Proceedings of The National Conference On Undergraduate Research (NCUR) 2012 Weber State University, Ogden Utah March 29 31, 2012 Fluvial Arsenic in Utah Valley, Salt Lake Valley and the Wasatch Range: Analogy with the Himalayan Range and the Ganges River Floodplain Gabriela Ferreira Department of Earth Sciences Utah Valley University Orem, UT 84058 Faculty Adviser: Steven H. Emerman Abstract Elevated arsenic in groundwater in the floodplain of the Ganges River has been well-documented over the last fifteen years. Measurements of arsenic in the Himalayan Range and Ganges floodplain found that dissolved arsenic was elevated in the Himalayan Range, but fell to undetectable in the Ganges floodplain. The sudden change in dissolved arsenic across the Himalayan-Ganges boundary was accounted for by the residence time in the vicinity of a sediment particle necessary for the large multivalent arsenate ion to adsorb onto sediment, so that arsenate can adsorb onto sediment only when the stream velocity drops. The result that dissolved arsenic falls to undetectable as a river passes from a steep mountain range onto a flat valley floor is so startling and has such major implications for understanding the arsenic cycle and its implications for global public health that the result must be tested in analogous geological environments, such as the Wasatch Range and the corresponding flat floors of Utah and Salt Lake Valleys. Twenty samples from Provo River were collected and analyzed for As and associated transition elements. Upstream of the mouth of Provo canyon, As is low (0.011-0.095 mg/l), and shows low correlation with most transition elements, except Cu (R 2 = 0.38). At the mouth of the canyon, As values increase significantly (0.367 0.436 mg/l) and are also moderately correlated with Cu (R 2 = 0.48); As continues to be elevated until nearing Utah Lake, where As values drop again (0.048-0.052 mg/l). The sharp increase in As may be due to historic mine tailings piling up at the mouth of the canyon and the formation of chalcopyrite co-precipitating Cu and As. The subsequent drop near Utah Lake may be where stream velocity actually decreases. Keywords: Arsenic, Utah Valley, Floodplain 1. Background 1.1 Arsenic Mobility Arsenic (As) is a widespread toxin with very harmful consequences for human populations exposed to Ascontaminated waters. Worldwide, as many as 60-100 million people may be at risk of exposure to excessive levels of As 1. Sources of As can be naturally occurring or anthropogenic, such as geothermal springs or mining drainage 2. Arsenic contamination in groundwater can influence surficial waters, so that the processes releasing As into groundwater can result in elevated fluvial As 3. The reverse can also be true in that losing As-contaminated rivers can increase the As concentration of groundwater 4. The methods of As mobilization in groundwater are still largely misunderstood, but several models have been proposed: ph desorption, organic complexation, anionic displacement or competition, sulfide oxidation, and reductive dissolution. Alkaline conditions can result in release of arsenic through the desorption of As from oxides. Metal oxides (especially Fe and Mn oxides) are thought to release As through desorption under alkaline conditions 2. As release
may be a result of desorption by hydroxide ions, as these ions can take up available positively-charged sites 1. As ph increases along with hydroxide ion concentration, As will desorb from oxide surfaces 2. Rubinos et al. 1 found that arsenic may bind to organic matter (OM) in soils by forming inner and outer-sphere complexes, but the adsorption maximum for As occurs under acidic conditions. In alkaline environments, organic matter can result in As mobilization 1. Also, dissolved organic matter can compete with As for sorption sites 1. Anion competition for sorption sites can also impact As loading and release As to the aqueous phase 2,3. Rubinos et al. 1 also found that anion competition, particularly by phosphate anions, resulted in increased As mobility. Phosphorous is more likely to compete with As, because both elements form (+5) oxidation states 1. Carbonate, natural organic matter, and other anions (Cl -, SO 2-4, and NO - 3 ) may also compete with As for sorption sites 3. A major competing model is the sulfide-oxidation model. According to this model, As co-precipitates with sulfides of the transition elements, and oxidation of these sulfides can result in the release of As. Mining