APPLICATION OF TEMPERATURE AND VEGETATION INDEX FOR SOIL MOISTURE DOWNSCALING FOR AGRICULTURE

Abstract

Information on the soil moisture content in the soil is important for planning both for farmers and agricultural water managers. The availability of this information is however limited owing to the expensive nature of in-situ soil moisture measurements and cloud conditions that limits satellite derivation of soil moisture and the availability of satellite derived soil moisture at coarse resolutions limiting their application at local scale. This study applies existing downscaling technique which makes use of vegetation indices and land surface temperature. The study compounds the technique with Harmonic Analysis of Time Series (HANTS) to provide continuous high-resolution soil moisture. The case study area is the state of Nebraska and the low-resolution soil moisture used is advanced Scatterometer (ASCAT). The use of HANTS successfully produced daily soil moisture estimates without degrading the accuracy of the soil moisture at 0.09m3/m3. The Precision however, was observed to be slightly degraded owing to seasonal variations and the different rooting depths within which the vegetation draws soil moisture from. Whereas the study has been conducted in Nebraska, the same can be replicated in regions with limited soil moisture measurements. The information is paramount for crop production forecasting, drought assessment and crop water management since it provides near real time information. This study shows that soil moisture downscaling can improve the resolution without degrading the accuracy. This accuracy was found to be within the threshold required for global climate observation systems (GCOS) which is set at 20% of the saturated water content