Assessment of Gayo agroforestry coffee characteristics and carbon stock potential in Mumuger social forestry area, Central Aceh Regency

Abstract

Coffee agroforestry has become a nature-based solution for controlling climate change impacts while, providing access to sustainable forest utilization for rural farmers, especially for the governance of social forestry policies in Indonesia. Ecosystem services established in coffee agroforestry provide high-carbon stocks that can reduce greenhouse gas emissions to the atmosphere. Statistical and spatial information on carbon stocks in coffee agroforestry in the Sumatran tropical forest region, especially above ground carbon (AGC), is still very limited. Therefore, this study aims to assess the available carbon stocks in Gayo coffee agroforestry in Mumuger Social Forestry Area, Central Aceh Regency, with the help of combining multi-source data (Landsat-Sentinel-NICFI imagery) and involving machine learning algorithms in estimation modelling. The agroforestry carbon stock distributed in the study area has 72.31 ± 48.46 Mg C ha^-1, which is dominated by the Leucaena-coffee agroforestry combination. There are 13 species at the overstory level that contribute carbon stock values up to 198 Mg C ha^-1. Based on modelling tests of carbon stock estimation using 37 predictors, the two best machine-learning algorithms were RF and SVM, with R² reaching 0.83 and 0.85. Carbon stock quantification information and remote sensing machine learning approaches play a strategic role in studying the impacts of agroforestry systems and as a policy evaluation in social forestry governance that can contribute to climate change mitigation.