Bibliography: leaves 137-151. / xiii, 151 leaves : ill. ; 30 cm. / Title page, contents and abstract only. The complete thesis in print form is available from the University Library. / Ten soils from northern New South Wales, Australia, were sampled and assessed for their capacity to sorb arsenate (Asv) in relation to soil properties. / Thesis (Ph.D.)--University of Adelaide, Dept. of Soil Science, 1998

Bibliography: leaves 106-114.

Arsenic contamination in drinking water is a severe problem worldwide. The best way to prevent hazardous diseases from chronic arsenic exposure is to remove the exposure. Efforts to remediate arsenic in drinking water have taken two tracks. One is to provide surface or shallow well water sources as an alternative to the arsenic contaminated deep wells. Another approach is to remove arsenic from the contaminated water. Different removal technologies like oxidation, chemical coagulation, precipitation, adsorption and others are available. There are problems and benefits associated with each of these approaches that can be related to cultural, socio-economic and engineering influences. The method proposed in this research is adsorption of arsenic to iron coated limestone. Different iron coated limestone samples were prepared. Standard solutions of 100ppb arsenic were prepared and batch and kinetic experiments were conducted. The final solution concentrations were analyzed by Graphite Furnace Atomic Adsorption Spectroscopy (GFAAs) and the results showed that iron coated limestone removed arsenic below 10ppb with 5 grams of material. Variations in iron coverage impacted efficiency of arsenic removal.

Graphite furnace and hydride generation atomic absorption, GFAA and HGAA, techniques have been developed and applied to the determination of arsenic concentrations in some marine invertebrates, mainly bivalves and gastropods, of British Columbia. Total arsenic concentrations in bivalves vary with species, ranging from 0.6-9.1 μg g⁻¹ (wet weight basis). Arsenic concentrations in the bivalve shells show a wider range of 0.1 to 26.3 μg g⁻¹ (dry weight basis). Gastropods show relatively higher arsenic concentrations in the soft tissues, 17.3-48.4 fig μg g⁻¹, and concentrations in the shells range from 1.4 to 16.3 fig μg g⁻¹. There is no correlation between arsenic levels in the soft-tissues and shells. There is also no correlation between arsenic levels in the organisms and the surrounding sediments and sediment pore waters. HPLC-GFAA techniques have been developed and used for the separation and quantitation of-arsenite, arsenate, methylarsonic acid, dimethylar-sinic acid, arsenobetaine, arsenocholine iodide and tetramethylarsonium iodide. This technique together with TLC, NMR, FAB and thermospray LCMS were employed for the detection of water-soluble arsenic compounds in 5 species of clams - Butter clam Saxidomus giganteus. Horse clam Schizothoerus nuttalli. Soft-shelled clam Mva arenaria. Native-littleneck clam Protothaca staminea and Manila clam Venerupis laponica. Varying amounts of arsenobetaine and tetramethylarsonium ion are found in all the clams. Butter clams show the pres ence of a third compound which appears to be trimethylarsine oxide. Small amounts of unknown arsenic containing compounds are present which are yet to be characterized . Arsenic speciation in 3 gastropods was also examined. The Northwest neptune Neptunea lvrata. the Thick-ribbed whelk Berinpius crebriscotata and Phoenician whelk Neptunea phoenicius all contain arsenobetaine and at least two unidentified arsenicals. / Science, Faculty of / Chemistry, Department of / Graduate

Critical water resources, such as groundwater, are undergoing a period of intense and global environmental change, driven by climate change, anthropogenic impacts and exploitation, and perturbations to interactions of fundamental processes that are affected by hydrological, mineralogical and biogeochemical factors. Arsenic contamination is a significant threat to these water resources and the populations who depend on them, yet there are few studies directly linking water quality with changes in hydrology and geochemistry in sediments on varying scales. My research explores environmental variability in hydrology and redox processes that regulate soluble arsenic concentrations