The influence of geology and land use on arsenic in stream sediments and ground waters in New England, USA

The influence of geology and land use on arsenic in stream sediments and ground waters in New England, USA
of 16
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  The influence of geology and land use on arsenic instream sediments and ground waters in New England, USA Gilpin R. Robinson Jr.  a,* , Joseph D. Ayotte  b a US Geological Survey, 954 National Center, Reston, VA 20192, United States b US Geological Survey, 361 Commerce Way, Pembroke, NH 03275-3719, United States Received 14 March 2005; accepted 11 May 2006Editorial handling by A.H. WelchAvailable online 25 July 2006 Abstract Population statistics for As concentrations in rocks, sediments and ground water differ by geology and land use featuresin the New England region, USA. Significant sources of As in the surficial environment include both natural weathering of rocks and anthropogenic sources such as arsenical pesticides that were commonly applied to apple, blueberry and potatocrops during the first half of the 20th century in the region. The variation of As in bedrock ground water wells has a strongpositive correlation with geologic features at the geologic province, lithology group, and bedrock map unit levels. The var-iation of As in bedrock ground water wells also has a positive correlation with elevated stream sediment and rock As chem-istry. Elevated As concentrations in bedrock wells do not correlate with past agricultural areas that used arsenicalpesticides on crops. Stream sediments, which integrate both natural and anthropogenic sources, have a strong positive cor-relation of As concentrations with rock chemistry, geologic provinces and ground water chemistry, and a weaker positivecorrelation with past agricultural land use. Although correlation is not sufficient to demonstrate cause-and-effect, the sta-tistics favor rock-based As as the dominant regional source of the element in stream sediments and ground water in NewEngland. The distribution of bedrock geology features at the geologic province, lithology group and map unit level closelycorrelate with areas of elevated As in ground water, stream sediments, and rocks.   2006 Elsevier Ltd. All rights reserved. 1. Introduction National and regional studies of As occurrence indrinking water (Welch et al., 2000; Peters et al.,1999; Ayotte et al., 1999, 2003) have identified areaswithin New England, USA, where ground waterwells have As concentrations that frequently exceedthe maximum contaminant level (MCL) value of 0.01 mg/L for drinking water (US EnvironmentalProtection Agency, 2002; Smith et al., 2002). Watersamples from drilled wells in bedrock have the high-est As concentrations, and it is estimated that asmany as 10 to 40% of bedrock ground water wellsin some areas exceed the MCL standard (Peterset al., 1999; Ayotte et al., 1999, 2003; Montgomeryet al., 2003). Most water samples from ground waterwells in surficial sediments have As concentrationsbelow the MCL (Ayotte et al., 1999). 0883-2927/$ - see front matter    2006 Elsevier Ltd. All rights reserved.doi:10.1016/j.apgeochem.2006.05.004 * Corresponding author. Fax: +1 703 648 6383. E-mail address:  grobinso@usgs.gov (G.R. Robinson Jr.).Applied Geochemistry 21 (2006) 1482–1497 www.elsevier.com/locate/apgeochem  Applied Geochemistry   Previous local and regional studies have identi-fied correlations between areas with elevated groundwater As and geologic features and units (Peterset al., 1999; Ayotte et al., 1999, 2003; Montgomeryet al., 2003). In some studies, the possibility thatanthropogenic activity can affect ground water Asconcentrations have been acknowledged but corre-lations have not been found (Marvinney et al.,1994; Boudette et al., 1985; Zeuna and Keane,1985). The widespread use of arsenical pesticidesand herbicides on crops and shrubs in the regionhas been suggested as the dominant source of anthropogenic As to surficial soils and sediments(D’Angelo et al., 1996; Marvinney et al., 1994; Per-yea, 1998; Nriagu and Pacyna, 1988). Stream sedi-ment chemistry in the New England regionappears to be influenced by both geologic and arsen-ical pesticide sources (Chormann, 1985; Robinsonand Ayuso, 2004). The objective of this study is tocompare As distributions and covariation map pat-terns to evaluate spatial associations between Asconcentrations in rock, sediment and water withgeology, lithology and land-use features. 