Assessing and Modelling Soil Organic Carbon Dynamics in a Mining Spill Restoration Area from 1998

Reliable quantification of soil organic carbon (SOC) change is essential for evaluating long-term restoration trajectories and constraining the potential contribution of restored soils to atmospheric CO2 mitigation. However, long-term SOC monitoring based exclusively on field measurements is often limited by data scarcity, spatial heterogeneity and the high cost of repeated sampling. This study assesses the potential of combining field observations, remote sensing-derived proxies and process-based modelling to simulate SOC dynamics in the Guadiamar Green Corridor, Seville, Spain, a restored area affected by the 1998 Aznalcóllar mining spill. SOC dynamics were simulated using the RothC model under a set of boundary condition scenarios, in which field-measured or locally observed input variables were progressively replaced by alternative data sources. The reference scenario was based mainly on field SOC and clay measurements and local meteorological records, while the remaining scenarios tested the use of NDVI-derived carbon inputs, satellite-derived and reanalysis-based climate variables, and SoilGrids-derived estimates of SOC and clay content. The simulations were evaluated against field observations collected over the restoration period. Overall, RothC reproduced the general direction of the surface-layer SOC decline observed in the monitoring plots, although most scenarios did not fully capture the magnitude of SOC loss measured in the field. Differences among boundary conditions highlighted the sensitivity of RothC to specific input variables, particularly initial SOC, clay content and evapotranspiration. Scenarios based on alternative soil datasets showed that global soil products can provide useful information where field data are limited, but their performance depends strongly on their agreement with local soil conditions. Similarly, the replacement of local climatic data with remote sensing or reanalysis products affected model outputs, especially when evapotranspiration was substituted. These results indicate that satellite-derived products, climate reanalysis and open-access environmental datasets can complement field measurements in SOC modelling, particularly in data-scarce restoration contexts. However, their use requires careful validation, as model accuracy depends on the variable being replaced, the quality of the proxy dataset and the calibration of RothC to plot-specific conditions.