at the pore scale (µm to mm), and develops methods of upscaling these mechanistic studies to understand heterogeneity in groundwater arsenic levels and their impacts on public health at larger scales (a couple of meters to hundreds of kilometers). Specifically, my research examines the interaction of redox processes in the Earth’s subsurface that drive the release of arsenic into groundwater. Naturally-occurring, or geogenic, arsenic contamination is the main source of arsenic release into groundwater that affects human health, with possible anthropogenic exacerbation of this natural contamination. Throughout this dissertation, I have developed a suite of data-driven approaches to understand and quantify the highly variable factors that underlie the mechanisms of geogenic arsenic release into groundwater and its migration in the environment. In Chapter 1, I investigate the effects of hydrologic perturbations on formerly uncontaminated aquifers that release arsenic due to increased groundwater pumping in the Red River Delta, Vietnam. To compare the effect of hydrologic processes to measured groundwater arsenic concentrations, I used Monte Carlo simulations in an end-member mixing model and quantified fraction of different recharge sources into an aquifer based on stable water isotopes. I find that changing flow patterns due to groundwater abstraction have increased the extent of arsenic release into groundwater and also changed the location of where arsenic contamination originates. In Chapter 2, I characterize iron mineralogy associated with arsenic release through sampling of sediment cores across a lateral redox gradient in Vietnam with extensive spectroscopy measurements. Through hierarchical cluster analysis on this data set of X-ray absorption spectroscopy (XAS) measurements of borehole cuttings paired with dissolved groundwater measurements, I reveal signatures of iron mineral reduction that could cause or exacerbate arsenic release. This was upscaled to other deltaic aquifers in South and Southeast Asia based on groundwater data to identify aquifers at risk of arsenic release. I showed that the extent of older and previously pristine aquifers that have been contaminated may have been misclassified and thus underrepresented in deltaic aquifers throughout South and Southeast Asia, disrupting the assumption that older and deeper aquifers are oxidized and thus guarded against arsenic release. In Chapter 3, I use process-based reactive transport modeling of a laboratory-scale experiment to mechanistically explain the infiltration of contaminated water into uncontaminated aquifers and find that arsenic contamination cannot be explained by the commonly invoked mechanism of iron reducing bacteria only, but instead relies on sulfate reduction and complexation of aqueous arsenic in solution. The role of sulfate reduction in mobilizing arsenic in groundwater is in stark contrast to and undermines the previous use of sulfate reduction as strategy for arsenic remediation. Finally, in Chapter 4, I quantitatively examine the processes that release arsenic across different arsenic-impacted aquifers, based on the relationships between redox status of iron and arsenic mineralogy and groundwater concentrations. Synthesis of X-ray absorption spectra of the deltaic aquifers of Southeast Asia and the glacial aquifer system in the Northern United States shows that arsenic release occurs in similar geochemical environments in both systems, and is highly generalizable via statistical and unsupervised machine learning approaches. This dissertation demonstrates that common assumptions behind geogenic arsenic release must be tested: from which aquifers are low in arsenic to the commonly assumed mechanism of arsenic release by iron reducing bacteria. These findings also reveal that the extent of anthropogenic impact on geogenic arsenic contamination is detectable: from changes in recharge sources to changes in mineralogy that affect arsenic concentrations and human health. The next step is to use these data driven and machine learning approaches to quantify the vulnerability of affected aquifers, to mitigate the risk of those currently reliant on contaminated groundwater, to reduce the risks of future contamination and, ultimately, to protect human health.

