Present and historical accumulation of mercury in Ontario lake sediments

Abstract

Mercury (Hg) is a persistent contaminant present naturally in our environment but, as a result of anthropogenic activity, has now accumulated in lake sediments at concentrations 2.2 times greater than those found in pre-industrial sediments (pre-1850) (Chapter 2). However, present spatial patterns of mercury accumulation do not match patterns in natural mercury accumulation (Chapter 4). Once spatially uniform across south-central Ontario, mercury concentrations in present-day lake sediment display strong spatial correlation. Further, the broad scale spatial patterns (>500 km) which correlate with mean annual precipitation (MAP) found to significantly predict pre-industrial sediment Hg, has now been replaced with finer scale spatial patterns (<120km). Rather than one significant spatial scale explaining sediment Hg, as during the pre-industrial era (p = 0.01), present sediment Hg are predicted by more than one spatial scale (p < 0.001, p < 0.001) both of which are highly correlated with lake pH. While MAP and lake pH explain much variance in mercury accumulation, their significance is spatially dependent. In addition to differences in spatial trends, models which suggest that pre-industrial sediment Hg was in steady state with watershed accumulation do not apply to present day Hg accumulation (Chapter 2). Instead, Hg enrichment in lake sediment decreased as a function of drainage ratio (R2 = 0.458, p = 0.0001), suggesting that Hg export from watersheds may be lagging behind atmospheric Hg deposition. The enrichment of Hg since the pre-industrial era is also influenced by the amount of open water (r = 0.91, p = 0.035), mining activity (r = 0.94, p = 0.019) and organic deposits within surficial geology (r = -0.91, p = 0.034) which lead to local hot and cold spots (Chapter 4). There is a close link between the present-day sediment Hg concentrations and that in dorsal muscle of smallmouth bass (Micropterus dolomieul) (r = 0.92, p = 0.0002) (Chapter 3). However, the discrepancies that do exist between these two components are determined by landcover, longitude, and dissolved organic carbon. An examination of methylmercury in two catchments revealed that the concentration of methylmercury in stream water from ten watersheds is predicted from a time sensitive (lagged regression) interaction among sulfate (tpartial = 9.60, p < 0.001), dissolved organic carbon (tpartial = 7.06, p < 0.001), and temperature (tpartial = 4.44, p < 0.001) (Appendix A). The error of which decreases exponentially with increasing wetland size (R2 = 0.838, p = 0.0105). The difference between methylmercury inflow and outflow (MeHgNet) within these two catchments (a dystrophic and an oligotrophic lake) was seasonally dependent (Chapter 5). However, seasonal accumulation and loss of methylmercury was significantly larger in magnitude within the dystrophic lake. Further, a net production of methylmercury was found in the dystrophic lake alone (1.9 +/- 0.3 mg·d -1). This study examines Hg contamination in a relatively pristine, remote environment which like other regions, has been challenged by contaminants originating from far distances.