WEASD
                                     -----

WEASD? What is WEASD?  It's NCEP-ese for snow depth.  Even though Reanalysis
outputs snow depth, it should be used with caution.  


                                   Background

By using satellite imagery, one can determine the snow cover.  NESDIS creates
such a weekly product.  Nominally this analysis is valid for a single time 
except that cloud-covered regions are based on earlier satellite imagery.  
However, the forecast model simulates process such as snow melting and 
accumulation and requires the snow depth.  Obviously one cannot derive the 
snow depth from the snow cover.  So the snow depth is parameterized as a 
linear function of the surface temperature in regions of observed snow.  
This has the nice property that relatively warm regions have less snow and 
are more likely to lose their snow cover than cold regions which have more 
parameterized snow.  Of course, the model can generate or remove snow cover; 
however, it only has 6 hours to generate the snow before the snow cover gets 
reinitiallied by the analysis system.


                                   Suggestion


Consider the snow (WEASD) to be a indicator of snow cover (1/0) rather than
the amount of snow.  Since WEASD is a 6-hour forecast, it is a combination
of model forecast and observed data (NESDIS weekly analysis).  For the
period predating the NESDIS snow analysis, it hasn't been decided whether
WEASD will be predicted by the model or based on a climatology. 


7/99:  Some anomalies in the T2m temperatures were noted in the spring 
of 1999.  These anomalies appear to be related to the new NESDIS
snow analysis that was used.  BTW the new analysis is global.