Integrating parsimonious models of hydrological connectivity and soil biogeochemistry to simulate stream DOC dynamics

To improve understanding and prediction of dissolved organic carbon (DOC) sources and fluxes in northern peat-dominated catchments, we present the development and application of a parsimonious tracer-aided rainfall-runoff model coupled with a biogeochemistry subroutine able to concurrently simulate streamflow and DOC dynamics. The modeling approach which included quantitative assessment of associated uncertainties was conditioned by geochemical tracers which discriminate dominant water sources. Integration of DOC was predicated on statistical time series models which identified air temperature and streamflow as the key proxies that capture DOC supply and transport processes in two upland catchments in Scotland, UK. Conceptualizing the nonlinear partitioning of quick near-surface and slower groundwater runoff sources in combination with a DOC mass balance resulted in a coupled, low-parameter mechanistic model. Model tests showed mostly sensitive parameters and reasonable simulation results with seasonally controlled DOC supply and event-based DOC transport. Transport is facilitated even for smaller events by overland flow from saturated histosols connected to the stream network. However, during prolonged dry periods, near-surface runoff switches off and stream DOC is dominated by low concentration groundwaters. Furthermore, the model was able to explain subtle differences in DOC dynamics between the two catchments mainly reflecting the distribution of saturated soils and available storage. We conclude that tracers and statistical time series models can successfully guide the development of parsimonious yet structurally consistent water quality models. Parsimonious models provide tools for estimating DOC dynamics and loads with reduced uncertainty and potentially greater transferability.