How to run netcdf.defVarFill for a variable defined as 'float'

Hi,

I am using Matlab 7.14.0.739 (R2012a) and I need to create forcing data for my model in NetCDF that should look something like this:

 variables:

 float T2D(Time, south_north, west_east) ;
           T2D:layer_top = 0.f ;
           T2D:layer_bottom = 0.1f ;
           T2D:units = "K" ;
           T2D:_FillValue = -1.e+36f ;

The NetCDF file is created like this:

 ncid = netcdf.create(filenc,'NETCDF4');

... and this is how I am currently writing my Matlab script:

 % Surface temperature
 varid_t2d =  netcdf.defVar(ncid,'T2D','float',dimid_west_east,dimid_south_north,dimid_Time]);
 netcdf.putAtt(ncid,varid_t2d,'layer_top',0);
 netcdf.putAtt(ncid,varid_t2d,'layer_bottom',0.1);
 netcdf.putAtt(ncid,varid_t2d,'units','K');
 netcdf.defVarFill(ncid,varid_t2d,false,-1.e+36);

This is the error message I get when running it:

 Error using netcdflib
 The fill value datatype class 'double' does not match the netCDF datatype for T2D.

 Error in netcdf.defVarFill (line 32)
 netcdflib('defVarFill',ncid,varid,noFillMode,fillvalue);

 Error in noah2d_text_to_netcdf (line 164)
       netcdf.defVarFill(ncid,varid_t2d,false,-1.e+36);

If I change "netcdf.defVar" line to 'double':

 varid_t2d = netcdf.defVar(ncid,'T2D','double',...)]

the script runs fine!

Is there a way I can run "netcdf.defVarFill" for a variable which is originally defined as 'float'?

Thanks

Answer 1

[John]

If your netCDF variable is single precision, then your fill value must also be given as single precision, so

    netcdf.defVarFill(ncid,varid_t2d,false,single(-1e+36));

[Rafael]

That did the trick! Thank you.