Hi all,
After a few more experiments, Andrew and I found that because his
rectilinear data is on a global grid and the curvilinear data only a
subset of the grid, the rect_to_curv function does something odd with
the modulo wrapping. That needs some more investigation, but the
solution to get a good regridding is to restrict the variable on the
rectangular grid to a region matching what's in the curvilinear
coordinates. Something along these lines:
! Get limits for the curvilinear longitudes and latitudes
yes? stat lon
yes? stat lat
yes? let var_curv = rect_to_curv(rvar[x=-36:44,y=33:73], lon, lat, 2)
yes? shade var_curv, lon, lat
Andrew Ferrone wrote:
Dear ferreters
I'm working with a regional climate model, which is defined in
curvilinear coordinates. For the moment I need to interpolate some
globale data on a 1°x1° regular grid onto my curvilinear grid
(0.2°x0.2°). For this purpose I tried the external function
rect_to_curv in ferret (the lon and lat were used from the output of
the regional model). The output I produced was not the one I attended.
I've attached the original data (on the region I am interested in) in
CLM_inerpol2.tiff and the result in CLM_inerpol2.tiff. There is clearly
a problem when the interoplation crosses the 0° meridian. Has anyone
experienced some similar problems with this function and knows hoa to
resolve it?
Any help and suggestion is welcome!
Thanks a lot
Andrew
______________________________________________________________
Andrew FERRONE
Physicist, PhD student
Regional climate modelling, Climate Impacts of Aviation
Université catholique de Louvain (UCL)
Institut d'astronomie et de géophysique G. Lemaître (ASTR)
Chemin du Cyclotron, 2, B-1348 Louvain-la-Neuve (Belgium)
Phone: +32 (0)10 473365 Fax: +32 (0)10 474722
E-mail: andrew.ferrone@xxxxxxxxxxxx
|