EVENT DETECTION OF HYDROLOGICAL PROCESSES WITH PASSIVE L-BAND DATA FROM SMOS

Since it's launch, the ESA's Soil Moisture and Ocean Salinity (SMOS) satellite, is delivering new data from its L-Band 1.4Ghz 2D interferometer [1]. The observations from SMOS are used to retrieve soil moisture in the first centimeters and ocean salinity at the surface of the water. The observations are multi-angular with a 3 days maximum revisit time. The spatial resolution of SMOS data is 40km. In this paper we present on event detection algorithm implemented at CATDS (Centre Aval de Traitement des Donnees SMOS) the CNES level 3 and level 4 SMOS enter. This algorithm is a three stage change detection algorithm. At stage one the possibility/probability of occurrence of the event is evaluated. This is done via spatio-temporal constraints maps. These maps are obtained from the analysis of NSIDC's freezing index products over the last century. Climate data from ancillary files are tested will taking into consideration the uncertainty of the data. Some selected retrieved variables are also tested. At stage two a time series analysis is applied. In the current version of the algorithm a direct change detection algorithm is used. The tests make use of available variables of polarization index, retrieved soil moisture. Finally at stage three a simple fuzzy logic approach is used to decide if the event occurred. This approaches takes into consideration the separation time of the data. Ascending and descending orbits are taken into consideration. In this study freezing detection is presented over central CONUS. The temporal and angular signature of SMOS will be presented. Comparison is done with the SCAN network