Event-Based Wet Deposition Study of Mercury

Program Background

The study and regulation of mercury (Hg) is a national and North American priority. Of the 189 chemicals required for study by the 1990 Clean Air Act Amendments, only mercury was singled out for intense study. The EPA recently submitted a comprehensive report to Congress on mercury ("Mercury Study Report to Congress" EPA-452/R-97-003, Dec. 1997). A North American Regional Action Plan (NARAP) for mercury was completed by the Commission on Environmental Cooperation in 1999 (CEC - US, Canada and Mexico). Exposure to wildlife and humans is primarily through the aquatic food chain where mercury bioaccumalates by up to factors of a million. The primary source of mercury to sensitive aquatic ecosystems is via atmospheric transport and deposition. The EPA Mercury Report states that "of the estimated 144 Megagrams of mercury emitted annually in to the atmosphere by anthropogenic sources in the United States, approximately 87% is from combustion point sources and 10% from manufacturing point sources. Four specific source categories account for approximately 80% of the total anthropogenic emissions: coal-fired utility boilers, municipal waste combustion, commercial/industrial boilers, and medical waste incinerators.

The EPA Mercury Report suggests that there is "…a plausible link between mercury emission from anthropogenic combustion and industrial sources and mercury concentrations in air, soil, water and sediments." However, they acknowledge that "…an apportionment between sources of mercury and mercury in environmental media and biota cannot be described in quantitative terms with the current scientific understanding of the environmental fate and transport of this pollutant." One reason that a quantitative linkage cannot be made is due the "lack of adequate mercury data near mercury sources." Mercury measurement data are needed to assess how well the modeled data predict actual mercury concentrations in different environmental media at a variety of geographic locations. Missing data include measured mercury deposition rates and measured concentrations in the atmosphere, soils, water bodies and biota. The development of atmospheric fate and transport models will require comprehensive field investigations to determine the important atmospheric transformation pathways.

Coal and municipal waste incineration is known to emit significant fractions of total mercury in the water-soluble oxidized form of mercury (Hg II) (Prestbo and Bloom, 1995). Furthermore, as a plume cools it is likely that emitted inorganic mercury (Hg II) will condense on co-emitted particulate matter (Imhoff, 1996). Rain falling through a coal combustion and/or municipal waste incinerator plume is expected to washout inorganic mercury (Hg II) and particulate mercury efficiently and close to the source (Pai et al., 1996; Seignuer, 1996). Assuming the following: 1) 46 grams Hg/emission stack/day, 2) sampling time = 1 day, 3) 2 cm of rainfall/day, 4) down wind grid size of 250,000 m², 5) 1% mercury washout efficiency within the downwind grid. We would expect to see a 10:1 difference in upwind verses downwind mercury deposition signal. However, even if a 10:1 difference in upwind verses downwind mercury deposition signal is observed, we expect the calculated mercury flux to be lower than estimates made in the December, 1997 EPA report to Congress. Recent Static Plume Dilution Chamber (SPDC) studies indicate that emitted inorganic mercury (Hg II) is rapidly dry deposited to the walls. During simulated rainwater SPDC runs, the inorganic mercury (Hg II) fraction present in the fluegas was readily removed to the aqueous phase. However, there appeared to be a trend toward conversion of inorganic mercury (Hg II) to elemental mercury (Hg⁰) over time, which would be expected to decrease the fraction of mercury, removed locally. Furthermore, the above assumptions about mercury speciation emissions and washout may not hold true. For example, there may be reduction of inorganic mercury (Hg II) to elemental mercury (Hg^o) in the plume (Munthe, 1991) and scavenging of inorganic mercury (Hg II) may be diffusion limited (Campbell and Zankel, 1993). In this case, no statistical difference in upwind verses downwind signal may be observed. This result would also be immensely important.The absolute amount of local mercury deposition by direct plume washout was limited to the absolute amount of time that it rained locally. Thus, even if event-based measurements showed that downwind mercury deposition were statistically greater than upwind sites, to determine the significance of this over a integrated time period of a year will require the application of a sophisticated plume model that takes into account the number of rain events, plume directional changes and other important factors. This has yet to be done with these data.

In order to do event-based deposition sampling on a set grid, historical a nd real-time meteorological data were employed. Sampling only occurred after prior, sufficient air mass ventilation and an expected incoming cyclonic frontal system that produced constant rain over a large area for a short time (<24 hours). There are excellent historical reports (PPRP) and internet meteorological databases (e.g. www.arl.noaa.gov/ready/hysplit4.html) that were used to locate sampling sites and forecast when samplers were deployed. The siting plan presented below was optimized, as new meteorological data became available.