activities can exacerbate mineralization of sulfides, particularly arsenopyrite 2,3. The dominant accepted model for As contamination of groundwater in South Asia is the reductive-dissolution model, which claims that As is mobilized in highly reducing conditions as a result of rapidly accumulating and buried sediments 1-3. Under this model, a major mechanism is the reductive desorption and dissolution of As from oxides and clays, particularly Fe oxides 2. Most important is the reduction of SO 4 and Fe(III), so that the reduced sulfur can then react with Fe to produce pyrite, FeS 2. This reduction process is mediated by microorganisms 2,5. Release of As can also occur by the desorption of As from Fe oxides 2. In this model, As is strongly correlated with Fe and Mn 2. Perhaps this model is widely accepted because the areas of the world most affected by As contamination (Bangladesh and India) coincide with a reducing environment 3. A recent study conducted by Emerman et al. 5 found that stream velocity may have a strong influence on the ability of fluvial sediments to demobilize dissolved arsenic. According to Emerman et al. 5, measurements of fluvial As in the Himalayan Range and Ganges River floodplain found that dissolved As was elevated in the Himalayan Range, but fell to undetectable in the floodplain of the Ganges River 5. The sudden change in dissolved As across the Himalayan-Ganges boundary was explained by the long residence time in the vicinity of a sediment particle necessary for the large multivalent arsenate oxyanion (AsO -3 4 ) to adsorb onto sediment; because of this, the arsenate oxyanion can adsorb onto sediment only when the stream velocity drops 5. The sediment and adsorbed As can then settle out of the contaminated water, thereby reducing As contamination. The result that dissolved As falls to undetectable as a river passes from a steep mountain range onto a flat valley floor is so startling and has such major implications for understanding the arsenic cycle and its implications for global public health that the results are being tested in the analogous geological environment of the Wasatch Range of Utah with its corresponding flat valley floors of Utah and Salt Lake Valleys. The purpose of this study is to examine the Provo and American Fork Rivers, which traverse the Wasatch Range and Utah Valley to drain into Utah Lake, as well as Little Cottonwood Creek which also originates in the Wasatch Range and later meets with Jordan River to drain into the Great Salt Lake (see Fig. 1), in order to determine the possible mechanisms controlling As release. Fig. 1: Locations of Little Cottonwood Creek, American Fork River, and Provo River; 1427
1.2 Geologic Setting Sample locations for American Fork and Provo Rivers marked with black circles Utah Valley and Salt Lake Valley are part of the Basin and Range Province with the Wasatch Mountains forming the western boundary (see Fig. 1). According to a USGS hydrological survey 6, the largest source of water for groundwater in Utah County is seepage from surficial sources, including creeks, rivers, and canals. In Utah County, Provo and American Fork Rivers are among the eight rivers that seep into basin-fill deposits, where groundwater is stored. High infiltration occurs at the mountain front, where the mountain streams flow into the valley 6. Near where the rivers discharge into Utah Lake, they become gaining streams. Provo River is by far the largest drainage basin; seepage is higher when it s in the mountain (39,400 acre-ft/yr) than when in the valley (30,000 acre-ft/yr) 6. A recent study by Lachmar, Burk, and Kolesar 7 found that metals and As in American Fork River were coming from groundwater, which was contaminated from Pacific Mine mine tailings. Groundwater added metals and As to the river even during months when the river was losing water overall to groundwater, but still receiving enough contaminated groundwater to have an overall positive flux of metals and As into the river. A negative net flux, meaning that the river is losing metals and As, occurs because the dissolved metals adsorb onto ferric oxides. Lachmar, Burk, and Kolesar 7 concluded that the most important factor in As and metal concentration is the amount of hydrous ferric oxides present in the water. 