Background: Environmental exposure to inorganic arsenic (InAs) is a considerable worldwide problem, and over 57 million people in Bangladesh have been chronically exposed to arsenic-contaminated drinking water. Ingested inorganic arsenic (InAs) undergoes hepatic methylation generating monomethyl- (MMAs) and dimethyl- (DMAs) arsenic species in a process that facilitates urinary As (uAs) elimination. Of these three metabolites (InAs, MMA and DMA), MMA the most toxic, InAs is the second most toxic, and DMA is the least toxic. Consequently, increased MMAs is associated with increased risk of As-related adverse health outcomes. One-carbon metabolism (OCM), the biochemical pathway that provides methyl groups for As methylation, is influenced by folate and B12. A growing body of research, including cell-culture, animal-model, and epidemiological studies, have demonstrated the role of OCM-related micronutrients in As methylation. While folate supplementation is known to increase As methylation and lowers blood As (bAs) in adults, little data is available for adolescents. OCM also supports nucleotide and amino acid synthesis, particularly during periods of rapid growth, such as adolescence. In Bangladesh, deficiencies in folate and vitamin B12 are widespread. These micronutrients are essential for OCM and As methylation, and they are also critical for neural development, since they are necessary for the synthesis of neurotransmitters and myelin, and critical for generation of S-adenosyl-methionine (SAM) via OCM. Deficiencies in folate and B12 have been shown to negatively influence cognitive function in adults and children, but this has not been adequately characterized in adolescents. Individuals living in Bangladesh, facing As exposure and nutritional deficiency, also experience environmental exposure to elevated levels of cadmium (Cd), manganese (Mn), and lead (Pb). These metals have been linked to adverse neurocognitive outcomes in adults and children, though their effects on adolescents are not yet fully characterized. Additionally, previous studies have linked selenium (Se) levels to protective effects against toxicity of these other metals, as Se is an essential nutrient. Metal mixtures are also understudied in adolescents, and more research is needed. Objectives: Firstly, the research presented in this dissertation will examine the previously published evidence that nutritional status and nutritional interventions can influence the metabolism and toxicity of As, with a primary focus on folate. Secondly, the associations between OCM-related micronutrients and As methylation in Bangladeshi adolescents chronically exposed to As-contaminated drinking water will be studied. Thirdly, this dissertation will investigate the associations between folate and B12 nutritional status, homocysteine, and cognitive function as measured by two different test instruments in Bangladeshi adolescents. Finally, the associations between mixed metals exposure and cognitive function in Bangladeshi adolescents will be examined. Methods: The Metals, Arsenic, & Nutrition in Adolescents study (MANAs) is a cross-sectional study of 738 Bangladeshi adolescents aged 14-16 years, whose parents were enrolled in our group’s previous Health Effects of Arsenic Longitudinal Study (HEALS). Venous blood samples were collected from adolescent participants for measurement of plasma folate, red blood cell (RBC) folate, plasma B12, plasma homocysteine (Hcys), blood As (bAs), blood cadmium (bCd), blood manganese (bMn), blood lead (bPb), and blood selenium (bSe). Urine samples were collected for measurement of urinary arsenic and urinary arsenic metabolites (InAs, MMA, and DMA) expressed as a percentage of total urinary As: %InAs, %MMAs, %DMAs. Additionally, participants completed a modified version of the Wechsler Intelligence Scale for Children-Fourth Edition (WISC-IV) and an abbreviated version of the Cambridge Neuropsychological Test Automated Battery (CANTAB). Associations between predictors and outcomes were assessed using linear regression analyses. Associations for the mixture of metals were also examined via Bayesian Kernel Machine Regression (BKMR), which assessed the effects of the metals mixture in addition to examining the effects each individual metal component. Results: In the linear regression analyses examining associations between nutritional status and arsenic methylation profiles controlling for water As and BMI, in girls we observed that RBC folate was inversely associated with bAs, plasma B12 was inversely associated with uAs, and, somewhat unexpectedly, plasma Hcys was inversely associated with %MMA. Among boys, we saw that plasma folate was inversely associated with %InAs and positively associated with %DMA, RBC folate was inversely associated with %InAs and positively associated with %MMA, while Hcys was positively associated with %InAs. In linear regression analyses examining the associations between nutritional status and cognitive function controlling for covariates, we found no significant associations between nutritional status and cognitive function as measured by WISC. For cognitive function as measured by CANTAB, we observed positive associations between plasma folate and spatial recognition memory, and between plasma B12 and spatial working memory. In our investigation of the associations between metal exposures and cognitive outcomes, linear regression analysis revealed negative associations between exposure to As and Mn and spatial working memory. Negative associations were also seen between bCd and Spatial Recognition Memory, and between bPb and Delayed Match to Sample. Finally, a positive association was seen between bSe and Spatial Span Length. Our BKMR