Area wise very high resolution soil moisture availability estimation using heterogeneous remote sensing

This paper proposes and investigates a novel approach of using data-driven machine learning technique to overcome the drawbacks of conventional water balance modelling approaches due to the fact that the spatial and temporal resolution are limited as well as relying highly on condition-specified field experiments as a control group. The proposed approach integrates various independent remote sensing data sources of low to very high resolution as a single knowledge base, and then conducts a two-stage analysis of the data. Result from the research study shows that large area water availability could be modelled and estimated at a 50cm spatial and daily temporal resolution.