Improvement and extension of a radar forest backscattering model

Radar modeling of mangal forest stands, in the Sundarbans area of Southern Bangladesh, was developed. The modeling employs radar system parameters such as wavelength, polarization, and incidence angle, with forest data on tree height, spacing, biomass, species combinations, and water content (including slightly conductive water) both in leaves and trunks of the mangal. For Sundri and Gewa tropical mangal forests, five model components are proposed, which are required to explain the contributions of various forest species combinations in the attenuation and scattering of mangal vegetated nonflooded or flooded surfaces. Statistical data of simulated images (HH components only) were compared with those of SIR-B images both to refine the modeling procedures and to appropriately characterize the model output. The possibility of delineation of flooded or non-flooded boundaries is discussed.