The emerging role of drought as a regulator of dissolved organic carbon in boreal landscapes

Abstract

One likely consequence of global climate change is an increased frequency and intensity of droughts at high latitudes. Here we use a 17-year record from 13 nested boreal streams to examine direct and lagged effects of summer drought on the quantity and quality of dissolved organic carbon (DOC) inputs from catchment soils. Protracted periods of drought reduced DOC concentrations in all catchments but also led to large stream DOC pulses upon rewetting. Concurrent changes in DOC optical properties and chemical character suggest that seasonal drying and rewetting trigger soil processes that alter the forms of carbon supplied to streams. Contrary to expectations, clearest drought effects were observed in larger watersheds, whereas responses were most muted in smaller, peatland-dominated catchments. Collectively, our results indicate that summer drought causes a fundamental shift in the seasonal distribution of DOC concentrations and character, which together operate as primary controls over the ecological and biogeochemical functioning of northern aquatic ecosystems.

Fig. 1. Inter-annual variation in summer discharge in relation to the 0.1 mm per day low flow threshold.

The jitter dots are actual average daily discharge values from site C7 in the Krycklan catchment from which the ridgeline distribution curves were estimated for each summer between 2003 and 2019.

Fig. 2. Stream chemical responses to summer drought severity across a boreal stream network.

a percent change in dissolved organic carbon (DOC) (r² range 0.30–0.65, p < 0.05), b low molecular weight DOC (LMW DOC) (r² range 0.20–0.54, p < 0.05), c carbon to nitrogen ratio (C/N ratio) (r² range 0.20–0.47 p < 0.05), and d percent change in specific UV absorbance at 254 nm (SUVA₂₅₄) (r² range 0.24–0.56, p < 0.05), all plotted against the number of low flow days. Points represent the changes for each summer relative to the long-term average, and regression lines are fitted to the data for individual catchments. The red horizontal line indicates zero change while the vertical gray line indicates the average number of low flow days (18 days). Note differences in y-axis scales.

Fig. 3. Effects of rewetting in relation to antecedent summer drought in a boreal stream network.

a percent change in dissolved organic carbon DOC (r² range 0.20–0.67, p < 0.05), b low molecular weight DOC (LMW DOC) (r² range 0.21–0.64, p < 0.05), c carbon to nitrogen ratio (C/N ratio) (r² range 0.2–0.59, p < 0.05), d percent change in specific UV absorbance at 254 nm SUVA₂₅₄ (r² range 0.21–0.64, p < 0.05 respectively), all plotted against the number of low flow days in the previous summer. Points represent the changes for each summer relative to the pre-drought (June) averages, and regression lines are fitted to the data for individual catchments. Data for the C/N ratio for the driest summer (2006) are missing because sampling for nitrogen started in 2007. Additionally, for some of the larger sites (C10, C12, C14, C15), sampling stopped in 2017, hence data for 2018 and 2019 were unavailable.

Fig. 4. Monthly variation in DOC from 2003 to 2019 expressed as a percentage of long-term averages for the Krycklan sub-catchment during wet, dry, and normal hydrological conditions.

Shown are the long-term dissolved organic carbon DOC averages (black line), the years with high numbers of summer low flow days (above the 90th percentile) (red dots), years with the lowest number of low flow days (blue dots), and the years with the average number of low flow days (gray dots) in Krycklan. The loess regression curves show the average DOC change in years with the highest number of low flow days (red line) and the lowest number of low flow days (blue line).

Fig. 5. Conceptual responses of dissolved organic carbon (DOC), low molecular weight DOC (LMW DOC), carbon to nitrogen ratio (C/N ratio), and specific UV absorbance at 254 nm (SUVA₂₅₄) to summer low flow conditions.

Different time intervals represent our predictions for a drought effects during summer and b the post-drought effects after the first rewetting. In all cases, the curves represent predicted changes in relation to long-term averages normalized to seasons.