PRELIMINARY INVESTIGATIONS ON SOIL MOISTURE RETRIEVAL THROUGH MICROWAVE REMOTE SENSING

Abstract

Soil moisture is an environmental descriptor that integrates much of surface hydrology and is the interface between solid earth surface and the atmosphere. Knowledge of spatial and temporal distribution of soil moisture over land surface is an important input to agricultural, hydrological' and meteorological applications. In this report emphasis is given on the retrieval of soil moisture using active microwave remote sensing. The ERS-1/2 SAR from the European Space Agency, the JERS-1 SAR from the Japanese and the Canadian Radarsat are the operational active microwave satellite sensors with frequencies suitable for estimating soil moisture. Significant work has been done around the world towards understanding target/sensor interaction for developing models for soil moisture retrieval from microwave remote sensing data. The unique sensitivity of microwave sensors for soil moisture variation has been examined by various investigators using ground based scatterometer systems. The objective of such studies was to identify optimum sensor parameters having maximum sensitivity to soil moisture variations and minimum effect of noise parameters like crop cover and surface roughness. In the present thesis a review of such studies is made, which shows the dependency of soil moisture on noise parameters and on sensor parameters. Review of, literature also show the models, which can be developed between soil moisture and backscattering coefficient of the radar. In the present study, C-band ERS-2 SAR data is studied to detect near surface (0-10 cm depth) soil moisture over parts of Solani catchment around Roorkee. ERS-2 SAR data of iii 23rd July, 2001 (orbit No. 32710) was evaluated for this purpose. Empirical relationships have been developed between the volumetric soil moisture, which is obtained from field observations and the bacicscattering coefficient, which is extracted for the ERS-2 SAR PRI image for bare and vegetative soils. Digital number of the ERS-2 SAR PRI image is also correlated with the volumetric soil moisture. These models can be used with an approximate error of ± 20% and shall prove to be of immense potential to agriculturists, hydrologists and administrators for identification of soil and water conservation measures, irrigation scheduling etc.