INFORMATION TO USERS This was produced from a copy of a document sent to us for microfilming. While the most advanced technological means to photograph and reproduce this document have been used, the quality is heavily dependent upon the quality of the material submitted. The following explanation of techniques is provided to help you understand markings or notations which may appear on this reproduction. 1.The sign or “target” for pages apparently lacking from the document photographed is “Missing Page(s)”. If it was possible to obtain the missing page(s) or section, they are spliced into the film along with adjacent pages. This may have necessitated cutting through an image and duplicating adjacent pages to assure you of complete continuity. 2. When an image on the film is obliterated with a round black mark it is an indication that the film inspector noticed either blurred copy because of movement during exposure, or duplicate copy. Unless we meant to delete copyrighted materials that should not have been filmed, you will find a good image of the page in the adjacent frame. 3. When a map, drawing or chart, etc., is part of the material being photo­ graphed the photographer has followed a definite method in “sectioning” the material. It is customary to begin filming at the upper left hand comer of a large sheet and to continue from left to right in equal sections with small overlaps. If necessary, sectioning is continued again—beginning below the first row and.continuing on until complete. 4. For any illustrations that cannot be reproduced satisfactorily by xerography, photographic prints can be purchased at additional cost and tipped into your xerographic copy. Requests can be made to our Dissertations Customer Services Department. 5. Some pages in any document may have indistinct print. In all cases we have filmed the best available copy. University Microfilms International 3 0 0 N. Z E E B RO A D , A N N A R B O R , Ml 4 8 1 0 6 18 B E D F O R D ROW, L O N D O N WC1 R 4 E J , E N G L A N D 8106466 Ba t t e r s o n , T e d R a n d a l l ARSENIC IN LAKE LANSING, MICHIGAN Michigan State University University Microfilms International Ph.D. 300 K Zeeb Road, Ann Arbor, MI 48106 1980 PLEASE NOTE: In all cases this material has been filmed 1n the best possible way from the available copy. Problems encountered with this document have been Identified here with a check mark v" . 1. Glossy photographs ______ 2. Colored Illustrations ______ 3. Photographs with dark background ______ 4. Illustrations are poor copy ______ 5. Dr1nt shows through as there 1s text on both sides of page _____ 6. Indistinct, broken or small print on several pages ^ 7. Tightly bound copy with print lost in spine _ _ _ _ _ _ 8. Computer printout pages with indistinct print 9. Page(s) _ _ lacking when material received, and not available from school or author 10. _ Page(s) _ _ _ _ _ _ _ seem to be missing in numbering only as text follows 11. Poor carbon copy ______ 12. Not original copy, several pages with blurred type ______ 13. Appendix pages are poor copy ______ 14. Original copy with light type ______ 15. Curling and wrinkled pages ______ 16. Other_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ________ _ University M iadnlms Internationa] 300 \ Z = = £ 3 0 . ANN ARBOR Ml >*8106 '3131 761-4700 ARSENIC IN LAKE LANSING, MICHIGAN By Ted Randall Batterson A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Fisheries and Wildlife 1980 ABSTRACT ARSENIC IN LAKE LANSING, MICHIGAN By Ted Randall Batterson Lake Lansing, Michigan was treated with sodium arsenite for control of aquatic macrophytes in 1957. Two 2.5 m sedi­ ment cores from deep portions of the lake basin were analyzed for total As in 5 cm increments. Seventeen to 20 ug g”^ dry weight occurred in lower portions of the cores, and this was taken as background. Both cores had maxima of 330-340 ug g-1 at depth interval 0.15-0.30 m. These peaks were taken to re­ present contamination from weed treatment in 1957. The rate of decrease in recently deposited sediments predicted that concentrations near background would exist in surficial sedi­ ments in the deep portions of the basin by 1989. One hundred ten samples of the upper 7 cm of littoral sediments were taken along transects. Concentrations two to six times background occurred over 85-90% of the sediment surface. An arsenic mass balance budget was constructed for the lake for the interval June 1978 to June 1979. This showed that the lake lost more arsenic than it received annually. Ted Randall Batterson Internal loading of the water column from the sediments was occurring. A laboratory experiment was conducted to obtain predic­ tions of arsenic expected in the lake as a result of interac­ tion between contaminated sediments and the water column. Since sediments in the well aerated littoral, as well as those from the periodically anaerobic hypolimnion were con­ taminated, aerobic and anoxic treatments were compared. When lake water over sediments was aerated, total As of 5-25 ug per liter were observed over an 89 day experimental period. Arsenic (V) averaged greater than As(III); 8.6 and 7.4 ug I”1 respectively. Total phosphorus was monitored, and remained at 15-41 ug l- 1 . In anoxic treatments, total arsenic rose in the water to a maximum of 117 ug l-^ in 35 days. Although As(V) was present (<32 ug 1"^), the increase was due princi­ pally to As (III). It reached a maximum of 75 ug I- '*’ on day 35. Sulfide was detected in the anoxic chambers after day 35. In its presence, As(III) steadily decreased from its maximum, presumably precipitated as a sulfide. With this de­ crease in progress, the anoxic chambers were aerated on day 58. Arsenic (III) rose, while As(V) dropped over a 24 hour period. Oxidation of sulfide was implicated in the former case; oxidation of iron and subsequent precipitation of As(V) with compounds of ferric iron was proposed to explain the latter. Following this initial response to aeration, As(III) steadily declined. It was apparently oxidized to As(V) and Ted Randall Batterson removed with ferric iron. At the end of 89 days, As(III), A s ( V ) , and total As approached the initial concentrations and relative proportions. anaerobic conditions. Phosphorus rose with the onset of Not responding to the presence of sul­ fide, it continued to increase beyond day 35, and until the chambers were aerated. levels. It then declined toward pretreatment Since classical studies have shown such a decline to be associated with ferric iron, it appears that As(V) and phosphate compete for iron as anoxic systems are aerated. Concentrations of arsenic observed in the lake fit ex­ pectations from the experimental work. Aerated littoral and epilimnetic water was found to be 5-25 ug total As 1"^. Con­ centrations on the order of 100 ug 1“ ^ were observed in an­ aerobic hypolimnetic waters; however, concentrations that high were not the rule. Currents associated with summer storms periodically erode the metalimnion of this shallow lake. Ex­ cept for infrequent summer periods, sulfide or ferric iron tend to limit total As in the hypolimnion to a concentration of 20-30 ug I” 1 - DEDICATION To my parents ACKNOWLEDGMENTS I would like to express my most sincere thanks to Professor Clarence D. McNabb who unceasingly offered guid­ ance and inspiration during this endeavor. Thanks are also extended to my other committee mem­ bers, Drs. Niles R. Kevern, Kenneth W. Cummins, and Peter G. Murphy. My fellow graduate students, G. Douglas Pullman, Robert P. Glandon, Frederick C. Payne, and George W. Knoecklein de­ serve thanks for their many contributions to this work. Special thanks are due my friend, John R. Craig, Limnological Research Laboratory Director, who spent many long hours in helping me to complete this undertaking. Lynn Leonik's long, arduous task of sample preparation is greatly appreciated. James T. Carrick, Nuclear Reactor Supervisor, is sin­ cerely thanked for providing the availability of his facil­ ities as well as his invaluable assistance. My deepest appreciation goes to my wife, Kathleen and children, Cassandra and Benjamin, for without their support this work would not have been accomplished. This study was supported by funds provided by the U.S. Environmental Protection Agency, Clean Lakes Program, under Grant, No. R 80504601, the Michigan Agricultural Experiment Station at Michigan State University, and the Herbert and Grace A. Dow Foundation. TABLE OF CONTENTS Page LIST OF TABLES LIST OF FIGURES .......................................... ......................................... VH INTRODUCTION ............................................ 1 THE STUDY SITE .......................................... 6 MATERIALS AND METHODS RESULTS .................................. 23 .................................................. 35 DISCUSSION .............................................. 54 LITERATURE CITED ........................................ 68 APPENDIX ................................................. 73 V LIST OF TABLES Table 1. 2. 3. 4. A-l. A-2. A - 3. A —4. A - 5. A-6 . Page Seasonal aspects of the annual hydrologic bud­ get of Lake Lansing, Michigan ................... 41 Seasonal aspects of the annual arsenic budget of Lake Lansing, Michigan ....................... 44 Concentrations of total arsenic (yg As 1™^) in Lake Lansing, Michigan .......................... 53 Sediment arsenic concentrations with depth and the number of years since the sodium arsenite treatment in 1957. Years since treatment based on sedimentation rates of 1.1 cm yr” l for the north basin and 0.7 cm yr-1 for the south basin ............................................ 57 Arsenic inputs from marsh drains during 6/20/78 to 6/18/79 ...................................... 74 Arsenic inputs from street drains during 6/20/78 to 6/18/79 ............................. 75 Arsenic losses from the outflow during 6/20/78 to 6/18/79 ...................................... 76 Arsenic inputs to the lake surface from atmo­ spheric bulk fallout ........................... 77 Net gain or loss of arsenic due to seepage during 6/20/78 to 6/18/79 ...................... 78 The various forms of arsenic in the water of the experimental units (yg As 1"1) .............. 79 vi LIST OF FIGURES Figure 1. 2. 3. Page Morphometric map of Lake Lansing, Michigan showing transects, inflows, outflow, and areas treated with sodium arsenite in June 1957 (stippled) ................................. 7 Depth-area curves for the upper 5 m of Lake Lansing, Michigan above and for the two deep holes of the lake b e l o w .............. ......... 