1.1. Occurrence and distribution of arsenic in thesurficial environment1.1.1. Rocks, soils, and sediments Although the average concentration of As in theearth’s crust is low, on the order of 1.7 mg/kg (Wed-epohl, 1995), As is widely distributed and com-monly found in many rock types, sediments, andsoils at concentrations near or exceeding this level(Smedley and Kinniburgh, 2002). Arsenic concen-tration in rocks varies by mineralogy, rock type,and geologic setting. In its reduced form, As is com-monly concentrated in sulfide and sulfosalt minerals(Rose et al., 1979). Pyrite, a common sulfide mineralin many rock types, is a prevalent host of As in mostrocks (Kolker and Nordstrom, 2001). Weatheringof As-containing rocks is considered to be the dom-inant source of natural As in the environment(Tamaki and Frankenberger, 1992). The baselineconcentration of As in stream sediments and soilsis in the order of 5–8 mg/kg (Smedley and Kinni-burgh, 2002); soils in the USA. have an average con-centration of approximately 7.4 mg/kg (Shackletteet al., 1974). In sediments, soils, and other oxidizedweathered environments, As has an affinity for sorp-tion and concentration with hydrous Fe oxide min-erals and mineral coatings (Goldberg, 2002;Stollenwerk, 2003; Smedley and Kinniburgh,2002). In similar weathering environments, whereAs concentrations are high, it also occurs in metal-arsenate–phosphate phases (Ayuso and Foley,2002; Wauchope, 1975; Yan-Chu, 1994; Woolson,1977). Stream sediments and soils integrate As fromboth natural and anthropogenic sources. The dom-inant natural source of As in sediments and soils isgeologic and is dependent upon the nature of theweathering environment and the concentrationand mineral form of As in the parent rock material(Tamaki and Frankenberger, 1992). 1.1.2. Anthropogenic sources of environmental arsenic The most significant anthropogenic source of Asin the region is believed to be from cumulativeapplications of arsenical pesticides and herbicidesthat were used in New England from the late1800s until the late 1960s (Peryea, 1998; D’Angeloet al., 1996). The application of arsenical pesticidesin New England predates systematic record-keepingand the exact locations and amounts of pesticideapplications are unknown, but it is estimated, basedon cultivation practices and history, that cumulativeapplication rates of As could have been 22 g/m 2 insome cultivated areas in the region (equivalent to200 lbs of elemental As per acre of cultivation, asreported by D’Angelo et al. (1996)). Widespreaduse of arsenical agricultural compounds occurredduring the 1920s to 1950s on agricultural lands con-taining potato fields, apple orchards and blueberryfields (D’Angelo et al., 1996); cultivation data forthese crops during this time period (US Departmentof Agriculture, 1935-1997) indicate, where arsenicalpesticides and herbicides were used most extensivelyin the region. 1.1.3. Ground water wells used for drinking water Arsenic is found at low levels in many naturalwaters (Smedley and Kinniburgh, 2002) and typicalwater concentrations are low in relation to the typ-ical abundance of As in the rocks, mineral coatingsand sediments that are in contact with ground waterin most settings. A very small amount of dissolutionor desorption of bound As from these materials canlead to high concentrations of As in local groundwater, explaining why many ground water areashigh in dissolved As are found in areas with nearaverage As concentrations in rocks and soils (Welchet al., 2000; Stollenwerk, 2003).Arsenic concentrations in ground water wellstypically exhibit a high degree of variability on a G.R. Robinson Jr., J.D. Ayotte / Applied Geochemistry 21 (2006) 1482–1497   1483  well-to-well basis, even within close proximity toone another. It is difficult to predict the As concen-tration in a particular well based on the As concen-trations in neighboring wells (Ayotte et al., 2003;Welch et al., 2000). 2. Data used in spatial analysis  2.1. Sample chemistry data sets 2.1.1. Rock arsenic chemistry Rock chemistry data from 149 samples collectedand analyzed for this study and geochemical datafor 1125 rock samples from the New England regionpreviously analyzed by US Geological Survey ana-lytical laboratories were used in the study. The rocksamples collected during this study were analyzedfor As using hydride-generation atomic absorptionspectrometry (HGAA) at a single laboratory facil-ity, XRAL Laboratories (Canada). Analyses of ref-erence materials and standards were used tostandardize and calibrate the equipment. TheHGAA method is described in Taggart (2002).Additional analytical data for As concentrationsin representative rock samples from the New Eng-land region were retrieved from the PLUTO data-base archive of geochemical data determined byUS Geological Survey laboratories (Baedeckeret al., 1998). Geochemical data was included onlyif the analytical technique reported a minimumdetection level of 0.3 mg/kg or lower for As. Rockanalysis data used in this study were restricted torepresentative rock samples from outcrop settings,where the rock type is identified. The rock samplesare from a large variety of igneous and metamor-phic rock types, but are dominated by samples of granite and felsic volcanics. Rock samples collectedfrom the vicinity of mines, prospects, mine dumps,or otherwise mineralized settings were omitted.The distribution of 1274 rock samples that meetthe above criteria cluster along two NE-trendingtransects from (1) southern Connecticut to northernVermont near the Vermont-New Hampshire borderand (2) eastern Connecticut to coastal Maine. Otherrock samples are clustered in areas across Maine(Fig. 1a).  