Planting date estimation in semi-arid environments based on Ku-band radar scatterometer data

A method to determine planting dates in semi-arid regions is presented, based on K_u-band spaceborne scatterometer data. The planting date analysis was performed for Mali, a region with a broad range of vegetation cover with tropical forest in the south and desert in the north. The Ku-band data was acquired by the Sea Winds scatterometer onboard the QuikSCAT satellite during the time from January 2000 to December 2003. Climate data from meteorological stations was compared with scatterometer time series of data colocated from a circular area of a specific size. The comparison shows that the evolution of the backscatter signal is highly correlated with the vegetation cycle triggered in turn by the rain season. An accurate date for the onset of the growing season and therefore a basic planting date can be determined from noise-filtered backscatter time series using a simple threshold method. The temporal variations of the backscatter time series are mainly caused by vegetation growth and changes of surface soil moisture. An increased backscatter signal indicates therefore more and more sufficient growing conditions. For the estimation of the contribution of surface soil moisture, the backscatter was additionally compared with in situ data from test sites within the Duero basin in Spain, covered by the soil moisture measurement network of the University of Salamanca. The comparison showed a significant influence of surface soil moisture on the microwave backscatter