[Met_help] Re: learning MET with WRF and TRMM

[John Halley Gotway]

Erik,

I'm not aware what other users in your situation have done.  But it seems to me that you have two choices.

(1) You could treat those observations as point observations and use the Point-Stat tool.  To do so, you could reformat the ASCII version of the HDF files into the ASCII point observation format that
the ASCII2NC tool accepts.  Then run the point observations through ASCII2NC and use that output in the Point-Stat tool.

(2) You could treat those observations as gridded observation and use the Grid-Stat tool.  To do so, you could reformat the HDF files into a NetCDF format that MET will accept.  Exactly how to do this
 may be a bit tricky - the forecast and observation files would ultimately need to be on the same grid.

Here's a link I found of someone trying to convert HDF to NetCDF: http://www.pmel.noaa.gov/maillists/tmap/ferret_users/fu_2002/msg00501.html

But once you have them in NetCDF, you may need to rename some variables and add some projection information for MET to use.  And since the HDF domain likely changes in time, the projection information
would change too!  So the first option would probably be easier.

If you send me one of the ASCII TRMM files, I could take a look and see if I have any other suggestions about reformatting it.

Hope this helps,
John

Erik Noble wrote:
> Dear John,
> I am still trying to use MET to compare WRF ARW model output (rain
> variables RAINC and RAINNC) with TRMM satellite rain observations only
> available in ascii or HDF. What does one do when data is only
> available in this format?
> -Erik