Science of the Total Environment 472 (2014) 530–537
Contents lists available at ScienceDirect
Science of the Total Environment
j ourna l homepage: www.e lsev ie r .com/ locate /sc i totenv
Arsenic and lead distribution and mobility in lake sediments in thesouth-central Puget Sound watershed: The long-term impact of a metalsmelter in Ruston, Washington, USA
James E. Gawel a,⁎, Jessica A. Asplund a, Sarah Burdick a, Michelle Miller a, Shawna M. Peterson a,Amanda Tollefson b, Kara Ziegler a
a University of Washington Tacoma, Environmental Science and Studies, Interdisciplinary Arts and Sciences Program, 1900 Commerce St., Box 358436, Tacoma, WA 98402, USAb Bellarmine Preparatory School, 2300 S Washington St., Tacoma, WA 98405-1399, USA
H I G H L I G H T S G R A P H I C A L A B S T R A C T
• We analyzed As and Pb in sediment andbottom water of 26 lakes near ASARCOsmelter.
• As and Pb levels in surface sedimentsdownwind are significantly elevated.
• As and Pb levels are correlated in cores,and consistent with smelter operation.
• Maximum dissolved As in bottom wateris correlated with As in surface sedi-ments.
• Sediment As and Pb in 83% of down-wind lakes exceed probable effects con-centration.
Article history:Received 18 June 2013Received in revised form 28 October 2013Accepted 1 November 2013Available online xxxx
Keywords:ArsenicLeadSedimentSmelterLakeWashington
The American Smelting and Refining Company (ASARCO) smelter in Ruston, Washington, contaminated thesouth-central Puget Sound region with heavy metals, including arsenic and lead. Arsenic and lead distributionin surface sediments of 26 lakes is significantly correlated with atmospheric model predictions of contaminantdeposition spatially, with concentrations reaching 208 mg/kg As and 1375 mg/kg Pb. The temporal distributionof thesemetals in sediment cores is consistentwith the years of operation of the ASARCO smelter. In several lakesarsenic and lead levels are highest at the surface, suggesting ongoing inputs or redistribution of contaminants.Moreover, this study finds that arsenic is highly mobile in these urban lakes, with maximum dissolved arsenicconcentrations proportional to surface sediment levels and reaching almost 90 μg/L As. With 83% of thelakes in the deposition zone having surface sediments exceeding published “probable effects concentrations”for arsenic and lead, this study provides evidence for possible ongoing environmental health concerns.
531J.E. Gawel et al. / Science of the Total Environment 472 (2014) 530–537
1. Introduction
Lake sediments act as reservoirs of metal contaminants and long-term sources to overlying waters (Senn et al., 2007; Tanner andClayton, 1990). Arsenic is of particular concern due to its potentialtoxicity to humans (Kazi et al., 2008) and other aquatic biota (Knaueret al., 1999; Lee et al., 1991; Wängberg et al., 1991), and its particularsusceptibility to remobilization in urban areas where organic matterand nutrient inputs feed bacterial respiration in sediments and bottomwaters (Lattanzi et al., 2007). The sediment pool may continue to bea significant source of arsenic to surface waters long after ongoingmetal deposition has been abated (Couture et al., 2010; Senn et al.,2007; Tanner and Clayton, 1990).
Arsenic chemistry in aquatic systems is often dominated by theredox status of the system (Ferguson and Gavis, 1972). In many naturalwaters, the primary inorganic arsenic species of environmental signifi-cance are arsenate, H2AsO4
−:As(+V), and arsenite, H3AsO3:As(+III).Organic forms of arsenic, such as the monomethyl- and dimethyl-forms, may be significant under certain conditions as well. Arsenatebinds to iron oxide solids under oxidizing conditions and precipitatesout of the water column (Ferguson and Gavis, 1972). Over time thisprecipitation results in an accumulation of arsenic in lake sediments.In many urban lakes during summer stratification, bacterial respirationresults in anoxic conditions in the hypolimnion as a result of culturaleutrophication. During anoxia, Fe(III)-oxides in the surface sedimentsare reduced to dissolved Fe2+:Fe(+II) (Ferguson and Gavis, 1972).In addition, direct reduction of As(+V) to As(+III) may occur underanoxic conditions (Harrington et al., 1998). Both mechanisms releasearsenic to overlying waters, increasing the potential for downstreamtransport, biotic exposure and toxicity (Aggett and O'Brien, 1985;Batterson and McNabb, 1983).
