Optimising carbon fixation through agroforestry: Estimation of aboveground biomass using multi-sensor data synergy and machine learning

• DOI: 10.1016/j.ecoinf.2023.102408
• Publication Date: 2023-12-03T18:40:45Z
• Authors: R.K. Singh; C.M. Biradar; M.D. Behera; A.J. Prakash; P. Das; M.R. Mohanta; G. Krishna; A. Dogra; S.K. Dhyani; J. Rizvi
• Keywords: Enhanced vegetation index

As agricultural land expansion is the primary driver of deforestation, agroforestry could be an optimal use strategy for climate change mitigation and reducing pressure on forests. Agroforestry a promising method carbon sequestration. With recent advancements in geospatial data science technology, ability to predict aboveground biomass (AGB) assess ecosystem services rapidly expanding. This study was conducted Belpada Block Balangir, Odisha, forest-dominated region eastern India. We recorded species occurrence measured plant parameters, including Circumference at Breast Height (CBH), height, geolocation, 196 plots (0.09 ha) intervention sites noted tree species. used Sentinel-1 Sentinel-2 multi sensor achieve synergy AGB estimation. Three machine learning models were used: Random Forest (RF), Support Vector Machine (SVM), Artificial Neural Network (ANN). The RF model exhibited highest level prediction accuracy (R2 = 0.69 RMSE 17.07 Mg/ha), followed by ANN 0.63 19.35 SVM 0.54, 21.97 Mg/ha. spectral vegetation indices that are (Normalized Difference Vegetation Index (NDVI), Soil-Adjusted (SAVI), Enhanced (EVI), Modified Simple Ratio (MSR), (MSAVI), (DVI), SAR backscatter values, found important variables prediction. findings revealed interventions plantations resulted average stock increase 15 Mg/ha over five years area. Plant Value (PVI), which indicates importance local economy storage, showed Tectona grandis dominant with PVI value (88.35), Eucalyptus globulus (56.87), Mangifera indica (53.75), Azadirachta (15.45). approach enables monitoring efforts systems, thereby promoting effective management strategies.