9 Depth-volume for Lake Lansing, Michigan with tabled volumes for strata of the two deep holes and the lake as a whole ........................ 11 Depth-time diagram of dissolved oxygen isopleths (mg 1~ 1) from the north basin deep hole, Lake Lansing, Michigan, 1978 .................... 14 Depth-time diagram of dissolved oxygen isopleths (mg 1” 1) from the south basin deep hole, Lake Lansing, Michigan, 1978 .................... 16 6 . Depth-time diagram of dissolved oxygen isopleths (mg 1” ^) from the north basin deep hole, Lake Lansing, Michigan, 1979 .................... 18 4. 5. 7. Depth-time diagram of dissolved oxygen isopleths (mg l- 1 ) from the south basin deep hole, Lake Lansing, Michigan, 1979 .................... 8 . The distribution of arsenic in a sediment core from 9 m in the north basin of Lake Lansing, Michigan, July 13, 1979 ....................... 9. 10. 20 36 The distribution of arsenic in a sediment core from 7 m in the south basin of Lake Lansing, Michigan, July 13, 1979 ....................... 38 Concentrations of arsenic (ug g"* dry weight) in the surficial sediments of Lake Lansing, Michigan, June 1979 ............................ 46 vii Figure 11. 12. Page Variation in the chemical composition of water overlying deep water sediments from the south basin of Lake Lansing for 89 days in experi­ mental tanks. Lefthand series: aerated treat­ ment; Righthand series: anoxic treatment, aer­ ated on day 58. A, arsenic concentrations for the (III), (V) , and total forms in yg l- 1 ; B, pH and concentrations of total phosphorus in yg 1~ 3- ........................................ 49 A proposed cycle for arsenic in a stratified lake (modified from Ferguson and Gavis, 1972; Wood, 1974) .................................... 60 INTRODUCTION Arsenic is a metalloid that exists in a variety of chem­ ical, biochemical, physical, and geochemical forms. It is widely distributed in nature and can be found in all environ­ mental substrates. It ranks 47th in the order of occurrence for elements in the earth's crust, constituting 5 x 10”^ per­ cent {Kipling, 1977) . Though it occasionally occurs in the free state, most arsenic is found in nature in a combined form. Minerals of which arsenic is a major constituent are arsenides, sulfarsenides or oxides of heavy metals such as silver, copper, nickel, lead, gold, and iron. CPeAsS) is the most abundant ore mineral Arsenopyrite (Onishi, 1969). Orpiment (AS2S3) and realgar (AsS) are other important miner­ als and are products of volcanic sublimation or deposits from hot springs. Onishi and Sandell (1955) discussed the geo­ chemical cycling of arsenic and conclude that volcanic exhala­ tions and hot springs have been the main source of the element now present in sediments and sedimentary rocks. Soil values, resulting from the weathering of parent rock material, range from 1 to 40 ppm As (Vinogradov, 1953 reported in Woolson, 1977); most soils average 6 ppm As (Bowen, 1966). The arsen­ ic content of coals varies greatly, probably reflecting the difference in where it was formed. 1 Coal ash from the 2 Virginias averaged 140 ppm As while elsewhere 8000 ppm As has (Onishi, 1969; Kipling, 1977) . been reported Seawater values ranged from 0.15 to 6.0 ug As 1”^ with an average value of 2 ug As 1“*. Various river waters ranged in value from 0.2 to 25.0 ug As 1“^ (Onishi, 1969) . Hot springs that are rich in bicarbonates frequently contain ele­ vated arsenic levels, ranging from 130 to 1000 ug As 1~^ (Woolson, 19 75) . Ritchie (19 61) reports that hot springs in New Zealand had values of arsenic as high as 8500 ug As 1“^. Lake waters display a wide range of values; from 0.1 to 243,000 ug As 1” ! (Onishi, 1969; Woolson, 1975; Schroeder and Balassa, 1966) . Lakes typically have values less than 10 ug As I-1; Durum, et al. (1971) report that 79% of the 727 samples they collected from surface waters of the United States were below this level. Excluding areas of volcanic activity or thermal springs, high concentrations of this element in either the water or sediments are usually the result of anthropogenic activity. The application of arsenical herbicides and pesticides, smelting and mining operations, and burning of fossil fuels have been the major sources of contamination Kobayashi and Lee, 1978; Lis and Hopke, (Shapiro, 1971; 1973; Walsh and Keeney, 1975; Crecelius, 1975; Wagemann, et a l . , 1978; Aston, et al., 1975). In the upper Great Lakes States, man has pur­ posely introduced arsenicals into aquatic systems as a means of controlling aquatic macrophytes. Minnesota's records show that between 1956 and 1969 over 408,233 kg of sodium arsenite 3 were added to lakes to kill rooted aquatic plants. In New York State, about 38,555 kg of this compound was used from 1961 to 1966 (Shapiro, 1971). In a twenty year period be­ ginning in 1950, Wisconsin's public waters received 741,495 kilograms of arsenic (Lueschow, 1972). Ferguson and Gavis (1972) have proposed that an arsenic cycle may exist in stratified lakes. In those basins that have accumulated a substantial arsenic burden from human sources, this cycle could be quite dramatic. Arsenic in aquatic systems has an interesting and unusually complex chemistry? oxidation-reduction, ligand exchange, adsorptiondesorption, and precipitation reactions can all take place (Ferguson and Gavis, 1972). It can occur in numerous oxida­ tion states (+5,+3,+1,0,-1,-3) and in inorganic and organic configurations. Exchanges between compartments of the water column and sediments are affected by diffusion, mixing by currents, and biological activity (Ferguson and Gavis, 1972; Wood, 1974). The dynamics of anthropogenic arsenic in lakes have not been well documented. A recent study by Kobayashi and Lee (1978) reported on the accumulation of arsenic in sediments of five Wisconsin lakes treated extensively with sodium arsenite. A maximum of 659 yg As g"1 dry weight was found in surface sediments. Core samples showed a progressive de­ crease in concentration in lower sediment layers. Ten parts per million As by dry weight were found in deep portions of cores. They took this to be background. They did not report 4 data for the overlying water columns. studied Chautauqua Lake, New York. Lis and Hopke (1973) They reported elevated levels (43.4 yg As 1“^) of dissolved arsenic in the water column. They hypothesized that these resulted from the slow release of arsenic from sediments that became enriched as a result of large-scale sodium arsenite treatment during the period 1955 to 1963. In a later report on the concentration and distribution of arsenic in the sediments of this lake, Ruppert, et al. (1974) attempted to support this contention. However, the hypothesis remained unsubstantiated since they did not measure inputs of arsenic from streams, seepage, and atmospheric fallout. Crecelius (1975) described geochemical cycling of arsenic in Lake Washington. He found elevated sediment concentrations; greater than 200 ppm As dry weight. However, the water column was at the low end of the freshwater range of concentrations, averaging 1.6 ppb As. Arsenic con­ tamination was attributed to a copper smelter 35 km upwind of the lake. Lake Washington does not typify most arsenic treated lakes of the Great Lakes States since its deep water does not become anaerobic and strongly reducing during the year. Lake Lansing, Michigan was chosen for this study. a productive, shallow lake of glacial origin. It is An historical record of treatment with sodium arsenite for weed control was available. The lake had received a single treatment in 1957. The hypolimnion of the lake was known to become anaerobic and reducing in nature during the summer. The objectives of the 5 investigation were: (1) to establish the historical perspec­ tive of the arsenic treatment by studying sediment cores, (2 ) to construct hydrologic and arsenic budgets to evaluate the impact of treatment on today's arsenic cycle, (3) to de­ scribe the surficial sediment distribution of arsenic, and (4) to conduct a laboratory experiment that would predict the arsenic concentration in the water column as a result of sediment-water exchange. THE STUDY SITE Lake Lansing is located in Ingham County, Michigan. It was created by the natural processes of glacial scouring and recession. The lake lies in the LaGrange moraine of the gla­ cial front known as the Saginaw Lobe (Martin, 1955). Depos­ its of a sand-gravel-clay soil were left during the retreat of Pleistocene glaciers (U.S. Army Corps of Engineers, The lake has a low relief watershed of 842 hectares. 1970). It has a surface area of 1816 x 103 m 2 , volume of 4124 x 103 m 3 , mean depth of 2.3 m, and a maximum depth of 10 m. Bathymetry is presented in Figure 1 showing the division of the lake in­ to a north and south basin. Depth-area and depth-volume curves are presented in Figures 2 and 3 respectively. The littoral zone of the lake extended to the 3 m depth contour. Seventy-nine percent of the surface area of the lake lies over this zone, while the pelagial region constitutes the remaining 21%. The lake has a single outlet that discharges only in the spring. Besides precipitation on the surface, water enters the lake via six intermittent surface streams and three street drains (Figure 1). The retention time has been calculated to be 19.5 y e a r s . During the 1 9 4 0 's and 1950's, fishing quality in Lake Lansing was progressively declining apparently due to 6 7 Figure 1 Morphometric map of Lake Lansing, Michigan showing transects, inflows, outflow, and areas treated with sodium arsenite in June 1957 (stippled). O utflow 9 Figure 2 Depth-area curves for the upper 5 m of Lake Lansing, Michigan above and for the two deep holes of the lake below. A r e a (m2 x IO '3) 200 400 600 0 800 1000 1400 1200 1600 1800 2000 L ake os a Whole 2 3 5 Depth (m) 4 0 25 50 75 100 125 ISO 175 200 225 250 o S o u th 10J Basin N o rth B asin 11 Figure 3 Depth-volume curve for Lake Lansing, Michigan with tabled volumes for strata of the two deep holes and the lake as a whole. Volume 2000 IOOO Volume 4- S tra tu m S outh Basin o-i (nrrxlO ) N orth Basin Whole Lake — 1600 67- 1-2 1108 2-3 604 3-4 49 269 336 4-5 32 198 230 123 141 10 -1 5 -6 6-7 8 61 69 7 -6 I 24 25 8-9 (m3 xlO~3) 8 3000 4000 13 increased biomass of aquatic vegetation (Roelofs, 1958). To reverse or halt this declining trend, the lake was treated to erradicate dense growths of submersed macrophytes. Sodi­ um arsenite was selected because of its relative low cost, effectiveness in eliminating nuisance weeds, its apparent harmlessness to either large or small fish, and its failure to exterminate or seriously diminish the supply of