SMOS L2 retrieval results over the American continent and comparisons with independent data sources

摘要

This paper shows results obtained by using the SMOS retrieval algorithm over forests at the prototype level. In each SMOS node, the algorithm estimates the soil moisture and the vegetation optical depth. For the optical depth, values retrieved in July 2011 in all forests of the American continent are shown and compared against forest height estimated by GLAS LIDAR of ICESAT satellite. A significant correlation between the two variables is observed. For each forest height estimated by LIDAR, the standard deviation of optical depth is slightly higher than 0.1. For soil moisture, 30 nodes of the SCAN/SNOTEL network have been considered. Over one year of data, retrieved values are compared against ground measurements. Overall, the rms error is of the order of 0.1 m3/m3. In general better results are obtained in the Eastern deciduous forest. The algorithm was run using different versions, corresponding to different initial guesses of soil permittivity and Leaf Area Index, but variations in the retrieved values are moderate.