There is an increasing trend in the use of real-time sensor technology to remotely monitor aquatic ecosystems. Commercially available probes, however, are currently not able to measure aqueous selenium (Se) concentrations. Because of the well-described bioaccumulation potential and associated toxicity of Se in oviparous vertebrates, it is crucial to monitor Se concentrations at sites receiving continuous effluent Se input. This study aimed to estimate Se concentrations in a boreal lake (McClean Lake) downstream from a Saskatchewan uranium mill using real-time electrical conductivity (EC) data measured by autonomous sensors. Additionally, this study aimed to derive a site-specific total aqueous Se (TSe) threshold based on Se concentrations in periphyton and benthic macroinvertebrates sampled from the same lake. To characterize effluent distribution within the lake, eight Smart Water (Libelium) sensor units were programmed to report EC and temperature for five and seven consecutive weeks in 2018 and 2019, respectively. In parallel, periphyton and benthic macroinvertebrates were sampled with Hester-Dendy's artificial substrate samplers (n = 4) at the same sites and subsequently analyzed for Se concentrations. Electrical conductivity was measured with a handheld field meter for sensor data validation and adjusted to the median lake water temperature (13 °C) registered for the deployment periods. Results demonstrated good accuracy of sensor readings relative to handheld field meter readings and the successful use of real-time EC in estimating TSe exposure (r = 0.87; r 2 = 0.84). Linear regression equations derived for Se in detritivores versus Se in periphyton and Se in periphyton versus sensor-estimated TSe were used to estimate a site-specific TSe threshold of 0.7 µg/L (±0.2). Moreover, mean Se concentrations in periphyton (16.7 ± 4.4 µg/g dry weight [d.w.]) and benthic detritivores (6.0 ± 0.4 µg/g d.w.) from one of the exposure sites helped identify an area with potential for high Se bioaccumulation and toxicity in aquatic organisms in McClean Lake.

No abstract

There is increasing interest in using autonomous sensor technology to monitor aquatic ecosystems in real time and in employing such monitoring data to perform better ecological risk assessments. At seven locations in McClean Lake in northern Saskatchewan (Canada) that received diluted uranium milling effluent, we deployed sensor units to track effluent distribution and help predict potential biological effects on aquatic invertebrates. Water was also collected from each location on multiple occasions to measure major ions, dissolved metals, and routine water quality, and sediment was sampled to analyze total metals. The ecotoxicological risk to aquatic invertebrates was estimated using hazard quotients (HQs). The cumulative risk was estimated by summing the individual HQs, and the major ions risk was based on total osmolarity. The results indicated temporal and spatial variations in effluent exposure based on sensor electrical conductivity (EC) measurements in the McClean Lake East Basin. Individual HQs for water ranged from “moderate” (0.40–0.69) to “very high” (greater than 1) for silver, cadmium, arsenic, selenium, mercury, iron, and thallium. At all sites, major ions risk was less than 1. Individual HQs for sediment were “moderate” (0.40–0.69), “high” (0.7–0.99), and “very high” (greater than 1) for vanadium and cadmium. The cumulative risk in water and sediment for all metals combined was greater than 1 at some sites in Vulture Lake (which discharged into McClean Lake) and in McClean Lake itself. A more detailed estimation of the risks for aqueous selenium and arsenic (the only two metals that had good correlation with sensor EC data) indicated that their 90th percentile HQ values were less than 1 in McClean Lake, suggesting that these contaminants of concern do not represent a significant direct risk to aquatic invertebrate communities. Environ Toxicol Chem 2022;41:1765–1777. © 2022 SETAC