Dissolved organic carbon in a drinking water catchment in the western Ore Mountains, Germany: How much, Where from, When and Why – first insights

Increasing concentrations of dissolved organic carbon (DOC) in tributaries threaten the water quality of drinking water reservoirs in Europe and North America. Understanding the key factors influencing DOC dynamics in streams is essential for effective water resource management. This study is part of a concerted effort to determine the major sources of DOC entering a reservoir and to identify the key biogeochemical processes within the terrestrial-aquatic continuum that affect DOC concentrations.

We conducted a four-year multi-scale observational study in a small, heterogeneous catchment (8.5 km²) in the western Ore Mountains, Germany. The research design combined low-resolution (biweekly) measurements of soil water variables (e.g., DOC, pH, Al, Fe) with high-resolution (15-minute) sensor-based monitoring of environmental variables (e.g., temperature, precipitation, soil water content) at representative locations within the catchment. End-member mixing analysis (EMMA) quantified the contributions of peat, forest floor, and mineral soil horizons as sources of DOC, based on previous findings by Charamba et al. (2024), who qualitatively identified these sources within the catchment. In addition, relationships between DOC concentrations and potential explanatory variables were analyzed using Spearman correlations and Random Forest modeling.

In total, 16.5 kg DOC/ha*a were exported from the catchment to the reservoir. EMMA showed that peat soils contributed to about 85 % of the DOC in a tributary adjacent to these soils, corresponding to the highest area-related DOC load of 53 kg/ha*a. Nevertheless, across the entire catchment, mineral soils were the dominant source of DOC, contributing the most to the total DOC load exported to the reservoir (78 %; 13 - 18 kg/ha*a), while forest floors made the smallest contribution. At the temporal level, the contribution of the forest floor to DOC runoff increased under high flow conditions, highlighting the dynamic nature of DOC translocation from different soil sources to stream. Preliminary results of the correlation analysis highlight the influence of soil water chemistry, particularly Al and pH in C-rich horizons, on stream water DOC concentrations. Environmental variables such as precipitation and soil moisture were only moderately correlated with DOC concentrations. Random Forest analysis provided limited insights into key predictors, highlighting the complexity of the catchment and the processes underlying DOC production and translocation. Our results suggest that even bi-weekly sampling intervals may be insufficient to capture the temporal variations in soil processes affecting stream DOC concentrations. The variable time lag between soil processes and their hydrological expression poses a significant analytical challenge. Future research should focus on integrating high-resolution sensor data of DOC concentrations and water fluxes from hydrological monitoring stations. To address the limitations of Random Forest, we will use structural equation modelling (SEM) to refine conceptual models and identify causal relationships. Significant Spearman correlations between DOC and environmental and soil water parameters guide variable selection. The refinement of our conceptual model by SEM will be the basis for process-based modeling to predict the future development of DOC concentrations and fluxes in heterogeneous catchments.