finidat and run-time error in clm45_Me

[rp336@cornell_edu]

Hi!I am trying to run clm offline (0.9x1.25 resolution and I_2000_CLM45_CN_4Me compset) using clm4_0_60. But I have got the “dimension mismatch error”. It may be likely due to old spun-up initial file (clmi.BCN.2000-01-01_0.9x1.25_gx1v6_simyr2000_c100303.nc) that were used for CLM4.As suggested by Sam, I set up finitdat = ' ' and ran the clm for a short period. Then, I used 'interpinic' to interpolate from the 0.9x1.25 native resolution file. However, there are many variables that were not matched (for example, soil1c, soil2c, soil1n, soil2n ..... are soil1c_vr, soil2c_vr, soil1n_vr and so on). Interpinic could not write for many variables because of differing variable names when I tried to get mapped from a spun-up file. Does anyone have a spun-up file for 1.9x2.5 or 0.9x1.25 resolution that can be used for I_2000_CLM45_CN_4Me compset? -Rajendra

 

[slevis]

If you have access to yellowstone/geyser, you might want to try a restart file that I'm using:
/glade/p/work/slevis/clm4_0_60/models/lnd/clm/tools/clm4_5/interpinic/out/clm45_4060_cruncepfillqian.clm2.r.fromclm45_4060_20th.1886.nc
for this case:
/glade/u/home/slevis/clm45_4060_cruncepfillqian

This case will run 1850-2010 at 1.9x2.5 horizontal resolution, so this restart file is appropriate for starting at 1850 and probably not at 2000. Maybe with interpinic you could fix that though.

Sam Levis

 

[slevis]

Hi Sam,

Thanks a lot for the swift response of my question in CGD Forum. I have two quick questions
1)Is the restart file is fully spun-up for CN? (/glade/p/work/slevis/clm4_0_60/models/lnd/clm/tools/clm4_5/interpinic/out/clm45_4060_cruncepfillqian.clm2.r.fromclm45_4060_20th.1886.nc) It worked perfectly to start for 1850 run.
2)I interpolated your output file (/glade/scratch/slevis/clm45sci12clm4054ctrl/run/clm45sci12clm4054ctrl.clm2.h0.0072-12.nc) from 360x720_cruncep_360x720_cruncep resolution to 1.9x2.5 by using interpinic. It looks like you have fully spun-up the CN. It works fine for clm45_4060. I am just wondering if I had lose any critical information by interpolating from such a coarse to fine resolution.

Once again, thank you.

Best,
Rajendra


Dear Rajendra,

1) Not necessarily "fully" spun-up because I used interpinic to put it together from a spun-up 1/2-degree simulation.

2) Interpinic usually works well. It uses nearest neighbor mapping, so you will likely lose information in some cases.

Sam