Estimation of above-ground biomass using machine learning approaches with InSAR and LiDAR data in tropical peat swamp forest of Brunei Darussalam

Forest above-ground biomass (AGB) is one of the critical measures of forest resources. Therefore, it is crucial to identify a reliable method to estimate the AGB, especially in the tropics, where forest ecosystems are exposed to several depleting factors, including deforestation, climate change and replacing natural forests with palm oil tree plantations. We investigated the digital elevation data over the forest and uses an artificial intelligence-based approach to develop a method for quick and cost-effective assessment of the AGB. The study was conducted in the tropical peatland rainforest of Brunei Darussalam. The Shuttle Radar Topography Mission (SRTM) elevation data product and Light Detection and Ranging (LiDAR) digital elevation data were used. A linear regression (LR) model and three different machine learning (ML) algorithms, i.e., Random Forest (RF), Artificial Neural Network (ANN) and Support Vector Machines (SVM), were tested and compared. As model inputs, the SRTM elevation and distance from the peat dome’s center, a feature of a peatland swamp forest, were used. ML methods were trained on the samples taken from the LiDAR elevations. The validation results showed that the SVM was the best method to predict AGB in the study area with R2 = 0.70, RMSE = 83.65 Mg ha-1, and MAE = 74.43 Mg ha-1, which in relative terms corresponds to approximately 6% of the AGB of the forest of interests. This study demonstrated the potential of ML algorithms in AGB estimation based on canopy height derived from the InSAR-based DEM in tropical forests.