Naturally occurring aqueous arsenic and seawater intrusion on Lummi Island, WA

Two different types of groundwater contamination may be present in the aquifers on northern Lummi Island, Washington: naturally occurring arsenic and seawater intrusion. Freshwater on northern Lummi Island is stored in bedrock and unconsolidated glacial sediments. The naturally occurring arsenic, sourced from an undetermined stratigraphic layer, varies spatially throughout the island. Additionally, seawater may be intruding into the groundwater supply, which is the primary source of drinking water for the residents of the island. The process of mobilization of the naturally occurring arsenic and the extent of the seawater intrusion has not been fully explored. The purpose of my study was to determine the geochemical, physical, and seasonal influences on concentrations of arsenic and major ions on Lummi Island. I collected water samples and made in situ measurements from wells distributed throughout Lummi Island for geochemical analysis. Statistical analysis was used to test for a relationship between arsenic concentrations and geochemical factors or season. The speciation of arsenic in the groundwater was determined by plotting pH and redox potential measurements on an arsenic species stability diagram. Whole-rock chemical analysis was used to investigate the bedrock source of the arsenic. The extent of the seawater intrusion was determined using major ion analysis, and the source of the ions was interpreted using Piper diagrams. The relationship between aquifers, major ions, and seasonality was explored using multivariate statistical analysis. Whole rock analysis indicated that the highest arsenic concentration was in the sample taken from the Chuckanut conglomerate. When Eh and pH field measurements were plotted on an arsenic stability diagram, arsenate was revealed as the dominant species in the groundwater. Speciation calculations in PHREEQC supported the conclusion that arsenate was the dominant species in most water samples. No wells indicated seawater intrusion and some plotted in the freshening region of the Piper diagram. Wells that plotted in the freshening area of the Piper diagram were more likely to have higher arsenic concentrations. Bivariate analysis, principal component analysis, non-metric clustering and Piper plots failed to show a difference in the measured variables between the April and August samples. A positive correlation was found between specific conductance, Na+, Cl- and total alkalinity and dissolved arsenic, and a negative correlation was found between Ca2+ and Mg2+ and dissolved arsenic. No correlation was observed between dissolved arsenic and Fe or Mn. Multivariate statistics indicated a correlation between the presence of major ions and the dissolved arsenic concentrations. The positive correlation between alkalinity and dissolved arsenic, negative correlations between Ca2+ and Mg2+ and dissolved arsenic, and no correlations with Fe or Mn is consistent with an arsenic release through a desorption process. The presence of dissolved carbonate and bicarbonate is indicative of a chemical weathering process, which could lead to arsenic desorption, and the charge on Ca2+ and Mg2+ ions can facilitate the adsorption and desorption of dissolved arsenic. Since the Chuckanut sandstone had the highest dissolved arsenic concentrations, a chemical weathering process is most likely occurring within this stratigraphic layer. No wells in this study exceeded the SMCL (Secondary Maximum Contaminant Level), nor did any wells experience a statistically significant fluctuation in chlorides between the April and August sampling seasons. When the major ions were plotted on a Piper diagram, all of the wells plotted in either the 'fresh' or the 'freshening' part of the diagram; none of the samples plotted in the 'intruding' or 'intruded' area. Because there was no evidence that the wells in my study were experiencing seawater intrusion, the salts must be released from another source. This relationship between major ions and dissolved arsenic was supported by the multivariate statistical tests principal component analysis and linear discriminant analysis. The principal component analysis successfully classified arsenic into high and low groups, and once trained with a subset of the data, the linear discriminant analysis divided arsenic into high or low categories. The relationship between the major ions and dissolved arsenic can be interpreted from a Piper diagram when the high dissolved arsenic concentrations ([As] >0.07 mg/L) is color coded. These water samples all plotted in the freshening region of the Piper diagram. Because chlorides and dissolved arsenic were positively related, specific conductance, used as a proxy for chlorides, could be used as a rough indicator for arsenic.