Sponsors

Volunteers

S7 Multi-scale modeling of mercury transport and fate in river ecosystems

Wednesday, 27 July, 2011

WS7-O1 — 8:30-8:45
MODELING MERCURY EXPOSURE AT DIFFERENT SCALES IN THE MCTIER CREEK WATERSHED AND EDISTO RIVER BASIN, SC, USA
Authors: KNIGHTES, Christopher1, GOLDEN, Heather E.1, BRADLEY, Paul2, DAVIS, Gary1, JOURNEY, Celeste2, CONRADS, Paul1
(1) USEPA, Knightes.chris@epa.gov; (2) USGS;

Mercury is the toxicant responsible for the largest number of fish advisories across the United States, with 1.25 million miles of rivers under advisory. The processes governing fate, transport, and transformation of mercury in lotic ecosystems are not well-understood, in large part because these systems are intimately linked with their surrounding watersheds. To understand the mercury exposure concentrations within streams and rivers, mercury fate and transport within the watershed must be understood and linked to the in-stream fate and transport processes. However, current advancements in the science of watershed Hg processing typically develop at distinct spatial scales and lack an understanding of linkages between Hg fate and transport processes among a variety of system sizes. This presentation serves to model the fate and transport of mercury at three different scales: a focused reach study (0.11 km2), a watershed scale (79 km2) and a basin scale (7,071 km2). These three systems were represented by linking a watershed hydrology and biogeochemical cycling (Hg, N and C) model (VELMA, Visualizing Ecosystems for Land Management Assessment) with a surface water mercury fate and transport model (WASP 7, Water Quality analysis Simulation Program). Both the watershed and water body models simulate three mercury species (Hg(0), Hg(II), and MeHg) and hydrology. The focused reach study used VELMA to parameterize and understand the processes governing Hg fate and transport, tracking mercury deposition and precipitation on the land surface, Hg transformation reactions and loss processes, and subsequent transport to the receiving stream. The process understanding and parameterization of the focus study were then applied at the watershed-scale to understand and represent mercury fate and transport by modeling individual sub-watersheds and linking them using WASP 7. This work was then further expanded to simulate mercury fate and transport at the basin scale, where state and national management strategies, such as TMDLs and emissions regulations, are applied.

WS7-O2 — 8:45-9:00
MODELING MERCURY EXPOSURE AT DIFFERENT SCALES IN THE MCTIER CREEK WATERSHED AND EDISTO RIVER BASIN, SC, USA
Authors: KNIGHTES, Christopher1, GOLDEN, Heather E.1, BRADLEY, Paul2, DAVIS, Gary1, JOURNEY, Celeste2, CONRADS, Paul1
(1) USEPA, Knightes.chris@epa.gov; (2) USGS;

Mercury is the toxicant responsible for the largest number of fish advisories across the United States, with 1.25 million miles of rivers under advisory. The processes governing fate, transport, and transformation of mercury in lotic ecosystems are not well-understood, in large part because these systems are intimately linked with their surrounding watersheds. To understand the mercury exposure concentrations within streams and rivers, mercury fate and transport within the watershed must be understood and linked to the in-stream fate and transport processes. However, current advancements in the science of watershed Hg processing typically develop at distinct spatial scales and lack an understanding of linkages between Hg fate and transport processes among a variety of system sizes. This presentation serves to model the fate and transport of mercury at three different scales: a focused reach study (0.11 km2), a watershed scale (79 km2) and a basin scale (7,071 km2). These three systems were represented by linking a watershed hydrology and biogeochemical cycling (Hg, N and C) model (VELMA, Visualizing Ecosystems for Land Management Assessment) with a surface water mercury fate and transport model (WASP 7, Water Quality analysis Simulation Program). Both the watershed and water body models simulate three mercury species (Hg(0), Hg(II), and MeHg) and hydrology. The focused reach study used VELMA to parameterize and understand the processes governing Hg fate and transport, tracking mercury deposition and precipitation on the land surface, Hg transformation reactions and loss processes, and subsequent transport to the receiving stream. The process understanding and parameterization of the focus study were then applied at the watershed-scale to understand and represent mercury fate and transport by modeling individual sub-watersheds and linking them using WASP 7. This work was then further expanded to simulate mercury fate and transport at the basin scale, where state and national management strategies, such as TMDLs and emissions regulations, are applied.

