Can the heterogeneity in stream dissolved organic carbon be explained by contributing landscape elements?
- 1Department of Forest Ecology and Management, Swedish University of Agricultural Sciences, S901 83, Umeå, Sweden
- 2Department of Biological Sciences and Department of Geography, University of Cincinnati, 312 College Drive, Cincinnati, OH 45221, Ohio, USA
- 3Centre for Ecology and Hydrology, Deiniol Road, Bangor, LL57 2UP, UK
Abstract. The controls on stream dissolved organic carbon (DOC) concentrations were investigated in a 68 km2 catchment by applying a landscape-mixing model to test if downstream concentrations could be predicted from contributing landscape elements. The landscape-mixing model reproduced the DOC concentration well throughout the stream network during times of high and intermediate discharge. The landscape-mixing model approach is conceptually simple and easy to apply, requiring relatively few field measurements and minimal parameterisation. Our interpretation is that the higher degree of hydrological connectivity during high flows, combined with shorter stream residence times, increased the predictive power of this whole watershed-based mixing model. The model was also useful for providing a baseline for residual analysis, which highlighted areas for further conceptual model development. The residual analysis indicated areas of the stream network that were not well represented by simple mixing of headwaters, as well as flow conditions during which simple mixing based on headwater watershed characteristics did not apply. Specifically, we found that during periods of baseflow the larger valley streams had much lower DOC concentrations than would be predicted by simple mixing. Longer stream residence times during baseflow and changing hydrological flow paths were suggested as potential reasons for this pattern. This study highlights how a simple landscape-mixing model can be used for predictions as well as providing a baseline for residual analysis, which suggest potential mechanisms to be further explored using more focused field and process-based modelling studies.