RADIOMETER AND RADAR REMOTE SENSING OF SOIL MOISTURE AND VOD OF CROPS WITH ROW STRUCTURES WITH NUMERICAL SOLUTIONS OF 3D MAXWELL EQUATIONS

T. Liao¹, Z. Huang², E. Koc², L. Tsang², J. Jeong³

¹Department of Electrical Engineering, National Taipei University of Technology, Taipei, Taiwan, ²Radiation Laboratory, Department of Electrical Engineering and Computer Science, University of Michigan, Ann Arbor, MI, USA, ³Temasek Laboratories, National University of Singapore, Singapore, Singapore

The integration of passive radiometer data from the Soil Moisture Active Passive (SMAP) mission and active radar data from the recently launched NASA-ISRO Synthetic Aperture Radar (NISAR) mission enables new opportunities for advancing vegetation and soil moisture studies. SMAP continues to provide vegetation optical depth (VOD) retrievals alongside soil moisture estimates, while NISAR offers dual-polarization (HH and HV) L-band radar observations at 10 m resolution, yielding an unprecedented dataset for capturing vegetation structure and dynamics. Together, these complementary missions enhance sensitivity to soil moisture, canopy properties, and scattering mechanisms.

Vegetation scattering is strongly influenced by the distribution and orientation of plant components. In agricultural fields, stalks are generally more uniformly distributed, while leaves exhibit greater variability in space and orientation. The row structure of crops such as corn further enhances anisotropy, complicating retrieval algorithms. Traditional incoherent models emphasize stalks as the dominant contributors, especially at VV polarization. However, recent analyses reveal that leaves, particularly under conditions of high vegetation water content, significantly affect the scattering response at horizontal polarization (HH). This polarization sensitivity underscores the need to better represent leaf contributions in VOD retrieval frameworks.

Over the past several decades, radiative transfer (RT) models have been widely applied, typically representing branches as cylinders and leaves as disks. However, these RT models rely on several ad hoc assumptions, including incoherent independent scattering and constant densities for the cylinders and disks. In recent years, we have pursued numerical 3D solutions of Maxwell’s equations using the Fast Hybrid Method (FHM). The accuracy of FHM has been validated through comparisons with results from commercial software packages such as FEKO, which are widely used in electromagnetic simulations. Importantly, FHM has demonstrated far greater efficiency, requiring up to two orders of magnitude less CPU time and memory than commercial software. The method integrates full-wave electromagnetic principles with computational efficiency by representing each plant as a single scatterer, extracting plant-level T-matrices with commercial solvers, and applying a hybrid multiple scattering scheme for fields with row structures. This approach captures detailed scattering, absorption, and bistatic responses while reducing CPU and memory requirements by more than two orders of magnitude compared to conventional solvers.

Crops such as corn and rice exhibit row structures in which the spacing along the row direction is much larger than that in the orthogonal direction. As a result, crops with row structures display anisotropic vegetation optical depth (VOD). Applied to corn fields, three representative scenarios—stalk-only, leaves-only, and combined plants—show that while stalks dominate at VV polarization, leaves contribute substantially at HH, particularly in bistatic scattering. These results demonstrate the critical roles of crop geometry, leaf distribution, and polarization sensitivity in shaping remote sensing signals, confirming the importance of integrating SMAP and NISAR observations for improved vegetation and soil moisture retrievals.