results showed no overall effect of the mixture but further characterized the associations for individual metals within the mixture. BKMR analyses indicate that bPb has a negative association to Delayed Match to Sample, and that there are positive associations between bSe and Planning, Reaction Time, and Spatial Span. BKMR also showed higher concentrations of bCd to be negatively associated with Spatial Recognition Memory. Posterior inclusion probability consistently rated Se, which has a protective effect, as the most influential component of the mixture. Conclusions: These findings suggest that associations between OCM nutritional status, blood arsenic, and distribution of urinary As metabolites in adolescents are similar to previously reported observations in adults and in children. The inverse association between Hcys and %MMA in girls is unexpected since Hcys is known to be an indicator of impaired OCM and low folate/B12 in adults. These findings also suggest that associations between nutritional status and cognitive function in adolescents are consistent with previously reported associations in adults and children. The observation of positive associations between folate and spatial recognition memory and between B12 and spatial working memory are consistent with findings in animal models, children, and adults, all of which link deficiencies in these two micronutrients to memory deficits. The findings of the metals investigation are also consistent with previously reported observations in adults and children, and these results also suggest agreement between linear regression and BKMR analysis of the mixed metal exposure, with the BKMR further demonstrating associations seen in the linear regression analysis. Generally, Se had a protective effect for cognitive outcomes, whereas Mn and As were linked to poorer working memory, and Cd and Pb were linked to poorer visual recognition and memory. BKMR reinforced and further characterized results of the linear regression analyses. Overall, these results indicate that the associations between OCM-related micronutrients, arsenic methylation, metal exposure, and cognitive function in adolescents are generally similar to prior findings in adults and children. However, additional studies are needed to evaluate the impact of OCM and As methylation on As-related adverse health outcomes (such as cancer and cardiovascular disease) in people exposed to As during adolescence. These results also suggest that further investigation into the associations between nutritional status and measures of cognitive function in adolescents is merited, and that further exploration of homocysteine’s role in adolescent physiology is needed. Most importantly, these findings suggest that both nutritional interventions involving B12 and folate supplementation and metal exposure mitigation efforts may have a positive impact on overall health and well-being for individuals facing these environmental exposures. In the long term, interventions to reduce exposure to metals and nutritional deficiency, and interventions to attenuate the effects of these exposures have the potential to influence lifelong cognitive function, possibly influencing professional achievement and economic outcomes in regions having a high prevalence of nutritional deficiencies, arsenic exposure, and environmental exposure to metals mixtures.

Exposure to arsenic (As) is a major public health concern globally. Inorganic As (InAs) undergoes hepatic methylation to form monomethyl (MMAs)- and dimethyl (DMAs)-arsenical species, facilitating urinary As elimination. MMAsIII is considerably more toxic than either InAsIII or DMAsV, and a higher proportion of MMAs in urine has been associated with risk for a wide range of adverse health outcomes. One-carbon metabolism (OCM) is a biochemical pathway that provides methyl groups for the methylation of As, and is influenced by folate and other micronutrients, such as vitamin B12, choline, betaine and creatine. A growing body of evidence has demonstrated that OCM-related micronutrients play a critical role in As methylation. To analyze the impact of As exposure, it needs to be properly quantitated. Urinary As (uAs) is a biomarker of As exposure. Urinary creatinine (uCr) or specific gravity (SG) are used to correct uAs for urine dilution. However, uCr is correlated with As methylation, whereas SG has limitations in individuals with kidney damage. Therefore, it is important to determine which urine dilution proxy is appropriate in As-related research. In Chapter 2 we conducted a review that summarized observational epidemiological studies, interventions, and relevant experimental evidence examining the role that OCM-related micronutrients have on As methylation, toxicity of As, and risk for associated adverse health-related outcomes. People with higher relative percentage of MMAs (%MMAs) in urine (inefficient As methylation), have been shown to have a higher risk of cardiovascular disease and several cancers but appear to have a lower risk of diabetes and obesity in populations from the US, Mexico, and Taiwan. It is unknown if this opposite pattern with obesity is present in Bangladesh, a country with lower adiposity and higher As exposure in drinking water. Efficiency of As methylation differs substantially between species, between individuals, and across populations. In Chapter 3, we aimed to evaluate which urine dilution correction methods for uAs most accurately predicted blood As (bAs). We used data from the Folic Acid and Creatine Trial (FACT; N = 541) and Folate and Oxidative Stress (FOX; N = 343) study in Bangladesh. Three linear regression models were assessed using uAs (1) adjusted for uCr or SG as separate