natural foods (Mackenthun, 1950). Approximately 20 hectares (11% of the surface area) of the lake, were selected for treat­ ment (cf. Figure 1). Those areas were treated in June 1957 by personnel from the Fish Division of the Michigan Depart­ ment of Conservation. They used 3785 liters of sodium ar­ senite (NaAs02 ) which contained 1 kg I-1 arsenic trioxide (AS2O 3) . This treatment resulted in an arsenic input of 2920 kg. This is the only documented application of arseni- cals to this lake. Lake Lansing is typically covered with ice from early December to mid-March. Anaerobic conditions have not been observed in the lake during recent winters. Dissolved oxy­ gen during the growing seasons of 1978 and 1979 is shown in Figures 4 through 7. Stratification was established in May 1978 and persisted through September of that year. In 1979, stratification set-up temporarily in the south basin in May but did not persist until June. Summer stratification of this lake is typically disrupted by high winds. These tend to come from the southwest and west, often in the company of rainstorms. The fetch of the south basin in relation to 14 Figure 4 Depth-time diagram of dissolved oxygen isopleths (mg 1” 1) from the north basin deep hole, Lake Lansing, Michigan, 1978. DL DEPTH (m) pv in MAY JUNE JULY AUGUST SEPTEMBER OCTOBER NOVEMBER 16 Figure 5 Depth-time diagram of dissolved oxygen isopleths (mg 1~1) from the south basin deep hole, Lake Lansing, Michigan, 1978. DEPTH (m) MAY JUNE JULY AUGUST SEPTEMBER OCTOBER NOVEMBER 18 Figure 6 Depth-time diagram of dissolved oxygen isopleths (mg 1”^) from the north basin deep hole, Lake Lansing, Michigan, 1979. D E PT H (m ) 10 9 MAY JUNE JULY AUGUST SEPTEMBER 20 Figure 7 Depth-time diagram of dissolved oxygen isopleths (mg 1"1) from the south basin deep hole, Lake Lansing, Michigan, 1979. 9 10 MAY JUNE JULY AUGUST SEPTEMBER 22 these winds is much shorter than that of the north basin. Hence, wind-related disruption of stratification is more frequent in the north basin. These impacts of winds can be seen in Figures 4 through 7 as the downward displacement of oxygen isopleths. With this pattern of circulation, hypolimnetic water temperature is relatively high in Lake Lansing. It ranged between 13° and 17° C during the sum­ mers of 1978 and 1979. Surface water temperatures were in the range of 22° to 26° C during that time. MATERIALS AND METHODS Sediment cores from the deep portions of Lake Lansing were taken to establish the degree of arsenic contamination due to weed treatment in 1957. sampling. Two sites were selected for They were from the deepest portion of the north and south basins. unconsolidated. These sediments were extremely loose and They were sampled by freezing the sediments onto the exterior surface of tubing which extended from the water surface and penetrated the sediments a known distance. Lengths of two inch o.d. thin-walled aluminum electrical conduit were used which were threaded and joined by couplings. Once at the site from which a sediment sample was to be taken, the water depth was measured. length of sediment desired. Added to this was the Sections of tubing were then selected which would exceed that length by several feet to provide for excess tubing above the water. It was imperative that all joints were water-tight; silicone sealant was ap­ plied to the threads to accomplish this. was stoppered and lowered into the water. The first section Additional lengths were added until the stoppered end was just above the sedi­ ment surface. The last section was attached and then the lower end was carefully pushed into the sediments to the ap­ propriate depth. After insertion into the sediments, 23 24 pelletized dry ice was added to the tube from the end extend­ ing above the water surface. The amount added was enough to freeze the sediments as well as a small portion of water above the sediment-water interface. Replenishment of dry ice was maintained at a rate to offset sublimation. Thirty min­ utes after the initial addition of dry ice the samples were retrieved. As the tube was raised out of the water the sec­ tions were uncoupled down to the frozen sample. That por­ tion was then lifted out of the water and placed in a verti­ cal postion. As the sample was removed from the sediments# the amorphous zone between frozen and unfrozen portions was smeared# thus disrupting the sediment's original position along the length of the tube. Once out of the water# this outer disrupted layer was stripped away. After that# the sample was wrapped in plastic and the tube repacked with dry ice for transportation back to the laboratory. Once there, the dry ice was removed from the tube and replaced with warm water. Conducted heat melted the sediments that were contig­ uous with the exterior of the tubing allowing for the tube to be pulled free. sections. The frozen sample was then cut into 5 cm These doughnut-shaped pieces were rinsed thorough­ ly with ion-free water and then individually placed in label­ ed plastic bags. There were two reasons for the rinsing: to wash away any contamination that might be the result of the sectioning process or the sediments being in contact with the aluminum tube# and to remove internal and external smear­ ed layers. 25 The frozen samples were then dried in a Napco Model 630 forced air drying oven at 70° C for 72 hours. The sample was then ground by mortar and pestle to a powder. From each of the well-mixed ground samples approximately one gram of sediment was removed and dried at 105° C for 24 hours. The sample was then introduced into a tared 2 dram polyvial and weighed. These vials had previously been acid washed. After weighing, the polyvials were heat-sealed and taken to Michi­ gan State University's nuclear reactor facility for neutron activation analysis. For each group of samples that was ir­ radiated there was included three standards for quantifying the analyses. Two of the standards were obtained from the National Bureau of Standards and prepared for introduction to the polyvials per their instructions. ard Reference Material 1645 chard Leaves). These were Stand­ (River Sediment) and 1571 (Or­ The other standard was a 2.0 ml solution containing 150 micrograms As ml- *-. A Triga Mark I nuclear reactor was used for irradiation. Twenty sediment samples and three standards were introduced into a 40 position specimen rack that was rotated during ir­ radiation to establish uniform flux for all sample positions. Sixteen to twenty hours following irradiation (allowing for the partial decay of ^ N a activity), the samples were count­ ed for 1000 seconds live-time with a 76.2 cm-* active vol­ ume Ge(Li) detector having a relative efficiency of 15% and an energy resolution of 1.8 keV FWHM at the 1.333 MeV photo­ peak of ®®Co. The source-to-detector geometry was kept 26 constant for all counts and the detector resolution was suf­ ficient to completely resolve the ^®As peak the adjacent peak of 82Br (554 k e V ) . (559 keV) and The gamma-ray spectrum from each sample and standard was analyzed by a Canberra Series 80 multi-channel analyzer. This analyzer computes the peak net area which is the number of counts in a peak that are above an average background level. Corrections for decay between counting time of the samples and stand­ ards were made. The mass of arsenic in the samples was derived from the time-corrected counts of the standards. Surficial sediment samples were collected from varying water depths along six transects (cf. Figure 1). These per­ manent transects were selected such that they crossed over the major sediment types of the lake. Transects 1 and 6 were located over fibrous peat; transects 2 and 3 over fine organic ooze; and transects 4 and 5 over sand mixed with fine organic particles (Siami, 1979). Since most of these sediments are loose and flocculent, typical dredging devices were not employed. Instead, a 5.7 cm i.d. clear acrylic tube that was 7 cm long was used for sampling. The samples were obtained by cautiously swimming down to the sediment surface, slowly inserting the open-ended tube into the sediments, stoppering first the top and then the bottom, and then re­ turning to the water surface. Samples were then dispensed into labeled zip-locked plastic bags and stored in a cooler prior to transport to the laboratory. Five samples from four water depths were taken along each of the six transects. 27 These surficial sediment samples were analyzed by neutron activation analysis according to the procedures described above. An annual hydrologic budget for the lake was construct­ ed from measurements of inputs (marsh drains and street drains of Figure 1), outflow, precipitation and evaporation, and net seepage. This budget covered the period June 20, 1978 to June 18, 1979. The methods used at Lake Lansing for measuring discharge from marsh drains and street drains have been described by Glandon, et al. (in press). Discharge of the outfall was measured three times per week from the on­ set of flow (April 27,1979) to cessation (June 1, 1979). Wa­ ter leaving the lake passed through a concrete control struc­ ture. Discharge was calculated by measuring stage height and applying this to a U.S.G.S. approximation formula for a rec­ tangular weir (U.S.D.I., 1967). Stage height measurements were recorded for the lake throughout the study. These were measured in relation to the southeast edge of the sill of the dam which had been measured to be at 259.6 m above mean sea level. At this stage height, the surface area of the lake was calculated to be 1816 x 10^ m^. When the elevation exceeded 259.6 m, natural banks of the lake and retaining walls re­ sulted in a nearly vertical rise (rather than lateral) of the water. Therefore, the surface area of the lake remained un­ changed for stage heights that exceeded this elevation. For those stage heights less than 259.6 m, the surface area was calculated from the depth-area curve of Figure 2. 