008 Answer changes -------------- Changes answers relative to baseline: NO Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): none =============================================================== =============================================================== Tag name: ctsm1.0.dev008 Originator(s): erik (Erik Kluzek) Date: Tue Aug 14 10:25:12 MDT 2018 One-line Summary: Update 1850 ndep file and last year for streams for Historical transient cases Purpose of changes ------------------ Bring in changes from release-clm5.0.05. Update to latest Nitrogen Deposition file from simulations with WACCM for 1850. Also fix an issue with the last year for historical transient cases. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): 461 #461 -- increase last year in streams for transient Known bugs found since the previous tag (include github issue ID): [If none, remove this line] #478 -- Bare soil g1 should be missing value or zero Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] clm4_5 [ ] clm4_0 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): None Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): None Changes made to namelist defaults (e.g., changed parameter values): Last year extended for transient datasets Changes to the datasets (e.g., parameter, surface or initial files): New ndep dataset for 1850 Substantial timing or memory changes: Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in .CTSMTrunkChecklist as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): None Changes to tests or testing: Lengthen some tests Code reviewed by: self CTSM testing: regular [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - PASS (11 show differences for 1850_control and 20thC_transient) unit-tests (components/clm/src): cheyenne - PASS regular tests (aux_clm): cheyenne_intel ---- OK cheyenne_gnu ------ OK hobart_nag -------- OK hobart_pgi -------- OK hobart_intel ------ OK CTSM tag used for the baseline comparisons: ctsm1.0.dev007 Answer changes -------------- Changes answers relative to baseline: Yes! (ndep files, and steams years) Summarize any changes to answers, i.e., - what code configurations: 1850_control or 20thC_transient for Clm50 - what platforms/compilers: all - nature of change: similar climate Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): None Pull Requests that document the changes (include PR ids): (https://github.com/ESCOMP/ctsm/pull) #477 -- Move changes from release-clm5.0.05 onto master =============================================================== =============================================================== Tag name: ctsm1.0.dev007 Originator(s): sacks (Bill Sacks) Date: Sun Aug 5 21:03:28 MDT 2018 One-line Summary: Avoid glacier dynamic landunit adjustments in first time step Purpose of changes ------------------ Always avoid generating dynamic landunit adjustments for glacier area changes in the first timestep of a startup or hybrid run - not just for cold start or interpolated start. See the detailed discussion in https://github.com/ESCOMP/ctsm/issues/340 for rationale. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Fixes #340 (Avoid generating dynamic landunit adjustment fluxes for glacier changes in the first timestep) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] clm4_5 [ ] clm4_0 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in .CTSMTrunkChecklist as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: none Code reviewed by: Erik Kluzek CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run unit-tests (components/clm/src): cheyenne - not run tools-tests (components/clm/test/tools): cheyenne - not run PTCLM testing (components/clm/tools/shared/PTCLM/test): cheyenne - not run regular tests (aux_clm): cheyenne_intel ---- ok cheyenne_gnu ------ pass hobart_nag -------- pass hobart_pgi -------- pass hobart_intel ------ pass ok means tests pass; baseline failures as expected in these two tests that do not use init_interp: FAIL ERI_N2_Ld9.f19_g17.I2000Clm50BgcCrop.cheyenne_intel.clm-default BASELINE ctsm1.0.dev006 FAIL SMS_Lm13.f19_g17.I2000Clm50BgcCrop.cheyenne_intel.clm-cropMonthOutput BASELINE ctsm1.0.dev006 (The glacier area on the finidat file used in these tests - clmi.I2000Clm50BgcCrop.2011-01-01.1.9x2.5_gx1v7_gl4_simyr2000_c180715.nc - seems to have come from an finidat_interp_dest file, and thus does not match the glacier area from CISM.) CTSM tag used for the baseline comparisons: ctsm1.0.dev006 Answer changes -------------- Changes answers relative to baseline: YES (small) Summarize any changes to answers, i.e., - what code configurations: Startup / Hybrid runs that include CISM (even NOEVOLVE) and do not use init_interp, due to: - different PE layout in new case vs. the one that generated the finidat (roundoff-level diffs) - different glacier area on finidat file vs. what's in CISM (greater than roundoff-level diffs) - what platforms/compilers: all - nature of change (roundoff; larger than roundoff/same climate; new climate): Roundoff-level for different PE layout, larger than roundoff but should be same climate when there is different glacier area on the finidat file vs. what's in CISM The differences arise from dynamic landunit adjustments of the below-ground C and N states in the old code, as CTSM adjusts its areas to match CISM's. In the new code, there are no C and N adjustments in the first time step from this adjustment. Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): - #470 (Avoid glacier dynamic landunit adjustments in first time step) =============================================================== =============================================================== Tag name: ctsm1.0.dev006 Originator(s): sacks (Bill Sacks) Date: Sat Aug 4 07:48:09 MDT 2018 One-line Summary: Minor bug fixes, cleanup, documentation and enhancements Purpose of changes ------------------ A collection of minor bug fixes, code cleanup, documentation and enhancements, all bit-for-bit. See list of issues fixed below for details. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Fixes #24 (ncd_io_1d_log_glob is broken) - Fixes #120 (Incorrect comments in Biogeophysics1Mod.F90 - Fixes #217 (Change some cheyenne_gnu tests to cheyenne_intel) - Fixes #245 (Put all .gitignore entries in top-level file) - Fixes #272 (Code should error on missing mxsoil_color when SOIL_COLOR is used) - Fixes #283 (Add more helpful message about need to do init_interp with wrong number of vertical layers) - Fixes #367 (For cmip6 runs: Turn on cpl hist output needed to drive a TG compset) - Fixes #412 (Fix documentation of init_interp_method) - Fixes #419 (Do not allow SOYFIXN diagnostic field with FUN) - Fixes #464 (Some lines longer than 132 characters) - Fixes #465 (Remove backwards compatibility check for snw_rds) - Fixes #467 (Increase wallclock limit for test) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] clm4_5 [ ] clm4_0 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in .CTSMTrunkChecklist as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): none Changes to tests or testing: - Changed a single-point test from gnu to intel to get better debugging information (SMS_D_Ly6_Mmpi-serial.1x1_smallvilleIA.IHistClm45BgcCropQianGs.cheyenne_intel.clm-cropMonthOutput) - For some single-point tests on both gnu and intel: removed one compiler, keeping the tests on just one Code reviewed by: self CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run unit-tests (components/clm/src): cheyenne - pass tools-tests (components/clm/test/tools): cheyenne - not run PTCLM testing (components/clm/tools/shared/PTCLM/test): cheyenne - not run regular tests (aux_clm): cheyenne_intel ---- pass cheyenne_gnu ------ pass hobart_nag -------- pass hobart_pgi -------- pass hobart_intel ------ pass NLCOMP failures as expected for tests with the cmip6 test mod Some long single-point gnu tests failed due to running out of wallclock time; I have removed these, just keeping them on intel (along with removing a few other tests after I ran the testing - see commit 5c226bdf): FAIL ERS_Lm54_Mmpi-serial.1x1_numaIA.I2000Clm50BgcCropGs.cheyenne_gnu.clm-cropMonthOutput RUN time=4803 FAIL ERS_Ly6_Mmpi-serial.1x1_smallvilleIA.IHistClm50BgcCropGs.cheyenne_gnu.clm-cropMonthOutput RUN time=6024 CTSM tag used for the baseline comparisons: ctsm1.0.dev005 Answer changes -------------- Changes answers relative to baseline: NO - bit-for-bit Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): - #392 (Remove old code for snw_rds restart that looks no longer needed) - (Other changes not documented in PRs) =============================================================== =============================================================== Tag name: ctsm1.0.dev005 Originator(s): sacks (Bill Sacks), mvr (Mathew Rothstein) Date: Fri Aug 3 07:54:59 MDT 2018 One-line Summary: Rework water data types to accommodate isotopes and other tracers Purpose of changes ------------------ This tag reworks the various water data types to allow having multiple instances of variables that are needed for isotopes and other water tracers. Specific changes include: (1) Separated "water state" variables into state, diagnostic and balance check-related variables. This separation was not essential for the work here, but was desired by Martyn Clark and others. (2) For each of water state, diagnostic and flux variables, separated variables into those needed for both bulk and tracers vs. those only needed for bulk. This way, we can have multiple instances of the variables needed by tracers, but only a single instance of variables that only apply to bulk water. This follows the design laid out in https://github.com/escomp/ctsm/pull/395. The separation was based largely on what was done in the old water isotope branch; we didn't put a lot of thought into this, because the new design allows us to easily migrate variables between bulk-only and bulk-and-tracer as needed. (3) Moved water fluxes that were defined in science modules back into waterflux_type or waterfluxbulk_type. This