INTEGRATION OF LANDSAT 8 IMAGERY AND CART MODEL FOR ESTIMATING SOIL ORGANIC CARBON IN DAK LAK PROVINCE

The storage potential of Soil Organic Matter (SOM) is critical for reducing CO2 emissions. Recent advancements in remote sensing and machine learning have enabled significantly more precise prediction of SOM compared to traditional soil surveys. This study aims to examine the integration of Landsat 8 imagery and Classification And Regression Tree (CART) in estimating SOM in Dak Lak province. Landsat 8 imagery is utilized to extract spectral indices covariables that relate to SOM. The CART model was then applied to estimate SOM based on the covariables. A representative dataset of soil samples from various sites across the province was divided into training and validation subsets to evaluate the performance of the CART-based prediction. The validation result of the CART indicates that the RMSE and standard error of the model are 1.323 and 0.165, respectively. The estimation result indicates that the total amount of soil organic carbon is approximately 70.22 million tonnes of carbon in the topsoil of the province. The study provides baseline information for future estimates and carbon monitoring efforts in the topsoil of the province.