Vegetation/Forest Effects in Microwave Remote Sensing of Soil Moisture

Abstract:

The mapping of soil moisture is important for hydrologic modelling, climate prediction, etc. The ESA’s SMOS and NASA’s SMAP are currently in orbit, and a new satellite NISAR will be launched shortly. All three satellites carry L-band sensors for soil moisture retrieval.  However, much of the earth surface are coved by vegetation/forests. Thus, electromagnetic modelling of vegetation/forests at microwave frequencies is important for remote sensing of soil moisture. My early work was on the analytical models for vegetated surfaces, using radiative transfer equation (RTE), distorted Born approximation (DBA), an improved coherent model and a unified model for combined active and passive, which were used for soil moisture retrieval in SMAP. Since 2017, I have been pursuing full wave simulations of Numerical Maxwell Model in 3D (NMM3D) of vegetation/forests. An efficient hybrid method is developed which combines off-the-shelf techniques for single objects and newly developed techniques to solve the Foldy-Lax multiple scattering equations with vector spheroidal/cylindrical wave expansions. The NMM3D results are illustrated at P-, L- and C-band.  NMM3D simulations predict transmissions several times larger than those of RTE/DBA. Much larger transmissions means the microwave can better penetrate vegetation/forest canopy, making it possible to retrieve soil moisture even for large VWC.

Chair: Professor Leung Tsang