Geospatial multi-model Carbon Sequestration and Green House Gas Emission framework

Abstract

Climate Change mitigation efforts require accurate assessment of local GHG emissions and carbon sequestration (CS) at less than 1km2 spatial resolution with flexibility to adjust based on local model and user input. Current approaches of GHG emissions rely on generic and coarse models not amenable to capture the local variations necessary to make these data useful for operational decisions. We demonstrate a geospatial framework where CS in soil and forest is combined with emission of CO2, CH4 and N2O from land management. The framework uses machine learning techniques on satellite and ground measurement data for land-cover classification, data imputation and multi-model validation, allowing estimation of CS and GHG emissions at a farm level.