Analysis of the effect of crown structure changes on backscattering coefficient using modeling and SAR data

Abstract

Stand level forest canopy structure such as the size, density, and distribution of the branches and leaves may have a strong effect on radar backscatter. In this study, several broad-leaf (birch) stands and needle stands (larch) with different growing stages and different canopy structures are established using parametric and stochastic L-system. Stands with 10, 30, 50 years and with 10, 40, 100 years, which correspond to young, mid-age and mature birch and larch stands respectively, are simulated according to field measurements. To stands with the same age, the above ground biomass is almost the same. Then different 3D birch and larch crown architectures faithful to the real stand are generated using L-system, which provide realistic and detailed canopy biometric data for radar model. The radar model used here is an improved 3D forest radar backscatter model based on Radiative Transfer Theory, which considers tree crown distribution and multiple scattering from canopy during backscattering calculation In this paper, total 60 stands with 30×30m area, namely three stand ages, ten canopy structures and two species, are simulated and analyzed at C-, L-band with different polarizations. Simulation results show that the backscatter coefficient is sensitive to the canopy structure, particularly at C-band and L-band HV polarization. The discrepancy between birch and larch stands with the same tree age is distinct. The crown structure effect to the C-band is more obvious than L-band because of its short wavelength. Then the simulation results of L-HH, C-HH and C-HV polarizations are compared with JERS-1, ASAR data of Changqing forest farm located at DaXinAnLing, northeast of China, which shows good correspondence.