Quantifying aboveground biomass of small urban remnant forest in South-East Queensland from global ecosystem dynamic investigation laser footprints and sentinel-2 imagery

Abstract

Estimating the biomass of urban forests is vital for evaluating their contribution to global climate mitigation efforts. Field-based sampling methods are unsuitable due to the costs and labour involved. In addition, regular field-based expeditions into urban forests could disturb these ecosystems. Although LiDAR sensors on drones could provide an alternative for capturing structural information about the forests, the proximity of these forests to urban populations makes drone surveys unsuitable. In the current research, we assess the potential of combining Global Ecosystems Dynamics Investigation (GEDI)-estimated aboveground biomass density (AGBD) with satellite-derived spectral indices, including Normalized Difference Vegetation Index (NDVI), Normalized Difference Built-up Index (NDBI), Normalized Difference Red Edge Index (NDRE) and Bare Soil Index (BSI), together with Digital Surface Model and slope, to predict aboveground biomass (AGB) for a remnant urban forest in South-East Queensland, Australia. We developed a Random Forest Regression model that accurately predicted AGB, with R² value of 0.81 and Root Mean Square Error of 46.75Mg/ha. We further estimated the total biomass of the forest to be approximately 35,981.6 Mg/ha for a 136-hectare study area, which is 95% of a recent estimate based on field-based allometric modelling. The global availability of the GEDI product and satellite-derived data makes this method applicable to many forests worldwide.