Multivariable modelling and carbon stored estimation in peatland: Constraints from peat properties, organic petrology, and remote sensing analysis in the Kapuas Wetland, Indonesia

Carbon stores in peatlands are correspondence of peat distribution and amount of total carbon; however, various interacting factors are involved affecting this process seem to be more complicated and difficult to generalize. This study improved the estimation of the amount of carbon stored in Kapuas Wetland, West Kalimantan using combined remote sensing analysis (UAV, Landsat 8 and 9 imageries and DEM) constraints with in-situ peat thickness measurements, peat properties and organic petrology. We demonstrated a novel finding on distribution and topography of domed peatland as well as their recent conditions: pristine and altered peatlands. Two types of peats were identified, sapric and hemic peats. Peat types correspond to their properties and control the amount of total carbon. The total carbon stored in the Kapuas Wetland contributes to a carbon pool of 221.08 MtC. The difference between two multivariable modelling approaches: a) elevation and peat types-based model, and b) peat thickness and river-based model in this study is only 1.59%, showing a good level of precision, while single variable modelling approach shows difference around 6.42% - 8.14%, compared to multivariable modelling approaches. Thus, this study presents an estimated modelling error of ±7%, making total carbon stored ranging from 205.6 MtC – 236.56 MtC, with 221.08 MtC as the best estimation. Our study suggests that each peat type has unique characteristics in both physical, chemical and biological properties of the peat. These differences will influence the amount of carbon, which should be considered for estimating the amount of carbon stored. • Surface and subsurface mapping on peatland to estimate carbon stored • Peat characteristics control the amount of total carbon • Identification of degraded and preserved peatland using remote sensing analysis • Proposed carbon pool estimation and volumetric modelling showing a good level of precision