The continuing mobilization of arsenic in lakes contaminated byhistoric sources creates the potential for human exposure and toxicity(Ferguson and Gavis, 1972). Food chain transfer to humans can occurin arsenic-contaminated lakes where fish consumption and irrigationof crops using lake water occur (Arain et al., 2009). In WashingtonState, there is increasing attention to the potential for contaminanttoxicity in particularly susceptible populations consuming large quanti-ties of fish. The Washington State Department of Ecology has stated,“The best current science now indicates that our present fish consump-tion rates do not accurately reflect how much of our state’s fish andshellfish Washingtonians actually eat” (Washington State Departmentof Ecology, 2013a,b). This review of fish consumption rates is a precur-sor to review of sediment andwater quality standards to protect humanhealth on the basis of realistic fish consumption rates. To inform thisreview of arsenic standards for freshwater sediments and water re-sources, it is necessary to examine the spatial extent of arsenic contam-ination and the magnitude of arsenic bioavailability in fish habitat,including lakes.
The south-central Puget Sound region in Washington State hasbeen heavily impacted by a century of metal emissions from theAmerican Smelting and Refining Company (ASARCO) smelter inRuston, Washington (Glass, 2003). Long-term emissions from theASARCO smelter have resulted in the accumulation of toxic metals,including arsenic and lead, in surface soils and sediments of thisregion (Crecelius and Piper, 1973; Peterson and Carpenter, 1986;Seattle and King County Public Health and Glass, 2000). The smelter,in operation from 1890 to 1986, specialized in the smelting of leadand then copper ores containing high concentrations of arsenic; thisfacility separated and concentrated arsenic – along with the primarytarget lead or copper – from these ores for commercial sale. Arsenicand lead contamination is now widespread throughout the regionvia a combination of atmospheric transport from the emissions stack(at one time the world's tallest) and the use of smelter slag for roadballast (Mariner et al., 1997; Seattle and King County Public Healthand Glass, 2000).
While arsenic and lead contamination of marine sediments andsoils in the south-central Puget Sound region has been studied quite ex-tensively (Crecelius et al., 1975), the distribution of arsenic and lead insediments and arsenic mobility in the region's lacustrine systems hasreceived relatively little attention from the research and regulatorycommunities to date. This region of Washington is home to a myriadof lakes, most of which are popular recreational resources for fishing,swimming and boating, and yet very few of these lakes have beenexamined for arsenic contamination and mobility, with almost no stud-ies conducted after the closure of the ASARCO smelter.
In this study we describe our investigation of the temporal andspatial distribution of arsenic and lead in the sediments of lakes in thesouth-central Puget Sound region, as well as the subsequent release ofdissolved arsenic to the overlying waters 20 years after the closure ofthe smelter.
2. Materials and methods
2.1. Site description
Lake sampling took place from 2003 to 2007. In total 26 differentlakes (Fig. 1 and Table 1) were sampled for sediment metal concentra-tions (arsenic and lead) using surface grab samples or sediment cores.A subset of 8 lake sediment cores were dated using 210Pb and a subsetof 17 lakes were monitored for water column arsenic concentrations.The choice of lakes sampled was designed to include lakes within thedominant wind direction zones (from East to Northeast and South toSouthwest) and outside these zones, but sampling was also influencedby the availability and ease of public boat access and the distributionof lakes in the region.
2.2. Sediment sampling
Surface sediments were collected in 2003 from 9 freshwaterlakes (Table 1) within a 20 mile radius of the ASARCO smelter emis-sions stack. The 20-mile radius was chosen based on data from theVashon/Maury Island Soil Study 1999–2000 (Seattle and KingCounty Public Health and Glass, 2000) which showed elevated soilarsenic concentrations were limited mostly to distances of 20 milesor less. Duplicate surface sediment samples (approximately the top10–20 cm of sediment) were collected from separate casts from aboat with a petit Ponar dredge (Wildlife Supply Company) from nearthe deepest point in each lake (based on available bathymetric mapsor local knowledge). Sediments were homogenized with a plasticspoon, and placed in acid-washed Nalgene jars in a cooler for transportto the laboratory.
Sediment coreswere collected in 2004, 2005 and 2007 from a subsetof 17 lakes (Table 1). Cores were taken from near the deepest point ineach lake using a gravity-driven, stainless steel corer with a plasticcore catcher for sediment retention during retrieval (K-B corer,WildlifeSupply Company). Separate cellulose acetate butyrate core sleeveswereused for each core. The corer was deployed and retrieved by hand from aboat and coreswere capped and stored upright on ice for transport to thelaboratory. The soft sediments were extruded and separated into 2 cmsections and placed into acid-washed Nalgene jars using a clear-PVCcore extruder constructed generally following the design of Kornijow(2013). All surface sediment samples and core sections were dried for3 days at 80 °C, homogenized in a Wiley mill, and stored in Whirl-pakbags prior to metals analysis.
