Satellite data is one of the major sources of information on snow cover areal extent and seasonal variability. At NOAA/NESDIS global monitoring of snow cover is conducted using microwave measurements from DMSP SSM/I. Snow maps are generated daily since 1987 at a resolution of 30 km; monthly statistics on snow cover frequency is derived for 1 x 1 degree grid cells. For the Northern Hemisphere several more satellite-based snow products are available operationally, which explains the fact that microwave-based snow maps currently find limited use in climatological studies and numerical applications. In the Southern Hemisphere, and in particular over South America, where seasonal changes of snow cover are most substantial, DMSP SSM/I observations remain virtually the sole operational source of data on snow cover areal extent and spatial distribution. Fourteen years time series of DMSP SSM/I observations contain indispensable information on seasonal and interannual variability of snow cover in the Southern Hemisphere and can be used for climatological analysis after a proper validation of snow retrievals is made.

A new automated system has been developed at NOAA/NESDIS for monitoring snow cover over South America. The technique implemented in the system is based on measurements in the visible, middle-infrared and infrared spectral bands from the Imager instrument of the Geostationary Environmental Operational Satellite (GOES). Snow maps are generated daily since May 2000 at a spatial resolution of 4 km. The major focus of the new system is the midlatitude portion of South America, which receives a considerable amount of wintertime snowfall.

The purpose of the study was to assess the accuracy of microwave-based snow cover retrievals and their applicability for an analysis of long-term changes of the seasonal snow cover in South America. We compared one-year time series of GOES and DMSP-based daily snow maps as well as weekly and monthly snow cover frequency estimated with the two techniques. The two products exhibited the closest agreement over a limited area of approximately 300,000 km² in the southern Argentina. For this area, the fourteen-year time series of microwave-based monthly snow-cover frequency was derived and analyzed. In an attempt to determine possible mechanisms affecting the interannual variability of the South American snow cover, we related variations of the wintertime snow cover to the changes in the Southern Oscillation index (SOI). The highest correlation with SOI equal to 0.61 was found for the mid-winter (July) monthly average snow cover extent.