The sampling grid included 18 wet-deposition samplers: 3 upwind (>5 km from sites), 12 downwind (0.2-3 km from sites), and 3 duplicates, 5 throughfall samplers: 2 upwind and 3 downwind, and 3 dry-deposition IX membrane plate samplers (1 upwind, 2 downwind) (Figure 1). The total number of collectors per event was 23 wet or 3 dry depending on the type of event sampled. The throughfall sampling sites were designed to evaluate whether there were upwind verses downwind differences in an integrated wet and dry deposition signal. In addition, two wet-deposit field blanks per event (10% of total samples) and one dry-deposit field blank were collected per event. Field blanks were generated by pouring double deionized water through the funnels in the field or by loading the blank filter to the filter plate in the field. Sampling sites were selected based on a preliminary site visit. Based on historical meteorological data, storms were expected to arrive from the south, southwest during the springtime. Therefore, actual location of the sites depended on the initial site visit, and refined meteorological information.

Wet-deposition samples were collected in ultra-trace metal free polyethylene funnels and 1-liter Teflon bottles mounted in a vertical PVC pipe (as illustrated in Figure 2.). The bottle was charged with 20 ml of a 1% solution (v/v) of HCl to preserve both MMHg and total Hg (Vermette et al., 1995). As shown, the bottle is enclosed in the PVC tube to minimize any photochemical reactions and external contamination of the sampler. The funnel and 1 liter bottle combination allow for a maximum of 16 cm of rain per event to be collected. Dry deposition samples were collected using a passive system consisting of an ion-exchange membrane mounted on an acrylic holder.

The major ion collectors consisted of a poly funnel (bottle with bottom removed) connected to a 1 L poly bottle. The bottle was wedged into a PVC pipe which acted as a stand. The design was similar to that used in PPRP projects in western Maryland. The major ion funnels and bottles were cleaned in a multi-step protocol that was similar to that used in western Maryland, and at other sites around the Chesapeake Bay. Major ion concentrations in water samples were analyzed according to EPA laboratory methods appropriate for monitoring surface water quality in acid deposition studies (USEPA 1986). Closed (for stream water) and open pH (for wet deposition and throughfall) were measured with an Orion (Model 611) pH meter, using a two point (4 and 7) calibration, and quality control check (QQCS) solutions (pH = 4 and/or 5) to verify calibration. All major ions, except Ca⁺² and Mg⁺², were measured using a Dionex DX-500 ion chromatograph, equipped with electronic conductivity suppressor and a computer-based data acquisition system. Calcium and Mg⁺² were analyzed by inductively-coupled plasma atomic emission spectrophotometry, or ICP-MS, if necessary.

Samples were analyzed for total mercury, methyl mercury, cadmium, lead, nickel, zinc, chromium, arsenic, and selenium. Samples received the following attention at the laboratory:

• Upon arrival the samples were checked for accuracy against the chain of custody or equivalent paperwork.
• Bottle weights were measured in order to determine total sample volume.
• The samples were split (by weighing) into fractions for analysis.
• Total mercury and methyl mercury were analyzed by cold vapor atomic fluorescence detection (CVAFS) after appropriate sample pretreatment (EPA Methods 1631 and 1630, Bloom, 1989; Horvat et al., 1993).
• The trace metals cadmium, lead, nickel, zinc and chromium were analyzed by ICP-MS using a HF/HNO₃ digestion for total recoverable metals.
• In order to achieve the lowest detection limit, arsenic and selenium were analyzed by hydride-generation AFS in the HF/HNO₃ digest.
• Duplicate samples at three sites were collected to determine overall precision for each analyte in the laboratory, SRMs, matrix spikes, reagent blanks, system blanks and duplicate analyses were performed. A sufficient number of quality assurance measurements were made in order to provide a high level of confidence in estimates of accuracy and precision.

The focus of the experiment was to measure the difference between upwind and downwind mercury wet-deposition during distinct rain events near combustion point sources. In addition, by measuring throughfall at a few sites, integrated wet+dry mercury deposition differences may be quantified. In order to help interpret and support the mercury data, additional specific trace metals in the rainwater samples were determined. Table 1 outlines the number of events, the analytes measured for each event and the analytical method description.

Table 1: Schedule of analytes verses events for wet-deposition (3 upwind, 12 downwind, 3 duplicate), dry deposition (1 upwind, 2 downwind) and 5 throughfall sampling site locations.

As shown in Table 1, the first two events measured only total mercury. The results from these first two events were generated from the laboratory using a quick turn around time in order to evaluate remaining sampling events. The last three events will include analyses for mercury, trace metals and major anions.

A full report of this study is not yet available. Data presented on this website represent fully QA’ed data that have not yet been used to determine the outcome of this study.