covariates, (2) standardized for uCr or SG, i.e., uAs/uCr, and (3) adjusted for residual corrected uCr or SG following adjustment for age, sex and BMI. Median uAs/bAs for FACT and FOX were 114/8.4 and 140/12.3 µg/L. In FACT, two-fold increases in uAs adjusted for uCr or SG were related to 34% and 22% increases in bAs, respectively, with similar patterns in FOX. In Chapter 4, we investigated the effects of folic acid (FA) and/or creatine supplementation on the concentrations of As species and primary (PMI: MMAs/InAs) and secondary (SMI: DMAs/MMAs) methylation indices in blood in Bangladeshi adults having a wide range of folate status. In a randomized, double-blinded, placebo-controlled trial, 622 participants were assigned to FA (400 or 800 μg/day), 3 g creatine/day, 3 g creatine + 400 μg FA/day, or placebo for 12 weeks. For the following 12 weeks, half of the FA participants were randomly switched to receive placebo. All participants received As-removal water filters at baseline. Blood As species were measured at baseline, and weeks 1, 12, and 24. In all groups, blood As species concentrations decreased due to filter use. After 1 week, the mean within-person increase in SMI for the creatine + 400FA group was greater than that of the placebo group (p = 0.05). The mean percent decrease (95% CI) in blood concentrations of MMAs (bMMAs) between baseline and week 12 was greater for all treatment groups compared to the placebo group [400FA: -10.3 (-11.9, -8.8); 800FA: -9.5 (-11.1, -8.0); creatine: -5.9 (-8.6, -3.0); creatine + 400FA: -8.4 (-10.0, -6.9); placebo: -2.0 (-4.0, 0.0)], and the percent increase in blood DMAs (bDMAs) concentrations for the FA treated groups all significantly exceeded that of placebo [400 FA: 12.8 (10.5, 15.2); 800 FA: 11.3 (8.90, 13.8); creatine + 400 FA: 7.40 (5.20, 9.70); placebo: -0.10 (-2.80, 2.60)]. The mean decrease in PMI and increase in SMI in all FA groups significantly exceeded placebo (p < 0.05). Data from week 24 showed evidence of a reversal of treatment effects on As species from week 12 in those who switched from 800FA to placebo, with significant decreases in SMI [-9.0% (-3.5, -14.8)] and bDMAs [-5.9% (-1.8, -10.2)] in those who switched from 800FA to placebo, whereas for those who remained on 800FA, PMI and bMMAs concentrations continued to decline [-7.2% (-0.5, -14.3) and -3.1% (-0.1, -6.2), respectively] for those who remained on 800FA supplementation. This trial was registered at https://clinicaltrials.gov as NCT01050556. In Chapter 5, we characterized the association between body mass index (BMI) and As methylation in Bangladeshi adults and adolescents participating in the FACT; FOX; and Metals, Arsenic, and Nutrition in Adolescents Study (MANAS). Arsenic species (InAs, MMAs, DMAs) were measured in urine and blood. Height and weight were measured to calculate BMI. The associations between concurrent BMI with urine and blood As species were analyzed using linear regression models, adjusting for nutrients involved in OCM such as choline. In FACT, we also evaluated the prospective association between weight change and As species. Mean BMIs were 19.2/20.4, 19.8/21.0, and 17.7/18.7 kg/m2 in males/females in FACT, FOX, and MANAS, respectively. BMI was associated with As species in female but not in male participants. In females, after adjustment for total urine As, age, and plasma folate, the adjusted mean differences (95% confidence) in urinary %MMAs and %DMAs for a 5 kg/m2 difference in BMI were -1.21 (-1.96, -0.45) and 2.47 (1.13, 3.81), respectively in FACT, -0.66 (-1.56, 0.25) and 1.43 (-0.23, 3.09) in FOX, and -0.59 (-1.19, 0.02) and 1.58 (-0.15, 3.30) in MANAS. The associations were attenuated after adjustment for choline. Similar associations were observed with blood As species. In FACT, a 1-kg of weight increase over 2 to 10 (mean 5.4) years in males/ females was prospectively associated with mean %DMAs that was 0.16%/0.19% higher. BMI was negatively associated with %MMAs and positively associated with %DMAs in females but not males in Bangladesh; associations were attenuated after plasma choline adjustment. In conclusion, we found that FA supplementation lowers bMMAs and increases bDMAs in a sample of primarily folate-replete adults, while creatine supplementation lowers bMMAs. Evidence of the reversal of treatment effects on As species following FA cessation suggests short-term benefits of supplementation and underscores the importance of long-term interventions such as FA fortification. Additionally, there is fairly robust evidence supporting the impact of folate on As methylation, and some evidence from case-control studies indicating that folate nutritional status influences risk for As-induced skin lesions and bladder cancer. However, the potential for folate to be protective for other As-related health outcomes, adverse health risks of high folate/FA levels (particularly in areas where folate supplements are common), and beneficial effects of other OCM-related micronutrients on As methylation and risk for health outcomes are not as well studied and warrant additional research. We also found that the role of body fat on estrogen levels that may influence OCM, e.g. by increasing choline synthesis. Research is needed to determine whether the associations between BMI and As species are causal and their influence on As-related health outcomes. Finally, we found that in assessing urine dilution correction approaches, models with uCr consistently had lower AIC values than SG across methods. The uAs associations with bAs were stronger after adjustment for uCr vs. SG. Decisions regarding urine dilution methods should consider whether the study outcomes are influenced by factors such as methylation or medical conditions.

