G-RSM Output and GRIB format


PWAT (kpds5=54) is precipitable water. The unit is kg/m^2.
In RSM PWAT is written out at two different stages in the code. The two fields should be practically the same. We will remove one of them from the code in the future.

PRATE (kpds5=59) is precipitation rate. The unit is kg/m^2/s. In order to get the total precipitation you have to multiply it by the forecast duration. I don't think total precipitation is a standard output in the current G-RSM.

HGTsfc (kpds5=7) is surface height or topography (m).

Omega: The unit of omega in r_pgb or pgb is Pa/sec. The corresponding code is src/fcst/omega.F (not omegas.F). I also checked that an instantaneous value can be 3~6 Pa/s (100~200 hPa/h) in both GSM and RSM, but in long term average, it converges about appropriate range. In a high resolution experiment, such as 10km, the omega value tends to be very large due to steeper topography. (contributed by Kei)


In order to translate variable names, see:
libs/etc/grib1.kpds5.vsn22
libs/etc/grib1.kpds6.vsn21
These tables are used with libs/etc/grmap .

When using wgrib or grib2ctl, the following table from NCAR is used:
http://dss.ucar.edu/datasets/ds090.0/docs/variablelists/all-vars-file


In order to convert GRIB to netCDF try this tool lats4d:

http://gmao.gsfc.nasa.gov/software/lats4d/


comparison.doc - Comparison of forecasts provided by Sai Ming

Discussions

hko

G-RSM output

hko 09 November 2006 15:35:55

Hi Hideki,
There is a variable PWAT in GSM's output gribs. The unit is kg/m^2. Is this the same as total precipitation? I guess not. In RSM's output gribs, there are two PWAT records and one PRATE record. The variable PRATE has a unit of kg/m^2/s. Do I need to multiply this variable by the forecast duration, say 6 hr for 6-hrly gribs, to get the total precipitation? In our RSM97 output gribs, there is a variable APCP with unit kg/m^2, which is the total precipitation. I check that this variable also exists in NCEP's GFS output gribs. Please advise how I can get the total precipitation forecast from GSM and RSM output gribs.
Thanks.
Sai Ming

khideki
khideki 09 November 2006 17:43:40

Please see the main page of G-RSM output.

hko

Output GRIB files of GSM

hko 08 May 2007 05:51:50

Hi,
I run GSM for a whole season and get global output GRIB files at 24 hr intervals. I have the following setting of environment variables:
ENDHOUR=2208
INCHOUR=$ENDHOUR
INTSFCX=24
INTSIG=6
INTSFC=6
INTPOST=24
INTERSTART=24
Is it possible to generate the intermediate 6-hourly GRIB files without running the model over again?

Sai Ming

khideki
khideki 08 May 2007 16:05:32

I don't think it is possible. You saved GRIB 24 hours and your restart files (sig, sigit, sfc files) are also created every 24 hours (which may or may not be saved, depending on script setting). Hideki

hko

RSM grib files

hko 05 May 2008 08:41:36

Hi,
In the rsm script, I set INTPOST=1 and INTPGB=1. Hourly grib files r_pgb.ftxxx are generated. In some of the records of the grib file, say r_pgb.ft255, the variable is described as "252-255hr ave" (this is part of the output of wgrib -V). But for forecast hour >=256, say r_pgb.ft256, the description is "256hr fcst". May I ask why the gribs are coded in this way, or could it be just a minor bug in coding the description of the grib records?
Sai Ming

hko
hko 06 May 2008 09:13:02

Hi,
I just want to make sure that I am doing the right thing. I want to calculate daily rainfall from hourly gribs produced by RSM.
1. Is that the grmean script and program simply calculate the average of input gribs, regardless of that some of the contents are 2-hourly average, some are 5-hourly average, some are forecast for a particular hour?
2. Let's say I want to calculate total rainfall for Day 3. In my part of the world, Day 3 is from ft=65 to ft=88. The PRATE record in the r_pgb.ftxxx file is an average over a variable period depending on ft, e.g. when ft=71, it is 66-71hr ave; when ft=72, it is 72-72hr ave; when ft=84, it is 78-84hr ave. (I get this information by wgrib -V.) So, to get the total rainfall for Day 3, I should select the appropriate PRATE records and multiply the rainfall rate by the appropriate period of time, and then do a summation. Is this the right way to do it?
Sai Ming