Hi @oleson , I'm currently checking the NetCDF file for the anomaly forcing. I found that the fill value for precipitation is set to 1 (with units of mm/s). The file path for the anomaly forcing I’ve checked is /inputdata/atm/datm7/anomaly_forcing/CMIP6-SSP5-8.5/af.allvars.CESM.SSP5-8.5.2015-2100_c20220628_classic.nc. Additionally, I noticed that there are no values less than 0. Given this, I wanted to confirm: does this mean that drought conditions were not considered in this scenario?
A file with that name doesn't exist in our inputdata on glade or in the repository here:
There is a file with a similar name: af.allvars.CESM.SSP5-8.5.2015-2100_c20220628.nc
Precipitation in that file has a fill value of 1.e36:
double pr(time, lat, lon) ;
pr:_FillValue = 1.e+36 ;
pr:units = "mm/s" ;
pr:mode = "time-dependent" ;
pr:long_name = "atmospheric rain, after rain/snow repartitioning based on temperature scale factor" ;
pr:historical_source_files = "b.e21.BHIST.f09_g17.CMIP6-historical.004.clm2.h0.RAIN.185001-201412.nc" ;
pr:scenario_source_files = "b.e21.BSSP585cmip6.f09_g17.CMIP6-SSP5-8.5.103.clm2.h0.RAIN.201501-206412.ncb.e21.BSSP585cmip6.f09_g17.CMIP6-SSP5-8.5.103.clm2.h0.RAIN.206501-210012.nc"
The units are misnamed, they should be unitless as in the issue I filed here:
github.com
It is a multiplicative factor, so there shouldn't be any negative values.
There is a file with a similar name: af.allvars.CESM.SSP5-8.5.2015-2100_c20220628.nc
Precipitation in that file has a fill value of 1.e36:
double pr(time, lat, lon) ;
pr:_FillValue = 1.e+36 ;
pr:units = "mm/s" ;
pr:mode = "time-dependent" ;
pr:long_name = "atmospheric rain, after rain/snow repartitioning based on temperature scale factor" ;
pr:historical_source_files = "b.e21.BHIST.f09_g17.CMIP6-historical.004.clm2.h0.RAIN.185001-201412.nc" ;
pr:scenario_source_files = "b.e21.BSSP585cmip6.f09_g17.CMIP6-SSP5-8.5.103.clm2.h0.RAIN.201501-206412.ncb.e21.BSSP585cmip6.f09_g17.CMIP6-SSP5-8.5.103.clm2.h0.RAIN.206501-210012.nc"
The units are misnamed, they should be unitless as in the issue I filed here:
Units attribute on anomaly forcing precipitation file is incorrect · Issue #2858 · ESCOMP/CTSM
The units attribute for the pr variable on the anomaly forcing files is incorrect. It says mm/s, but I think it should be unitless. E.g. this file: /glade/campaign/cesm/cesmdata/cseg/inputdata/atm/...
github.com
It is a multiplicative factor, so there shouldn't be any negative values.
Hi Olson, I am not sure the files are the same one. The below post is the story that I got the file from the pub/olson. I am sure the _classic files has the fill value with 1 rather than 1.e36, could you please check it again?A file with that name doesn't exist in our inputdata on glade or in the repository here:
There is a file with a similar name: af.allvars.CESM.SSP5-8.5.2015-2100_c20220628.nc
Precipitation in that file has a fill value of 1.e36:
double pr(time, lat, lon) ;
pr:_FillValue = 1.e+36 ;
pr:units = "mm/s" ;
pr:mode = "time-dependent" ;
pr:long_name = "atmospheric rain, after rain/snow repartitioning based on temperature scale factor" ;
pr:historical_source_files = "b.e21.BHIST.f09_g17.CMIP6-historical.004.clm2.h0.RAIN.185001-201412.nc" ;
pr:scenario_source_files = "b.e21.BSSP585cmip6.f09_g17.CMIP6-SSP5-8.5.103.clm2.h0.RAIN.201501-206412.ncb.e21.BSSP585cmip6.f09_g17.CMIP6-SSP5-8.5.103.clm2.h0.RAIN.206501-210012.nc"
The units are misnamed, they should be unitless as in the issue I filed here:
Units attribute on anomaly forcing precipitation file is incorrect · Issue #2858 · ESCOMP/CTSM
The units attribute for the pr variable on the anomaly forcing files is incorrect. It says mm/s, but I think it should be unitless. E.g. this file: /glade/campaign/cesm/cesmdata/cseg/inputdata/atm/...
It is a multiplicative factor, so there shouldn't be any negative values.
Anomaly forcing run adds extra streams at runtime, causing an error
Oh, you are talking about a file I created back in June. You indicated that the file was ok, but now it is not? Those files are long gone.
So, I recreated it and put it on the ftp site. It checks out fine here:
double pr(time, lat, lon) ;
pr:_FillValue = 1.e+36 ;
pr:units = "mm/s" ;
pr:mode = "time-dependent" ;
pr:long_name = "atmospheric rain, after rain/snow repartitioning based on temperature scale factor" ;
pr:historical_source_files = "b.e21.BHIST.f09_g17.CMIP6-historical.004.clm2.h0.RAIN.185001-201412.nc" ;
pr:scenario_source_files = "b.e21.BSSP585cmip6.f09_g17.CMIP6-SSP5-8.5.103.clm2.h0.RAIN.201501-206412.ncb.e21.BSSP585cmip6.f09_g17.CMIP6-SSP5-8.5.103.clm2.h0.RAIN.206501-210012.nc"
So, I recreated it and put it on the ftp site. It checks out fine here:
double pr(time, lat, lon) ;
pr:_FillValue = 1.e+36 ;
pr:units = "mm/s" ;
pr:mode = "time-dependent" ;
pr:long_name = "atmospheric rain, after rain/snow repartitioning based on temperature scale factor" ;
pr:historical_source_files = "b.e21.BHIST.f09_g17.CMIP6-historical.004.clm2.h0.RAIN.185001-201412.nc" ;
pr:scenario_source_files = "b.e21.BSSP585cmip6.f09_g17.CMIP6-SSP5-8.5.103.clm2.h0.RAIN.201501-206412.ncb.e21.BSSP585cmip6.f09_g17.CMIP6-SSP5-8.5.103.clm2.h0.RAIN.206501-210012.nc"
Thanks, Oleson. I think those files are still good. At first, I thought the unit was the mm/s, and the fill value with 1 could cause some issues, but now I realize the unitless precipitation scale factor with 1 as the fill value is okay. I know the _fillvalue is 1e36, but when I read it into the Matlab matrix, the value in the ocean or somewhere that should be NaN is filled with 1. If I am not misunderstanding, is the precipitation parameter times the GSWPv3 precipitation rather than plus? And for the temperature and humidity parameters, is it a + rather than *?
- Status