One hundred and eighty-six soil samples from Northwest Oregon were tested for arsenic content. The highest values measured were 13.9 ppm in the A horizon (site C4) and 20.4 ppm in the B horizon (Site P4). Arsenic was not detected in 28 A horizon samples and 23 B horizon samples. Data are grouped based on the age and rock type of underlying bedrock. Lithologic groups with six or more data points were compared statistically to ascertain if groups are distinct. Analysis of Variance (ANOVA) multiple comparison tests indicate that the arsenic content of the Marine Sediments and Sedimentary Rocks group samples is distinguishable from the Quaternary Basalts group in the A horizon and all other groups in the B horizon. Kruskal-Wallis multiple comparison tests indicate that the arsenic content of the Marine Sediments and Sedimentary Rocks group is distinguishable from the Quaternary Basalts, Quaternary/ Tertiary Sediments and Sedimentary Rocks and Volcanic Sediments groups in both the A and B soil horizons. The ANOVA and Kruskal-Wallis tests compared A and B horizon data by lithologic group. The ANOVA shows the Marine Sediments and Sedimentary Rocks group in the A horizon is distinct from the Quaternary Basalts in the A and B horizon. The Kruskal-Wallis test yielded the same result. Per the ANOVA, the Marine Sediments and Sedimentary Rocks in the B horizon are distinct from all other tested groups. The Kruskal-Wallis test shows the Marine Sediments and Sedimentary Rocks group in the B horizon as distinct from the Quaternary Basalts, Quaternary/ Tertiary Sediments, and Volcanic Sediments groups in the A and B horizon. A K-means cluster analysis was used to group all available data independent of underlying bedrock. Three, four, and five group analyses were conducted, and the results of these tests were compared to the data grouped by underlying rock type. No correlation between the groups resulting from the K-means cluster analysis and groups based on underlying lithology was found. This analysis supports the creation of a map distinguishing arsenic content in the soils above Marine Sediments and Sedimentary Rocks group units from arsenic content in all other tested lithologic groups. The mean and standard deviations of these groups (in ppm) are: A horizon: Marine Sediments (6.09 ±2.66); other groups (3.10 ±3.19); B horizon: Marine Sediments (10.26 ±4.65); other groups (3.13, ±2.52). This analysis indicates that geologic context must be taken into account when determining background levels of naturally occurring arsenic in soils.

This study examines the natural background concentrations of arsenic in the soils of southwest Oregon, using new samples in addition to data collected from previous theses (Khandoker, 1997 and Douglas, 1999). The original 213 samples were run by ICP-AES with a reporting limit of 20 ppm, and only three samples had detected values. The original samples were tested again (2013) at a lower reporting limit of 0.2 ppm by ICP-MS, as were 42 new samples (2013), to better ascertain the natural levels of arsenic in undisturbed soils. The aim is to add to the existing DEQ data set, which has been used to establish new regulatory levels based on natural levels in the environment that are both safer and more economically viable than the former risk-based remediation levels (DEQ, 2013). The maximum and mean concentrations, respectively, for each province (with high formation map unit) are 85.4 and 21.99 ppm for South Willamette Valley (Tfee), 45.4 and 5.42 ppm for the Klamath Mountains (Jub), 11.9 and 2.76 ppm for the Cascade Range (Tbaa), 10.6 and 5.15 ppm for the Coast Range (Ty), 2.32 and 1.29 ppm for the Basin and Range (Qba) and 1.5 and 1.20 ppm for the High Lava Plains (Tmv). In addition, the distribution and variance of arsenic in the A and B soil horizons is assessed in this study by comparing deviation at a single site, and also by comparing A and B horizons of 119 PSU sites. One of 18 new sites sampled for this study (distinguished with the HH prefix), site HH11, was randomly chosen to evaluate differences at a single location. Site HH11 is an Inceptisol soil above volcanic rock (KJdv map unit) located at 275 meters elevation in Douglas County within the Klamath province. Five samples were taken from the A and from the B horizons at site HH11. The means and standard deviations were 3.74 ± 0.44 for the A horizon and 4.53 ± 0.39 for the B horizon. The consistency and low deviation within each horizon indicate that a single sample within a horizon is a good