A subset of eight of the sediment cores (Table 1) collected in 2004 and2007 were dated in the Nittrouer lab at the University of WashingtonSeattle using excess 210Pb activity. Excess 210Pb activity was measuredin each core section to determine sediment age and section-to-sectionsediment accumulation rate (Nittrouer et al., 1979). A dried, homoge-nized sample was spiked with 209Po as a yield indicator. 210Pb activitywas determined by alpha spectrometry of granddaughter isotope 210Po
Fig. 1.Map of study lakes showing proximity and direction in relation to ASARCO smelter.
532 J.E. Gawel et al. / Science of the Total Environment 472 (2014) 530–537
released from the sediment by successive digestions with concentratedHNO3 and HCl. After centrifugation the 209Po and 210Po in solution wereplated onto silver discs and then counted for alpha decay of 209Po and210Po isotopes using a silicon surface barrier detector.
For the purpose of comparing surface sediment samples collectedusing differentmethods (dredge or gravity corer), metal concentrationsin the top 12 cm of each core were averaged to create one surfacesediment concentration for those lakes for analysis of spatial depositionpatterns. This depth is comparable to penetration depth of the Ponardredge used. Sediments from nine of the sampled lakes were collectedby both methods: surface dredge and sediment core. The mean metalconcentrations in the surface sediments using the two different samplingmethods on these nine lakes were highly correlated (Pearson correla-tion coefficient 0.91, with p b 0.001), supporting the use of these twomethods as comparable for our analysis.
2.3. Water column sampling
Water column measurements were conducted and water samplescollected in 2004, 2005 and 2007 from the same set of 17 lakes wheresediment cores were collected (Table 1). Water column profiles oftemperature, dissolved oxygen, specific conductivity, and pH were col-lectedusing a real-timewater quality probe (HydroLabQuanta) deployedfromaboat. The probe sensorswere calibrateddaily.Water sampleswerecollected at specific depths using an onboard, battery-operated peristalticpump and acid-washed tubing.
Total dissolved arsenic was determined by filtering water samplesin the field through a 0.4 μm syringe filter and placing the filtratein acid-washed Nalgene bottles on ice. Upon return to the laboratory,all water samples for metal analysis were acidified with 1% v/v trace-metal grade concentrated nitric acid.
2.4. Metal analysis
For analysis of total arsenic and lead in sediments, sampleswere digested in concentrated nitric acid in pressurized vessels ina programmablemicrowave digestion unit (MARS5, CEMCorp.) follow-ing standard methods (EPA method 3051A). Following digestion,solutions were diluted with 18 MΩ water and stored in acid-washedNalgene bottles.
Arsenic in water column samples and arsenic and lead in digestedsediment samples were quantified by graphite furnace atomic absorp-tion spectrometry (Shimadzu model AA-6800) using platform graphitetubes and a palladium chloride matrix modifier for arsenic. The diges-tion procedure was checked using NIST standard reference material(#2711a Montana II Soil) with recoveries of 85% or greater.
3. Results
Due to constricted airflow caused in part by the Olympic Mountainsto the west and the Cascade Mountains to the east, the dominant winddirection with the highest wind speeds is NE/ENE, with lighter, lessfrequent winds to the SSW/SW (Glass, 2004; Washington StateDepartment of Ecology, 2002). Surface sediment concentrations ofarsenic and lead in lakes in this region (Figs. 2 and 3) are significantlycorrelated with each other (Spearman Rank Probability p b 0.05) andwith direction from the ASARCO smelter (Spearman Rank Probabilityp b 0.01) with the highest values found in lakes in the downwind areafrom NE to E in direction. These results are consistent with singlestorm arsenic deposition measured while the smelter was in operation(Larson et al., 1975), as well as model predictions for surface soilcontamination using averagewind data (Fig. 4, Area-Wide Soil Contam-ination Task Force, 2003). This depositionmodel predicts elevated arse-nic concentrations in the following sampled lakes: Angle, Killarney,
Table 1List of lakes included in this study with comparative morphometric and chemical information. Lake morphometric data provided by sources shown. Evidence of bottom water anoxia,or lowest dissolved oxygen (DO) value (in mg/L) recorded from the authors' monitoring work.
Lake name Year collected Sampling method Maximum depth(m)
a sediment core dated using 210Pb.b Bortleson et al., 1976.c Mathieu and Friese, 2012.d Sumioka and Dion, 1985.e Wolcott, 1973.f Heller, P., personal communication, Surface Water Management Technician, City of Federal Way.g King County, 2013.h McConnell et al., 1976.i Entranco Engineers, Inc., 1989.j Tetra Tech, Inc., 2008.
