I finally ran the recent versions to produce counterfactuals for ERA5-Land Europe. Here an output
As you see, there are issues with rsds and rlds variables. I think this is related to the _FillValue handling in attrici. You addressed this in
05f6d78 but it is seemingly not enough. I checked the input, timeseries and output files for rsds.
Attributes in input files:
int time(time) ;
time:standard_name = "time" ;
time:long_name = "Time" ;
time:units = "days since 1860-1-1 00:00:00" ;
time:calendar = "proleptic_gregorian" ;
time:axis = "T" ;
float lon(lon) ;
lon:standard_name = "longitude" ;
lon:long_name = "Longitude" ;
lon:units = "degrees_east" ;
lon:axis = "X" ;
float lat(lat) ;
lat:standard_name = "latitude" ;
lat:long_name = "Latitude" ;
lat:units = "degrees_north" ;
lat:axis = "Y" ;
float rsds(time, lat, lon) ;
rsds:standard_name = "surface_downwelling_shortwave_flux_in_air" ;
rsds:long_name = "Surface Downwelling Shortwave Radiation" ;
rsds:units = "W m-2" ;
rsds:_FillValue = 1.e+20f ;
rsds:missing_value = 1.e+20f ;
Attributes in timeseries files:
float y(time, lat, lon) ;
y:_FillValue = NaNf ;
y:standard_name = "surface_downwelling_shortwave_flux_in_air" ;
y:long_name = "Surface Downwelling Shortwave Radiation" ;
y:units = "W m-2" ;
double cfact(time, lat, lon) ;
cfact:_FillValue = NaN ;
cfact:standard_name = "surface_downwelling_shortwave_flux_in_air" ;
cfact:long_name = "Surface Downwelling Shortwave Radiation" ;
cfact:units = "W m-2" ;
double logp(lat, lon) ;
logp:_FillValue = NaN ;
int time(time) ;
time:standard_name = "time" ;
time:long_name = "Time" ;
time:axis = "T" ;
time:units = "days since 1860-01-01" ;
time:calendar = "proleptic_gregorian" ;
float lat(lat) ;
lat:_FillValue = NaNf ;
float lon(lon) ;
lon:_FillValue = NaNf ;
Attributes in merged output files:
double cfact(time, lat, lon) ;
cfact:standard_name = "surface_downwelling_shortwave_flux_in_air" ;
cfact:long_name = "Surface Downwelling Shortwave Radiation" ;
cfact:units = "W m-2" ;
float lat(lat) ;
double logp(lat, lon) ;
float lon(lon) ;
int time(time) ;
time:standard_name = "time" ;
time:long_name = "Time" ;
time:axis = "T" ;
time:units = "days since 1860-01-01" ;
time:calendar = "proleptic_gregorian" ;
float y(time, lat, lon) ;
y:standard_name = "surface_downwelling_shortwave_flux_in_air" ;
y:long_name = "Surface Downwelling Shortwave Radiation" ;
y:units = "W m-2" ;
So in timeseries there are incorrect fill values (NaN and NaNf), and in merged output files, they are absent. Maps show fine for rsds in ncview, so I am quite sure that this is the issue leading to the above plot.
I finally ran the recent versions to produce counterfactuals for ERA5-Land Europe. Here an output
As you see, there are issues with
rsdsandrldsvariables. I think this is related to the _FillValue handling in attrici. You addressed this in05f6d78 but it is seemingly not enough. I checked the input, timeseries and output files for rsds.
Attributes in input files:
Attributes in timeseries files:
Attributes in merged output files:
So in timeseries there are incorrect fill values (NaN and NaNf), and in merged output files, they are absent. Maps show fine for rsds in ncview, so I am quite sure that this is the issue leading to the above plot.