Hi Robert:
I have seen WRF files where the projection parameters are incorrectly
specified. No one noticed because they only use the lat/lon (XLAT,XLON?)
variables. However, interpreting the parameters correctly is difficult. Can
you send me the file (or http/ftp URL) and i'll have a look.
Robert B. Schmunk wrote:
John,
To follow up on my e-mail of a week ago, I have been able to use
the NJ22 libraries to get into several sample WRF datasets and
not only extract the data but also the coordinate transform
information. Things seem to work beautifully with three different
datasets that use Lambert conformal conic grids (centered at
different locations), and I've even been able to doublecheck that
by comparing plot made with my own code for the L.c.c. projection.
However, one problem I am still encountering is with a dataset
that is on a transverse Mercator grid. The data extract and plot
easily enough, but an eyeball check of how certain variables
compare to coastline locations suggests that something is awry.
See for example attached image of HGT (height) and notice how the
Mediterranean coastline is a mess. Unfortunately, I have no idea
whether this indicates a potential bug in NJ22 or if the dataset
is just not gridded as the parameters claim it is gridded. Are
you aware of any specific issues between NJ22 and WRF data on
a transverse Mercator grid?
rbs
------------------------------------------------------------------------
--
Robert B. Schmunk, rschmunk@xxxxxxxxxxxxx
NASA Goddard Institute for Space Studies, 2880 Broadway, New York, NY 10025