2.1.2. Stream sediment arsenic chemistry The stream sediment As data used in this studyare a subset of 1597 stream sediment samples(Fig. 1b) that were re-analyzed by using similarmethods to those described in Section 3 and witha reporting limit of 0.3 mg/kg. The samples wererandomly selected from an archive of approximately Fig. 1. (a) Rock chemistry sample locations in New England, showing As concentration ranges. The symbol legend for the Asconcentration data is based on percentiles for the ranked As concentrations. Symbol transitions occur at 0.5 (1.1 mg/kg), 0.8 (6.5 mg/kg),and 0.95 (25 mg/kg) percentile values. (b) Stream sediment chemistry sample locations in New England, showing As concentration ranges.The symbol legend for the As concentration data is based on percentiles for the ranked As concentrations. Symbol transitions occur at 0.5(2.8 mg/kg) and 0.8 (7.3 mg/kg) percentile values. (c) Water chemistry sample locations from public-supply bedrock wells in New England,showing As concentration ranges. The symbol legend for the As concentration data is based on percentiles for the ranked Asconcentrations. Symbol transitions occur at 0.5 (0.005 mg/L), 0.8 (0.01 mg/L), and 0.95 (0.02 mg/L) percentile values.1484  G.R. Robinson Jr., J.D. Ayotte / Applied Geochemistry 21 (2006) 1482–1497   7900 stream sediment samples distributed through-out New England collected from 1977 to 1980 bythe National Uranium Resource Evaluation(NURE) Program conducted by the Departmentof Energy. The NURE program did not sampledrainages in Maine north of 45   latitude. The sedi-ment samples processed under the NURE programwere sieved to below 100 mesh (<150  l m grain size).Information on the NURE sample-site attributes isgiven in Smith (2001a,b) and the reanalyzed analyt-ical data are provided in Robinson et al. (2004).  2.1.3. Bedrock well arsenic chemistry Arsenic data for ground water wells that pene-trate bedrock in New England were obtained fromstate records on public-water supply wells collectedfor compliance with the Safe Drinking Water Actduring the years of 1992–1999 (Fig. 1c). The dataand criteria for selection are described in Ayotteet al. (1999, 2003). Sources of data include public-water supply records collected by the MaineDepartment of Health; the New Hampshire Depart-ment of Environmental Services, Water Division;the Massachusetts Department of EnvironmentalProtection, Bureau of Resource Protection, Drink-ing Water Program; the Rhode Island Departmentof Health; and the Vermont Agency of NaturalResources. The laboratory reporting level (LRL)for the ground water samples used in this studywas variable but not higher than 0.005 mg/L.Approximately 80% of the wells in the databasehave As concentrations that are below the highestLRL limit of 0.005 mg/L. For all censored anduncensored wells, where the reported analytical val-ues for As are below 0.005 mg/L, ranks wereassigned as tie values in the nonparametric statisti-cal procedures and statistics.  2.2. Explanatory variable data sets 2.2.1. Bedrock geology and lithology spatial datalayers The geologic data sets were compiled from state-wide maps of bedrock geology for Connecticut(Rogers, 1985), Maine (Osberg et al., 1985), Massa- chusetts (Zen et al., 1983), New Hampshire (Lyons et al., 1997), Rhode Island (Hermes et al., 1994), and Vermont (Doll et al., 1961). Over 1200 individ-ualmapunitsarenamed andportrayedinthesestatebedrock geology maps. Map units that were the siteof 5 or more sediment or ground water samples wereevaluated for differences in As distribution. Foranalysis of differences in As distributions, the bed-rock map units were grouped into lithology and geo-logic province categories using map unitdescriptions and other geologic information pro-vided with the geologic maps. These lithology groupcategories (Robinson and Kapo, 2003) are shown inFig. 2a. The bedrock map units have been dividedinto geologic provinces (Fig. 2b), using the geologicprovince categories presented in Robinson andKapo (2003). Each province group shares commonfeatures of (1) lithology, (2) age of formation, (3)geologic setting, and (4) tectonic history. The prov-ince groups generally occur as NE trending beltsthat follow the structural fabric of the Appalachianfoldbelt and faults in New England. Each provincegroup was evaluated for differences in As distribu-tion. The geologic province categories that are adja-cent and have similar As distributions for groundwater, sediment and rocks were combined to sim-plify the discussion and presentation of results.  