WS7-O3 — 9:00-9:15
DYNAMIC MODELING OF LANDSCAPE CONTROLS ON THE MOBILIZATION OF MERCURY AND DISSOLVED ORGANIC CARBON IN A FORESTED NORTH-EASTERN USA WATERSHED
Authors: SCHELKER, Jakob1, BURNS, Douglas A.2, WEILER, Markus 3, LAUDON, Hjalmar1
(1) Swedish University of Agricultural Sciences, Umeå, Sweden, jakob.schelker@slu.se; (2) U.S. Geological Survey, New York Water Science Center, USA; (3) Albert-Ludwigs Universität, Freiburg, Germany;

The transport of mercury and dissolved organic carbon (DOC) during high flow conditions such as snowmelt is known to account for a major fraction of the annual total mercury (THg) and DOC loads in many northern catchments. We used a combined field and modeling approach to study the controls on THg and DOC transport in Fishing Brook, a headwater tributary in the Hudson River basin of the Adirondack Mountains, New York. Specifically we tested how the saturation-state of catchment soils and the hydrological connectivity of riparian wetlands and upland/wetland transition zones to the stream network effects the mobilization of Hg. Stream water THg concentrations typically vary strongly (mean = 2.25 ± 0.5 ng L-1) and show a counterclockwise hysteresis of concentration to discharge during snowmelt. Further, a large fraction (40-48%) of the yearly THg load is exported to downstream ecosystems during snowmelt. Methyl mercury (MeHg) concentrations also showed high values (0.02 to 0.26 ng L-1), but follow an inverse logarithmic relationship with discharge. Application of a dynamic hydrological model (TOPMODEL) based on a 10x10m elevation model of the catchment shows that two-thirds of the simulated saturated area at peak flow corresponded with wetland areas determined by remote sensing. Application of a multi-direction flow algorithm based on an elevation-above-creek (EAC) approach further suggests that most of the wetlands become well connected to the stream network during high flow conditions. The dynamics of TOPMODEL-simulated saturated area and soil storage deficit were able to explain a large part of the variation of THg concentrations in stream water (r2=0.53 and r2=0.72 for the non-growing and growing season, respectively). In contrast, THg concentrations were not strongly correlated to DOC concentrations, and the simulated saturation and soil storage deficit variables were not able to explain DOC variation at the catchment outlet. These results indicate that all three constituents (DOC, MeHg and THg) follow different mobilization patterns: (1) DOC varies by season and the variation cannot be explained by the simulated variables, (2) MeHg is diluted during high flow, suggesting flushing from a supply limited pool whereas (3) THg dynamics follow a pattern independent of DOC, but closely related to the simulated saturation state of the catchment. Overall the differing gain and/or loss processes of solutes within the Fishing Brook stream network make this study a unique example of how the processing and transport can vary over spatial scales.

WS7-O4 — 9:15-9:30
CHARACTERIZING MERCURY CONCENTRATIONS AND FLUX DYNAMICS IN A COASTAL PLAIN WATERSHED USING MULTIPLE MODELS
Authors: GOLDEN, Heather E.1, KNIGHTES, Christopher D.2, BRADLEY, Paul M.3, DAVIS, Gary4, FEASTER, Toby D.5, JOURNEY, Celeste A.3, BENEDICT, Stephen T.5, CONRADS, Paul A.3, BRIGHAM, Mark M.6
(1) U.S. Environmental Protection Agency, ORD, Ecological Exposure Research Division, Cincinnati, Ohio, USA, golden.heather@epa.gov; (2) U.S. Environmental Protection Agency, ORD, Ecosystem Research Division, Athens, Georgia, USA; (3) U.S. Geological Survey, Columbia, South Carolina, USA; (4) U.S. Environmental Protection Agency, ORD, Ecosystems Research Division, Athens, Georgia, USA; (5) U.S. Geological Survey, Clemson, South Carolina, USA; (6) U.S. Geological Survey, Mounds View, Minnesota, USA.