28 Discharge (m-*) measurements from the marsh drains, street drains, and outfall were converted to changes in lake level by dividing those values by the average surface area of the lake for the interval in question. Precipita­ tion and evaporation data were collected at the weather sta­ tion located on the South Farm of Michigan State University, approximately 9 km from the lake. The net atmospheric ef­ fect on the water budget was determined for intervals of interest by the difference between precipitation and evapo­ ration (mm). The net atmospheric effect for an interval was added to the change in stage height that could be accounted for in that interval by the discharge of input and the out­ flow. The residual change in lake level for the interval was taken to be the result of seepage. Multiplying that re­ sidual by the corrected surface area for that interval gave the net volume (m^) gain or loss due to seepage. The water budget thus developed was used with measurements of arsenic concentrations to estimate the annual mass arsenic budget for the lake. Glandon, et al. (in press) have described the methods used at Lake Lansing for collecting and compositing represent­ ative water samples from the discharge of marsh drains, street drains, and the lake's outflow. A discharge-proportional scheme was used to make up composite samples for arsenic anal­ yses. Composite samples from these sources were preserved with 2 ml concentrated H N O 3 per liter and were refrigerated until analyzed. 29 Atmospheric arsenic loading was determined by placing triplicate lexan containers in a fabricated stand one meter above the surface of the lake and 50 m from the western shore in the south basin. The containers were 23 cm deep and had a surface area of 0.26 m^ each. Prior to placement, the con­ tainers were successively washed and rinsed with 1:1 HNO3, 1:1 HCl, and ion-free water. Six liters of deionized water were then poured into each container, rinsing the inside sur­ face well. One liter from each container was withdrawn. These three liters were then mixed together; a one liter sample was withdrawn, preserved with 2 ml concentrated HNO3 and refrigerated until analysis. the beginning of an interval. This sample represented "Beginning" sample containers were then carried to the field and exchanged for "finishing11 containers, those that had been exposed for a period of time. By having water in the containers, the lake surface which would be receiving atmospheric fallout was simulated. The interval exposure was varied to maintain water in the con­ tainers, but to prevent overflowing, depending on the precipitation-evaporation balance. was two weeks. Typically the exposure period The "finishing" containers were returned to the laboratory where all internal surfaces were scraped down with an acid-washed rubber spatula. Water in the containers was washed around to rinse the surfaces. Water from all con­ tainers was combined, the volume measured, and then mixed. From the well-mixed composite, one liter of water was with­ drawn, preserved with acid and refrigerated. This sample 30 represented the end of the exposure period. The atmospheric arsenic input to the surface area of the lake was calculated from the following equation: { (Cf x V f) - {Cb x Vb ) } As = -------- i--------------®sa x L3a where As = mass of arsenic input during the interval in ques­ tion: C^, Cb = concentration of total arsenic in the "finishing" and "beginning" containers, respectively: Vj?, Vb = volume of water in the "finishing" and "beginning" containers, respectively; Saa * total surface area of the containers; Lsa = average corrected (for changes in stage height) surface area of the lake for that interval. Lake water for arsenic determinations was collected by a PVC Kemmerer bottle and composited into three samples; lit­ toral and upper pelagial water, and two lower pelagial waters, one representing the north basin and the other the south ba­ sin. The littoral and upper pelagial water was composited from samples taken at mid-depth from the 0.5, 1.5, 2.5, and 3.75 m depth contours on each of six transects (cf. Figure 1), and from those same depths over both basin's deep holes. The volumes of water for the composite, representing each of 32 sampling sites, were in proportion to the volume of water in the zones of the lake that the samples represented. The north basin lower pelagial water sample was composited from volumes of water from 5.5, 6.5, 7.5, and 8.5 m in proportion 31 to the volumes of the 5-6, 6-7, 7-8, and greater than 8.0 m strata of the north basin of the lake. The south basin lower pelagial water was composited from volumes of water from 5.5 and 6.5 m in proportion to the volumes of the 5-6 and greater than 6 m strata of the south basin of the lake. A one liter sample from each of the three composites was preserved with HNO3 and refrigerated until analyzed. All water samples were analyzed for total arsenic by atomic absorption spectrophotometry. The method employed was the gaseous hydride evolution technique as described in Standard Methods for the Examination of Water and Wastewater 14th Edition (APHA, 1976). Following analyses, estimates of arsenic in the seepage component of the annual budget were made. A net volume loss from the lake due to seepage in an interval was multiplied by the average littoral and upper pelagial arsenic concentration for that interval on the as­ sumption that seepage from the lake occurred through sediments in the shallows of the basin (cf. McBride and Pfannkuch, Dunst and Beauheim, 1979). 1975; If a net seepage gain occurred in the basin over an interval of the year, that volume was multiplied by the average arsenic concentrations in the marsh drains at low flow. The relief of the watershed of Lake Lansing predicts that seepage to the lake comes predominately through the marshes to the north and east of the basin {cf. Winter, 1978). A laboratory experiment was designed to evaluate the in­ teraction between the water column and sediments of the lake. 32 Sediments were collected from the deep portion of the south basin using an Ekman dredge. These were homogenized. Five, 38 liter capacity metal containers, lined with 3 mil poly­ bags were filled with the mixture to a depth of 5 cm. This resulted in a 835 cm2 sediment surface and a remaining ca­ pacity in the containers of approximately 35 liters. A sed­ iment sample was removed from each unit and analyzed for ar­ senic by neutron activation with techniques previously de­ scribed. The units with sediments were then exposed to the atmosphere for five days. After that, they were filled with water taken from near the surface of the lake over the deep portion of the south basin. The water was carefully layered over the sediment in an attempt to minimize the disruption and suspension of material. a 15° C water bath. All units were aerated and placed in This temperature was the summer time ave­ rage for water overlying deep basin sediments. The experimental units had lids that were slightly con­ vex, creating an air space between the top of the water and the lid. A gasket and silicone sealant afixed the lid to the container. Lids were fitted with small diameter inflow and outflow gas ports, and a large diameter sampling port. Sus­ pended from the inflow gas port was a diffusor stone that hung below the water surface. The other end of this port was connected to a manifold with various gas sources. One unit was randomly selected and aerated during the experiment. The other four units were subjected to a period of anoxia and then aerated. The sampling procedure consisted 33 of first withdrawing water for arsenic and phosphorus deter­ minations. After that, in situ measurements of temperature, dissolved oxygen, and pH were taken. Temperature and dis­ solved oxygen were measured with a YSI Model 54A D.O. meter and probe. Hydrogen ion concentration was determined by us­ ing a Corning Model 6 portable pH meter with an Orion Model 91-25 epoxy combination probe. After the initial sampling, the four anoxic units were purged with 99.9% N 2 . ports were sealed. Then all Following in situ measurements, the anoxic units were briefly purged with N 2 to expel any intru­ sion of O 2 that might have resulted during the sampling pro­ cedure. as well. This filled the head-space above the water with N 2 At the end of 58 days, the four anoxic units were aerated for the remainder of the experiment (89 days). Dur­ ing the period of aeration, sampling continued. Water samples were analyzed for total phosphorus, arsen­ ic (III), arsenic (V), and total arsenic. Total phosphorus concentrations were determined colorimetrically using the single reagent-ascorbic acid method on unfiltered, persulfate digested samples (APHA, 1976). Before the addition of the color reagent, all samples and standards were subjected to a reducing reagent to eliminate arsenate interference (Johnson, 1971). Water for arsenic determinations was preserved with 1 ml concentrated HC1 and analyzed by atomic absorption spec­ trophotometry. was employed. The gaseous hydride technique described above Speciation was accomplished by the fact that only As(III) is effectively converted to hydride (Aggett and 34 Aspell, 1976). Therefore, aliquots of the sample were treat­ ed differently. One was analyzed directly to determine the As (III) content. Another was subjected to KI, a mild reduc- tant, which converts As(V) to As(III). The difference be­ tween the first and second determination yields the amount of As(V) present. A third aliquot is digested by using H NO3 and H2SO4 , converting all forms of arsenic present to the As(V) state, then treated with KI. the total arsenic. This analysis represents Subtracting the sum of the As(III) and As(V) values from this final analysis yields the amount of organic arsenic present. RESULTS The arsenic concentrations for the deep sediment sam­ ples are presented in Figures 8 and 9. In the north basin, the maximum value occurred at a depth of 0.25-0.30 m and was 337 ug As g "*1 dry weight; in the south basin it occurred at 0.15-0.20 m and was 335 ug As g“ ^ dry weight. From these maximum concentrations there was a rapid decrease with depth, and then a gradual decline with some minor fluctuations to a concentration of 17 to 20 ug As g“ ^ dry weight. This range is taken as the natural background concentration for sedi­ ments in deep portions of this basin. The similarity in the shapes of the curves for the ar­ senic distribution with depth in the north and south basins is striking. The data lead to the conclusion that peak con­ centrations were the result of the 19 57 sodium arsenite treatment. Based on that assumption, the sedimentation rate in the north basin calculates to be 1.1 cm yr- 1 ; in the south basin the estimate is 0.7 cm yr- ^. The difference in sedimen tation rates could be a function of the ratio of catchment area to basin area. That ratio is markedly smaller for the south basin. An annual hydrologic budget was developed for Lake Lansing to serve as the basis for an arsenic mass balance 35 36 Figure 8 The distribution of arsenic in a sediment core from 9 m in the north basin of Lake Lansing, Michigan, July 13, 1979. 37 A rsenic C oncentration 100 (jug g"' dry w eight) 200 300 0.0 0.4 - Depth (m) 0.8 2.0 2.4 NORTH - BASIN 38 Figure 9 The distribution of arsenic in a sediment core from 7 m in the south basin of Lake Lansing, Michigan, July 13, 1979. 