Impact of wildfire on levels of mercury in forested watershed systems: Voyageurs National Park, Minnesota

Abstract

Atmospheric deposition of mercury to remote lakes in mid-continental and eastern North America has increased approximately threefold since the mid-1800s (Swain and others, 1992; Fitzgerald and others, 1998; Engstrom and others, 2007). As a result, concerns for human and wildlife health related to mercury contamination have become widespread. Despite an apparent recent decline in atmospheric deposition of mercury in many areas of the Upper Midwest (Engstrom and Swain, 1997; Engstrom and others, 2007), lakes in which fish contain levels of mercury deemed unacceptable for human consumption and possibly unacceptable for fish-consuming wildlife are being detected with increasing frequency. In northern Minnesota, Voyageurs National Park (VNP) (fig. 1) protects a series of southern boreal lakes and wetlands situated on bedrock of the Precambrian Canadian Shield. Mercury contamination has become a significant resource issue within VNP as high concentrations of mercury in loons, bald eagle eaglets, grebes, northern pike, and other species of wildlife and fish have been found. The two most mercury-contaminated lakes in Minnesota, measured as methylmercury in northern pike ( Esox lucius ), are in VNP. Recent multidisciplinary U.S. Geological Survey (USGS) research demonstrated that the bulk of the mercury in lake waters, soils, and fish in VNP results from atmospheric deposition (Wiener and others, 2006). The study by Wiener and others (2006) showed that the spatial distribution of mercury in watershed soils, lake waters, and age-1 yellow perch ( Perca flavescens ) within the Park was highly variable. The majority of factors correlated for this earlier study suggested that mercury concentrations in lake waters and age-1 yellow perch reflected the influence of ecosystem processes that affected within-lake microbial production and abundance of methylmercury (Wiener and others, 2006), while the distribution of mercury in watershed soils seemed to be partially dependent on forest disturbance, especially the historic forest fire pattern (Woodruff and Cannon, 2002). Forest fire has an essential role in the forest ecosystems of VNP (Heinselman, 1996). Because resource and land managers need to integrate both natural wildfire and prescribed fire in management plans, the potential influence of fire on an element as sensitive to the environment as mercury becomes a critical part of their decisionmaking. A number of recent studies have shown that while fire does have a significant impact on mercury at the landscape level, the observed effects of fire on aquatic environments are highly variable and unpredictable (Caldwell and others, 2000; Garcia and Carrigan, 2000; Kelly and others, 2006; Nelson and others, 2007). Caldwell and others (2000) described an increase in methylmercury in reservoir sediments resulting from mobilization and transport of charred vegetative matter following a fire in New Mexico. Krabbenhoft and Fink (2000) attributed increases in total mercury concentrations in young-of-the-year fish in the Florida Everglades to release of mercury resulting from peat oxidation following fires. A fivefold increase in whole-body mercury accumulation by rainbow trout ( Oncorhynchus mykiss ) following a fire in Alberta, Canada, apparently resulted from increased nutrient concentrations that enhanced productivity and restructured the food web of a lake within the fire’s burn footprint (Kelly and others, 2006). For this study, we determined the short-term effects of forest fire on mercury concentrations in terrestrial and aquatic environments in VNP by comparing and contrasting mercury concentrations in forest soils, lake waters, and age-1 yellow perch for a burned watershed and an adjacent lake, with similar samples from watersheds and lakes with no fire activity (control watersheds and lakes). The concentration of total mercury in whole, 1-year-old yellow perch serves as a good biological indicator for monitoring trends in methylmercury concentrations in food webs of lakes in North America (Wiener and others, 2007). With a limited gape, age-1 yellow perch that hatched the previous year and resided in a lake for 1 year feed largely on zooplankton and small benthic invertebrates. Thus, age-1 yellow perch provide a baseline for methylmercury concentrations for individual lakes that can be compared across spatial areas. The nine appendixes that accompany this report contain the complete datasets for soils, lake waters, and age-1 yellow perch collected for this study. This report uses data from these three media to provide a framework for evaluating short-term effects of fire on mercury in forested soils and possible effects of the mobilization of mercury from soils on lake water quality and aquatic health.