Aboveground biomass estimation using multi-sensor data synergy and machine learning algorithms in a dense tropical forest

Abstract

Forest aboveground biomass (AGB) is an important factor for tracking global carbon cycle to tackle the impact of climate change. Among all available remote sensing data and methods, Synthetic Aperture Radar (SAR) data in combination with decision tree based machine learning algorithms has produced favourable results in estimating higher biomass values. Suitability of this method for dense tropical forests has not been properly checked with an adequate number of studies. In this study, aboveground biomass was estimated for two major tree species, Shorea robusta, and Tectona grandis, of Katerniaghat Wildlife Sanctuary, a tropical forest situated in northern India. Aboveground biomass was estimated by combining C-band SAR data from Sentinel-1A satellite, texture images generated from Sentinel-1A data, vegetation indices produced using Sentinel-2A data and ground inventory plots. Decision tree-based machine learning algorithms were used in place of parametric regression models for establishing a relationship between fields measured values and remotely sensed parameters. Using random forest model with a combination of vegetation indices with SAR backscatter as predictor variables shows the best result for S. robusta forest, with a coefficient of determination value of 0.71 and an RMSE value of 105.027 t/ha. In T. grandis forest best result can be found in the same combination but for stochastic gradient boosted model with a coefficient of determination value of 0.6 and an RMSE value of 79.45 t/ha. This study shows that Sentinel series satellite data has exceptional capabilities in estimating dense forest AGB and machine learning algorithms can be very helpful to do so.