Push notifications in your browser are not yet configured.
You are not logged in, you may not see all content and functionalities. If you have an account, please login .
Question about Z0 and PLCOV parameters in input file
Dear colleagues,
I have a problem with running the nested simulation (cclm to cclm) for my domain (Western Russia). I have adapted the Training Course’s step-by-step scripts to my parameters (see the attached int2lm_nest.sh file). I have previously simulated the ordinary gcm-to-cclm simulation successfully. As a result, I have become output lffd……nc files in the out01 directory. I have used the lffd2012020100c.nc file as input external parameters file for my nesting simulation. But I have become an error (see the attached slurm…..out-file). I have suggested the lffd2012020100c.nc file doesn’t contain any Z0 and PLCOV data that are necessary for int2lm. Could you suggest me how to take these data? Maybe, I must simulate the gcm-to-cclm chain with any corrections in INPUT parameters?
Thank you very much for your help.
Your external data file very likely contains the MAX and MIN values for the vegetation not the 12 monthly ones.
For the MIN MAX values you have to set itype_ndvi=0 in your namelist.
In case of a nesting run (lm2lm=.TRUE.) you have to distinguish between two external data sets.
One for the coarse grid, and one for the new, higher resolved grid.
As external data for the coarse grid only HSURF and/or FIS (I recommend to have both) and FR_LAND are required. They should be in your lffd2012020100c.nc file. Check it!
As external data for the fine grid muche more variables are required, amongst others Z0 and PLCOV (see your CFO _00625pep.nc file).
And now the explanation of Burkhardt comes into the play.
However, I just saw, that the name of the external file for the fine grid in your slumr-321662.out file ( CFO _00625pep.nc) seems to be different from the name
in your int2lm_nest.sh script (mid-europe_00625.nc).
Hans-Juergen
Burkhardt, Hans-Juergen, thank you for suggestions.
My lffd2012020100.nc file has HSURF , FIS , FR_LAND, SOILTYP parameters and doesn’t have Z0 and PLCOV parameters. However, my fine grid file CFO _00625pep.nc has all these parameters.
About mid-europe_00625.nc – you can disregard it, I have fixed this mistake before running the simulation.
I have replaced itype_ndvi value to 0. The above-mentioned problem was fixed, but new ones appeared further (see the other attached slurm…..out file, please). Could you suggest me, what is it?
Thank you very much.
This may have something to do with the land sea distribution. Can you upload the files lffd2012020100c.nc and CFO _00625pep.nc, please. Then I can have a look.
I have uploaded files.
Thank you very much.
I guess the problem is that in the region of your small domain lakes are resolved which are not in that part of the large domain. Since the large domain is cut out in int2lm to just use the part that covers the small region, int2lm does not find water pixels which can be used for determining the lake surface temperature in the small region. I hope this problem will be solved in a future version of int2lm. Meanwhile you may enlarge the small region by about 20 pixel to the North. I hope this helps.
Burkhardt, thank you very much for suggestions! int2lm finished without problems. I hope, cclm task will complete successfully too. Therefore, I have a technical question. Does this problem with int2lm mean, that too small domains are not preferable? What is a minimum of domain size? Does it depend on model grid size? For example, if I want to simulate the model with the minimal resolution – about 2.8 km, does any limitations to domain size exist?
Sorry for a lot of questions.
Thank you.
Except of the single water or land box issue, the technically lower limit is 3×3 grid boxes, the upper limit is defined by your computer system (others may correct me). For scientifically upper and lower limits of RCM s I refer to the literature.