Fig. 2. Arsenic concentrations (mg As/kg dry weight) in surface sediments of study lakes. The threshold effects concentration (TEC) “below which harmful effects are unlikely to beobserved” is 9.79 mg/kg As, and the probable effects concentration (PEC) “above which harmful effects are likely to be observed” is 33 mg/kg As (“consensus-based sediment qualityguidelines for freshwater ecosystems” proposed by MacDonald et al., 2000).
533J.E. Gawel et al. / Science of the Total Environment 472 (2014) 530–537
Fig. 3. Lead concentrations (mg Pb/kg dryweight) in surface sediments of study lakes. The threshold effects concentration (TEC) “belowwhich harmful effects are unlikely to be observed”is 35.8 mg/kg Pb, and the probable effects concentration (PEC) “abovewhich harmful effects are likely to be observed” is 128 mg/kg Pb (“consensus-based sediment quality guidelines forfreshwater ecosystems” proposed by MacDonald et al., 2000).
534 J.E. Gawel et al. / Science of the Total Environment 472 (2014) 530–537
North, Brook, Dolloff, Geneva, Steel, Waughop, Snake, Fenwick, Bow,andMeridian. The nine lakeswith the lowest surface arsenic concentra-tions are located outside this predicted deposition zone in non-dominant wind directions, while the five lakes with the highest surfacearsenic levels are within this zone in the dominant wind direction. The
w
Fig. 4. Estimate of area affected by historical ASARCO smelter emissions based on data available as
lakes sampled within the predicted high arsenic deposition zone(n = 12, Area-Wide Soil Contamination Task Force, 2003) have signifi-cantly higher arsenic (p b 0.01; one-tailed, two-sample unequal variancet-test) and lead (p b 0.05) concentrations in surface sediments than lakesoutside that zone (n = 14). Moreover, we found no correlation between
of January 2003 (recreated from Figure I-1, Area-Wide Soil Contamination Task Force, 2003).
Fig. 5. Arsenic and lead concentrations (mg/kg dry weight) in sediment cores collected from select lakes: (a) Lake Killarney, (b) Angle Lake, (c) American Lake, (d) Bonney Lake, (e) LakeDolloff, (f) Waughop Lake.
Table 2Arsenic concentrations in surface sediments (mg As/kg dry weight) are significantlycorrelated (Pearson probability, p b 0.001) to maximum total dissolved arsenicconcentrations (μg As/L) in the bottom waters of 14 of the study lakes exhibitinghypolimnetic anoxia at the time of sampling. Variability indicated is one standarddeviation of the mean.
Lake name Mean As in sediment(mg As/kg dry wt)
Max dissolved As in water(μg As/L)
Steilacoom Lake 9.6 ± 2.5 0.9American Lake 32.1 ± 18.6 1.2Fivemile Lake 49.5 ± 25.1 2.0Bonney Lake 19.8 ± 2.7 2.1Lake Geneva 48.6 ± 6.0 2.1Spanaway Lake 16.5 ± 4.6 2.3Lake Tapps 28.0 ± 5.6 3.4Bow Lake 31.4 ± 4.8 3.7Waughop Lake 47.3 ± 10.6 3.9Lake Meridian 24.2 ± 5.7 4.8Lake Dolloff 50.9 ± 7.9 8.1Wapato Lake 19.8 ± 3.3 15.8North Lake 84.6 ± 4.6 17.5Angle Lake 208.2 ± 28.2 87.5
535J.E. Gawel et al. / Science of the Total Environment 472 (2014) 530–537
surface sediment arsenic or lead concentrations and lake surface area orvolume or drainage basin area. These results strongly indicate that theASARCO smelter is a primary source of arsenic and lead contaminationin the region's lakes.
Of those lakes where sediment cores were dated using 210Pb tech-niques, we were able to collect cores extending to pre-1890s dates fortwo of the sites, Lake Killarney and Angle Lake (Fig. 5a–b), with datesback to the 1910s–1920s for four other sites, American Lake, BonneyLake, Lake Dolloff, and Waughop Lake (Fig. 5c–f). The two older coresshow increasing arsenic and lead near the end of the 19th century,while the other four cores show similar patternswith increasing arsenicand lead beginning near the start of the 20th century. This is consistentwith the chronology of ASARCO smelter operations, with the originallead smelter brought online in 1890 and arsenic separation operationsbeginning in 1912 (Glass, 2003).