2.2.2. Agricultural lands with As-pesticide use data Robinson and Ayuso (2004) demonstrate thatelevated As concentrations in stream sediments cor-relate with former agricultural areas in New Eng-land that used arsenical pesticides. Previousstudies (Boudette et al., 1985; Marvinney et al.,1994; Zeuna and Keane, 1985) have suggested a linkbetween areas with elevated ground water As andagricultural areas with historic application of largeamounts of arsenical pesticides.Long-term site-specific application rates of arsen-ical pesticides are not known in the region but canbe inferred from regional crop production recordsand general pesticide application rates. This studyuses agricultural census data for apple, blueberry,and potato crops in New England from 1935 to1977 (US Department of Agriculture, 1935-1997)to define the location and estimate the relative inten-sity of arsenical pesticide applications in the region.D’Angelo et al. (1996) estimate that, on a per-acre-basis, the cumulative application rates for arsenicalpesticides applied to apple, blueberry, and potatocrops are comparable. The agricultural census datawere compiled by crop type at county level for mul-tiple agricultural census years. The GIRAS land-usedatabase (scale 1:250,000; Hitt, 1994) compiledfrom high-altitude aerial photographs from the late1960s to early 1970s was used to partition thecounty-scale agricultural census data into smallerunits, based on census tracts, where agriculturalland was located in each county (US Census G.R. Robinson Jr., J.D. Ayotte / Applied Geochemistry 21 (2006) 1482–1497   1485  Bureau, 2000a,b). The agricultural data from multi-ple census years were averaged on a census tractlevel to determine the area of apple, blueberry andpotato cultivation per tract area. The agricultural-index (Agr-index) value is the ratio of cultivationarea divided by tract area, given in percent units.These census-tract level agricultural cultivation esti-mates (areas on the order of a few tens of km 2 ) werecompared to stream sediment data that are sampledin drainages in the order of a few tens of km 2 . TheAgr-index is used as a proxy for the arsenical pesti-cide application rate in the area (Fig. 3).  2.2.3. Interpolation maps for stream sediment and  ground water well arsenic chemistry The chemical data for stream sediments, rocks,and ground waters has been interpolated so thatspatial comparisons could be made (Fig. 4a, b).The interpolation grids were produced using theGeoDAS System inverse-distance weighted (IDW)multifractal interpolation method (Cheng, 2003;Agterberg, 2001), with an output grid cell size of approximately 1 km 2 . IDW techniques have beenrecommended in studies comparing interpolationmethods (Weber and Englund, 1992, 1994; Englundet al., 1992). Multifractal IDW is a technique thatfits the source data accurately and preserves localanomalies in the interpolation grid. The interpola-tion grids of stream sediment and ground waterchemistry provide coverage over more than 70%of New England. Due to the unevenly scattered dis-tribution of rock samples, the interpolation grid forrock As concentrations is poorly constrained. Forthis study, the rock As interpolation grid wasrestricted to areas within 2 km of a rock sample.For the interpolation grid process, the censoreddata for ground water well As values below theLRL were assigned estimated values of one half theLRL value. After interpolation, all grid cell valueslessthantheupper LRLvalueof0.005 mg/LAswerereassigned a value of <0.005 for data comparisonevaluations. The interpolation grid process generatessmoothed geochemical gradients that eliminate someof the local variability inherent in the point sourcedata. Data smoothing is particularly evident in theinterpolation grid for bedrock well water when com-pared to the well data that displays a high degree of well-to-well variability in areas, where higher As con-centrations in well water are prevalent.All statistical analyses in this study were per-formed on categorical or rank transformed data;these transformations are not sensitive to differences Fig. 2. (a) Generalized lithology of bedrock geologic units in New England. The rock types have been grouped into seven generalcategories based on map unit descriptions and depositional setting information summarized in Robinson and Kapo (2003).(b) Generalized geologic provinces in New England. The provinces have been grouped into 6 general categories based on descriptionsand information summarized in Robinson and Kapo (2003).1486  G.R. Robinson Jr., J.D. Ayotte / Applied Geochemistry 21 (2006) 1482–1497 
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!