Search

The post-glacial adjustment of streams can be described as a battle between two geomorphic giants: the magnitude and resistance of the glacial imprint and the competence of subsequent fluvial flows. In northern Fennoscandia, which deglaciated ~9500 y BP, the outcome of this battle provides both insight into poorly studied high-latitude regions and a glimpse into the trajectory of currently deglaciating regions. The glacial signature of eroded bedrock and deposited till in this region have organized tributary stream networks into three distinct process domains: lakes, slow-flowing reaches (S0<0.1% in peat/ fine sediment), and rapids (S0: 0.1-7% with gravel to boulder beds). Thus, these post-glacial streams have two characteristics that complicate a process-based understanding of channel form and sediment transport: (1) they are typically semi-alluvial, in that they contain coarse glacial legacy sediment, and (2) mainstem lakes buffer sediment and water fluxes. Here, I present research to constrain controls on channel geometry and sediment transport in semi-alluvial rapids at three scales: (1) the particle-scale, quantifying sediment transport with a field tracer study, (2) the reach-scale, determining the potential for bedforms or sediment clusters through a flume experiment, and (3) the catchment scale, examining downstream variations in width using GIS-based analyses.Sediment particle dynamics within five semi-alluvial rapids (S0: 2.1-7%) were quantified using RFID-tagged tracer clasts. After a large snowmelt flood (>Q50), median transport distances for D10- to D50-sized clasts were <0.1 m, with no clear relationship between slope or shear stress and transport distance. Rather, as previous studies in boulder-bed mountain streams show, boulder protrusion and grain resistance likely play a greater role, increasing critical shear stress and reducing sediment transport. Based on a flume study mimicking reach-scale conditions of semi-alluvial rapids, it is clear that large clasts (>D84) are basically immobile and do not adjust to form step-pool features. But patterns of deposition and scour around boulders exist, with more deposition upstream of boulders at higher flows. Most geomorphic work is done at low and intermediate flows (Q2-Q10), with little additional transport at Q50 flows, which has a relatively low magnitude due to lake buffers. To determine whether channel width has adjusted to contemporary flow regimes at the catchment scale, downstream variation in bankfull width was analyzed in two medium-sized catchments. Regressions of width vs. drainage area (derived from 2m DEMs) have low explanatory power (R2: 0.05-0.5); coefficients (α) are low (<1) and regression exponents (β) show wide variation (0.26-0.75). Instead, spatial variability in till deposits and peat formation likely control channel width. While these semi-alluvial streams in stream-lake catchments are technically in equilibrium, as they are not aggrading or degrading, their stable channel geometry does not reflect the contemporary landscape sediment yield or hydroclimatic regime. Past studies have shown that bedrock river gorges in northern Sweden formed during deglaciation with little Holocene erosion. Similarly, tributary channels with coarse glacial sediment and mainstem lakes are frozen in time, reflecting catchment- and reach-scale morphologies shaped during high flows at deglaciation, with contemporary adjustment occurring only at the particle to sediment-cluster scales. [ABSTRACT FROM AUTHOR]

Recent global studies suggest that the total emission of CO2 from streams and rivers is highly significant and correspond to as much as 85% of the total oceanic CO2 uptake. To address this issue, a large boreal river in Northern Sweden (Indalsälven) was instrumented in April 2018 with water-side measurements combined with an atmospheric eddy-covariance system measuring the vertical fluxes of CO2. This is one of the first known studies of its kind where long term eddy covariance measurements are conducted in a river settingThe river Indalsälven was found to be almost constantly oversaturated with respect to CO2 and with high water-side pCO2 dynamics varying in the range of 400-4500 ppm. During the spring, the pCO2 in the water was in the range of 500-700 ppm and at the same time, large emissions of CO2 from the river were measured. The rapid snowmelt resulting in high water flow in combination with occasionally high winds resulted in high gas transfer velocities, k600. By the beginning of July, the pCO2 in water had increased to over 4000 ppm, likely due to high aquatic degradation of organic material. During the period of July to September, a diel cycle of pCO2 in water was observed with amplitudes of up to several thousand ppm. Simultaneously CO2 emissions in the range of 13-17 µmol m-2s-1 were measured. Wind speed showed a positive correlation with the measured CO2 emissions and k600, with the highest emissions associated to data from the wind sector where the upwind distance to land was greatest.Our data show that a boreal river act as a strong source of atmospheric CO2. The dynamic behaviour in pCO2 and emission clearly highlight the importance of continuous high-frequency CO2 measurements in order to correctly incorporate rivers in landscape carbon balances. [ABSTRACT FROM AUTHOR]

The export of terrestrial carbon (C) to inland waters is a key control on aquatic ecosystems as well as the large scale cycling of C. Historically, the temporal resolution of aquatic CO2 concentration observations have been too low to capture the spectrum of time scales over which the governing C processes occur. However, the development of new technologies to obtain high frequency observations using in-situ sensors has provided new possibilities to monitor and analyze the C cycling in inland waters. Here, we investigated CO2 concentration dynamics along an upslope-riparian-stream transect in a boreal headwater stream in the Krycklan Catchment, located in the northern part of Sweden. We utilized a set of spectral methods and analyzed a high-frequency (hourly resolution) dataset for aquatic C with the aim to identify spatiotemporal patterns in CO2 concentration fluctuations and soil-stream linkages. The spectral methodology decomposes the observed time series into a spectrum of periodicities and the analyses revealed a high spatiotemporal variability in predominant periodicities of CO2 concentration fluctuations across the soil-stream transect. In addition, the spectral coherence between time series of CO2 concentration and water levels were used to explore soil-stream linkages affecting the in-stream CO2 concentration. The hydro-chemical connectivity along the soil-stream transect was found to be far more complex and intermittent than the hydrological connectivity, which had a very high and consistent coherence across a wide spectrum of periods Moreover, the spectral coherence between the riparian groundwater level and in-stream CO2 concentration, highlighted a strong control of hydrology on in-stream CO2 dynamics. However, during specific hydrological events the simple concentration dilution relationship in the stream was obscured. We hypothesize that flow dependent riparian source areas as well as in-stream processes caused the variabilities in the hydrochemical connectivity between the stream and the riparian soil. This work emphasizes the relationship between stream CO2 and riparian soil hydrology and highlights the power of decomposing hydrochemical time series to test hypotheses about the connectivity and biogeochemical transformation along pathways that link aquatic and terrestrial systems. [ABSTRACT FROM AUTHOR]