Groundwater travel times predict DOC in streams and riparian soils across a heterogeneous boreal landscape.

Dissolved organic carbon (DOC) in surface waters is an important component of the boreal landscape carbon budget and a critical variable in water quality. A dominant terrestrial DOC source in the boreal landscape is the riparian zone. These near stream areas play a key role in regulating DOC transport between land and aquatic ecosystems. The groundwater dynamics at this interface have been considered a major controlling variable for DOC export to streams. This study focuses on the regulating role of groundwater levels and mean travel times (MTT) on riparian DOC concentrations and, subsequently, stream DOC. This is done by comparing them as explanatory variables to capture the spatial and intra-annual variability of the stream and riparian groundwater DOC. We used a physically based 3D hydrological model, Mike SHE, to simulate DOC concentrations of the riparian zones for 14 sub-catchments within the Krycklan catchment (Sweden). The model concept assumes that DOC concentrations will be higher in groundwater moving through shallow flow paths. In the model, this can be linked to the position of the groundwater table at a point of observation or the travel time, which will generally be shorter for water that has travelled through shallow and more conductive soil layers. We compared the results with both observed stream and groundwater concentrations. The analysis revealed that the correlation between modelled and observed annual averages of stream DOC increased from r = 0.08 to r = 0.87 by using MTT instead of groundwater level. MTT also better captured the observed spatial variability in riparian DOC concentrations and more successfully represented seasonal variability of stream DOC. We, therefore, suggest that MTT is a better predictor than groundwater level for riparian DOC concentration because it can capture a greater variety of catchment heterogeneities, such as variation in soil properties, catchment size, and input from deep groundwater sources.