Environmental Determinants of Peatland Degradation in Central Kalimantan: A GIS-Based Remote Sensing Analysis

Sammanfattning

Tropical peatlands in Central Kalimantan, Indonesia, represent important terrestrial carbon stores and support highly diverse tropical rainforest ecosystems with significant ecological, hydrological, and climatic functions. However, decades of drainage development, deforestation, plantation expansion, land-use change, and recurrent fire have contributed to extensive peatland degradation across the region. Large-scale hydrological disturbance following the Mega Rice Project initiated in the 1990s further accelerated peat drying, vegetation degradation, forest loss, and increased fire vulnerability. This study analysed peatland degradation in south-eastern Central Kalimantan between 2017 and 2025 using remote sensing and GIS-based spatial analysis. Satellite-derived indicators including the Normalized Difference Vegetation Index (NDVI), Normalized Difference Water Index (NDWI), VIIRS active fire detections, annual forest loss data, and supplementary Land Surface Temperature (LST) information were integrated with environmental variables related to hydrology, terrain, drainage infrastructure, and land use. Particular emphasis was placed on NDWI as a proxy indicator of peatland moisture conditions associated with hydrological disturbance. Spatial and statistical analyses were performed using Pearson correlation, hotspot analysis, Ordinary Least Squares (OLS), Geographically Weighted Regression (GWR), Random Forest (RF), and weighted spatial overlay techniques. The results demonstrated that peatland degradation is spatially heterogeneous, with major degradation hotspots concentrated in the southern and central peatland regions where vegetation loss, moisture reduction, forest loss, and fire occurrence overlapped. Areas located near drainage canals generally exhibited stronger drying trends and higher levels of disturbance than areas located further from drainage infrastructure. Pearson correlation, regression analyses, and Random Forest modelling consistently identified distance to drainage canals as one of the most influential variables associated with peatland degradation, whereas Topographic Wetness Index (TWI) showed weaker and less consistent relationships. GWR further revealed substantial spatial variation in the environmental controls of degradation, indicating that peatland responses differ across the landscape. Although vegetation and moisture indicators showed relatively strong relationships, fire occurrence appeared to be influenced by additional climatic and anthropogenic factors beyond those included in the analysis. Overall, the study highlights the critical role of hydrological disturbance in tropical peatland degradation and emphasises the importance of long-term hydrological restoration and sustainable peatland management, particularly in heavily drained peatland areas. The integration of remote sensing, GIS, and spatial modelling techniques provides a useful framework for identifying degradation hotspots and supporting future peatland restoration and conservation strategies in tropical regions.