SPATIAL CONTROLS ON TOTAL AND METHYL HG IN THE UPPER HUDSON RIVER BASIN, NEW YORK, USA

Surface water mercury concentrations, especially those of the neurotoxin methylmercury (MHg) can vary widely across regional landscapes. Approaches are needed to better predict spatial variation in Hg species concentrations across heterogeenous landscapes that include mountainous areas, wetlands, and open waters. Here, we used multivariate regression to develop models to best describe the spatial variation of total mercury (THg) and MHg concentrations across the 500 km² Upper Hudson River basin in the Adirondack Mountains of New York, an identified "hot spot" for fresh water Hg contamination. High spatial variation of about 6-fold for THg concentrations and 40-fold for MHg concentrations were present across this basin for samples collected during spring and summer of 2006 and 2008 at 27 sub-basins. THg, MHg, dissolved organic carbon (DOC) and iron concentrations were significantly greater in summer than spring reflecting the increased influence of riparian areas dominated by wetlands during summer. Hg species concentrations were significantly related to percent wetland area and to DOC concentrations, but these relations were weaker in summer, when only about one-third of spatial variation was accounted for by bivariate regression relations. In contrast, multivariate regression relations that included metrics of: (1) hydrogeomorphology such as slope and overland flow distance, (2) riparian/wetland area, and (3) open water, explained 66% to more than 90% of the spatial variation in these data for each Hg species in each of the seasons. These metrics reflect the combined influence of basin morphometry and riparian soils on Hg source and transport, and the role of open water as an Hg sink in this basin. Multivariate models based solely on these metrics, predicted Hg species concentrations as well as, or in some cases, better than models based on more expensive and time intensvie chemical and physical metrics such as DOC concentrations and specific ultraviolet absorbance. Metrics derived from a digital elevation model, land cover, and hydrography show great promise for identifying areas of expected high Hg concentrations in waters and biota in the Adirondack region, and are likely applicable in similar glaciated riverine landscapes dominated by mountains, riparian wetlands, and lakes/ponds.