Carbon Stock Assessment Using Sentinel-2 and Forest Inventory: A Case Study of Changa Manga Forest, Punjab

Abstract

Forests play a significant role in the adaptation and mitigation of climate change and act as a major carbon sink in the global carbon cycle. A successful mitigation strategy for protecting and conserving forests is necessary to reduce rising emissions. In the past, no work has been carried out on assessing irrigated forests. The objective of this study was to estimate soil organic carbon above & below ground biomass in the Changa Manga man-made irrigated forest, using forest inventory, remotely sensed data, and forest covers temporal analysis. Sentinal-2 satellite images were used for above-ground biomass (AGB) and carbon stock assessment, while Landsat imageries were used in forest cover. The change in forest cover was evaluated using Landsat satellite data and supervised classification using the Support Vector Machine (SVM) algorithm. Data from the forest inventory, including diameter at breast height (DBH), height, and soil samples, were collected through field surveys to determine AGB and carbon stock. These data sets were analyzed using allometric equations of different tree types. Systematic grid sampling was used to collect soil samples from mapping soil carbon. In the forest cover area, below-ground biomass (BGB) ranged from 121.72 t/ha to 31.64 t/ha, with a mean total biomass of 153.37 t/ha. The estimated mean above-ground and below-ground carbon stocks were 57.20 t/ha and 8.22 t/ha, respectively. The research area's total carbon content (AGC + BGC + Soil Carbon) was 4335.25 t/ha. Approximately 908.73 ha (17%) of the forest area has been deforested in the past 31 years, with an annual rate of deforestation of 3.22%. Sentinel-2 VIS were evaluated in the estimation of AGB, and a linear regression model was used to analyze the association between various VIS and AGB. Maps of anticipated biomass were created from regression models. Due to its high R-square value of 0.61, low RMSE value of 19.48 t/ha, and low p-value of less than 0.01, the study concluded that the AGB and EVI were the best. The findings demonstrate that the rapid estimation of AGB over a large, forested area can be managed using satellite remote sensing data.