Combining remote sensing and terrestrial photography in a snowmelt modeling framework to retrieve snow evolution in a semiarid region

Abstract

This work shows the strength of combining different data sources and modelling to generate long time map series of snow variables in Sierra Nevada (Spain), an alpine environment within a semiarid region in which the snowpack variability is greatly enhanced by the topographic gradients and the sea influence. Different control areas monitored by time-lapse terrestrial cameras validate a spectral mixture algorithm on Landsat TM data to derive snow cover fraction maps in the study site. These maps are the basis to calibrate an energy balance snow model on a distributed scale, from which long-term series of snow cover fraction and snow water equivalent maps are finally obtained. The results prove this methodological approach to be efficient to overpass to some extent the individual limitations of each data source/model, and generate distributed variables relevant for water resource assessment in these regions.