Add multi-year test with mid-year restart

[samsrabin]

I added a bunch of per-harvest and annual output variables as part of #1863, and @billsacks suggested:

We should add one or more tests that exercise the new options and output variables. I think it would be good to have a multi-year test with a mid-year restart to ensure that all of the necessary variables are being saved to the restart file.

Maybe something like ERS_Lm54.f10_f10_mg37.IHistClm51BgcCrop.cheyenne_intel.clm-cropMonthOutput?

[billsacks]

Thanks for pursuing this, @samsrabin . I think you have the best sense of what the most appropriate test would be, including the necessary length, but this seems like it's probably good. A couple of comments, though:

• Using a Hist compset makes sense if transient capability is important to test with this; otherwise, if the focus is just on making sure that a mid-year restart works for some of this crop logic, a present day (I2000) compset could be better since there will be more crop areas present in that one than in an IHist test.
• For a test of this length, we should explicitly specify a PE layout, like ERS_Lm54_P144x1, so that we use more than the default number of processors for this resolution, which will give us faster turnaround.