Assessing carbon and economic loss from oak decline in Iran’s Zagros forests using Sentinel-2 and machine learning

Over recent decades, the decline of oak trees, which has emerged as the most significant challenge to the Zagros forests (ZFs), poses serious threats to carbon storage (CS) and economic valuation, jeopardizing the ecological and economic sustainability of this vital ecosystem. This study utilizes remote sensing techniques to evaluate the impact of oak tree decline on CS and its economic value in the ZFs of Kuhdasht, Lorestan Province, Iran. We used Sentinel-2 data (15 June 2021). This study measures in situ forest above-ground biomass (FAGB) by analyzing 224 systematically selected 20 m × 20 m plots and recording the average crown diameter (ACD) of each tree. We employed multivariate linear regression (MLR), artificial neural networks (ANN), and K-nearest neighbors (KNN) to model FAGB. Next, we used 50% of the AGB map value to calculate the CS map. Subsequently, we considered 3.66 t of carbon dioxide (CO2) for each CS ton. Additionally, we calculated the economic value of CS using an average rate of €49/t of CO2 sequestered. Results indicate that MLR based on the normalized difference vegetation index (NDVI) (R2 = 0.73, rRMSE = 21.88) was the most effective model for the area studied. Moreover, Sentinel-2 imagery proved highly efficient in estimating FAGB in the degraded zones of ZFs. The economic valuation estimates the value of carbon absorption at €53 million for a healthy forest and €37 million for a severely declining one. The study found that as forest decline increases, its economic value decreases, with the highest values in healthy areas and the lowest in the most declining regions.