2. Methodology Twenty water samples were collected from both the American Fork and Provo Rivers, which traverse the Wasatch Range and Utah Valley to drain into Utah Lake (see Fig. 1). At the site of sampling, ph was measured with a Hach EC-10 ph Meter and electrical conductivity (EC) with a Hanna HI9033 Multirange Conductivity Meter. Location and elevation were measured with a Trimble Juno GPS Receiver at the site of sampling. Water samples were analyzed with a Hach DR-2700 Spectrophotometer for As and the transition elements normally associated with As (Fe, Cu, Ni, Co, Mn, Zn, Cr). The analysis method for As followed the silver diethyldithiocarbamate method (U.S. Environmental Protection Agency Standard Method 3500-As) with a threshold of 0.01 mg/l As. The spectrophotometer was calibrated after every 35 samples or 60 days, whichever came first, using three standard solutions with As concentrations of 0.020 mg/l, 0.040 mg/l and 0.200 mg/l. 3. Data Analysis of all twenty samples from Provo River is shown below (see Table 1). Nine water samples from upstream of the mouth of Provo Canyon show elevated dissolved As (As = 0.045-0.095 mg/l) at the highest elevations followed by a sharp downstream drop in dissolved As (As = 0.011-0.017 mg/l) (see Fig. 3). On the other hand, nine water samples downstream from the mouth of Provo Canyon showed dissolved As sharply rising to extremely elevated values (As = 0.376 0.436 mg/l). These As levels downstream of the mouth of the canyon far exceeded the EPA water standard for freshwater streams for chronic exposure (As = 0.150 mg/l) and for acute exposure (0.340 mg/l) 8. The last two samples closest to Utah Lake dropped significantly to 0.048-0.052 mg/l. Arsenic does not appear to follow a normal or lognormal distribution, but instead, seems to show a bimodal distribution (see Fig. 3). Any regression statistics were completed using normal values, although interpretation is cautious. 1428
Table 1: Provo River Geochemistry As Fe Co Mn Cr Ni Cu Zn PR01 0.095 0.010 0.000 0.2 0.01 0.000 0.12 0.1 PR02 0.045 0.010 0.000 0.6 0.01 0.013 0.14 0.13 PR03 0.011 0.010 0.030 0.3 0.01 0.000 0.11 0.09 PR04 0.013 0.020 0.001 0.4 0.01 0.000 0.02 0.011 PR05 0.017 0.000 0.025 0.5 0.00 0.014 0.01 0.09 PR06 0.012 0.010 0.003 0.1 0.00 0.025 0.09 0.10 PR07 0.013 0.010 0.002 0.2 0.02 0.011 0.03 0.06 PR08 0.011 0.040 0.000 0.1 0.01 0.006 0.00 0.07 PR09 0.011 0.010 0.008 0.2 0.03 0.048 0.02 0.09 PR10 0.367 0.010 0.003 0.7 0.04 0.019 0.02 0.06 PR11 0.408 0.010 0.051 0.1 0.02 0.003 0.00 0.09 PR12 0.406 0.000 0.025 1.3 0.03 0.001 0.04 0.11 PR13 0.382 0.000 0.045 0.1 0.00 0.000 0.00 0.09 PR14 0.406 0.000 0.010 1.2 0.02 0.000 0.03 0.11 PR15 0.412 0.000 0.033 2.6 0.04 0.012 0.05 0.10 PR16 0.430 0.020 0.011 0.1 0.01 0.006 0.06 0.25 PR17 0.436 0.060 0.054 0.3 0.00 0.008 0.16 0.27 PR18 0.418 0.030 0.005 0.7 0.03 0.004 0.04 0.26 PR19 0.048 0.020 0.036 0.3 0.02 0.005 0.12 0.19 PR20 0.052 0.030 0.052 0.1 0.02 0.005 0.14 0.15 Elevation (m) 1650 1600 1550 1500 1450 1400 1350 Elevation Copper Arsenic 1 0.1 Element concentration 1300 0.01 0 5 10 15 20 25 30 Distance from Deer Creek Resevoir (km) Figure 2: Arsenic and copper values as Provo River progresses downstream 1429
Cumulative Percentage 100 90 80 70 60 50 40 30 20 10 0 0 20 40 60 80 100 Cumulative Percentage of Equivalent Normal Distribution Figure 3: Arsenic values seem to follow bimodal distribution, rather than normal or lognormal distribution For the low values upstream of the mouth of the canyon, As was not correlated with slope (R 2 = 0.04), as measured by the change in elevation over a 500-m reach of river centered on the sample site. Arsenic did not show any correlation with ph (R 2 = 0.20) or EC (R 2 = 0.03). Arsenic values showed no correlations with any of the transition elements. Most of the transition elements showed weak correlations with As when separating values into upstream (above mouth of Provo Canyon) and downstream (below mouth of Provo Canyon) groups (R 2 = 0.0002-0.142). However, Cu showed a moderate correlation with the low As above the mouth of the canyon (R 2 = 0.38) and with the ten high values below the mouth of the canyon (R 2 = 0.48) (see Fig. 4). However, Cu does not experience a similar spike at the mouth of the canyon (see Fig. 3). 0.5 0.4 R² = 0.48 As 0.3 0.2 0.1 0 Wasatch Range Wasatch Valley Floor R² = 0.36 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 Cu Figure 4: Arsenic values show moderate correlation with copper 4. Tentative Conclusions Results thus far suggest that there are multiple competing factors that mobilize As in Provo River. The high values downstream of the mouth of the canyon suggest an accumulation of historic mine tailings at the mouth of the 1430