39 Arsenic Concentration 100 (jug g"1 dry weight) 200 300 0.0 - 0.8 - Depth (m) 0.4 2.0 - 2.4 - SOUTH BASIN 40 estimate for the lake. The data are summarized in Table 1. It shows that the lake did not discharge through the over­ flow structure during the summer, fall or winter. The only overland flow input during those seasons was from street drains as a result of rain or snow melt. The discharge from street drains was relatively small, constituting 0.5% of the annual overland discharge to the lake. Net losses to the atmosphere and seepage were the major components of the water budget in summer and fall. Evapotranspiration from extensive lake-side marshes undoubtedly contributed to high rates of net seepage loss in the dry summer and fall of 1978. Net losses of water from the lake in summer and fall resulted in a drop of the lake surface below the elevation of the outlet dam. Winter during the year of study was relatively wet. Precipitation on the watershed in this season was held as snow; there was virtually no melt in this particular winter. Seepage loss appears to have continued to recharge the defi­ cit in the groundwater accrued from the previous summer and fall. With the onset of spring, overland flow from marsh drains, and seepage, and to a lesser extent discharges from street drains, gradually brought the surface of the lake to the sill of the outlet dam. By late April, outflow occurred and continued to early June. The seasonal patterns of overland flow and changes in lake level measured in this study were, by our observation, typical for the basin during recent years. Lake level Table 1. Seasonal aspects of the annual hydrologic budget of Lake Lansing, Michigan. Tabled values are in units of m 3. Atmosphere Interval Precipitation minus Evaporation3 Overland Flow Marsh Drain Discharge Street Drain Discharge Seepage Outlet Discharge Net Gain or Loss Summer 6/20/78-8/28/78 -251849 0 + 543 0 -131620 Fall 8/28/78-11/6/78 - 6969 0 + 630 0 - 58018 Winter 11/6/78-2/8/79 +194290 0 + 178 0 - 72631 Spring Runoff 2/8/79-4/23/79 + 5390 +402909 + 639 0 +199440 Runoff-Outflow 4/23/79-6/18/79 -127119 +145429 +1011 -2124912 - -186257 +548338 +3001 -212491 - 66461 TOTALS 1. 3632 Evaporation was calculated using a pan coefficient of 0.7. During certain per­ iods of the year (late fall and early spring) water in the evaporation pan froze while the lake surface remained open. Values for those periods were estimated by extrapolating data from just prior to or immediately following that phenomenon. During the winter, precipitation was taken as the water content of snow or ice. Water losses through sublimation from ice and snow were taken to equal water gained from condensation; net evaporation was taken as zero. Table 1. Continued. 2. Outflow began in the interval 4/25-4/27/79 and stopped in the interval 5/30-6/1/79. 43 records of the U.S.G.S. suggest that the seasonal patterns of 1978-79 have existed for at least three decades. From the water budget, the turnover-time for Lake Lansing (lake volume divided by annual discharge) was calculated to be 19.5 years. The long residence time promotes retention in the basin of arsenic introduced during weed treatment of the lake in 1957. The hydrologic budget was used to generate an annual arsenic budget for Lake Lansing. Whether this contaminated lake was functioning as a sink or a source of arsenic was the question of primary interest. Arsenic in the discharge of marsh drains, street drains, and the outflow from the lake were measured and applied to discharge data such that Appendix Tables 1, 2, and 3 resulted. Measurements for at­ mospheric loading are given in Table 4 of the Appendix. Appendix Table 5 presents data on seepage. A summary of the annual arsenic budget is given in Table 2. Atmospheric loading and the discharge of drains contri­ buted nearly the same amounts of arsenic to the lake annually. Street drains, while discharging 0.5% of the water, carried 2 .5% of the arsenic flowing overland to the basin in a year. The totals of Table 2 show that the annual arsenic loading was less than discharge to the downstream environment in the spring of the year. In addition, there was a loss of arsenic from the lake by seepage. Thus, Lake Lansing serves as a source rather than a sink for arsenic. This can only occur as a result of internal loading of the water column with arsenic from surficial sediments. Table 2. Seasonal aspects of the annual arsenic budget of Lake Lansing, Michigan. Tabled values are in units of g total arsenic. Overland Flow Atmosphere Loading Interval Marsh Drain Discharge Street Drain Discharge Seepage Outlet Discharge Net Gain or Loss Summer 6/20/78-8/28/78 121 -2598 Pall 8/28/78-11/6/78 122 -1474 Winter 11/6/78-2/8/79 <1 - 899 645 6 + 339 199 248 9 2426 - 542 828 893 22 2426 -5174 216 0 Spring Runoff 2/8/79-4/23/79 170 Runoff-Outflow 4/23/79-6/18/79 TOTALS 45 A study was undertaken to establish the amounts of ar­ senic in the surficial sediments of the basin. are given in Figure 10. The results Values for the two deepest portions of the lake are from the 0.00-0.05 m layer of the sediment cores portrayed in Figures 8 and 9. The remaining values represent the mean of five samples taken at each site. The standard deviation of these averaged 15% of the mean. It can be seen from Figure 10 that only the wave washed northeast shore of the lake had concentrations that could be considered in the range of background values. The level of arsenic elsewhere was two to six times higher than the pre­ weed treatment concentrations observed in deep portions of the cores shown in Figures 8 and 9. These data demonstrate that arsenic contamination occurs in 85-90% of the surficial sediments. The highest concentrations were found in the deep portions of the north and south basins. A laboratory experiment was conducted to obtain an ap­ proximation of the concentration of arsenic to be expected in the water column over contaminated sediments of the lake. The sediments used in the experimental units were obtained from the deep portion of the south basin. The mean sedi­ ment value for all units was 288 yg As g - ^ dry weight. The dissolved oxygen in the aerobic treatment averaged 9.0 mg l- ^? in the anoxic treatment it averaged 0.4 mg 1” ^ for the first 58 days and 8.7 m g l” 1 after aeration. The mean temperature in all units for the duration of the experiment was 14.8° C. The data for arsenic in the water columns of the experimental units is given in Table 6 of the Appendix. 46 Figure 10. Concentrations of arsenic (yg g"1 dry weight) in the surficial sediments of Lake Lansing/ Michigan, June 1979. 47 * 58 100 00 99- -87 72 79 72 30 130 96 87 38 250 m 125 103 •69 46, 48 Figure 11 shows the results of the experiment; indivi­ dual points for the righthand series represent the mean of four units. The aerobic treatment (lefthand series) only minor changes for the parameters measured. values 13.9-30.0, 21.7; arsenic (III), 3.4-12.0, 7.4; (V), 2.9-15.0, 8 .6 . Total phosphorus ranged from 15-41, with a mean value of 29. from 7.3 to 8.1. tions. Arsenic (ug As 1“^) with the range followed by the mean were: total arsenic, arsenic shows (yg PO4-P 1“ ^) The pH ranged The anoxic treatment showed marked varia­ Total arsenic began at 25.5 yg As l- ^ and increased slowly for the first seven days. Between day seven and 21 it showed the greatest rate of increase; thereafter the rate declined. on day 35. The maximum value of 117.0 yg As 1“ ^ was reached From that peak concentration, there was a steady decrease for 14 days; this continued at approximately the same rate for an additional 14 days after aeration. The greatest decrease occurred during the last 16 days of the experiment. yg As 1~^. The final total arsenic concentration was 35.0 Arsenic (III) displayed the same general trend as total arsenic during the anoxic portion of the experiment. The initial As(III) value was 2.5 yg As l” 1 and maximized at 74.8 yg As 1“ ^ on day 35. ing aeration, As(III) After an initial increase follow­ showed a fairly constant decrease for the remainder of the experiment, ending at 8.1 yg As 1~1. Arsenic (V) began at 7.1 yg As l- ^ and increased slowly dur­ ing the first seven days. Between days seven and 21 it in­ creased the greatest amount, reached its maximum on day 49 Figure 11. Variation in the chemical composition of water overlying deep water sediments from the south basin of Lake Lansing for 89 days in experimen­ tal tanks. Lefthand eeriest aerated treatment; Righthand eeriest anoxic treatment, aerated on day 58. A, arsenic concentrations for the (III), (V), and total forms in yg l”1 ; B, pH and con­ centrations of total phosphorus in yg l'1 . 120 A Total As 100 80 As(III) 60 40 As(V) 20 0 300 in o 250 Phosphors 200 m a b e £ 150 a m o 100 50 0 i_____ I--------1______i------- 1------ 0 15 30 45 60 1 75 i 90 I______ I______ I______ I______ I------ 1------ 1 0 15 30 45 60 75 90 51 21 (31.1 yg As l- ^), then it remained relatively constant until aeration. As(V). Aeration caused a significant decline in It slowly increased thereafter, reaching a final value of 17.4 yg As 1“ ^. Total phosphorus behaved differ­ ently from arsenic by continually increasing throughout the low oxygen period. Phosphorus began at 44 and reached 271 yg P O 4-P 1"! on day 58. Following aeration, there was ini­ tially a rapid decrease which lessened with time. the total phosphorus was 81 yg PO4-P 1” ^-. On day 89 The pH rose dur­ ing the first week of anoxic conditions; daily N 2 purging during this interval drove off free CO2 formed by respira­ tion. After that, the units were purged only at sampling times. The pH declined to 7 by day 28 and remained near that value until day 58. Aeration caused a significant in­ crease in pH, perhaps by the liberation of free C02 . The experimental sediments contained a somewhat elevated arsenic concentration as compared to the surficial sediments of Figure 10. Nevertheless, the results of the laboratory experiment suggest that concentrations of <30 yg As 1“ ^ would be expected in the water column of the lake, if the water column over contaminated sediments were predominately aerobic. Additionally, concentrations on the order of three to four times higher would be expected if the water column over those same sediments were anoxic for three or more weeks. These expectations for the lake would hold if external load­ ing to the water column (stream discharge, atmospheric load­ ing, and seepage) were small or negligible. That was the 52 case for Lake Lansing. The maximum external loading occurred during the interval 2/8/79 to 4/23/79, delivering 821 g of arsenic. Mixed and held in the volume of the lake, this a- mount of arsenic would result in a concentration increase of 0.2 yg As l” 1 . Loading in other intervals of the year had a substantially smaller impact on concentration in the water column. Data for the total arsenic in the different compartments of Lake Lansing are presented in Table 3. A range of concen­ trations on the order of 5-25 yg As 1” ^ was observed for all water strata excluding 7/31/78. On that date, total arsenic was 87 and 115 yg As l” 1 for the lower pelagial water in the north and south basins, respectively. 