Arsenic and lead are significantly positively correlated (p b 0.01;Pearson's probability) in sediment cores from 13 of the 17 lakes cored,with arsenic and lead correlated in Lake Tapps at 95% confidence, andno significant correlation between these two metals in Steilacoom,Spanaway and Brook lakes (see Supplementary Data). As none ofthese three lakes were dated, it is difficult to surmise what may beresponsible for the difference between these lakes and the otherswhere arsenic and lead in the cores are significantly correlated.
Measured maximum near-bottom, dissolved arsenic concentrationsin the study lakes where anoxia prevailed during sampling (n = 14)
were significantly correlated (Pearson probability, p b 0.001) to surfacesediment arsenic concentrations across the region (Table 2). Althoughthis correlation is heavily weighted by the high results in Angle Lake,the correlation remains significant (p b 0.08) if this high value is not
536 J.E. Gawel et al. / Science of the Total Environment 472 (2014) 530–537
considered. Dissolved arsenic concentrations in the water column ofAngle Lake reach as high as 88 μg/L, and maximum dissolved arsenicconcentrations in all study lakes exceeded the National ToxicsRule freshwater criterion of 0.018 μg/L for the consumption of fish(Washington State Department of Ecology, 2013a,b), although the crite-rion is established for average water column concentrations.
4. Discussion
The Washington State Department of Ecology analyzed arsenic andlead concentrations in surface sediments of Washington Lakes in 1989and 2002, including four reference lakeswith no known significant con-tamination sources (Jack, 2003; Johnson andNorton, 1990). The averagereference values determined in these studies were 6 mg/kg As and11.5 mg/kg Pb. Sediment studies in Lake Washington, another lake inthe same region as this study, also show relatively lowbackground arse-nic and lead concentrations in pre-development sediments collected incores: 6 mg/kg As (Yake, 2001) and 25 mg/kg Pb (Crecelius and Piper,1973). Moreover, concentrations of both metals are below 10 μg/kg indeep core sections of our study lakes where dating of the core indicatespre-ASARCO deposition (Fig. 5 a–c; American, Angle, Killarney).
In comparison, surface sediment concentrations in our study lakesranged from a low of 5 mg/kg As in Wye Lake and 3.3 mg/kg Pb inLake Tapps to over 200 mg/kg As in Angle Lake and over 1370 mg/kgPb in Steel Lake. A few other lake-specific studies have collected surfacesediment metal concentrations from lakes also sampled in this study:American, Steilacoom and Angle Lakes. The average concentrations of ar-senic and leadmeasured in surface sediments in these studies agree gen-erallywith our results (Figs. 2 and3): 11.9 mg/kgAs and196 mg/kg Pb inAmerican Lake (Johnson andNorton, 1990), 22 mg/kg As and 207 mg/kgPb in Steilacoom Lake (Bennett and Cubbage, 1992), and 434 mg/kg Pbin Angle Lake (arsenic not measured, Mathieu and Friese, 2012).
Although there are no current actionable sediment quality criteria inWashington State for freshwater sediments, final rule-making is inprogress in the state that is scheduled to create criteria by the end of2013. However, comparative guidelines already exist that have beenused for discussion by the Washington State Department of Ecology inpublished documents (“consensus-based sediment quality guidelinesfor freshwater ecosystems” proposed by MacDonald et al., 2000). Thethreshold effects concentration (TEC) these authors propose, “belowwhich harmful effects are unlikely to be observed,” is 9.79 mg/kg Asand 35.8 mg/kg Pb. The probable effects concentration (PEC) proposed,“above which harmful effects are likely to be observed,” is 33 mg/kgAs and 128 mg/kg Pb.
Ten of the twelve lakes in the predicted arsenic deposition zone(Area-Wide Soil Contamination Task Force, 2003) have surface sedimentconcentrations that exceed the PEC for arsenic,with the highest two lakeshaving arsenic levels over six times the PEC (Angle and Killarney Lakes),while only three of the fourteen lakes outside that zone exceed the PEC(Fig. 2). Ten of the twelve lakes in the predicted arsenic deposition zonealso exceed the PEC for lead, while only two of the fourteen outsidethat zone exceed this amount (Fig. 3).
Although other sources of arsenic and lead, including general urbanstormwater runoff and leaded gasoline, may also have contributed tothe contamination of these lake sediments, several pieces of evidencesuggest that the ASARCO smelter played a dominant role. First, arsenicand lead concentrations are significantly correlated in the sedimentcores (p b 0.01) from 14 of the 17 lakes studied. This correlationbetween co-contaminants is indicative of ASARCO smelter emissions(Mariner et al., 1997), whereas leaded gasoline has not been shown tobe a significant source of arsenic contamination.
