Inquiry about MOSART 8th/100 degree resolution input data

[kabirtam@msu_edu]

Hi all,
I am trying to do a regional simulation using CLM5.0 at 0.05-degree spatial resolution. The extent of the area is 90 to 110 degrees east and 8 to 35 degrees north (Mekong River Basin). I subset MOSART 8th/100 degree resolution input data from global file /glade/p/cesmdata/cseg/inputdata/rof/mosart/MOSART_Global_8th_20191007.nc to get my regional mosart input file. I updated ID and dnID accordingly. I created all necessary files (domain, mapping from/to lnd to rof and rof to lnd) and managed to run the model.

Generated runoff from the land model seemed accurate for the entire area. This is to mention that the MOSART output is also correct in the upstream areas(17 to 35 North). After a certain latitude (in the region of 17 to 8 degree north) there is no storage in the tributaries (VOLR is zero). I don't have any clue what can cause such a problem.

I tried with coupling CLM and 50 km MOSART grid (subsetting data from global 50 Km MOSART input file ) /glade/p/cesmdata/cseg/inputdata/rof/mosart/MOSART_routing_Global_0.5x0.5_c170601.nc) which worked well with reasonal output.

I am unsure why 8th/100 degree resolution input data didn't work for the entire study area where half-degree resolution worked without any anomaly. I just wanted to make sure I am using the correct input data. Is there anyone who used the same data for global/regional simulation? Any suggestions on how to diagnose the problem?

I appreciate any help in this regard.

Thanks
Tamanna

 

[swensosc]

Two things that I would try: 1) do a run with the global mosart file. You can run a regional clm -> global mosart simulation; it will just get inputs over the clm region and route those as usual. If that works, it may indicate an issue with the mapping or regional input file. 2) activate some of the runoff variables in
components/mosart/src/riverroute/RtmHistFlds.F90 (e.g. the QSUR_ AND QSUB_ history fields). That will confirm that the clm runoff is actually being passed correctly to mosart.

 

[kabirtam@msu_edu]

Thanks, @swensosc for your suggestion. I tried your second suggestion first and it seems like CLM runoff is not being passed into Mozart correctly. The lower half of the grids have zero runoff.

About regional CLM but Global MOSART simulation:

1. Do I need to use a global MOSART domain file ??
2. As I have a 0.05-degree clm grid but a 0.125-degree MOSART grid, Do I need a mapping file to/from lnd/rof similar to regional MOSART simulation???

Also, I am struggling with using the mapping file created by area conservative remapping option by "gen_cesm_maps.sh"
I was having the following error using a mapping file created by area conservative remapping.

272:(seq_frac_check) [rof init] ERROR aborting
272: ERROR: Unknown error submitted to shr_abort_abort

3. How can I deal with this mapping file issue? Could you help with any specific suggestions? I attached my cesm and cpl log file.

I really appreciate your help.
Thanks!

 

[swensosc]

I don't think you need to do a global mosart simulation; it seems that something is not correct with your regional mapping. It could be related to how you are generating your mapping files. First, have you tried the tool cime/tools/mapping/check_maps/check_map.sh? It may tell you whether is a problem. Second, sometimes the order of the vertices in the scripgrid file is important. You could try this script to generate scripgrid files that can then be used to create mapping files: /glade/scratch/swensosc/mapping/create_scripgrid.py.

 

[kabirtam@msu_edu]

@swensosc You are absolutely right. The problem was due to the incorrect mapping files. I created the regional mapping files again. Seems like MOSART is getting correct runoff from clm and the problem with the simulation(incorrect VOLR) has been solved.

Thanks, Swenson for your help!
I Appreciate it.