Kalman Filter Retrieval of Surface Temperature and Emissivity from Seviri Observations and Comparision with Iasi and Modis Products

In this paper, a suitable form of the Kalman filter is applied for the simultaneous retrieval of emissivity and surface temperature from SEVIRI observations. Exploiting the Kalman filter dynamical properties, this novel retrieval approach is capable of specifically taking into account for the time continuity of SEVIRI observations. SEVIRI results are compared to surface emissivity-temperature products obtained using IASI observations. The comparison has been performed considering a Sahara Desert target area, which is located within the Ourgla province (Algeria). Satellite observations for this area have been analysed for the month of July 2010. Surface temperature and emissivity retrieved from SEVIRI show a good agreement with IASI products. Comparison with space-time collocated ECMWF surface temperature shows that the ECMWF model has a negative bias during daytime. The bias can reach values as large as 10 K. A better agreement is found at night-time. For daytime observations, a negative bias is also found when we compare SEVIRI retrievals with MODIS surface temperature products.