The main point of this testmod is to exercise the case where nlevgrnd < nlevurb (connected with https://github.com/ESCOMP/CTSM/issues/674). If/when we have an out-of-the-box configuration with nlevgrnd < nlevurb (such as the NWP configuration once nlevurb = 10 is the standard) then we can remove this test. A secondary consideration is having a test that interpolates from a configuration with nlevgrnd > nlevurb onto one with lower nlevgrnd and nlevgrnd < nlevurb. But I don't think it's critical to maintain this test just for that purpose - e.g., if we start having an out-of-the-box configuration with nlevgrnd < nlevurb but where the finidat file also has nlevgrnd < nlevurb, then I feel it will be okay to remove this test.