53 Table 3. Concentrations of total arsenic (ug As 1"^) in Lake Lansing, Michigan. Littoral & Upper Pelagial Water Date Lower Pelagial Water North Basin South Basin 7/17/78 17.0 24.0 14.0 7/31 26.0 87.0 115.0 8/14 19.0 21.8 14.7 8/28 22.0 24.0 19.0 10/9 19.6 18.7 18.5 10/23 17.5 17.6 15.5 11/6 14.9 13.9 13.1 1/24/79 11.0 11.0 2/8 11.7 11.0 11.3 9.0 8.9 7.2 5/7 10.0 10.5 11.2 5/21 16.3 3.4 14.2 6/18 10.7 12.0 11.2 7/2 23.0 5.6 11.1 7/16 18.4 19.2 22.2 7/30 23.0 23.9 22.5 8/13 11.2 16.2 19.4 8/27 5.4 17.0 21.0 9/17 10.0 8.0 8.7 4/23 DISCUSSION The amount of data on the concentrations of arsenic in sediments of lakes is rather extensive. Seydel (1972) re­ ports Lake Superior surface sediments as ranging from 2.85.4 ppm As while Lake Michigan sediments ranged from 7.228.8 ppm As. She used Lake Superior as a comparison to Lake Michigan assuming the former to be relatively unpolluted. Walters, et al. (1974) found a core from Lake Erie to be greatly enriched with mercury and copper but not with arsenic they reported a maximum of 2.5 ppm. Galloway and Likens (1979) reported on atmospheric enhancement of metal deposi­ tion to sediments of Woodhull Lake, New York. They do not include arsenic as one of those being enriched; the core had values of 3.12 to 10.4 mg As kg- 1 . Ruppert, et al. (1974) took 98 sediment grab samples from Chautauqua Lake, New York and compared those to two bedrock and soil samples from that area. They concluded that the arsenic values of the lake sediments were not from natural sources but were the result of sodium arsenite treatment during the period 1955-1963. Arsenic concentrations for 96 of the samples ranged from <0.5-58.75 ppm, while two other samples had values of 140 and 306 ppm As. The mean for all samples was 22.1 ppm. Wagemann, et al. (1978) report sediment arsenic concentra­ tions for five lakes in the vicinity of Yellowknife, 54 55 Northwest Territories, Canada. Two of those lakes, Kam and Keg, were known to be contaminated with arsenic as a conse­ quence of gold mining activities. Ten sediment grab samples (top 0-20 cm) ranged from 6 to 3500 ppm As by dry weight. Three other lakes in the same area were chosen as reference lakes for the study; they had sediment concentrations which ranged from 19-105 ppm As. Kobayashi and Lee (1978) studied 15 sediment cores from five Wisconsin lakes that had been ex­ tensively treated with sodium arsenite for weed control. Surface sediment concentrations (to 5 cm depth) ranged from 10 yg As g - ^- in Lake Mendota to 659 yg As g “ ^ in the south bay of Big Cedar Lake. than 300 ppm As. Typically concentrations were less In all of the cores, arsenic concentration fell to 10 ppm or less at depths below anthropogenic influ­ ence. Arsenic concentrations for 14 surface sediments (0- 1 cm) from Lake Washington ranged from 15-210 ppm As; these averaged 81 ppm (Crecelius, 1975). Five sediment cores showed generally higher concentrations of arsenic near the surface. It decreased with depth to the background concen­ tration of about 10 ppm As. From the above, it is apparent that the concentration that constitutes anthropogenic arsenic contamination in lake sediments is site specific. The data for the cores in Lake Lansing show background to be in the range of 17-20 yg As g“ l dry weight. Contamination of the lake in 1957 was re­ flected in the elevated arsenic concentrations in the upper 0.5 m of c o r e s ; the maximum concentrations which occur there (over 330 ppm As) are 18 times greater than background. In 56 the depth interval between background and peak concentrations associated with weed treatment, there was a slight but dis­ tinct elevation of arsenic content in both cores that were studied. While this increase {cf. Figures 8 and 9) may be due to the downward migration of arsenic that was applied to the lake, it should be noted that the initial rise from back­ ground can be correlated in time with early immigration and commercial development of the drainage basin. In particular, coal-burning locomotives began to move along the rail route less than 100 m from the south shore of the lake in 1878 (Raphael, 1958). regular stop. The watering-station located there was a Implicating arsenic from the coal of these trains as the primary cause for the rise from background would require a mean sedimentation rate of approximately 1 cm per year; based on the data, this is not an unrealistic esti­ mate. Further, sediments in the south basin of the lake rose from background to higher concentrations at that time than did sediments in the more remote north basin. Superimposed on that arsenic that may have fallen out from early coalfueled locomotives, was whatever contribution that may have been made from the volcanic ash of Krakatoa that erupted in 1883. The data from the deep sediment cores can be used to predict how much longer it will take for the surficial sedi­ ments of those areas to once again reach background levels (17-20 ppm A s ) . Using the calculated sedimentation rates for the two basins, Table 4 was constructed. From these data linear regression equations were derived for each of the 57 Table 4. Sediment arsenic concentrations with depth and the number of years since the sodium arsenite treatment in 1957. Years since treatment based on sedimenta­ tion rates of 1.1 cm yr"l for the north basin and 0.7 cm yr-1 for the south basin. North Basin Depth (cm) South Basin Total As (lig g"1) Years since treatment Total As (ug g“ ) Years since treatment 0-5 111 22 125 22 5-10 160 18 175 15 10-15 181 14 251 7 15-20 226 10 335 0 20-25 284 6 25-30 337 0 58 basins. North basin: p = 337 - 10.3 (x); r 2 = 0.99 South basin: p = 326 - 9.5(x); r2 = 0.99 where p is the arsenic concentration in ug As g- -*- and x is the number of years since treatment. The predicted time to reach background levels would be 31 years in the north basin and 32 years in the south basin tion rates). (assuming constant sedimenta­ This means that by the late 1980's the contami­ nation from treatment would be ameliorated in those deep water areas. However, this does not imply that the water column will no longer be affected by the herbicide contami­ nation after this date. This prediction can not be applied to the surficial shallow water sediments since the sedimenta­ tion rate in these areas is unknown. These contaminated sed­ iments could continue to serve as a source of arsenic to the water column long after its amelioration in the deep basins. Surficial sediments of the lake, excluding the near northeast shore, showed arsenic burdens of two to six times the background level. They also showed, in nearly all cases, an increase in concentration with increasing depth of the lake water. These samples were taken from a variety of sediment types. The lowest concentrations were found in the shallow region of the lake's northeast shore. Prevailing winds dur­ ing the open-water season arise from the southwest, blowing across the long axis of the lake. Fine-grained organic par­ ticles are swept by wave action away from the northeast shoreline leaving an area of sand. The greatest flux of seepage is thought to occur in this highly permeable area 59 as well. This in conjunction with wave-generated currents could be important in moving arsenic from the shore. Crecelius, et al. (1975) and Ruppert, et al. (1974) found the highest concentrations of arsenic in sediments that were predominated by sand. However, they both conclude that arsenic is more highly correlated with smaller particle sizes than sand. Those disparities as well as the findings here suggest that arsenic concentrations are not necessarily a function of sediment type. In Lake Lansing, the distribu­ tion of arsenic in surficial sediments appears to be a func­ tion of where it was placed during weed treatment and its transport by currents to deeper portions of the lake. Annual mass balance budgets for arsenic in lakes are scarce in the literature. Using procedures for input-out­ put estimates similar to the procedures used in this study, Crecelius (1975) concluded that the sediments of Lake Wash­ ington trapped 55% of the annual arsenic input. served as a sink for the element. The lake Conversely, the annual budget for Lake Lansing showed that internal loading of the water column from the sediments resulted in an annual loss from the lake that was greater than the sum of the inputs. Concentrations of arsenic in water exiting Lake Lansing, controlled by rates of sediment-water column exchange, were an order of magnitude higher than concentrations in input water thus accounting for the imbalance. A proposed cycle for arsenic in a stratified lake is presented in Figure 12 (Ferguson and Gavis, 1972; Wood, 1974). Transformations include oxidation-reduction and ligand 60 Figure 12. A proposed cycle for arsenic in a stratified lake (modified from Ferguson and Gavis, 1972; Wood, 1974) . CM, HO- Af OH CM, CM, Al HO ■A*' CH, CH. OlHCrHttARSINC CH, ’----- H OH HO OH HYAOMMNKM TftM CTHYLA ASlNE mccihtathm m AK O nU TK IM ON COMC1A1TATION A W N H T l M WITH I TATKN AMO < M M mm on f t a s u i N N M M M tU C M T M T IO H AMD b CH| HO- P*i 0*3 O M C T H V L M S IM C ACIO OH I H T H » lA tlO ll HO OH OH m c t h v l a h u m ic ACID HtTW fLATIQH OH NOCWTE MPMCTION H0< 62 exchange. Transfers from solution to solid phases and vice- versa are shown. The lefthand side of the figure depicts organoarsenical pathways; the righthand side shows inorganic interconversions. Methylation of arsenic compounds is ther­ modynamically unfavorable in water and can only occur by bio­ logical mediation (Ferguson and Gavis, 1972). However, the presence of these forms is not precluded by biological ac­ tivity since both methylarsenic acid (MAA) and dimethylarsenic acid (DMA) are synthesized pesticides and may be in natural waters as a result of agricultural and home use. Measurements of ambient concentrations of organoarsenicals have been limited by the lack of appropriate speciation meth­ odologies that are economically