Second, leaded gasoline was not heavily used in the United Statesuntil the 1920s (Newell and Rogers, 2003).While the tall ASARCO emis-sions stack was not constructed until 1917 (Glass, 2003), thereforepotentially resulting in the coincidence of lead increases from leaded gas-oline and the ASARCO smelter in sediments, concentrations of arsenic
and lead in sediment cores (Fig. 5a–b) from the most impacted lakes(e.g. Angle and Killarney) show clear increases in these metals prior tothis time, and coinciding with the start of smelter operations at theASARCO smelter in the 1890s.
Finally, arsenic concentrations in at least some of the dated sedimentcores (Figures c–f; American, Bonney, Dolloff, andWaughop) begin to de-crease around the time of reduced ASARCO production and smelter shut-down in the 1970s and early 1980s. A “Meteorological CurtailmentProgram”was put in place in the early 1970swhich effectively decreasedemissions when winds blew toward the SW (toward American andWaughop Lakes), and the ASARCO smelter ceased operations completelyin 1986 (Glass, 2003).
Although arsenic and lead contaminants were distributed in the re-gion by both slag disposal and reuse and atmospheric transport (Glass,2003), it is likely that aerial deposition played a significant role in thisdispersal. Crecelius (1975) measured atmospheric deposition andother inputs to Lake Washington along with arsenic concentrations inthe sediments while the ASARCO smelter was in operation. His mea-surements showed that dry andwet atmospheric deposition accountedfor half of the arsenic inputs to Lake Washington, and that arsenicconcentrations in dry deposition samples were ten times greater whenwinds were blowing from the direction of the ASARCO smelter thanwhen winds blew from the north (Crecelius, 1975). Our own studyfound a significant correlation between surface sediment arsenic concen-trations in lakes and direction from the ASARCO smelter (Spearman RankProbability p b 0.01) with the highest values found in lakes in the down-wind area from NE to E in direction.
More importantly, cores from Killarney and Angle Lakes (Fig. 5a–b)show greater than 200 mg/kg As in the top 2 cmof the core, the highestvalues in each core. This suggests some combination of continuingarsenic inputs from the watershed and/or vertical or horizontal arsenicmigration in the sediments (Senn et al., 2007). Thus, surface sedimentsin lakes affected by ASARCO continue to be an environmental healthconcern two decades after the smelter ceased operations. High arsenicand lead concentrations persist, with the potential for direct impactson benthic organisms. Moreover, arsenic mobility and transport fromthe sediments into the overlying water column may also pose a signifi-cant threat to human health through fish consumption.
In our study the hypolimnetic waters of fourteen of the seventeenlakes are anoxic at least periodically during late summer (Table 1),potentially leading to iron oxide reductive dissolution and the releaseof bound arsenic. This leads to elevated arsenic concentrations in thebottomwaters of lakes in the region (Table 2),with the potential for up-ward mixing into oxygenated waters where fish reside. The mean ratioof the relationship in our study (0.42) between themaximum dissolvedarsenic in the water column (μg/L) and total arsenic in the surfacesediments (mg/kg) of lakes exhibiting hypolimnetic anoxia at thetime of sampling is similar to the relationship found by Peterson andCarpenter (1986) in Lake Washington comparing the maximumdissolved arsenic in porewater and total arsenic in sediments (0.46after unit conversion). In a study of multiple lakes in Massachusetts,the average ratio of maximum dissolved arsenic in the water columnto total arsenic in surface sediments was 0.37, with most of theselakes also anoxic in the summer (Lattanzi et al., 2007).
4.1. Conclusions
Thus, lakes in urban areas contaminated with arsenic may be particu-larly important to monitor as possible threats to human health throughfish consumption. Cultural eutrophication in urban areas decreases dis-solved oxygen levels in the bottom waters of lakes, leading to reducingconditions that contribute to arsenic mobility. Our results show clear ev-idence for increased arsenic exposure in urban lakes in the depositionzone impacted by the ASARCO smelter, pointing to the need to furtherinvestigate arsenic fate and transport in these lakes.
537J.E. Gawel et al. / Science of the Total Environment 472 (2014) 530–537
Acknowledgments
The authors wish to thank Patti Sandvik, Lindsay Tuttle, JoeChynoweth and Brian Rurik for early contributions to this research,and Dena Reaugh for creation of all GIS-based figures in this paper.Funding for this researchwas provided by theUniversity ofWashingtonTacoma Founders Endowment, with writing assistance provided bythe Helen R. Whiteley Center at the University of Washington's FridayHarbor Laboratories.
Appendix A. Supplementary data
Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.scitotenv.2013.11.004.
