Monitoring of Alpine snow using satellite radiometers and artificial neural networks

Abstract

Abstract The Alps represent an extremely complex environment in which snow properties suffer dramatic spatial variations that cannot easily be followed by space-borne microwave radiometers, due to their coarse spatial resolution: some studies demonstrated that the algorithms developed for global scale monitoring of the snow depth (SD) are unable to retrieve this parameter with a satisfactory accuracy on mountainous areas. An improved method for monitoring the Snow Depth (SD) on Alpine areas is presented here. Equivalent Brightness Temperature Tbeq at an enhanced spatial resolution, corrected for the effects of orography and forest coverage, were computed from the AMSR-E measurements by using ancillary information on land use, surface temperature, and a digital elevation model (DEM). These equivalent Tbeq values were used instead of the original AMSR-E measurements as inputs of an algorithm that estimates SD on a global scale basing on and Artificial Neural Network (ANN) techniques from AMSR-E brightness temperatures at X-, Ku- and Ka-bands, V-polarization. The improvement in the retrieval accuracy using these Tbeq equivalent values was evaluated using data collected during the winters between 2002 and 2011 on a test area located in the eastern part of the Italian Alps.