Discriminating runoff source areas in a small forested catchment using a multi-tracer approach

Rising trends in the concentrations of dissolved organic carbon (DOC) exported from catchments can be observed in many places around the world, including in the catchment of the river Große Ohe (19 km²) in the Bavarian Forest National Park (Germany). During flood events, DOC is mainly exported via hydrological pathways due to the activation of loaded pre-event water, affecting for example aquatic ecosystems and the use of surface water for drinking water supply.

In previous work, the uppermost soil horizon, enriched with DOC through microbial decomposition of organic matter, was identified as the main source of DOC. However, more recent data show that soil properties play an important role in flood genesis, but are minor for DOC export. More important are the contributions of existing wet areas in the catchment, which provide good conditions for the solution of organic carbon. During rain events, these wet areas become connected to stream runoff and are the main source of DOC exports from the catchment.

We sampled and analysed three flood events with different hydrological conditions and used the results for multi-tracer flow separation and end ember mixing analyses. To do this, we collected water samples from the stream runoff, precipitation water, groundwater and soil water from various locations in the catchment and analysed them for DOC, SiO₂, K⁺, Fe²⁺ and other cations and anions. Observations of the fingerprints of DOC and K⁺ show significant differences between water from the uppermost soil horizons (high DOC- and K⁺-concentrations) and the wet areas (high DOC-, low K⁺-concentrations). K⁺ results from the decomposition of organic matter and is mainly present in stemflow and water from the uppermost soil horizons.

During the flood event, we observe a significant correlation between instream DOC-concentration and runoff, but no correlation between instream K⁺-concentration and runoff. Nevertheless, instream DOC- and K⁺-concentrations increase at the beginning of the event, leading to the assumption that rapid surface runoff contributes to stream runoff in the first phase of the flood event. As the event progresses and the discharge continues to rise, K⁺ decreases rapidly, while DOC mainly follows the pattern of the hydrograph. In addition, we consistently observed a delay between the peaks of discharge and DOC concentration of approx. 1.5 - 2 hours. Both facts indicate that these are the fingerprints of the wet areas, which, depending on the hydrological conditions, take some time to “fill up and overflow”. The water of these wet areas then gets connected to the stream and leads to a delayed DOC release.

The findings are used to develop a predictive model for runoff generation and DOC mobilization. Since the identified processes are closely linked to the characteristics of the catchment (topography, hydrotopes, ...), the results have to be compared with other catchments in order to generalize them and ensure the transferability of the model to similar catchments. In the event of success, such forecast models are important tools, e.g. for drinking water suppliers who have to react quickly to changing DOC concentrations to ensure water quality standards.