What causes rising DOC concentrations in streams from peat-affected catchments? Insights with high-resolution water quality analysis

Peatlands are an important natural terrestrial carbon store. Any impacts on the drivers of hydro-biogeochemical processes in these ecosystems can be particularly severe. Climate change and degradation through drainages and ditches are changing peatlands dramatically. Degraded peats turn from powerful carbon sinks to emitters. They can also threaten drinking water supplies, as (heavy) metals can be leached from degraded peats along with dissolved organic carbon (DOC). However, quantifying DOC discharges from terrestrial to aquatic ecosystems is challenging. The hydro-biogeochemical processes occurring at the soil-aquatic interface are not only complex but also occur at different spatial and temporal scales. These processes depend on a variety of constantly changing external conditions such as temperature, nutrition- as well as oxygen availability. On top, there is no sensor available, which can measure the DOC concentrations of streams in situ and directly.

Here we investigated the DOC concentration in two nested catchments of two adjacent streams in the Ore Mountains of southern Saxony in Germany. One stream is dominated by mineral soils, while the other is dominated by (degraded) peat soils. Each of the four sites is equipped with YSI-EXO fDOM sensors. Further data comprise discharge, water temperature, turbidity and electric conductivity. A machine-learning algorithm (Random Forest) was trained to predict DOC concentration from the available data set (validation r² between 0.85 and 0.98). We investigated the gained 15-minute resolution DOC data on potential driving factors. Interestingly, the area-specific loads of the peat-dominated catchment with 3.5 mg C m^-2 a^-1 did not differ significantly from that of the mineral soil-dominated catchment with 3.1 mg C m^-2 a^-1. However, the loads over the year were almost twice as high as previously detected from data collected on a monthly basis. With the high-resolution DOC data, we can detect the drivers of extreme DOC concentrations (up to 40 mg l^-1) after heavy rainfall events in summer and constant high-level DOC concentrations of 20 mg l^-1 during snowmelt in winter. By applying the algorithm on DOC:DON ratios, we were further able to quantify the different sources of plant-based material from the peat soils and microbial-degraded material from the mineral soil-dominated catchment.

Previous DOC measurements, mostly based on 2-week to monthly measurements, likely greatly underestimate the contribution of DOC to C fluxes in ecosystems. For C-rich ecosystems such as Peatlands, this is particularly significant.