Modelling the Space-Time Dynamics of Soil Organic Carbon Stock: From Point to Landscape Scale by Yue zhou

Soil organic carbon (SOC) is a key component of ecosystems, influencing soil health, fertility, and the carbon cycle. Intensive agricultural practices have led to SOC losses, increasing atmospheric carbon levels. This study explores the spatio-temporal variability and sequestration potential of SOC in cropland under different managementpractices. Our hypothesis is that remote sensing proxies can serve as input for process-based models, identifying management practices across fields and estimating SOC dynamics. The study aims to: (1) provide spatial soil data, (2) develop management practice datasets, (3) evaluate the effectiveness of readily available data by comparing it with long-term experimental data, and (4) apply these methods to landscape case study. Our findings show that incorporating human activity-related covariates improves SOC distribution predictions. We also successfully predicted agricultural management practices (e.g. cover crop, tillage types) via remote sensing. We integrated all spatial data into the RothC model running at a landscape level. Effective management strategies, such as reduced tillage and longterm cover crops, have the potential to transform fields from carbon sources into carbon sinks. Overall, this thesis develops a comprehensive approach to simulate SOC variability and sequestration potential in cropland and contributes to understanding carbon stocks in the Walloon region of Belgium. The methodology also establishes carbon baselines essential for greenhouse gas trading schemes, creating a scalable, reliable system for SOC monitoring, reporting, and verification.