Application of Neural Network With Error Correlation and Time Evolution for Retrieval of Soil Moisture and Other Vegetation Variables

Present paper utilizes the time evolution for estimating the soil moisture and vegetation parameter with radar remote sensing data. For this purpose, vegetation ladyflnger has been taken as a test fleld and experimental observations have been taken by bistatic scatterometer at X-band in the regular interval of 10 days for both like polarizations (i.e., Horizontal-Horizontal, HH-; Vertical-Vertical, VV-) and at difierent incidence angles. At this interval, all the vegetation parameters and scattering coe-cient have been recorded and computed. Three similar types of fleld of size 5 £ 5m have been especially prepared for this purpose. The observed data is critically analyzed to understand the efiect of incidence angle and polarization efiect on scattering coe-cient of the ladyflnger. It is observed that VV-polarization gives better result than HH-polarization and incidence angle 55 - is the best suited to observe composite efiect of vegetation ladyflnger biomass (Bm) and vegetation covered soil moisture at X- band. This analysis is further used for retrieval of soil moisture and