Characterization of forests in Western Sayani Mountains, Siberia from SAR data

Abstract

This paper investigated the possibility of using SIR-C data to map forest types and logging in the mountainous Western Sayani area in Siberia. L and C band HH, HV, and VV polarized images were used in the study. Band ratio and DEM data were used to reduce the topographic effects on radar images. Classification accuracy was estimated for both training and testing sites. The correlation between forest biomass and radar backscattering (LHV) were investigated. A model-based slope correction was applied to LHV backscattering image, and then biomass map was produced from this image using a regression model. The results indicate that multi-channel and multi-polarization SAR data can be used to detect different forest types and logged areas, and provide estimation of total above-ground biomass of forest stands.