canyon. The correlation between Cu and As suggests that As may be derived from Cu sulfides, such as chalcopyrite. A spike in one of the transition elements is expected, but not seen, perhaps because only some Cu deposits have coprecipitated As. Low correlation of As with slope above the mouth of the canyon suggests that even before the accumulation of mine tailings at the mouth, stream velocity is not impacting As desorption, as was seen at the Himalaya-Ganges Boundary. Arsenic values are consistently high for a long stretch of Provo River, with a steep drop as the river approaches Utah Lake. Since Provo River is a constructed channel in the valley, it s possible that a significant drop in stream velocity does not occur at the mouth of the canyon, but rather, only once the river nears Utah Lake. This is unlike the results found at the Himalayan-Ganges boundary, where As dropped to undetectable concentrations immediately as the river entered the valley, suggesting that the factors related to As contamination in Provo River are more complex. The artificial redirection of Provo River may impact stream velocity and As values, as well as the impact of mining in the area. 5. Future Research Ten water samples will be collected from Utah Lake, as well as twenty water samples from Little Cottonwood Creek, which traverses the Wasatch Range and Salt Lake Valley to drain into Great Salt Lake. Forty sediment samples will also be collected from the three rivers and Utah Lake to test the hypothesis that dissolved As adsorbs onto sediment at the base of the Wasatch Range. Sediment samples will be analyzed for As by adding 10 g of airdried sediment to 250 ml of 1 M HCl and stirring for 60 minutes. The mixture will be filtered and then measured for As following the same method as for water samples. The effect of mine tailings may act as a confounding factor, so an additional river unaffected by mining will be added to this study. Utah's State Geographic Information Database contains information on historic mining in Utah, which can be used to determine whether mining has affected Provo River, and also, whether there are any mountain streams that empty out into a flat valley that are unaffected by mining. Further research in Kathmandu Valley, a flat valley floor at the base of the steep Shivapuri Range, will also be useful to continue examining As mobilization. The Bagmati River is geologically analogous in that it starts as a steep, rapid mountain stream and empties out in a flat valley floor. We will collect water and sediment samples from the Shivapuri summit (the source of the Bagmati River) to the Chobhar Gorge, where the Bagmati River exits Kathmandu Valley and heads toward the Ganges River. 6. References 1. David Rubinos, Luz Iglesias, Rosa Devesa-Rey, Francisco Diaz-Fierros,, and Maria T. Barral, Arsenic Release from River Sediments in a Gold-Mining Area (Anllons River Basin, Spain); Effect of Time, ph and Phosphorous Concentration, European Journal of Mineralogy 22 (2010): 665. 2. P. L. Smedley and D. G. Kinniburgh, A Review of the Source, Behavior, and Distribution of Arsenic in Natural Waters, Applied Geochemistry 17 (2002): 517. 3. Scott Fendorf and Benjamin D. Kocar, Biogeochemical Processes Controlling the Fate and Transport of Arsenic: Implications for South and Southeast Asia, Journal of Agronomy, 104 (2009): 137. 4. Steven H. Emerman, Tista Prasai, Ryan B. Anderson, and Mallory A. Palmer, Arsenic Contamination of Groundwater in the Kathmandu Valley, Nepal, as a Consequence of Rapid Erosion, Journal of Nepal Geological Society 40 (2010): 49. 5. Steven H. Emerman, Ryan B. Anderson, Sushmita Bhandari, Roshan R. Bhattarai, Mallory A. Palmer, Tara N. Bhattarai, and Michael P. Bunds, Arsenic and other Heavy Metals in the Sunkoshi and Saptakoshi Rivers, Eastern Nepal, Journal of Nepal Geological Society 43 (2011): 101. 6. Jay R. Cederberg, P. M. Gardner, and S. A. Thiros, Hydrology of Northern Utah Valley, Utah County, Utah, 1975-2005, (U.S. Geological Survey Scientific Investigations Report 2008-5197, 2009). 1431
7. Thomas E. Lachmer, Neil I. Burk, and Peter T. Kolesar, Groundwater Contribution of Metals from an Abandoned Mine to the North Fork of the American Fork River, Utah, Water, Air, and Soil Pollution 173 (2006): 103. 8. U.S. Environmental Protection Agency, National Recommended Water Quality Criteria, Section 304(a) Clean Water Act, http://water.epa.gov/scitech/swguidance/standards/current/index.cfm#cmc 1432