The AMSR-E snow water equivalent product: status and future development

Passive microwave observations from Advanced Microwave Scanning Radiometer - EOS (AMSR-E) launched in 2002 aboard NASA's Aqua satellite are used to estimate daily, five day and monthly global snow water equivalent (SWE). In this paper we show how the standard daily SWE product has evolved from a static semi-empirical algorithm to a more dynamic algorithm that incorporates estimates of snow properties in the retrieval process. The basis of the algorithm is a brightness temperature difference approach that uses the 10 GHz, 18 GHz and 36 GHz vertically polarized channels. Separate retrievals for forested and un-forested fractions of the instantaneous field of views are combined and we demonstrate how the 10GHz response is less affected by the snowpack itself than is the standard 18GHz response. For deeper snowpacks, we show how the scattering signal of the 18 GHz channel, in combination with the 10 GHz channel, cab be used to improve retrievals. The approach also uses 89 GHz channel measurements to detect shallow snow. AMSR-E snow retrievals are tested using several data sets. First, we compare AMSR-E estimates with GEWEX (Global Energy and Water Experiment) Asian Monsoon Experiment (GAME) and Coordinated Enhanced Observing Period (CEOP) data. Seven acoustic snow gauge sites recorded daily snow depth within a 100 km x 100 km domain located near Yakutsk in Siberia. These point snow depth measurements span a period between October 2002 to February 2003 and October 2003 to April 2004. Additionally, Cold Land Processes Experiment data from Colorado in 2002 and 2003 are also used to test AMSR-E estimates. Finally, snow depth data from WMO Global Surface Observation of the Day measurement sites are used to test the AMSR-E estimates. Future developments of the product are outlined especially with respect to algorithm developments and the estimation of generalized error statistics on a per-sample basis that can be incorporated directly into the product processing stream.