Redox driven mobilisation of DOC from riparian wetlands in Krycklan (N Sweden) after a drought experiment

Abstract

Riparian wetlands are major sources of dissolved organic carbon (DOC) to streams. Increasing DOC concentrations were observed for many northern streams during the last decades, with potential implications for carbon (C) storages of wetland soils and streamwater quality. Drivers behind these trends, and particularly the significance of redox processes in wetland soils, are still incompletely understood. In soils, organic C is often associated with or bound to iron (oxy) hydroxides. These associations of iron (Fe) and organic C may immobilise and protect soil organic matter from mineralisation under oxic conditions. However, organic C can be remobilised if ferric Fe is reduced under anoxic conditions, a process which also increases pH further enhancing DOC solubility. Redox processes are therefore presumably important drivers of DOC dynamics in both wetland soils and the adjacent streams. We hypothesised that in-stream DOC concentrations are mainly driven by redox conditions within riparian organic soils, where DOC mobilisation is controlled by reduction of DOC associated Fe. We further propose that these DOC mobilising redox processes are particularly relevant for periods of rewetting of riparian soils, e.g. in autumn. In this study, were used monitoring data following a drought experiment conducted in summer 2017 in a sub-catchment of Krycklan in northern Sweden. For the experiment, a drought was simulated for a sub-catchment in Krycklan by damming a lake outlet that feeds a small stream. For the rewetting period after the drought experiment, daily time series of discharge, DOC and Fe feeding into the manipulated stream section were calculated from data measured at the top and the bottom of the stream section. Discharge was measured by flumes. From discharge time series, baseflow feeding into the stream section was computed using a baseflow separation filter. Time series of baseflow was assumed to represent average watertable dynamics in riparian wetlands. Both Fe and DOC concentrations were obtained from absorbances measured across different wavelength using a portable ultraviolet–visible probe. Adsorbances were converted into aquatic concentrations using a partial least-squares regression model calibrated on Fe and DOC concentrations measured in the laboratory. Furthermore, concentration time series were corrected for discharge and in-stream retention (for DOC only). We found that Fe increased with increasing baseflow with a time lag of 5d indicating delayed iron reduction in riparian areas in response to elevated watertables. Dynamics of DOC were weaker related to baseflow than to Fe, but DOC was significantly correlated to Fe. From rules to obtain directed acyclic graphs it can be inferred that changing baseflow - as a proxy of watertables in riparian wetlands - caused changing discharge corrected Fe concentrations in the stream, which can be understood as a proxy of Fe concentrations in riparian wetlands. Changing Fe concentrations caused changes of computed in-stream DOC concentrations, which can be seen to represent mobilised DOC pools in riparian wetlands. It is concluded that redox driven mobilisation of DOC is a plausible process for rainfall periods in autumn, when riparian soils are rewetted after summer.