Joint Study of Above Ground Biomass and Soil Organic Carbon for Total Carbon Estimation using Satellite Imagery in Scotland

Abstract

Land Carbon verification has long been a challenge in the carbon credit market. Carbon verification methods currently available are expensive, and may generate low-quality credit. Scalable and accurate remote sensing techniques enable new approaches to monitor changes in Above Ground Biomass (AGB) and Soil Organic Carbon (SOC). The majority of state-of-the-art research employs remote sensing on AGB and SOC separately, although some studies indicate a positive correlation between the two. We intend to combine the two domains in our research to improve state-of-the-art total carbon estimation and to provide insight into the voluntary carbon trading market. We begin by establishing baseline model in our study area in Scotland, using state-of-the-art methodologies in the SOC and AGB domains. The effects of feature engineering techniques such as variance inflation factor and feature selection on machine learning models are then investigated. This is extended by combining predictor variables from the two domains. Finally, we leverage the possible correlation between AGB and SOC to establish a relationship between the two and propose novel models in an attempt outperform the state-of-the-art results. We compared three machine learning techniques, boosted regression tree, random forest, and xgboost. These techniques have been demonstrated to be the most effective in both domains. This research makes three contributions: i) Including Digital Elevation Map (DEM) as a predictor variable in the AGB model improves the model result by 13.5 % on average across the three machine learning techniques experimented, implying that DEM should be considered for AGB estimation as well, despite the fact that it has previously been used exclusively for SOC estimation. ii) Using SOC and SOC Density improves the prediction of the AGB model by a significant 14.2% on average compared to the state-of-the-art baseline, which strengthens our experiment results and suggests a future research direction of combining AGB and SOC as a joint study domain. iii) Including AGB as a predictor variable for SOC improves model performance for Random Forest, but reduced performance for Boosted Regression tree and XG Boost, indicating that the results are specific to ML models and more research is required on the feature space and modelling techniques. Additionally, we propose a method for estimating total carbon using data from Sentinel 1, Sentinel 2, Landsat 8, Digital Elevation, and the Forest Inventory.