The GMI bytemap format is identical to that of RSS  AMSR-2, TMI, and AMSR-E data.
We are in the process of consolidating read routines for these consistently processed microwave instruments.

IDL users please can use the read routines provided here:
ftp.remss.com/support/read_routines/idl/

In the mean time, we continue to offer instrument specific read routines for Fortran, Matlab and IDL with GMI information provided are located in: ftp.remss.com/gmi/support_v08.2/

These IDL, FORTRAN 90, Python, Matlab and C++ routines are available 
to aid in reading the RSS bytemap daily and time-averaged data files.  
The read routines have been tested and work correctly within 
our PC Windows environment. We do not guarantee that they work perfectly 
in a different environment with different compilers.  If portability 
is a problem, we suggest using the Python code.

We provide data from the GMI instrument on GPM and
process the data using the RSS Version-8 algorithms into 
ocean products of sea surface temperature, ocean surface 
winds (10 meters height), atmospheric water vapor, cloud liquid 
water, and rain rate.


The FORTRAN subroutines are located in read_gmi_subroutines_v8.f
read_gmi_day returns a 1440x720x7x2 real array called gmi_data.
A description of the data within this array is at the top of the 
subroutine.  read_gmi_averaged returns a 1440x720x6 real array
called gmi_data.  The 3-day, weekly and monthly binary files
can all be read with this one routine.  Just supply the correct
filename with path.  These routines have been tested with 
Intel Visual Fortran 90. Data files must be unzipped prior to 
using these routines. Use a file unzipper of your choice.


The IDL read_gmi_day_v8.pro routine requires a full path filename
and returns seven 1440x720x2 real arrays.  The time-averaged
data files are read using read_gmi_averaged_v8.pro.  This routine
returns six 1440x720 real arrays.  A description is provided 
within the routine.  These routines have been tested with IDL 8.4
Data files do not need to be unzipped when using the /compress keyword
in the read call.


Matlab read routines, read_gmi_day_v8.m and read_gmi_averaged_v8.m, 
function similar to the IDL routines listed above and have been tested 
with Matlab 7.1   Data files must be unzipped before using the matlab 
routines.


The Python code consists of gmi_daily_v8.py and gmi_averaged_v8.py.  
Both require the use of bytemaps.py.  The example_usage.py code provides a main
program that you can use to test the data or adapt to your needs.
Description of file contents is provided at the top of each routine.

C++ read routines are also available and have been tested 
with the GNU compiler and Intel compiler.


Once you have further developed one of these skeleton programs
to suit your processing needs, you may wish to use the verify file 
to confirm that the data files are still being read correctly.
Please check that you have the correct verify file.

If you have any questions regarding these programs, 
or the RSS binary data files, contact our support desk at
at support@remss.com