Try:

NCL: http://www.ncl.ucar.edu/

GrADS: http://www.iges.org/grads/

Ferret: http://www.ferret.noaa.gov/Ferret/

They are free.

Guilherme.

Here is a sample grib file.

The Climate Data Operators (CDO) is very easy to use. This tool is powefull to extract and manipulate datasets.

Link: https://code.zmaw.de/projects/cdo

Guilherme.

Dear Saeed,

Using R (http://www.r-project.org/) is quite straightforward using the raster library. Reading a grib is as simpler as this two lines of code:

library(raster)

r

Hi Saeed,

I think in grads u can use efficiently.

u can check in this link..

http://www.iges.org/grads/gadoc/grib.html

all the best.

Hi Xuanlong,

You are right. In my example, although netCDF support multiple bands, my input data was a list of single-band files. I totally agree with your point. When dealing with multiband is more efficient to use 'brick'.

Best regards,

David

Hello,

just in case it would be practical to proces the data in GIS (i.e. to make spatial crop of the data of a region of interest, to visualize/highlight changes in time, do simple raster cell computations, etc.), it is very well possible. As stated by Xuanlong Ma, gdal library supports this format, which means virtually any free & open-source GIS would be able import such dataset. I just tried your example file with QGIS (http://qgis.org) and it opened it as 47-band raster dataset without georeference just by dragging the file into the program window.

Best regards,

Tomas B.

correction: It probably has georeference, it seems it has a geograpic coordinate system, only the exact variant is not defined in the data or not recognized by the software.

Surprisingly all useful tools has be already mentioned before, my favourites are in this order wgrib (http://www.cpc.ncep.noaa.gov/products/wesley/wgrib.html), CDO (https://code.zmaw.de/projects/cdo) and of cause R-project. Together with wgrib you can also use GMT (http://gmt.soest.hawaii.edu/).

Last but not least you should consider to convert all GRIB files to netcdf (especially if you use GRIBs from different sources) there is a very interesting article from John Caron from Unidata/UCAR which highlights the problems of BUFR and GRIB (online available https://www.unidata.ucar.edu/staff/caron/papers/GRIBarchivals.pdf)

Regards Markus

If you want to do this in MATLAB you can use read.grib:

g=read_grib('HADCM3_A2_wind_1981-1990.grb',-1)

If you would like some example code for manipulating and plotting the data, please email me on [email protected].

I'd also recommend CDO (https://code.zmaw.de/projects/cdo) . It's a very powerful set of command-line routines, and you can easily use them for selecting time intervals and/or specific regions of your GRIB files.

I have used a lot of the software suggested but my last try was with R. You can find here info on how to read and plot GRIB files. http://stackoverflow.com/q/16669633/709777

NCL (NCAR Command Language) is very good at analyzing meteorological data in popular GCM model formats (grib, netcdf, hdf, etc.). NCL is freely available (after registration) from http://www.ncl.ucar.edu/Download/

NCL is not as capable as R (or Matlab) in advanced statistical analysis, but as far as basic analysis is concerned, there are lots of available subroutines and examples. If you are interested in basic analysis (averages, correlations, anomalies, trends, etc.) the chances are there is code available on the NCAR that can be used with minimal modifications.

Here is sample code to read grib data. NCL has capabilities to convert grib to netcdf (or hdf), and do high quality plotting and even if one decides to do the analysis in R(or Matlab), NCL is still a powerful tool in the overall workflow.

begin

;************************************

; create pointer to several types of

; files. Note the syntax is the same

;************************************

in = addfile("./uv300.nc","r") ; netcdf

h = addfile("./S2000759.hdf","r") ; hdf

g1 = addfile("./monthly_1979_1151.tar.grb","r") ; grib1

g2 = addfile("./monthly_1979_1151.tar.grb2","r") ; grib2

;************************************

; read variable from the file

; no space between in and the variable name

;************************************

u = in->U ; on RHS, case sensitive

end

And just to complete the list, python can do all this as well using the pygrib module (and you get a really nice scripting language as a bonus).

Hi Saeed. I would recommend the code of CDO (mentioned by Markus). This one includes in the mannual very useful commands to extract, convert and analyze the grib files.

+1 with python. Working with gripapi, matplotlib, gdal, numpy... in python, and all that tools together are great for plot custom graphs of local areas.

how to extract data from meteosat images( .grb) using matlab

(MATLAB/Octave)+Python+Java, I think. 

You can use the GrADS software, it's good