With climate change and constant declines in global soil health, it is imperative that regenerative land management practices are assessed for their capacity to curb trends of land degradation and desertification. This project will assess the ability of a variety of regenerative practices to sequester soil organic carbon (SOC). The regenerative practices selected will be abstracted into their functional properties to be integrated effectively into SOC models. SOC models will be programmatically adapted with machine learning methods, trained at least partially on remote sensing data to model drought affected soils regionally and/or globally. The models developed will be compared with already existing process-based SOC models and ground truth data from collaborating stakeholders to assess accuracy and applicability of the models.