and routinely feasible (Holm, et al., 1979). Detection of gaseous arsines is difficult since they can be rapidly oxidized; dimethylarsine burns spontaneously in the air (Peoples, 1975). Braman and Foreback (1973) were the first to report methylated arsenicals in natural waters. They reported values from four small lakes in and around Tampa, Florida that ranged from 0.5 to 0.22 ppb for MAA and 0.15 to 0.62 for DMA (cacodylic acid). They concluded that the MAA was generally present in smaller concentrations than DMA since it was only an intermediary in the methylation sequence. DMA, unless subjected to bacter­ ial oxidation (since it is extremely resistant to chemical oxidation), was thought to be very persistent in natural waters. Braman (1975) has reported a lake value of 0.14 ppb for trimethylarsine. The reactions and cycling of these or­ ganoarsenicals in freshwater systems has yet to be thoroughly 63 investigated. The inorganic interconversions shown in Figure 12 (Ferguson and Gavis, 1972) suggest that arsenite, As(III), tends to be oxidized to arsenate, As(V), in aerobic epilimnetic water. In this portion of a lake, arsenate is most likely to exist as the anion, H A s O ^ " 1972). (Ferguson and Gavis, Chemically similar to phosphate, it can be adsorbed, occluded or precipitated with hydrous ferric oxides. Thus ferric iron may control arsenate solubility in oxic portions of a lake basin. Wagemann (1978) has predicted that barium, chromium, and copper could also form insoluble metal com­ plexes with arsenic. He feels that barium is the most con­ vincing candidate capable of holding arsenic to rather low concentrations under pH and Eh conditions of aerated freshwaters. Presently, studies of arsenic and barium in natural waters are too few to provide an assessment of his theoreti­ cal predictions. Figure 12 suggests that turbulent dispersion and convec­ tion can transport some of the arsenate and metal complexes into an oxygen-depleted hypolimnion. Once there, reduction of these compounds is likely to take place in the water or on the surface of anaerobic sediments. Depending on the pH, Eh, iron and sulfur concentrations within these lower strata, arsenite, insoluble arsenic sulfides or ferrous arsenic sul­ fides could result. Ferguson and Anderson (1974) report that at low Eh in the presence of sulfide (S 2— ), As(III) should be effectively removed from the water column as insoluble 64 sulfides. Arsenic (V) is not similarly affected. Mortimer (1941,1942) has demonstrated the importance of the aerated microzone at the sediraent-water interface in con­ trolling the release of phosphorus, iron, and sulfur to the water. In his experiment, the release of iron and phosphorus increased markedly as oxygen was depleted at the sediment surface and the redox potential decreased. This was explain­ ed by the reduction of ferric iron in complexes holding phos­ phate. Ferrous iron and phosphate appeared in the water col­ umn simultaneously. As the Eh continued to fall, sulfate concentrations decreased. Sulfate is reduced to sulfide at a substantially lower Eh than the reduction of ferric to fer­ rous iron (Hutchinson, 1957). Once the Eh has fallen low enough for sulfide production, ferrous sulfide can be formed which is exceedingly insoluble at neutral or alkaline pH (Wetzel, 1975). It appears that considerable quantities of iron and sulfur were lost to the sediments as FeS in his sys­ tem. As these events proceeded, the concentration of phos­ phate continued to rise. A companion sediment-water system was aerated to provide a comparison in his experiment. His results are relevant to this study, since he demonstrated the control exerted by ferric iron on the migration of phosphate across an aerated sediment surface. The control of sulfides on ferrous iron concentration in anaerobic water overlying sediments was supported by his data. The model of Figure 12 and the work of Mortimer and others regarding the cycles of iron, sulfur, and phosphorus 65 can be used to develop an hypothesis regarding the results of the laboratory experiment obtained in this study. Arsenic here was dominated by inorganic species; an organic fraction was detected by the analytical procedures used. Arsenic and phosphorus did not increase over time in water above an aer­ ated sediment surface. creased. In anoxic treatment chambers both in­ The dramatic increase in total arsenic was due principally to an increase in As(III), derived either direct­ ly from the sediments or from the reduction of A s ( V ) . It is proposed that ferric iron controlled solubility of arsenic and phosphorus in the aerated water-sediment system; reduc­ tion to ferrous iron allowed release to the water column in the anoxic treatment. Sulfide accumulated in low oxygen chambers as anoxia persisted. noted by odor on day 35. Hydrogen sulfide was first The increase in sulfide concentra­ tion was expected to lag in time behind ferric iron reduction because of lower Eh optima for its accumulation. The arsenic maximum of day 35 was depressed thereafter by the formation of insoluble compounds of As(III) and sulfide. Predicted by the hypothesis, phosphorus continued to increase in water of the anoxic treatment units until they were aerated. Aeration after 58 days caused dramatic changes in the concentrations of inorganic arsenic species and phosphorus in anoxic chambers. The initial rise in As(III) is ascribed primarily to the oxidation of sulfides to sulfates and con­ comitant release of As(III). As(V) The initial rapid decline in is assigned to its complexing and precipitation with 66 newly formed hydrated ferric oxides. The oxidation of fer­ rous to ferric iron is extremely rapid in the near neutral pH range of these waters (Stumm and Morgan, 1970). Phospho­ rus was likely removed from the system by the same mechanism in the manner of the work of Mortimer (1941,1942). Wauchope (1975) studied the affinities of ferric hydroxides for arse­ nate and phosphate, and showed a greater affinity for the former. The total arsenic declined gradually as aeration continued; As(III) was slowly converted to As(V) and removed in combination with ferric iron. Phosphorus slowly declined by removal with this compound as well. Data from the laboratory experiment led to expectations regarding levels of arsenic to be observed in water of the lake. With periods of three to five weeks of anaerobiosis in the hypolimnion of the north and south basins, arsenic concentrations on the order of 100 yg 1“^ were predicted, depending on the availability of the sulfide ion. An abun­ dance of sulfide or periodic aeration of the hypolimnion during high winds in spring or summer were expected to de­ press the concentration. Ferric iron control on solubility was predicted for the latter case. Thermal and oxygen strat­ ification of the lake were periodically disrupted during the summers of this study. Data from the field are not suffi­ cient to sort out the proposed control on arsenic concentra­ tions in the hypolimnions due to sulfide and ferric iron. That conditions of stable anaerobiosis similar to those ob­ served in the laboratory did in fact exist in the 67 hypolimnions in July of 1978 was suggested by observed con­ centrations near 100 pg As l" 1 at the end of that month. The epilimnion-littoral region of Lake Lansing and the water column at overturn had arsenic concentrations on the order of 5-25 pg 1"^. Anaerobic conditions did not develop in the lake during the winter of this study; the arsenic con­ centration was in the mid-portion of the range at that time. These results were expected from the laboratory work with sediments held under aeration. However, the substantial por­ tion of As(III) in the aerobic treatment would not be pre­ dicted from the model in Figure 12. This result is similar to the findings in seawater, in which the thermodynamically unfavorable arsenite occurs at concentrations above those that would be expected (Johnson, 1972; Johnson and PiIson, 1975; Andreae, These findings have been explained as 1979). the consequence of biologically mediated reactions in which arsenate (V) is reduced to arsenite (III). Brunskill, et al. (in press) working in freshwater found that little or no reduction of arsenate occurred under aerobic conditions and rapid algal growth. This is contrary to the results of Johnson and Burke (1978), who found a reduction of arsenate during marine phytoplankton blooms. It is clear that this inconsistency and certain aspects of the proposed arsenic cycle in lakes awaits experimental results that will provide clarification. LITERATURE CITED Aggett, J. and A.C. Aspell. 1976. The determination of arsen ic (III) and total arsenic by atomic-absorption spec­ troscopy. Analyst. 101:341-347. American Public Health Association. 1976. Standard Methods for the Examination of Water and Wastewater, 14th ed. American Public Health Association, Washington, D.C. 1193 pp. Andreae, Meinrat O. 1979. Arsenic speciation in seawater and interstitial waters: The influence of biological-chemi­ cal interactions on the chemistry of a trace element. Limnol. Oceanogr. 24:440-452. Aston, S.R., I. Thornton, and S.J. Webb. 1975. Arsenic in stream sediments and waters of south west England. Sci. Total Environ. 4:347-358. Bowen, H.J.M. 1966. Trace Elements in Biochemistry. Academic Press, New York. 250 pp. Braman, Robert S. 1975. Arsenic in the environment. In E.A. Woolson, ed. Arsenical Pesticides (ACS Symposium Series No. 7). American Chemical Society, Washington, D.C., pp. 108-123. Braman, Robert S. and Craig C. Foreback. 19 73. Methylated forms of arsenic in the environment. Science. 182:12471249. Brunskill, G.J., B.W. Graham, and J.W.M. Rudd. In press. Experimental studies on the effect of arsenic on micro­ bial degradation of organic matter and algal growth. Crecelius, Eric A. 1975. The geochemical cycle of arsenic in Lake Washington and its relation to other elements. Limnol. Oceanogr. 20:441-451. Crecelius, Eric A., Michael H. Bothner, and Roy Carpenter. 1975. Geochemistries of arsenic, antimony, mercury, and related elements in sediments of Puget Sound. Environ. Sci. Technol. 9:325-333. 68 69 Dunst, R. and R. Beauheim. 1979. Effect of dredging and nu­ trient inactivation at Lilly Lake, Wisconsin. In S.A. Peterson, ed. Limnological and Socioeconomic Evaluation of Lake Restoration Projects. National Technical Infor­ mation Service, Springfield, Virginia, pp. 138-164. Durum, W.H., J.D. Hem, and S.G. Heidel. 1971. Reconnaissance of Selected Minor Elements in Surface Waters of the United States. Geological Survey Circular 643. USDI, Washington, D .C . Ferguson, John F. and Jerome Gavis. 