Development and evaluation of the AMSU-based snow water equivalent retrieval algorithm

The estimation of snow water equivalent (SWE) from passive microwave sensors remains a formidable challenge primarily due to the non-unique nature of microwave signatures over different snow-cover surfaces. There is now wide recognition shared by the microwave research community of the dominant role of snow structure parameters on the brightness temperatures observed by the satellite. Such a dominant influence of other snow properties relative to the variable of interest complicates the retrievals of SWE and deteriorates their accuracy.

Existing microwave algorithms typically use a dual, lower frequency channel approach to estimate SWE or snow depth, where one channel (19 GHz for the SSM/I instrument) is used as reference and the other (37 GHz for SSM/I) is used as primary retrieval channel. The objective of this paper is to introduce a new, multiple-frequency channel approach to the retrievals of SWE based on the microwave measurements from the AMSU instrument. These retrievals are being evaluated within an operational framework called the Microwave Surface and Precipitation Products System (MSPPS). The AMSU instrument offers a unique combination of millimeter wavelength channels over a wider frequency range from 23 to 150 GHz that are sensitive to both precipitation- and snow-cover sized particles. Another advantage of the AMSU sensor is its large swath width, which, when combined with other microwave sensors like SSM/I, offers more global coverage in time and space.

AMSU measurements over the 23-150 GHz frequency range are compared with SWE and snow depth estimates derived from in-situ observations, aircraft measurements and snow process model simulations at several locations in the US Great Plains. Specifically, several scattering indices, computed from linear combinations in the brightness temperature at 23, 31, 50, 89 and 150 GHz are analyzed for estimating SWE. The analysis includes a variety of snow-cover types and weather conditions so that the retrievals of SWE are robust and representative of a broader climatology. Evaluation of these retrievals is under way and is being made over North America and other regions of the world.