Soil organic carbon (SOC) mineralization plays a crucial role in the carbon cycle and in determining whether soil is a net carbon source or sink. Multiple factors are influencing SOC mineralization rates. Yet, their spatial and temporal patterns at landscape scale remain elusive so far, especially with respect to the role of subsoils. In the Carbon4D project, we apply a new approach in soil biogeochemistry to model SOC mineralization with a machine learning algorithm. The model will be based on remote sensing data (e.g. soil maps, land use), meteorological data, and measurements of SOC mineralization in different soil depth under respective in-situ temperature and moisture conditions over one year. With these data the algorithm will “learn” the relationship between measured SOC mineralization and various predictor variables. The goal of this project is to provide the first near real-time modelling framework of SOC mineralization rates in all 4 dimensions, which will provide new insights into controlling factors of spatial and seasonal patterns.
We will present first results from the fieldwork with a specific focus on the relationship between moisture, temperature and SOC mineralization as well as sub- versus topsoil.