Mercury-related fish-consumption advisories are widespread in the coastal plain of the southeastern U.S., where atmospherically deposited mercury interacts with an abundance of wetlands and high-dissolved organic carbon (DOC), acidic waters. Recent trends in decision-making processes require knowledge of mercury cycling at a variety of spatial scales (e.g., mesoscale watersheds, regions) and within diverse land cover types to better understand and manage the effects of this challenging water quality issue. Watershed models are primary tools for advancing questions related to such ecological exposure research. Spatially-explicit process-based watershed models can (1) improve spatial and temporal linkages between controls on environmental processes and subsequent water quality when observational studies are limited and (2) predict future changes in surface waters by using mathematical formulations. However, the science of spatially-explicit watershed scale mercury modeling is just beginning to emerge and most approaches have not been applied in mixed land cover, coastal plain watersheds. In response to this gap in current knowledge, we quantify total mercury (HgT) concentrations and fluxes from McTier Creek Watershed, South Carolina, USA, using three novel independently developed watershed mercury models (a grid based watershed mercury model (GBMM), the VELMA-Hg model, and the TOPMODEL-Hg model) and measured in-stream HgT concentrations and fluxes. The study watershed is located in an upper coastal plain landscape, an area with more diverse land cover, a larger drainage area, and a different geophysical setting than many previous sites of mercury research in North America, i.e. small forested headwater boreal or northern forested catchments. Therefore, we aim to improve the characterization of mercury cycling in coastal plain watersheds, identify important watershed processes influencing total mercury loadings to surface waters on daily and seasonal time scales, and advance the developing science of watershed-scale mercury modeling. Based upon our understanding of the diverse mercury dynamics represented within each model, simulated HgT fluxes at the watershed outlet using these models, and observed HgT data, this study moves toward our goals.

WS7-O5 — 9:30-9:45
IN SITU OPTICAL MEASUREMENTS AS PROXIES FOR AQUEOUS MERCURY
Authors: SHANLEY, James B.1, AIKEN, George R.1, PELLERIN, Brian A1, SARACENO, JohnFranco1, BERGAMASCHI, Brian A.1, DRISCOLL, Charles T.2, RISCASSI, Ami3
(1) U.S. Geological Survey, jshanley@usgs.gov; (2) Syracuse University; (3) University of Virginia.

Recent advancements in inexpensive in situ optical sensors now permit continuous measurements that reveal information about the quantity and quality of organic matter. Due to the close relation between organic matter and mercury transport, these measurements have been applied to generate continuous records of aqueous mercury (Hg) species by proxy. One example measurement is fluorescing dissolved organic matter (FDOM), measured by excitation at 370 +/- 10 nm and detecting emission at 460 +/- 60 nm. Typically only about 2% of organic matter fluoresces, but the measurement is highly sensitive and fairly specific to the fraction of DOM that is associated with Hg transport. We have applied unfiltered FDOM measurements in three northeastern USA forested research catchments, based on previously observed strong relations of dissolved Hg to dissolved organic carbon (DOC) and even stronger relations (r2>0.9) to laboratory measurements of the hydrophobic acid fraction of the DOC (HPOA) and filtered UV absorbance at 254 nm. Dissolved Hg correlations with the field measurements were somewhat lower due to occasional interference of FDOM by turbidity. However, with accurate and continuous turbidity measurements, one can compensate for this source of error. Optical turbidity measurements have the added advantage of serving as a proxy for particulate organic matter, which is often highly correlated to particulate Hg, the dominant form of Hg export in some streams. By combining these proxies, it is possible to quantify the majority of Hg flux in these streams. Because dissolved and particulate Hg both vary strongly but not always predictably with flow, these high frequency optical proxies greatly improve the accuracy of Hg flux calculations. The high temporal resolution also allows inference of in-stream dynamics and processes controlling DOC and Hg. These continuous data series will foster improvements to models of when and how inorganic ionic and particulate Hg moves to downstream methylation sites. We will also present examples in which optical measurements serve as proxy indicators for dissolved methylmercury.

WS7-O6 — 9:45-10:00
LANDSCAPE CONTROLS ON TOTAL AND METHYL MERCURY IN THE UPPER HUDSON RIVER BASIN OF NEW YORK STATE
Author: BURNS, Douglas A1
(1) U.S. Geological Survey, daburns@usgs.gov