representative of that horizon and supports the field methodology used in this study of taking only one sample in the A horizon and one sample in the B horizon. Wilcoxon Rank-Sum test determined that A and B horizons for the 119 sites that had data for both the A and B horizons were not statistically different (p-value 0.76). Arsenic concentration is not associated with a particular horizon for these sites. However, differentiation between soil horizons increases with age (Birkeland, 1999), as does accumulation of the iron oxides and sulfide minerals on clay surfaces (McLaren et al., 2006) which concentrate in the B horizon. These associations warrant further study to see how they relate to arsenic level, soil development and age in Oregon soils. Lastly, this study statistically examines six potentially important environmental predictors of naturally occurring arsenic in southwestern Oregon: site elevation, geomorphic province, mapped rock type and age, and sample soil order and color (redness). A Classification and Regression Tree Model (CART) determined soil order, elevation and rock type to be of significant importance in determining arsenic concentrations in the natural environment. According to the regression tree, arsenic concentrations are greater within Alfisol and Ultisol/Alfisol and Vertisol soil orders, at lower elevations below 1,207 meters, and within soils from sedimentary, mixed volcanic/sedimentary and unconsolidated rock types.

Volcanic rocks typically have only low to moderate arsenic concentrations, none-the-less, elevated levels of arsenic in ground waters have been associated with pyroclastic and volcaniclastic rocks and sediments in many parts of the world. The potential for arsenic leaching from these deposits is particularly problematic as they often comprise important water-bearing units in volcanic terrains. However, the role that chemical and mineralogical variations play in controlling the occurrence and mobility of arsenic from pyroclastic rocks is largely unexplored. This study uses chemical and X-ray diffraction data to characterize and classify 49 samples of ash-flow tuffs, and 11 samples of tuffaceous sediments. The samples exhibit a range of devitrification and chemical weathering. Total and partial digestion, and water extractions of samples are used to determine the total, environmentally available, and readily leachable fractions of arsenic present in all tuff samples. Leaching experiments were also performed with buffered solutions to determine the influence of elevated pH levels on arsenic mobility. The 49 tuff samples have a mean arsenic content of 7.5 mg kg-1, a geometric mean arsenic content of 4.8 mg kg-1, a median arsenic content of 5.2 mg kg-1, and a maximum arsenic concentration of 81 mg kg-1. The mean and median values are 2.8 - 4.4x the average crustal abundance of 1.7 mg kg-1 (Wedepohl, 1995), and consistent with previously reported values for volcanic glasses and felsic volcanic rocks (Onishi and Sandell, 1955; Wedepohl, 1995), although the maximum arsenic content is higher than previously reported (e.g., Casentini et al., 2010; Fiantis et al., 2010; Nobel et al., 2004). In addition, the arsenic concentrations of tuffs were found to be highly heterogenous, both between and within individual units, and in some cases, individual outcrops. Results of whole rock and leachate analyses indicate that there is no significant difference in the total arsenic content of tuffs as a result of devitrification or weathering, but both devitrified and weathered tuffs contain higher levels of environmentally available arsenic than unweathered glassy tuffs. Glassy tuffs did not produce any readily leachable arsenic, while individual devitrified and weathered tuffs both generated aqueous concentrations that exceeded regulatory limits after 18 hours. Leaching of weathered tuffs produced higher levels of arsenic at high (~9-11) pH than in tests conducted at circum-neutral pH. Devitrified and glassy tuffs showed no increase in leachable arsenic with increasing pH. The results of this study indicate that devitrification and weathering processes determine the host phases, degree of adsorption, and overall mobility of arsenic from ash-flow tuffs. Tuffs that have undergone different types of alteration are likely to have different host phases of arsenic, and different mechanisms that mobilize arsenic into the environment. Potential host phases and mobility mechanisms are discussed, and a conceptual model of arsenic behavior in ash-flow tuffs is proposed.