[ncl-talk] 2-D Kernel Convolution in NCL...

[Capehart, William J]

Hi All

I am working with a dataset with strong horizontal linear features in the data (created by download errors).  I'm working on a diagnostic to lock on to these horizontal discontinuities with mixed results (so far they're great with temperatures with high expected spatial correlation, but not so great for precipitation and other fields with lower expected spatial correlation).

I'm wanting to try a remote sensing/image processing solution using kernel filters with the window being a classic W = [[1,1,1],[-1,-1,-1]] or W = [[1,1,1],[0,0,0],[-1,-1,-1]] kernel window but I am not seeing an equivalent in NCL that we use in IDL with its convol() function.   

e.g., https://www.harrisgeospatial.com/docs/convol.html

Is there an option in NCL that lets me do this with little slowdown or drama or should I stick with a WRAPIT Fortran Function with OpenMP to speed it up?

Cheers-n-Thanks
Bill

---------------------------------------------------------------------------------------
Dr. Bill Capehart <William.Capehart at sdsmt.edu>
Director, Atmospheric and Environmental Sciences Program
Department of Civil and Environmental Engineering
201 Mineral Industries Building
South Dakota School of Mines and Technology
501 East St Joseph Street
Rapid City, SD 57701-3995
Ph: +1-605-394-1994   Mobile: +1-605-484-5692