Biomass Estimation Using Synergy of ALOS-PALSAR and Landsat Data in Tropical Forests of Brazil

Abstract

Satellite remote sensing technologies are currently tested and suggested as a tool in REDD+ (MRV, Measurement Reporting, and Verification). SAR (Synthetic Aperture Radar) has got an extensive application in the estimation of biomass due to its all-weather capabilities. L band radar signals penetrate the canopy more efficiently when compared to C band. Scientific biomass study using SAR has not been conducted in Para in spite of extensive field datasets being freely available under CMS (Carbon Monitoring System) project. This study aims in using various polarization combinations like HH + HV, HH − HV, HH + HV/HH − HV and vegetation index such as NDVI from the optical data. ALOS-PALSAR and Landsat 7 data acquired over Paragominas in Brazil, where field samples were collected in the form of transects. Regression analysis was performed using backscatter coefficients and field collected Above Ground Biomass (AGB). Semi-empirical model was developed to model AGB using various polarization combinations and NDVI as predictor variables. Combination gave higher R2 value of 0.657 for biomass prediction. Multiple linear regression using NDVI and HH + HV as variables yielded R2 of 0.73 during calibration and 0.363 during validation. There is future scope to use other vegetation indices such as RVI, EVI, etc., along with increased number of samples, which may yield more robust models with acceptable level of accuracy for practical application.