Development of a Time Series–Based Methodology for Estimation of Large-Area Soil Wetness over India Using IRS-P4 Microwave Radiometer Data

Abstract

Soil moisture is a very important boundary parameter in numerical weather prediction at different spatial and temporal scales. Satellite-based microwave radiometric observations are considered to be the best because of their high sensitivity to soil moisture, apart from possessing all-weather and day?night observation capabilities with high repetitousness. In the present study, 6.6-GHz horizontal-polarization brightness temperature data from the Multifrequency Scanning Microwave Radiometer (MSMR) onboard the Indian Remote Sensing Satellite IRS-P4 have been used for the estimation of large-area-averaged soil wetness. A methodology has been developed for the estimation of soil wetness for the period of June?July from the time series of MSMR brightness temperatures over India. Maximum and minimum brightness temperatures for each pixel are assigned to the driest and wettest periods, respectively. A daily soil wetness index over each pixel is computed by normalizing brightness temperature observations from these extreme values. This algorithm has the advantage that it takes into account the effect of time-invariant factors, such as vegetation, surface roughness, and soil characteristics, on soil wetness estimation. Weekly soil wetness maps compare well to corresponding weekly rainfall maps depicting clearly the regions of dry and wet soil conditions. Comparisons of MSMR-derived soil wetness with in situ observations show a high correlation (R > 0.75), with a standard error of the soil moisture estimate of less than 7% (volumetric unit) for the surface (0?5 cm) and subsurface (5?10 cm) soil moisture.