Accounting for seasonality in a soil moisture change detection algorithm for ASAR Wide Swath time series
- 1Laboratory of Hydrology and Water Management, Ghent University, Coupure links 653, Ghent, Belgium
- 2Department of Mathematical Modelling, Statistics and Bioinformatics, Ghent University, Coupure links 653, Ghent, Belgium
Abstract. A change detection algorithm is applied on a three year time series of ASAR Wide Swath images in VV polarization over Calabria, Italy, in order to derive information on temporal soil moisture dynamics. The algorithm, adapted from an algorithm originally developed for ERS scatterometer, was validated using a simple hydrological model incorporating meteorological and pedological data. Strong positive correlations between modelled soil moisture and ASAR soil moisture were observed over arable land, while the correlation became much weaker over more vegetated areas. In a second phase, an attempt was made to incorporate seasonality in the different model parameters. It was observed that seasonally changing surface properties mainly affected the multitemporal incidence angle normalization. When applying a seasonal angular normalization, correlation coefficients between modelled soil moisture and retrieved soil moisture increased overall. Attempts to account for seasonality in the other model parameters did not result in an improved performance.