Tracing total and dissolved material in a western Canadian basin using quality control samples to guide the selection of fingerprinting parameters for modelling

Abstract

The source dynamics of total and dissolved material in riverine systems are being affected by anthropogenic activities resulting in the degradation of waterways worldwide. Identifying the main sources of total and dissolved material is thus central to the management of increasingly scarce water resources. Here, we utilize data generated from water quality monitoring programs to investigate the sources of total and dissolved material in a large, semi-arid basin in western Canada. Our research focuses on the confluence of two major tributaries in the South Saskatchewan River Basin (SSRB) in the Province of Alberta: the Bow River (25,611 km 2 ) and the Oldman River (28,270 km 2 ). A tributary tracing technique coupled with a Deconvolutional-MixSIAR (D-MIXSIAR) modelling approach is used to estimate the potential source contributions of total and dissolved material from major tributary sites to target node sites on the main stem of the Bow River and Oldman River in addition to target nodes downstream of their confluence. In total, 812 samples were taken from 29 sites across the SSRB. A novel approach to selecting fingerprints for modelling is presented based on the analyses of additional quality control samples (146 duplicate and 172 blank samples). Overall, the Rocky Mountain headwater catchments were found to dominate the supply of material modelled using total recoverable (68%) and dissolved (76%) metals. There were seasonal fluctuations in source dynamics evident where the Bow River dominated the supply of total (69%) and dissolved (57%) material during the ice-covered season (November-March), and the Oldman River dominated the supply of total (73%) and dissolved (59%) material during the open water season (April-October). On the one hand, these seasonal dynamics are potentially the result of the extensive regulation of flow, particularly along the Bow River. On the other hand, the intensification of agriculture in the prairie/plain catchments may also facilitate the excess supply of total relative to dissolved material. For example, the Little Bow River, with ~70% agricultural land cover, contributed ~14 times more total material than anticipated based on discharge and 1.6 times more than anticipated based on unit area during the open water season. Overall, this research has improved our understanding of the source dynamics of total and dissolved material in the SSRB, providing the foundation for focussed studies targeting the main sources of total and dissolved material in this large, semi-arid basin in western Canada. In addition, our research highlights the potential of using existent data generated from water quality monitoring programs along with quality control best practices to help improve our understanding of the source dynamics of total and dissolved material in waterways around the world.

How to evaluate sediment fingerprinting source apportionments

“…This was performed because MixSIAR assumes that source tracer values are normally distributed, and the removal of non-normally distributed tracers may reduce model accuracy (Smith et al 2018). According to Laceby et al (2021), log transformations were able to provide a higher degree of normality for geochemical data used as input for MixSIAR, compared to other transformations.…”

Quantifying Meltwater Sources and Contaminant Fluxes from the Athabasca Glacier, Canada

“…Grab samples were collected at the hydrometric station following EPA Method 1669 . Samples were collected monthly on 12 separate occasions during the 2017 and 2018 open-water season, with two full sets of field duplicates collected in August 2017 and 2018, yielding a total sample set of 14 grab samples. Most of the parameters were analyzed for both filtered (i.e., dissolved) and unfiltered (i.e., total) fractions.…”

“…Samples intended for dissolved trace elements were kept at 4 °C and returned to the laboratory within 48 h then filtered and acidified. Data quality was assessed as part of the Government of Alberta Tributary Monitoring Network. , …”

A conceptual model-based sediment connectivity assessment for patchy agricultural catchments

“…Subsequently, we propagated the uncertainty in the regression fit by simulating posterior distributions of the model coefficients (β0, βk) with an informal Bayesian function of the R package 'arm' (Gelman and Hill, 2007), as in Batista et al (2021). The posterior distributions were used to simulate 1000 sediment concentration values for each day i.…”

“…However, the accelerated sediment supply from agricultural soils to riverine and lacustrine environments leads to negative off-site consequences. Specifically, nutrient-rich and pollutant-bound particulate matter from arable land is associated to the eutrophication and contamination of water courses (Krasa et al, 2019;Laceby et al, 2021). Extreme erosion events in agricultural fields are also linked to the occurrence of muddy floods (Boardman, 2020) and to damages to downstream infra-structure (Bauer et al, 2019).…”