1972. A review of the arsenic cycle in natural waters. Water Res. 6:1259-1274. Ferguson, John F. and Marc A. Anderson. 1974. Chemical forms of arsenic in water supplies and their removal. In Alan J. Rubin, ed. Chemistry of Water Supply, Treatment, and Distribution. Ann Arbor Science Publishers Inc., Ann Arbor, Michigan, pp. 137-158. Galloway, James N. and Gene E. Likens. 1979. Atmospheric en­ hancement of metal depositions in Adirondack lake sedi­ ments. Limnol. Oceanogr. 24:427-433. Glandon, R.P., F.C. Payne, C.D. McNabb, and T.R. Batterson. In press. A comparison of rain-related phosphorus and nitrogen loading from urban, wetland, and agricultural sources. Holm, Thomas R . , Marc A. Anderson, Dennis G. Iverson, and Robert S. Stanforth. 1979. Heterogeneous interactions of arsenic in aquatic systems. In Everett A. Jenne, ed. Chemical Modeling in Aqueous Systems (ACS Symposium Series, No. 93). American Chemical Society, Washington, D.C., pp. 711-736. Hutchinson, G. Evelyn. 1957. A Treatise on Limnology. I. Geo­ graphy, Physics, and Chemistry. John Wiley and Sons, Inc., New York. 1015 pp. Johnson, David L. 1971. Simultaneous determination of arsen­ ate and phosphate in natural waters. Environ. Sci. Technol. 5:411-414. Johnson, David L. 1972. Bacterial reduction of arsenate in sea water. Nature. 240:44-45. Johnson, David L. and Richard M. Burke. 1978. Biological me­ diation of chemical speciation. II. Arsenate reduction during marine phytoplankton blooms. Chemosphere. 8:645648. 70 Johnson, D.L. and M.E.Q. Pilson. 1975. The oxidation of ar­ senite in seawater. Environ. Letters. 8:157-171. Kipling, M.D. 1977. Arsenic. In John Lenihan and William W. Fletcher, eds. The Chemical Environment. Environment and Man, Volume Six. Academic Press, New York, pp. 93120 . Kobayashi, Shigeru and G. Fred Lee. 1978. Accumulation of arsenic in sediments of lakes treated with sodium ar­ senite. Environ. Sci. Technol. 12:1195-1200. Lis, Steven A. and Philip K. Hopke. 1973. Anomalous arsenic concentrations in Chautauqua Lake. Environ. Letters. 5:45-51. Lueschow, L.A. 1972. Biology and control of selected aquatic nuisances in recreational waters. Tech. Bull. No. 57. Department of Natural Resources, Madison, Wisconsin. Mackenthun, Kenneth M. 1950. Aquatic weed control with sodi­ um arsenite. Sewage Ind. Wastes. 22:1062-1067. Martin, H.M. 1955. Map of the Surface Formations of the Southern Peninsula of Michigan. Publication No. 49. Michigan Department of Conservation, Lansing, Michigan. McBride, M.S. and H.O. Pfannkuch. 1975. The distribution of seepage within lake beds. J. Res. USGS. 3:505-512. Mortimer, Clifford H. 1941. The exchange of dissolved sub­ stances between mud and water in lakes (Parts I and II). J. Ecol. 29:280-329. Mortimer, Clifford H. 1942. The exchange of dissolved sub­ stances between mud and water in lakes (Parts III, IV, summary and references). J. Ecol. 30:147-201. Onishi, H. 1969. Arsenic. In K.H. Wedepohl, ed. Handbook of Geochemistry, Volume II-l. Springer-Verlag Berlin, Heidelberg, Germany, pp. 33,B-33,0. Onishi, H. and E.B. Sandell. 1955. Geochemistry of arsenic. Geochim. Cosmochim. Acta. 7:1-33. Peoples, S.A. 1975. Review of arsenical pesticides. In E.A. Woolson, ed. Arsenical Pesticides (ACS Symposium Series, No. 7). American Chemical Society, Washington, D.C., pp. 1-1 2 . Raphael, Evelyn Huber. 1958. A History of the Haslett-Lake Lansing Area, Meridian Township, Ingham County, Michi­ gan. Edward Brothers, Ann Arbor, Michigan. 94 pp. 71 Ritchie, J.A. 1961. Arsenic and antimony in some New Zealand thermal waters. N.Z.J. Sci. 4:218-229. Roelofs, Eugene W. 1958. The effect of weed removal on fish and fishing in Lake Lansing. Department of Fisheries and Wildlife, Michigan State University, East Lansing, Michigan. 10 pp. Ruppert, D.F., P.K. Hopke, P. Clute, W. Metzger, and D. Crowley. 1974. Arsenic concentrations and distribu­ tion in Chautauqua Lake sediments. J. Radioanal. Chem. 23:159-169. Schroeder, Henry A. and Joseph J. Balassa. 1966. Abnormal trace metals in man: Arsenic. J. Chronic Dis. 19:85106. Seydel, U s e S. 1972. Distribution and circulation of arsenic through water, organisms and sediments of Lake Michigan. Arch. Hydrobiol. 71:17-30. Shapiro, Joseph. 1971. Arsenic and phosphate: Measured by various techniques. Science. 171:234. Siami, Mehdi. 1979. Distribution and abundance of benthic macro-invertebrates in Lake Lansing. M.S. thesis. Michigan State University, East Lansing, Michigan. 109 pp. Stumm, Werner and James J. Morgan. 1970. Aquatic Chemistry: An Introduction Emphasizing Chemical Equilibria in Natural Waters. Wiley-Interscience, New York. 583 pp. U.S. Army Corps of Engineers. 1970. Reconnaissance Report Eutrophication Problem Lake Lansing, Michigan. U.S. Army Corps of Engineers, Washington, D.C. 25 pp. U.S. Department of the Interior. 1967. Water Measurement Manual, 2nd ed. USDI, Washington, D.C. 329 pp. Wagemann, R. 1978. Some theoretical aspects of stability of inorganic arsenic in the freshwater environment. Water Res. 12:139-145. Wagemann, R . , N.B. Snow, D.M. Rosenberg, and A. Lutz. 1978. Arsenic in sediments, water and aquatic biota from lakes in the vicinity of Yellowknife, Northwest Territories, Canada. Arch. Environ. Contam. Toxicol. 7:169-191. Walsh, L.M. and D.R. Keeney. 1975. Behavior and phytotoxicity of inorganic arsenicals in soils. In E.A. Woolson, ed. Arsenical Pesticides (ACS Symposium Series, No. 7). American Chemical Society, Washington, D.C., pp. 35-52. 72 Walters, L.J., Jr., T.J. Wolery, and R.D. Myser. currence of As, Cd, Co, Cr, Cu, Fe, Hg, Ni, Zn in Lake Erie sediments. Proc. 17th Conf. Res. Internat. Assoc. Great Lakes Res., pp. 1974. Oc­ Sb, and Great Lakes 219-234. Wauchope, R.D. 1975. Fixation of arsenical herbicides, phos­ phate, and arsenate in alluvial soils. J. Environ. Qual. 4:355-358. Wetzel, Robert G. 1975. Limnology. W.B. Saunders Co., Phila­ delphia. 743 pp. Winter, Thomas C. 1978. Ground-water component of lake water and nutrient budgets. Verh. Int. Ver. Limnol. 20:4 38-444. Wood, J.M. 1974. Biological cycles for toxic elements in the environment. Science. 183:1049-1052. Woolson, E.A. 1975. Bioaccumulation of arsenicals. In E.A. Woolson, ed. Arsenical Pesticides (ACS Symposium Series, No. 7). American Chemical Society, Washington, D.C., pp. 97-107. Woolson, E.A. 1977. Fate of arsenicals in different environ­ mental substrates. Environ. Health Persp. 19:73-81. APPENDIX 73 Table A-l. Arsenic inputs from marsh drains during 6/20/78 to 6/18/79. Interval Discharge (m3) Total As con­ centration (g As m“ 3) Total As input for interval (g As) 6/20/78 - 2/8/79 0 2/8/79 - 4/23/79 402909 0.0016 644.7 17470 79886 0.0016 0.0020 28.0 159.8 17004 23734 0.0020 0.00086 34.0 20.4 7266 0 0.00086 6.2 0.0 - 0.0 4/23/79 - 5/7/79 4/23-4/27 4/27-5/7 5/7/79 - 5/21/79 5/7-5/11 5/11-5/21 5/21/79 - 6/18/79 5/21-6/8 6/8-6/18 75 Table A-2. Arsenic inputs from street drains during 6/20/78 to 6/18/79. Interval Discharge (m3) Total As con­ centration (g As m ”3) 6/20/78 - 7/5/78 176 0.00921 1.6 7/5/78 - 7/17/78 0 - 0.0 7/17/78 - 7/31/78 104 0.0092 1.0 7/31/78 - 8/14/78 118 0.0092 1.1 8/14/78 - 8/28/78 145 0.0092 1.3 8/28/78 - 10/9/78 479 0.00232 1.1 Total As input for interval (g As) 10/9/78 - 10/23/78 81 0.0023 0.2 10/23/78 - 11/6/78 70 0.0023 0.2 178 0.0023 0.4 11/6/78 - 1/24 79 - 0.0 1/24/79 - 2/8/79 0 2/8/79 - 4/23/79 639 0.0092 5.9 4/23/79 - 5/7/79 477 0.0092 4.4 5/7/79 - 5/21/79 292 0.0092 2.7 5/21/79 - 6/18/79 242 0.0092 2.2 1. Average of six storms that occurred during the spring and summer. 2. Average of six storms that occurred during the fall. 76 Table A-3. Arsenic losses from the outflow during 6/20/78 to 6/18/79. Interval 6/20/78 - 4/23/79 4/23/79 - 5/7/79 Discharge (m3) 0 Total As Con­ centration (g As m ” 3) - Total As out­ put for interval (g As) 0.0 106574 0.0111 1183.0 39104 61448 0.0111 0.0121 434.1 743.5 5365 0 0.0121 64.9 0.0 5/7/79 - 5/21/79 5/7-5/11 5/11-5/21 5/21/79 - 6/18/79 5/21-6/8 6/8-6/18 77 Table A-4. Arsenic inputs to the lake surface from atmo­ spheric bulk fallout. Interval Number of days in interval Rate of As Total As input input for interval (g As day"l) (g As) 6/20/78 - 7/5/78 15 1.7481 26.2 7/5/78 - 7/17/78 12 1.748 21.0 7/17/78 - 7/31/78 14 1.748 24.5 7/31/78 - 8/14/78 14 1.748 24.5 8/14/78 - 8/28/78 14 1.748 24.5 8/28/78 - 10/9/78 42 1.748 73.4 10/9/78 - 10/23/78 14 1.748 24.5 10/23/78 - 11/6/78 14 1.748 24.5 11/6/78 - 1/24/79 79 2 . 3022 1/24/79 - 2/8/79 15 2. 302 34. 5 2/8/79 - 4/23/79 74 2.302 170. 3 4/23/79 - 5/7/79 14 9.565 133.9 11 3 1.524 1.544 16.8 4.6 2 26 1.544 1.558 3.1 40.5 181.9 5/7/79 - 5/21/79 5/7-5/18 5/18-5/21 5/21/79 - 6/18/79 5/21-5/23 5/2 3-6/18 1. Average value for nine intervals from 6/18/79 to 11/14/79. 2. Average value for 13 intervals from 4/25/79 to 11/14/79. 78 Table A-5. Net gain or loss of arsenic due to seepage during 6/20/78 to 6/18/79. Interval Seepage (m3) Total As con­ centration! (g As m ” 3) Total As gain or loss for interval (g As) -41768 0.0195 -814.5 7/5/78 - 7/17/78 -21759 0.0195 -424.3 7/17/78 - 7/31/78 -39589 0.0195 -772.0 7/31/78 - 8/14/78 -30272 0.0195 -590.3 8/14/78 - 8/28/78 +1768 0.0017 +3.0 8/28/78 - 10/9/78 -72011 0.0208 -1497.8 10/9/78 - 10/23/78 +13933 0.0017 +23.7 10/23/78 - 11/6/78 0 0.0017 o • o 6/20/78 - 7/5/78 11/6/78 - 1/24/79 -44147 0.0130 -573.9 1/24/79 - 2/8/79 -28484 0.0114 -324.7 2/8/79 - 4/23/79 +199440 0.0017 +339.0 4/23/79 - 5/7/79 +14528 0.0017 +24.7 5/7/79 - 5/21/79 +27240 0.0017 +46.3 5/21/79 - 6/18/79 -45400 0.0135 -612.9 1. Since lake sampling did not begin until 7/17/78, an aver age value of six 1978 littoral and upper pelagial water column concentrations was applied to intervals from 6/20/78 to 8/14/78. When the lake was gaining water from seepage (a positive value), an average value from the total arsenic concentrations of the marsh drains was used. All other values represent the mean littoral and upper pelagial water column concentrations of the inter­ val in question. Table A-6. The various forms of arsenic in the water of the experimental units (pg As 1“^). Aevobia Treatment 1. Anoxia Treatment Day As(III) As (V) Organic Total As (III) As (V) Organic Total 0 4.5 2.9 6.5 13.9 2.5 7.1 16.0 25.5 7 3.4 5.3 10.3 19.0 6.7 9.6 13.5 29.8 21 7.6 6.1 4.3 18.0 53.5 31.3 17.3 102.0 35 9.0 11.0 0.0 20.0 74.8 29.8 12.5 117.0 58 6.0 12.0 4.0 22.0 51.0 27.3 8.3 86.5 591 8.0 10.0 7.0 25.0 62.5 14.5 5.5 82.5 61 12.0 6.0 7.0 25.0 53.3 15.3 12.0 80.5 65 9.4 15.0 5.6 30.0 48.0 17.8 12.7 78.5 73 8.6 12.0 6.4 27.0 31.3 19.3 27.3 77.8 89 4.6 12.0 1.4 18.0 8.1 17.4 9.6 35.0 After the sampling on day 58 the anoxic units were aerated for the remainder of the experiment.