Intra-lake response of Arcellinida (testate lobose amoebae) to gold mining-derived arsenic contamination in northern Canada: Implications for environmental monitoring

Abstract

Arcellinida (testate lobose amoebae) were examined from 40 near-surface sediment samples (top 0.5 cm) from two lakes impacted by arsenic (As) contamination associated with legacy gold mining in subarctic Canada. The objectives of the study are two folds: quantify the response of Arcellinida to intra-lake variability of As and other physicochemical controls, and evaluate whether the impact of As contamination derived from two former gold mines, Giant Mine (1938–2004) and Tundra Mine (1964–1968 and 1983–1986), on the Arcellinida distribution in both lakes is comparable or different. Cluster analysis and nonmetric multidimensional scaling (NMDS) were used to identify Arcellinida assemblages in both lakes, and redundancy analysis (RDA) was used to quantify the relationship between the assemblages, As, and other geochemical and sedimentological parameters. Cluster analysis and NMDS revealed four distinct arcellinidan assemblages in Frame Lake (assemblages 1–4) and two in Hambone Lake (assemblages 5 and 6): (1) Extreme As Contamination (EAC) Assemblage; (2) High calcium (HC) Assemblage; (3) Moderate As Contamination (MAC) assemblages; (4) High Nutrients (HN) Assemblage; (5) High Diversity (HD) Assemblage; and (6) Centropyxis aculeata (CA) Assemblage. RDA analysis showed that the faunal structure of the Frame Lake assemblages was controlled by five variables that explained 43.2% of the total faunal variance, with As (15.8%), Olsen phosphorous (Olsen-P; 10.5%), and Ca (9.5%) being the most statistically significant (p < 0.004). Stress-tolerant arcellinidan taxa were associated with elevated As concentrations (e.g., EAC and MAC; As concentrations range = 145.1–1336.6 mg kg−1; n = 11 samples), while stress-sensitive taxa thrived in relatively healthier assemblages found in substrates with lower As concentrations and higher concentrations of nutrients, such as Olsen-P and Ca (e.g., HC and HM; As concentrations range = 151.1–492.3 mg kg−1; n = 14 samples). In contrast, the impact of As on the arcellinidan distribution was not statistically significant in Hambone Lake (7.6%; p-value = 0.152), where the proportion of silt (24.4%; p-value = 0.005) and loss-on-ignition-determined minerogenic content (18.5%; p-value = 0.021) explained a higher proportion of the total faunal variance (58.4%). However, a notable decrease in arcellinidan species richness and abundance and increase in the proportions of stress-tolerant fauna near Hambone Lake’s outlet (e.g., CA samples) is consistent with a spatial gradient of higher sedimentary As concentration near the outlet, and suggests a lasting, albeit weak, As influence on Arcellinida distribution in the lake. We interpret differences in the influence of sedimentary As concentration on Arcellinida to differences in the predominant As mineralogy in each lake, which is in turn influenced by differences in ore-processing at the former Giant (roasting) and Tundra mines (free-milling).