Prior 1979, CORe used model snow that was reduced in mass by an
empirical factor.  https://www.cpc.ncep.noaa.gov/products/CORe/pns/eval/45cdpw-WEbisuzaki.pdf
This correction was run daily and would reduce the snow from the system./


From 1979-present CFS snow, based on 

   snow depth from AFWA->USAF 557th Weather Wing snow depth and 
   IMS snow cover from NESDIS->U.S. National Ice Center


from https://www.arcgis.com/home/item.html?id=d25319e15dfe4514a97ed398c276479e
Description
The USAF 557th Weather Wing, formerly known as the Air Force Weather Agency, produces global precipitation 
and temperature data sets used in drought and flood assessment, crop production and forecasting. The merged 
product estimates daily snow depth in centimeters. Prior to year 2009, snow depth was estimated daily by 
the Air Force Weather Agency (AFWA) by merging satellite-derived snow cover data with daily snow depth reports 
from ground stations. Snow depth reports were updated by additional snowfall data or decreased by calculated 
snow melt. If no new snow depth or accumulation information was available, surface temperature data was used 
to calculate a possible snow melt for temperatures above 32-degree Fahrenheit.
Terms of Use
No special restrictions or limitations on using the item's content have been provided."




   snow cover from NESDIS->U.S. National Ice Center (IMS)

Interactive Multisensor Snow and Ice Mapping System (IMS) 

https://usicecenter.gov/Resources/ImsInfo

NH snow covert based on satellite imagery, rad, models and ground station data.


The CFS snow depth product has AFWA snow depth and IMS snow cover inputs.  The
CFS snow depth is ingested by the "cycle" program of the forecast system.  "Cycle"
considers the snow depth analysis to be inaccurate when over 2 meters.  So cycle
sets snow depth to be mix of the observed snow depth and the model snow.  When 
the observations showed no snow, the model snow was eliminated with a loss of 
frozen water from the system.  

Whenever the model was forced with a new snow field, the depth was given assuming 
the snow density was 10" snow = 1" liquid water.

The CFS snow was created from satellite data, and early satellites yielded worse 
snow analyses than later analyses.  The variability of the pre-1979 which depended 
on model snow was smaller than later period which was observed.


-------------------------

Snow has a strong albedo affect.  Even a 1 mm of snow will act like a 
snow covered grid point. In real life, a 1 mm averge snow cover in a 
grid box will have mountains, hills, rivers and trees.  Only a fraction 
of the grid will be snow covered.

-------------------------

The CORe cycle only ingests the snow depth analysis when there is a new analysis.
Under normal circumstances, the snow depth analysis is updated daily and CORe
in ingesting the snow analysis every 24 hours.  

The CORe model and the observed snow analysis can be inconsistant.  The CORe grid point
is .7 degree size, and elevation of the grid point is the average elevationin the grid 
cell.  On the other hand, the snow analysis is done at a higher resolution and
the grid cell could contain a snow-covered mountain and snow-free valley.  So grid point
could be at a snow-free elevation and with a snow covered analysis.  (This was a problem
with the NCEP/NCAR Reanalysis.)