References
Aggett J, O'Brien GA. Detailed model for the mobility of arsenic in lacustrine sedimentsbased on measurements in Lake Ohakuri. Environ Sci Technol 1985;19:231–8.
Arain MB, Kazi TG, Baig JA, Jamali MK, Afridi HI, Shah AQ, et al. Determination of arseniclevels in lake water, sediment, and foodstuff from selected area of Sindh, Pakistan:estimation of daily dietary intake. Food Chem Toxicol 2009;47:242–8.
Area-Wide Soil Contamination Task Force. Area-Wide Soil Contamination Task ForceReport. Prepared for Washington State Department of Ecology, Olympia, WA; 2003.[http://www.ecy.wa.gov/programs/tcp/area_wide/Final-Report/PDF/TF-Report-final.pdf].
Batterson TR, McNabb CD. Arsenic in Lake Lansing, Michigan. Environ Toxicol Chem1983;2:1–17.
Bennett J, Cubbage J. Copper in sediments from Steilacoom Lake, Pierce County,Washington, Report 92-e00. Prepared for Sediment Management Unit, WashingtonState Department of Ecology, Olympia, Wash; 1992. [https://fortress.wa.gov/ecy/publications/summarypages/92e00.html].
Bortleson GC, Dion NP, McConnell JB, Nelson LM. Reconnaissance data on lakes inWashington, Volume 3: Kitsap, Mason, and Pierce Counties, Water Supply Bulletin43, Vol. 3. Olympia, WA: Washington State Department of Ecology; 1976.
Couture RM, Shafei B, Van Cappellen P, Tessier A, Gobeil C. Non-steady state modeling ofarsenic diagenesis in lake sediments. Environ Sci Technol 2010;44:197–203.
Crecelius EA. The geochemical cycle of arsenic in Lake Washington and its relation toother elements. Limnol Oceanogr 1975;20:441–51.
Crecelius EA, Piper DZ. Particulate lead contamination recorded in sedimentary coresfrom Lake Washington, Seattle. Environ Sci Technol 1973;7:1053–5.
Crecelius EA, BothnerMH, Carpenter R. Geochemistries of arsenic, antimony,mercury, andrelated elements in sediments of Puget Sound. Environ Sci Technol 1975;9:325–33.
Entranco Engineers, Inc. Snake Lake restoration: a Phase 1 diagnostic and restorationfeasibility study. Report prepared for the Metropolitant Park District of Tacoma,Tacoma, WA; 1989.
Ferguson JF, Gavis J. A review of the arsenic cycle in natural waters. Water Res 1972;6:1259–74.
Glass GL. Credible Evidence Report, the ASARCO Tacoma Smelter and Regional SoilContamination in Puget Sound, Final Report, Seattle, WA; 2003.
Glass GL. Pierce County Footprint Study, Soil Arsenic and Lead Contamination in WesternPierce County, Final Report, prepared for Tacoma-Pierce County Health Departmentand Washington State Department of Ecology, Seattle, WA; 2004. [http://www.ecy.wa.gov/programs/tcp/sites_brochure/tacoma_smelter/tsp_Pierce_county_studies/Footprint/XPC%20Foot%20entire%20report.pdf].
Harrington JM, Fendorf SE, Rosenzweig RF. Biotic generation of arsenic(III) inmetal(loid)-contaminated freshwater lake sediments. Environ Sci Technol 1998;32:2425–30.
Jack R. Investigation of background inorganic and organic arsenic in four Washingtonlakes, Report 03-03-024. Olympia, WA: Washington State Dept. of Ecology; 2003[https://test-fortress.wa.gov/ecy/testpublications/SummaryPages/0303024.html].
Johnson A, Norton D. 1989 lakes and reservoir water quality assessment program:survey of chemical contaminants in ten Washington lakes, Report 90-e35. Olympia,WA: Washington State Department of Ecology; 1990 [https://fortress.wa.gov/ecy/publications/summarypages/90e35.html].
Kazi T, Arain M, Baig J, Jamali M, Afridi H, Jalbani N, et al. The correlation of arsenic levelsin drinking water with the biological samples of skin disorders. Sci Total Environ2008;407:1019–26.
King County. King County Lake Stewardship. https://your.kingcounty.gov/dnrp/wlr/water-resources/small-lakes/data/default.aspx, 2013.
Knauer K, Behra R, Hemond H. Toxicity of inorganic and methylated arsenic to algal com-munities from lakes along an arsenic contamination gradient. Aquat Toxicol 1999;46:221–30.
Kornijow R. A new sediment slicer for rapid sectioning of the uppermost sediment coresfrom marine and freshwater habitats. J Paleolimnol 2013;49:301–4.