was needed for (2); there is more discussion on this in https://github.com/escomp/ctsm/pull/395 and the log message for commit 711e5cd7. (4) Introduced a top-level water_type that holds instances of all of the other water-related objects. This follows the design laid out in https://github.com/escomp/ctsm/pull/395. This is particularly valuable for the tracer instances: the logic related to number of tracers can be encapsulated in water_type, rather than infiltrating clm_instMod. (5) Added placeholders for water tracer instances (6) Added infrastructure to generate history / restart field names for the tracer instances. Eventually, the isotope class can also hold information specific to each isotope. This work was a joint effort between Mathew Rothstein and myself; Mat gets much of the credit for the actual refactoring done here. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Fixes #358 (Separate WaterStateType into multiple types) - Fixes #434 (Separate WaterFluxType into a base class and a class that just applies to bulk) - Fixes #359 (Set up infrastructure for multiple instances of WaterState and WaterFlux types) - Fixes #458 (Implement handling of history and restart variables for water tracers) Known bugs introduced in this tag (include github issue ID): - #464 (Some lines longer than 132 characters) - #456 (Remove backwards compatibility check for snw_rds) (not a newly-introduced issue, but it's more important to fix this with the separation of waterstate_type) Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] clm4_5 [ ] clm4_0 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): none Changes made to namelist defaults (e.g., changed parameter values): none Changes to the datasets (e.g., parameter, surface or initial files): none Substantial timing or memory changes: none Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in .CTSMTrunkChecklist as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): - Water tracers are not yet functional - they are just placeholders - Whether to use water tracers and which water tracers to use are currently hard-coded; eventually, this should be namelist-controlled Changes to tests or testing: none Code reviewed by: Mathew Rothstein and Bill Sacks worked together on these changes and reviewed each other's changes to some extent. The high-level design was also reviewed and co-developed by Mariana Vertenstein. Martyn Clark contributed substantially to the breakout of water state into state, diagnostic and balance check-related variables. CTSM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run unit-tests (components/clm/src): cheyenne - pass tools-tests (components/clm/test/tools): cheyenne - not run PTCLM testing (components/clm/tools/shared/PTCLM/test): cheyenne - not run regular tests (aux_clm): cheyenne_intel ---- pass cheyenne_gnu ------ pass hobart_nag -------- pass hobart_pgi -------- pass hobart_intel ------ pass CTSM tag used for the baseline comparisons: ctsm1.0.dev004 Answer changes -------------- Changes answers relative to baseline: NO - bit-for-bit Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): - https://github.com/escomp/ctsm/pull/395 (documents the high-level design) =============================================================== =============================================================== Tag name: ctsm1.0.dev004 Originator(s): erik (Erik Kluzek,UCAR/TSS,303-497-1326) Date: Wed Jul 18 02:10:23 MDT 2018 One-line Summary: Add some new diagnostic fields, fix a few issues, update cmip6 output Purpose of changes ------------------ Add some new diagnostic fields. Some needed for CMIP6. Update the CMIP6 user-mods output. Fix a couple issues. Get full list of history tapes working correctly. Check for valid range of CO2. New IC file interpolated from the previous one for f19_g17_gl4 for 2000 Clm50BgcCrop. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): $210, #427, #429 #210 -- increase number of history tapes $427 -- Check for zero CO2 #429 -- New IC for present day Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? No (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] clm4_5 [ ] clm4_0 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): none Changes to CTSM's user interface (e.g., new/renamed XML or namelist variables): New history fields TSKIN --- skin temperature TSL ----- temperature of near-surface soil layer GSSUNLN - sunlit leaf stomatal conductance at local noon GSSHALN - shaded leaf stomatal conductance at local noon Changes made to namelist defaults (e.g., changed parameter values): None Changes to the datasets (e.g., parameter, surface or initial files): New finidat file f19_g17_gl4 2000 Clm50BgcCrop New file is just interpolate from the previous file to a f19_g17_gl4 gridfor Clm50BgcCrop and 2000 simulation year Substantial timing or memory changes: None Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in .CLMTrunkChecklist as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): Changes to tests or testing: None Code reviewed by: self, oleson, abtawfik CTSM testing: regular [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - OK (95 are different because of finidat file change) unit-tests (components/clm/src): cheyenne - PASS regular tests (aux_clm): cheyenne_intel ---- OK cheyenne_gnu ------ OK hobart_nag -------- PASS hobart_pgi -------- FAIL hobart_intel ------ PASS CTSM tag used for the baseline comparisons: ctsm1.0.dev003 Answer changes -------------- Changes answers relative to baseline: Yes -- only 2000 cases because of new IC Summarize any changes to answers, i.e., - what code configurations: Just 2000 cases - what platforms/compilers: All - nature of change: same climate Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): None Pull Requests that document the changes (include PR ids): (https://github.com/ESCOMP/ctsm/pull) #338 -- Radtemp and other new diagnostic fields =============================================================== =============================================================== Tag name: ctsm1.0.dev003 Originator(s): erik (Erik Kluzek,UCAR/TSS,303-497-1326) Date: Sun Jul 15 00:26:46 MDT 2018 One-line Summary: Update cime/cism to work on upgraded hobart and with glade changes on cheyenne Purpose of changes ------------------ Update cime and cism so can work on upgraded hobart. Also fix an issue that the Nag6.2 compiler found in CTSM. Also get working with glade changes that happened on cheyenne, DIN_LOC_ROOT_CLMFORC was changed to a new directory. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): [If none, remove this line] #441 -- Changes in glade invalidate previous softlinks and data locations #438 -- Illegal argument aliasing caught by nag6.2 #433 -- with hobart upgrade CTSM not working Known bugs introduced in this tag (include github issue ID): #444 -- PGI fails on cheyenne and hobart Known bugs found since the previous tag (include github issue ID): #443 -- Wood C:N ratios #440 -- leafcn_max < target leafcn? Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? No (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [ ] clm5_0 [ ] clm4_5 [ ] clm4_0 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): None Changes to CLM's user interface (e.g., new/renamed XML or namelist variables): DIN_LOC_ROOT_CLMFORC on cheyenne DIN_LOC_ROOT_CLMFORC was change on cheyenne Changes made to namelist defaults (e.g., changed parameter values): None Changes to the datasets (e.g., parameter, surface or initial files): None Substantial timing or memory changes: None Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in .CTSMTrunkChecklist as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): None Changes to tests or testing: path for cheyenne specific test had to be changed Add some gnu tests on hobart Lengthen wallclock on a test Fix path for DA test_mod Code reviewed by: self CLM testing: regular [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - PASS unit-tests (components/clm/src): cheyenne - PASS tools-tests (components/clm/test/tools): cheyenne - PASS PTCLM testing (components/clm/tools/shared/PTCLM/test): cheyenne - OK regular tests (aux_clm): cheyenne_intel ---- OK cheyenne_gnu ------ OK hobart_nag -------- OK hobart_pgi -------- FAIL hobart_gnu -------- PASS hobart_intel ------ PASS CLM tag used for the baseline comparisons: ctsm1.0.dev002 Answer changes -------------- Changes answers relative to baseline: No bit-for-bit Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): cime, cism cism-wrapper to release-cesm2.0.04 cime to cime5.6.10 Pull Requests that document the changes (include PR ids): (https://github.com/ESCOMP/ctsm/pull) #444 -- Get working on upgraded hobart and with glade changes on cheyenne =============================================================== =============================================================== Tag name: ctsm1.0.dev002 Originator(s): erik (Erik Kluzek) Date: Fri Jul 6 16:33:03 MDT 2018 One-line Summary: Fix NFIX flux variables so special land-units are zeroed out, tools update, add some *_MAX fields on mksurfdata_map for transient cases Purpose of changes ------------------ Nitrogen Fixation flux arrays were being set to missing value over non-vegetated landunits. This sets them to zero everywhere and averages in the zero's at the gridcell level for history output. It also reads in restarts with missing values and converts them to zero's. There are also some tools updates, getting the tools working on the new upgraded hobart.cgd.ucar.edu. And mksurfdata_map is updated to add some *_MAX files on the landuse.timeseries files that will allow us to conserve memory for transient cases. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): #435, #436, #433 (for tools) #435 -- intel build for tools #426 -- Nitrogen Fixation flux variables #433 -- hobart broken, got it working for tools Known bugs found since the previous tag (include github issue ID): #433, #431 #433 -- with hobart upgrade, hobart build isn't working #431 -- When set_* options are used to hist_addfld be careful of ordering Significant changes to scientifically-supported configurations -------------------------------------------------------------- Does this tag change answers significantly for any of the following physics configurations? (Details of any changes will be given in the "Answer changes" section below.) [Put an [X] in the box for any configuration with significant answer changes.] [X] clm5_0 [ ] clm4_5 [ ] clm4_0 Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): None Changes to CLM's user interface (e.g., new/renamed XML or namelist variables): None Changes made to namelist defaults (e.g., changed parameter values): ndep file updated for clm5_0 for 1850 Changes to the datasets (e.g., parameter, surface or initial files): None Substantial timing or memory changes: None Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in .CLMTrunkChecklist as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): NFIX flux variables NFIX flux variables when read from restart have special-values (missing on non veg landunits) converted to zero's They are now initialized to zero everywhere. Changes to tests or testing: None Code reviewed by: self, wwieder, and lawrence, sacks (mksurfdata_map code) CLM testing: regular (although none on hobart!) [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - PASS (5 tests differ from previous version) unit-tests (components/clm/src): cheyenne - PASS tools-tests (components/clm/test/tools): cheyenne - PASS hobart --- PASS regular tests (aux_clm): cheyenne_intel ---- OK cheyenne_gnu ------ OK CLM tag used for the baseline comparisons: ctsm1.0.dev001 Answer changes -------------- Changes answers relative to baseline: Yes, clm5_0 BGC configurations with FUN and clm5_0 with BGC for new 1850 ndep file Summarize any changes to answers, i.e., - what code configurations: All clm5_0 BGC due to NFIX fields, and clm5_0 1850 BGC due to new ndep file - what platforms/compilers: All - nature of change: similar climate Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): None Pull Requests that document the changes (include PR ids): (https://github.com/ESCOMP/ctsm/pull) #437 -- ChangeLog template #432 -- nfix #331 -- _MAX fields on mksurfdata_map =============================================================== =============================================================== Tag name: ctsm1.0.dev001 Originator(s): sacks (Bill Sacks) Date: Fri Jun 22 07:31:43 MDT 2018 One-line Summary: Begin separating SoilHydrology flux calculations Note the new tag naming: Starting with this tag, we are naming tags "ctsm..." rather than "clm...". We are starting with ctsm version 1, which for now is nearly the same as clm version 5. We are moving to this new tag naming now because: (1) The changes in this tag represent the first step towards implementing the CTSM vision (separating the biogeophysics flux calculations from state updates, making it easier to plug in alternative parameterizations, etc.). (2) This tag changes answers relative to the CLM5 release. (We expect the climate to be the same, but we haven't tested this carefully yet.) Purpose of changes ------------------ First steps toward separating various flux calculations in the soil hydrology code. The focus here is on saturated surface runoff and infiltration excess runoff. The changes here separate flux calculations from state updates and extract various calculations into their own subroutines to facilitate swapping in alternative parameterizations. Most of the changes here are refactorings that are either bit-for-bit or just introduce roundoff-level differences. However, there are also a few larger answer changes, as described below. These are the greater-than-roundoff-level answer changes: (A1) Use full qflx_surf in BGC code. Previously, the subroutines ch4 and SoilBiogeochemNLeaching had been using a flux that excluded surface water runoff (qflx_h2osfc_surf). That was deemed to be incorrect (by Dave Lawrence and others), so these BGC subroutines have been changed to use the full surface runoff (saturated excess runoff + infiltration excess runoff + h2osfc runoff). Configurations affected: BGC compsets Magnitude of change: Larger than roundoff; expected to be same climate, but not investigated carefully (A2) VIC: Don't ever use the TOPModel formulation for SurfaceRunoff The code was using the TOPModel-based formulation for SurfaceRunoff if frost_table > zwt_perched. Sean Swenson felt this shouldn't be done, and refactoring will be easier if VIC always uses the vic-looking formulation, so I'm stopping applying this formulation when using VIC. Configurations affected: VIC compsets Magnitude of change: Larger than roundoff; not investigated carefully (A3) VIC: Remove infiltration excess runoff Martyn Clark reviewed the VIC implementation, and felt that the current implementation of infiltration excess runoff is inconsistent with the standard VIC implementation. It appears that what was being called VIC's infiltration excess runoff was actually