Data-driven management solutions for carbon sequestration in grasslands

Grasslands cover 20% of Flanders and have the potential to store over 100 tons of carbon per hectare in the soil, making them priority habitats for carbon capture and storage. However, the efficacy of grasslands in mitigating greenhouse gas emissions depends on their context and management. The high costs of carbon monitoring have left it largely unknown how much carbon is stored in grasslands in Flanders, and which grasslands function as a carbon sink or source, thus hindering their potential for climate-change mitigation. While sensors can offer a cost-effective means of data collection, issues of practicality in real-world scenarios and integrating sensor data into existing soil carbon dynamics models have impeded their implementation in soil carbon monitoring. To address this, my research aims to integrate sensor data with existing soil data layers into new Soil Organic Carbon (SOC) digital soil maps through data fusion and machine learning. Specifically, I aim to (1) evaluate if grasslands in Flanders are currently carbon sources or sinks, (2) optimize methods for assessing carbon sequestration in real-world scenarios and (3) evaluate two common management strategies for grasslands and their potential to maximize stable carbon sequestration in order to combat climate change