Larson TV, Charlson RJ, Knudson EJ, Christian GD, Harrison H. The influence of a sulfurdioxide point source on the rain chemistry of a single storm in the Puget Soundregion. Water Air Soil Pollut 1975;4:319–28.
Lattanzi PR, Senn DB, Jay JA, Monastra V, Regan KM, Durant JL. Persistence and remobili-zation of arsenic in Massachusetts (USA) lakes treated with arsenical herbicides. LakeReservoir Manage 2007;23:59–68.
Lee C, Low K, Hew N. Accumulation of arsenic by aquatic plants. Sci Total Environ1991;103:215–27.
MacDonald DD, Ingersoll CG, Berger TA. Development and evaluation of consensus-basedsediment quality guidelines for freshwater ecosystems. Arch Environ Contam Toxicol2000;39:20–31.
Mariner PE, Holzmer FJ, Jackson RE, White JE, Wahl AS, Wolf FG. Fingerprinting arseniccontamination in the sediments of the Hylebos Waterway, Commencement BaySuperfund site, Tacoma, Washington. Environ Eng Geosci 1997;3:359–68.
Mathieu C, Friese M. PBT chemical trends in Washington State determined fromage-dated lake sediment cores: 2011 sampling results, Report 12-03-045. Environ-mental Assessment Program, Washington State Department of Ecology, Olympia,WA; 2012 [https://fortress.wa.gov/ecy/publications/publications/1203045.pdf].
McConnell GC, Bortleson GC, Innes JK. Data on selected lakes in Washington, Part 4, WaterSupplyBulletin 42, Part 4.Olympia,WA:WashingtonStateDepartment of Ecology; 1976.
Newell RG, Rogers K. The U.S. experience with the phasedown of lead in gasoline.Washington, DC: Resources for the Future; 2003.
Nittrouer CA, Sternberg RW, Carpenter R, Bennett JT. The use of Pb-210 geochronology asa sedimentological tool: application to the Washington continental shelf. Mar Geol1979;31:297–316.
Peterson ML, Carpenter R. Arsenic distributions in porewaters and sediments of PugetSound, Lake Washington, the Washington coast and Saanich Inlet, B.C. GeochimCosmochim Acta 1986;50:353–69.
Seattle & King County Public Health, Glass GL. Final report: Vashon/Maury Islandsoil study 1999-2000. Seattle, WA: Seattle-King County Department of Public Health,Environmental Health Division; 2000 [http://www.kingcounty.gov/healthservices/health/ehs/toxic/VashonMaury.aspx].
Senn DB, Gawel JE, Jay JA, Hemond HF, Durant JL. Long-term fate of a pulse arsenic inputto a eutrophic lake. Environ Sci Technol 2007;41:3062–8.
Sumioka SS, Dion NP. Trophic classification of Washington lakes using reconnaissancedata, Water Supply Bulletin 57. Olympia, WA: Washington State Department ofEcology; 1985.
Tanner CC, Clayton JS. Persistence of arsenic 24 years after sodium arsenite herbicide appli-cation to Lake Rotoroa, Hamilton, New Zealand. N Z J Mar Freshw Res 1990;24:173–80.
Tetra Tech, Inc. Management of Wapato Lake quality. Report for Tacoma Metro Parks,Seattle, WA; 2008.
Wängberg S-Å, HeymanU, Blanck H. Long-term and short-term arsenate toxicity to fresh-water phytoplankton and periphyton in limnocorrals. Can J Fish Aquat Sci 1991;48:173–82.
Washington State Department of Ecology. Washington State Department of Ecology,Toxics Cleanup Program, Tacoma Smelter Plume Site, King County Mainland SoilStudy. Olympia, WA: Washington State Department of Ecology; 2002 [http://www.ecy.wa.gov/programs/tcp/sites_brochure/tacoma_smelter/executive_summary_020404/exec_summary].
Washington State Department of Ecology. Overview: reducing toxics in fish, sedimentsand water, reducing toxic threats. http://www.ecy.wa.gov/toxics/fish_overview.html, 2013.
Washington State Department of Ecology. National toxics rule human health criteriafor water and equivalent fish tissue concentrations. 2013. [http://www.ecy.wa.gov/programs/wq/swqs/NTRbyPriorityPollutantName.pdf].
Wolcott EE. Lakes of Washington, Volume 1: Western Washington. 3rd ed. Olympia, WA:Washington State Department of Ecology; 1973.
Yake B. The use of sediment cores to track persistent pollutants in Washington State: areview, Report 01-03-001. Olympia, WA: Washington State Department of Ecology;2001 [https://fortress.wa.gov/ecy/publications/summarypages/0103001.html].
