Probabilistic Mapping of Tropical Peatlands in the Lower Ogooué Delta in Gabon Using a Multicriteria Approach and Remote Sensing

Peatlands are major global carbon reservoirs and play a vital role in climate regulation and ecosystem functioning. However, their spatial distribution remains poorly understood in tropical regions, particularly in Central Africa, given the difficulties in accessing these sites and the high costs of large-scale field surveys. Based on the hypothesis that combining data from multiple sources can provide indications of peat presence to optimise research in the Lower Ogooué Delta basin, this study proposes a spatial modelling framework aimed at estimating the probability of peatland occurrence in the Lower Ogooué Delta in Gabon, using multi-source remote sensing data that integrates optical, radar, topographic, and structural information. Specifically, it aims to produce the explanatory variables likely to govern the presence of peat in the Lower Ogooué, as well as a probability map of its occurrence. Spectral indices derived from Landsat 8 (NDVI and NDWI), radar parameters from ALOSPALSAR (HH backscatter and RFDI) across different seasons, as well as topographic variables from the Digital Elevation Model (DEM), such as the Topographic Wetness Index (TWI), altitudes and slopes, as well as soil and chronostratigraphic data, were used as proxies in a multi-criteria analysis based on an Analytic Hierarchy Process (AHP). The resulting probability map highlights vast areas of 470,834.1 hectares of low-lying terrain, permanently waterlogged, with a high probability of peatland occurrence. This reproducible and transferable methodology constitutes a relevant tool and a milestone for the mapping of tropical peatlands and for decision-making regarding biodiversity conservation and climate change mitigation.