Hi,

   Found a bug in the wgrib code.  I was passing the address of
the bitmap section rather than the bit map itself to the decode
routine.  This will affect only the following operations:

	- converting from grib to (ieee, binary, text)
	  AND the grib field uses bitmaps.

	The vast majority of the fields do not use bitmaps
        as the data is defined on all points.  The fields
	produced by reanalysis that may use bitmaps are:
		gravity wave flux
		potential evaporation
		pressure, temperature, nearest sigma level of clouds
		runoff
		soil moisture, soil temperature
		and isentropic surfaces that go underground

	To determine if the file uses a bitmap, type

		% wgrib -V [file]

	and if you see the work 'bitmap', bingo! that field
	uses a bitmap.

You should upgrade to version 1.3.0 (or higher) ASAP.  The
location on the cray Y-MP is:
	/wd5/wd51/rl51we/bin/wgrib (executable)
	/wd2/cpc/cdrom/PROGRAMS/WGRIB/ALL.C (source code)

On the NMC workstations, the executables are on:
	~wd51we/bin.[type]/wgrib      where type is hp, sun, or sgi

The copy on nic.fb4.noaa.wwb will be updates ASAP.




On a non-wgrib matter.  The w3lib routines used to write some
of the grib files had a bug (monthly means, monthly variances, 
reduced time series, and derived time series).  This error occurs 
when (max_value - min_value) is a power of 2.  In screening of 3500 
monthly means, this error only cropped up once (Aug 88 U200 winds).  
Since once is too much, the monthly means and derived time series
will be recalculated.  (The reduced time series are NMC internal).


BTW: if you haven't heard, the SSMI winds were incorrectly used
in the reanalysis.  All data after July 87 should be deleted.
If you want more information, contact E. Kalnay at
wd23ek@sun1.wwb.noaa.gov


				Wesley Ebisuzaki

PS
    If you want to be on a mail list for wgrib bug fixes/upgrades
send your e-mail address to wd51we@sun1.wwb.noaa.gov