CDO

Hello:

So the command below runs well but I get an erroneous result:
cdo -P 20 remapcon2,c96_grid_May2019_region_0_to_180_no_time.nc glm_source_nan.nc glm_downscale_052019_con_nan.nc

This command is regridding a curvilinear 5424 x 5424 grid of GLM lightning frequency to a 128 x 162 lonlat grid.

The regridded file gives me very large values for lightning frequency. You will see these in glm_downscale_052019_con_nan.nc

From what I read and watched on youtube, since the downscaling is going from very fine to very coarse, conservative is the way to go. And also my understanding is that the 2nd order conservative regridding factors in the lon/lat gradients of the input grid.

I have attached an image of the fine grid and the result from the glm_downscale_052019_con_nan.nc grid.

The destination grid from the c96 coarse grid is attached.

Here is the ncdump -h glm_source_nan.nc as the file is too large to attach:

etcdf glm_source_nan {
dimensions:
nlat = 5424 ;
nlon = 5424 ;
grid_corners = 4 ;
variables:
float glm_ff(nlat, nlon) ;
glm_ff:_FillValue = NaNf ;
glm_ff:units = "Number of flashes 2km squared per seconds in a month" ;
glm_ff:standard_name = "GLM Flash Frequency" ;
glm_ff:coordinates = "lon lat" ;
float lat(nlat, nlon) ;
lat:units = "degrees_east" ;
lat:standard_name = "latitude" ;
lat:long_name = "latitude" ;
lat:bounds = "grid_corner_lat" ;
float lon(nlat, nlon) ;
lon:units = "degrees_east" ;
lon:standard_name = "longitude" ;
lon:long_name = "longitude" ;
lon:bounds = "grid_corner_lon" ;
float grid_corner_lat(nlat, nlon, grid_corners) ;
float grid_corner_lon(nlat, nlon, grid_corners) ;

// global attributes:
:description = "GLM Flash Frequency" ;
:created = "Created 23/03/22" ;
:history = "" ;
:coorninates = "lon lat" ;
}

Any help with figuring out whether or not remapcon2 is the right method or if I have left out anything that would get me comparable results to the observations would be great.

Thanks
Jonathan