just an attempt to give a better numerical approximation to the solution for saturated surface excess runoff. So deleting this leaves only a first-order approximation to VIC's saturated surface excess runoff. Eventually we may want to put in place a more accurate solution for VIC's saturated surface excess runoff. But Martyn's feeling is that this can come in with other changes we want to make regarding numerical solutions in CTSM. Configurations affected: VIC compsets Magnitude of change: Larger than roundoff; expected to be same climate, but not investigated carefully (A4) Change in QOVER diagnostic field: now includes QH2OSFC. These are the major refactorings (either bit-for-bit or just roundoff-level differences): (R1) Extract surface runoff to its own module, and other modularization related to what used to be subroutine SurfaceRunoff: extract a subroutine for each fsat method, move urban surface runoff code into different routines. (R2) Extract infiltration excess runoff to its own module, and other modularization related to what used to be subroutine Infiltration. NOTE: For a detailed breakdown of changes, including documentation of which changed changed answers, see the file ChangeLog_branch which was deleted from this branch shortly before it was merged to master. Bugs fixed or introduced ------------------------ Issues fixed (include CTSM Issue #): - Fixes #424 (decStart testmod is missing an `--append` in an xmlchange command) Notes of particular relevance for users --------------------------------------- Caveats for users (e.g., need to interpolate initial conditions): Changes to CLM's user interface (e.g., new/renamed XML or namelist variables): Changes made to namelist defaults (e.g., changed parameter values): Changes to the datasets (e.g., parameter, surface or initial files): Substantial timing or memory changes: Notes of particular relevance for developers: (including Code reviews and testing) --------------------------------------------- NOTE: Be sure to review the steps in .CLMTrunkChecklist as well as the coding style in the Developers Guide Caveats for developers (e.g., code that is duplicated that requires double maintenance): Changes to tests or testing: - Shortened a 25-month f19 test to 13-months (SMS_Lm13.f19_g17.I2000Clm50BgcCrop.cheyenne_intel.clm-cropMonthOutput): This test took a long time to run, and we already have long tests of this configuration at coarse-resolution. I suggested removing this test entirely, but Erik Kluzek wanted to maintain a shortened version of it. Erik's comments were: How about either flip it to f45 (because Rosie is using that resolution for science) -- or leave it at f19 and have it run for 9 months? The 25 months was to take it just past the 2-year spinup time. From spunup initial conditions 9-months pretty much does a full harvest cycle. f19 and f09 are the main configurations for science and it would be good to have something longer than just really short tests. I decided on Lm13, because this still completes in significantly less than 2-hours, and if we're running for 9 months anyway, it seems valuable to finish out the year and run a bit into the next year, since some crop stuff triggers at the end of the year. Code reviewed by: Various code reviews done last Fall, when these changes were first implemented. CLM testing: [PASS means all tests PASS and OK means tests PASS other than expected fails.] build-namelist tests: cheyenne - not run unit-tests (components/clm/src): cheyenne - PASS tools-tests (components/clm/test/tools): cheyenne - not run PTCLM testing (components/clm/tools/shared/PTCLM/test): cheyenne - not run regular tests (aux_clm): cheyenne_intel ---- OK cheyenne_gnu ------ OK hobart_nag -------- OK hobart_pgi -------- OK hobart_intel ------ OK OK means tests pass, answers change as expected. CLM tag used for the baseline comparisons: clm5.0.dev013 Answer changes -------------- Changes answers relative to baseline: YES (roundoff) Summarize any changes to answers, i.e., - what code configurations: Essentially all (except CLM4) - what platforms/compilers: All - nature of change (roundoff; larger than roundoff/same climate; new climate): Roundoff-level changes for nearly all configurations (except CLM4). Greater-than-roundoff-level changes for the following configurations: - BGC: see (A1) note above (expected to be same climate, but not investigated carefully) - VIC: see (A2) and (A3) notes above (magnitude of change not investigated carefully) - QOVER diagnostic field: see (A4) note above Detailed list of changes ------------------------ List any externals directories updated (cime, rtm, mosart, cism, fates, etc.): none Pull Requests that document the changes (include PR ids): - https://github.com/ESCOMP/ctsm/pull/244 - https://github.com/ESCOMP/ctsm/pull/190 - Other PRs that existed on the old NCAR/clm-ctsm repository (which has been deleted), which were not migrated to the new repository ===============================================================