High levels of mercury (Hg) in aquatic biota have been identified in surface waters of the Adirondacks of New York, and factors such as the prevalence of wetlands, extensive forest cover, and low productivity waters are associated with Hg bioaccumulation in this region. Past research in this region has focused on improving understanding of the Hg cycle in lake ecosystems. In the study described herein, landscape controls on total Hg (THg) and methyl Hg (MHg) concentrations in streams were explored through synoptic surveys of 27 sites in the upper Hudson River basin in the central Adirondacks. Stream were sampled and analyzed for THg, MHg, dissolved organic carbon (DOC), and ultraviolet absorbance (Abs254) during spring and summer of 2006 and 2008. Landscape indices including common land cover, hydrographic, and topographic-based measures were explored as predictors of Hg through multivariate linear regression. Multivariate models that included a wetland or riparian area-based metric, an index for open water area, and in some cases a topographic metric, explained 55 to 75 percent of the variation in MHg concentrations, and 60 to 80 percent of the variation in THg concentrations. An open water index (OWI) was developed as the surface area of ponds/lakes normalized for the outlet drainage area relative to the total basin area, and this index was inversely related to THg and MHg concentrations. The OWI was also inversely related to specific ultra-violet absorbance (Abs254*100/DOC), consistent with previous studies indicating that open water increases the influence of algal-derived aliphatic carbon on DOC, decreasing aromaticity, and decreasing the ability of DOC to bind Hg. The OWI was not significant in models for THg that also included Abs254 as a predictive variable, but remained significantly inversely related to MHg in similar models suggesting that open waters in addition to decreasing carbon aromaticity, limit MHg concentrations through additional processes in this river basin. These data are consistent with removal of MHg by photo-reduction and volatilization in open waters, but other processes such as particle settling and demethylation in bottom sediment may also contribute to the patterns observed. Study results confirm the importance of riparian wetlands as sources of Hg species to flowing waters, but also highlight the role of open water in limiting downstream transport of MHg in river networks. These results may be broadly applicable in northeastern North America and other settings where rivers consist of linked open water bodies.

WS7-O7 — 10:00-10:15
MODELING TRANSPORT AND TRANSFORMATION OF MERCURY FRACTIONS IN HEAVILY CONTAMINATED MOUNTAIN STREAMS BY COUPLING A GIS-BASED HYDROLOGICAL MODEL WITH A MERCURY CHEMISTRY MODEL
Authors: LIN, Yan1, LARSSEN, Thorjørn1
(1) Norwegian Institute for Water Research, yan.lin@niva.no

Many heavily polluted areas are located in remote regions that lack routine hydrologic monitoring. A modeling method that can produce scenarios of water chemistry trends for regions that lack hydrological data is therefore needed. The Wanshan mining area, in Guizhou province in south-western China, is such a region. This site is heavily polluted with mercury (Hg). The large majority of this Hg is associated with particles and is therefore susceptible to particle transport. To model Hg transport in a stream draining the Wanshan mining area, a Georaphic Information System (GIS) hydrologic model (HEC-HMS) is coupled with a simulation model for Hg fractions in water (WASP Hg). Hydrological variations in the stream flow are thereby simulated based on readily available precipitation data. The WASP 7 MERC Hg model is used for simulating variations in total Hg (THg), dissolved Hg (DHg) and methyl-Hg (MeHg) concentrations.

Modeled flow results show typical seasonal variations. During the winter season (Oct-Dec), flow drops below 1 m3 s-1. Flow increases beginning in February and reaches peak values of over 6 m3 s-1 in June and July. Particles transport is also modelled with a relative error of 16% and 20% for two flow regimes in September 2007 and August 2008, respectively.

The modeled Hg concentrations were tested against monitoring data from two sampling campaigns conducted in September 2007 and August 2008. The model produces reasonable simulation results for THg, DHg and MeHg. The method is also applied to three scenarios of reducing Hg mobilisation; it is shown that building a sedimentation pond can effectively reduce Hg transport downstream. For instance, a sedimentation pond with a surface area of 5000 m2 can reduce the THg concentration by 23% for one of the cases and shortens the distance of THg water concentrations above 100 ng L-1 by 1.5 km.

WS7-O8 — 10:15-10:30
SOURCE INVESTIGATION FOR MERCURY IN OTTAWA RIVER SEDIMENTS.
Authors: ETHIER, Adrienne1, SILKE, Renee1, DOLINAR, George1, MELCHER, Hilary2
(1) Atomic Energy of Canada Limited, ethiera@aecl.ca; (2) Trent University.

In 2001, Atomic Energy of Canada Limited (AECL) started evaluating sediment in the Ottawa River adjacent to Chalk River Laboratories site in Ontario in order to determine extent of contamination from operations and possible need for remedial action. Sediment mercury levels near the AECL Process Outfall were found exceed 5 µg/g. This mercury footprint overlapped with, but did not line up with, known radiological footprint. Other than mercury, no other heavy metals were found to be elevated adjacent to the site when compared with upstream and downstream sediment.

In order to identify the source of mercury footprint found in Ottawa River sediments near AECL, multiple natural and industrial sources were investigated (e.g., Process Outfall, Powerhouse stack, upstream paper mills, etc...). Results of this source investigation will be presented. A screening-level model (HERMES) was used to study historical and current mercury fate and flux patterns in the Ottawa River to assist with this source investigation and evaluation of remedial options (if required).

Wednesday, 27 July, 2011