Retrievals of land surface parameters from measured brightness temperatures at 1.4 and 10.65 GHz by a neural network

Field measurements are utilized to examine the sensitivity of H- and V-polarized radiometric signatures at 1.4 (L-band), 10.65 (X-band) GHz, and their combinations in terms of viewing angles and frequencies to the wheat biomass and soil moisture through an error propagation learning back propagation (EPLBP) neural network. The radiometric measurements were taken by the PORTOS radiometer over wheat fields through 3 month growth cycles in 1993 (PORTOS-93) and 1996 (PORTOS-96). Soil moisture content (SMC) profiles were measured to a depth of 10 cm in 1993, and 5 cm in 1996. The wheat was sampled daily in 1993 and twice a week in 1996 to obtain its dry and wet biomass. The EPLBP neural network is trained with observations randomly chosen from the PORTOS-93 data, and evaluated by the rest of the very same data set. Then, the trained neural network is further evaluated based on the PORTOS-96 data set for practical concern. During both field campaigns, the L-band radiometer measured brightness temperatures at a wide range of viewing angles from 0 to 50 degrees, while the X-band radiometer was operated at an incident angle of 50 degrees.