Question about masking out

[Matt Georgescu]

Hello fellow GrADS users.

I obtained some free hi-res (9km) AVHRR Pathfinder 4 sea surface
temperature data on-line.  The data however has many gaps, due to cloud
presence. It was up to me to fill in these gaps as best as I could.

My interpolation over the ocean seems realistic enough.  I include here a
plot of the first week's SSTs during  the month of July 1995:
http://cep.rutgers.edu/~matt1/sst_070195.gif

My question is the following.  In the image that you see I simply used
basemap.gs to maskout the land values.  But the land values, erroneous as
they are (I didn't include an image also containing temp values over land,
but they are clearly in the wrong as the entire continent was initially
undefined) are still there in the actual data file.  I need to make
further use of this data in a regional model, which right now is not
accepting this data (and I believe it is due to the temps over land).

Similarly to using basemap.gs, how can i write an alternative temperature
value (say 2-m monthly average air temp from another file) over the continent,
while still keeping the SSTs (over water) as is?  Is something like this
possible using GrADS and does someone have a script handy?

thanks,
matt