I’m trying to evaluate my daily WRF precipitation using the run_prism_comp.csh script. I ran a 5.5-day chunk of WRF with 12 hours of spinup and output into 12-hour files. Based on run_prism_comp.csh, it seems like the script expects a wrfout file beginning at 00z instead of 12z, and that the output files should be in hourly increments?
I keep getting the following error when I execute run_prism_comp.csh:
Reading PRISM monthly precip: /uufs/chpc.utah.edu/common/home/daq-group2/AMET_v14/obs/MET/prism_daily/2017/PRISM_ppt_stable_4kmD2_20170704_asc.asc
Reading WRF output: /uufs/chpc.utah.edu/common/home/daq-group2/AMET_v14/model_data/MET/June062021_run5_D02_wrf/wrfout_d02_2017-07-04_12:00:00
Reading WRF output: /uufs/chpc.utah.edu/common/home/daq-group2/AMET_v14/model_data/MET/June062021_run5_D02_wrf/wrfout_d02_2017-07-10_00:00:00
Error in Rsx_nc4_get_vara_double: NetCDF: Index exceeds dimension bound
Var: RAINC Ndims: 3 Start: 23,0,0 Count: 1,291,291
Error in ncvar_get_inner(ncid2use, varid2use, nc$var[[li]]$missval, addOffset, :
C function R_nc4_get_vara_double returned error
Calls: wrf_precip → ncvar_get → ncvar_get_inner
Execution halted
I’ve tried running the script for WRF simulations that have hourly output files, and for a MODEL_START wrf file at 00z, but I get the same error for RAINC. I added emphasis to the “reading PRISM monthly precip” because I’ve selected daily analysis (the loaded file prefix reflects this). I’ve also tried running a whole month to see what happens and I get the same RAINC errors. Has anyone had any luck with this?
I would need to see the index settings for more guidance. These are just the time index in specified WRF outputs that define the start and end period of precip accumulation (RAINC RAINNC). If the script says these are out of bounds, they are not set correct. If I had 12 hourly WRF outputs starting at 12 UTC, you obviously cannot calc the precip total for a day with one file. You have to provide two files.In your case above it seems incorrect to attempt to do precipitation for Jul 4 when that file starts at 12 UTC and is the 12 hour spin up. You’d need the previous 5.5. day run segment’s WRF outputs that ended on Jul 5 00 UTC for precip on the 4th.
Let’s say I want precipitation for Jul 5, 2017. You should have a Jul 5 00 UTC file and a Jul 6, 00 UTC file. Time index 1 for start file would be 00 UTC on Jul 5. That’s where you want to start. Time index 1 of end file would be 00 UTC on Jul 6. So you set TINDEX start and end to 1.
We made it flexible so really any size WRF output can be used as long as you have accumulated precip in these file and run is not reset between the output files. So you cannot mix files from one 5.5 day run segment to another unless you are using WRF restart files and do not reset the precipitation counter (reset_simulation_start = .false.)
I would add that 5.5 run segments are old school. It’s inefficient. We now prefer developing monthly input files and run continuously with output in 1 day files.
Hope this helps. Again, the start and stop time index settings have to match the start and stop WRF output files. If they are larger than the number of times in the file it will fail.