The Potential of Fully Polarized ALOS-2 Data for Estimating Forest Above-Ground Biomass

SAR data have a longer wavelength and stronger penetrating power compared with traditional optical remote sensing. Therefore, are more suitable for the estimation of above-ground biomass (AGB) forests. This study was aimed at evaluating sensitivity L-band full polarization to AGB. were improved estimate saturation point produced by AGB, found be estimating wide range extracted backscattering coefficients, decomposition variables, terrain factors. New parameters constructed from these their performance in predicting AGB evaluated. Significant variables added multivariate linear model. A statistical analysis showed presence multicollinearity between variables. ridge regression, random forest method (RF), principal component (PCA) introduced solve problem collinearity. In all three methods, regression model low, reaching it 150 t/ha. Better accuracy obtained RF No obvious incident detected established using analysis. could attributed low levels observed our area. provided accurate results (adjusted r2 = 0.90 rmse 14.24 t/ha), indicating that potential Additionally, models selected this study, being helpful combining various parameters. The achievement synergy among