Nonlinear and threshold‐dominated runoff generation controls DOC export in a small peat catchment

Abstract

AbstractWe used a relatively simple two‐layer, coupled hydrology‐biogeochemistry model to simultaneously simulate streamflow and stream dissolved organic carbon (DOC) concentrations in a small lead and arsenic contaminated upland peat catchment in northwestern Germany. The model procedure was informed by an initial data mining analysis, in combination with regression relationships of discharge, DOC, and element export. We assessed the internal model DOC processing based on stream DOC hysteresis patterns and 3‐hourly time step groundwater level and soil DOC data for two consecutive summer periods in 2013 and 2014. The parsimonious model (i.e., few calibrated parameters) showed the importance of nonlinear and rapid near‐surface runoff generation mechanisms that caused around 60% of simulated DOC load. The total load was high even though these pathways were only activated during storm events on average 30% of the monitoring time—as also shown by the experimental data. Overall, the drier period 2013 resulted in increased nonlinearity but exported less DOC (115 kg C ha−1 yr−1 ± 11 kg C ha−1 yr−1) compared to the equivalent but wetter period in 2014 (189 kg C ha−1 yr−1 ± 38 kg C ha−1 yr−1). The exceedance of a critical water table threshold (−10 cm) triggered a rapid near‐surface runoff response with associated higher DOC transport connecting all available DOC pools and subsequent dilution. We conclude that the combination of detailed experimental work with relatively simple, coupled hydrology‐biogeochemistry models not only allowed the model to be internally constrained but also provided important insight into how DOC and tightly coupled pollutants or trace elements are mobilized.