Mapping peat thickness and carbon stock of a degraded peatland in West Sumatra, Indonesia

Tropical peatlands store a large amount of carbon and act as carbon sinks, thus have crucial roles in regulating climate by preventing CO2 emission and enhancing carbon sequestration. However, the conversion of these peatlands for agricultural purposes, such as oil palm plantations, leads to carbon loss and increased CO2 emissions. To effectively protect and manage vulnerable areas, accurate mapping of peatland thickness and carbon stocks is essential, aligning with Indonesia's National Determined Contribution to climate action. This research article aims to assess the status of peatland in Agam-West Pasaman, Western Sumatra, which was cleared between 1990 and 2000 for oil palm cultivation. Digital Soil Mapping (DSM) approaches were employed to map peat thickness, carbon stock and estimate carbon loss resulting from land use change. The study area, spanning 54,000 hectares, was covered by peat-swamp forest in 1989. A grid of 2 km was used for a soil survey, resulting in 134 observation points of peat thickness, water level and subsidence. Various spatial prediction methods, including geostatistics, machine learning (ML) and their combination, were tested to map peat thickness, carbon stocks, subsidence rate and carbon loss. The covariates considered in the analysis were elevation, nearest distance from rivers and Sentinel 1a radar images. The results obtained through 10-fold cross-validations revealed that ordinary kriging exhibited the best performance, with an R2 of 0.44 for peat thickness and 0.39 for carbon stocks. The superior performance of ordinary kriging can be attributed to the severe impact of human activities in the area, which disrupted the clear relationship between peat parameters and environmental covariates. The estimated carbon stock of the area was 107 Mt C (std. dev. 0.143 Mt), while the carbon loss since the establishment of oil palm plantations was estimated to be 19.50 Mt C (std. dev. 0.017 Mt C) based on subsidence data. These findings provide insights into the degradation of the peatland and the magnitude of carbon loss over the past three decades. This information supports informed decision-making and contributes to efforts aimed at preserving and restoring peatlands.