Grand Pré is a rural community located on the southern shore of the Minas Basin in Kings County, Nova Scotia. The town site is located within Canada’s first designated Historic Rural District and is also a candidate for UNESCO’s World Heritage List. Water needs are supplied by privately owned drilled wells. Residents and businesses in the area have expressed concerns regarding elevated concentrations of natural contaminants in drinking water, most notably uranium. Nova Scotia does not have a comprehensive program for monitoring private well water quality, and as a result private well owners are responsible for the management and treatment of their drinking water (Nova Scotia Department of Environment and Labour 2005). Because no previous groundwater quality study has been done in the Grand Pré region to investigate these concerns, this study was undertaken to gain an understanding of the controls on uranium concentrations and mobility within aquifers underlying Grand Pré.

The presence of anomalous concentrations of uranium, radium, and radon in the Carboniferous Horton Group and underlying basement rocks of Atlantic Canada has been known for many years (e.g., Ryan and O’Beirne-Ryan 2006). During the late 1970s and early 1980s exploration for uranium in Nova Scotia was undertaken by numerous companies and resulted in several significant discoveries. However, in 1982 public pressure resulted in a moratorium on all exploration and prospecting for uranium in Nova Scotia that continues to the present.

Recent research involving the occurrence of uranium in groundwater in the region includes the identification of uranium in surface and shallow subsurface water in the Avon Valley in central Nova Scotia by Kronfeld et al. (2004). In this study, uranium concentrations in ground and surface waters were found to be generally low, except immediately adjacent to weathering uranium occurrences. Kronfeld et al. (2004) suggested that relatively high precipitation rates in the region and the high permeabilities of the sandstone aquifers in the Horton Group might result in persistent flushing. In a study of water derived primarily from these Horton Group aquifers, Page (1999) found a strong positive correlation between dissolved uranium and radon. Research by O’Beirne-Ryan and Zentilli (2003, 2006) examined the weathered surfaces on the South Mountain Batholith and their contribution of uranium and radon to groundwater in the region. Areas of high uranium concentration in Nova Scotia are associated with intrusive rock types, principally the South Mountain batholith in southwestern Nova Scotia. Both studies recognized that natural contaminants are mobilized through water-rock interactions as groundwater flows from recharge zones in granite- and slate-dominated highlands to sedimentary lowlands (Kronfeld et al. 2004). Ryan and O’Beirne-Ryan (2006) noted that numerous uranium occurrences are located in the area near Windsor, Nova Scotia. These occurrences can be divided into two types: (1) carbon- and pyrite-related quartz-sandstone-hosted type and (2) hematite arkosic sandstone-hosted type. The aerial extent of Horton Group strata throughout eastern Canada establishes the potential that similar environments for uranium accumulation likely occur elsewhere (Ryan and O’Beirne-Ryan 2006).

This study focuses on determining the influence of the physical and chemical properties of the local aquifers in Grand Pré on the concentration and mobility of uranium. Triassic Wolfville Formation sandstone and Carboniferous Horton Bluff Formation sandstone and shale underlie the study site but their distribution is poorly constrained (Fig. 1). Elevated levels of uranium may pose a local health risk but no research has been conducted to determine the influence that local geology may have on ground water concentrations.

Slate of the Cambrian-Early Ordovician Halifax Group, carbon-rich shale and sandstone of the Late Devonian–Early Mississippian Horton Bluff Formation, and arenaceous sandstone of the Late Triassic Wolfville Formation all underlie the Grand Pré region (Fig.1). The Halifax Group is composed of slate, metasiltstone, and minor quartz metawacke deposited in a deep marine environment (Williams 1995) and is unconformably overlain by Horton Group sandstone and shale. The Horton Group is host to the best-documented occurrences of uranium in southwest Nova Scotia (Ryan and O’Beirne-Ryan 2006) and is comprised of the Horton Bluff Formation (HBF) and the overlying Cheverie Formation. The HBF was deposited in a fluvial deltaic and flood-plain environment; and coal measures are common (Martel and Gibling 1996). The lower unit of the HBF is composed of coarse sandstone and conglomerate and interbedded grey siltstone and mudstone. The middle unit is dominantly mud-stone, fissile shale and siltstone. A lacustrine influence in the middle and upper units is indicated by the presence of fresh water fish, plant debris, and thin pedogenic carbonates. The presence of thin horizons containing glauconite grains, marine-affinity bivalves, and crinoids suggests a distal marine influence at various times (Ryan 1998). The upper unit of the HBF and the overlying Cheverie Formation do not appear to be present in the Grand Pré region (Ferguson 1983). Instead, an unconformable contact between the middle unit of the HBF and the overlying Triassic Wolfville Formation occurs in the Grand Pré area (Moore and Ferguson 1986).

The Wolfville Formation is composed of interbedded and stratified sandstone, pebbly conglomerate, and minor mudstone, and was deposited in an arid braided river–aeolian environment (Leleu et al. 2009). At and near its base are pebbly and conglomeratic units with clasts derived from the metamorphic and granitic highlands nearby, and from material eroded from adjacent Early Carboniferous clastic sedimentary strata. The Wolfville Formation contains both calcite and silica cement and commonly displays well developed porosity (> 20 %).

The type and distribution of surficial sediments in the study area is a reflection of the glacial processes that dominated the area during the Late Wisconsinan (ca. 22 ka–12 ka). Both regional and local ice sheets were active and resulted in a complex ice-flow history. The last regional ice advance in western Nova Scotia was most likely towards the north (Scotia Phase; Stea et al. 1992), although drumlin orientation indicates that an earlier westerly advance may have had a strong influence on sediment deposition and deformation. Drumlins have relief of up to 20 m and are composed of clay-rich Lawrencetown Till derived from Paleozoic sediments to the east. Till thickness is highly variable; in low areas between drumlins, bedrock can be within 2 m of the surface, whereas elsewhere up to 20 m of till has been recorded in well logs. A very thin veneer (<1 m) of coarser grained ablation till is also common in the study area and is characterized by angular, locally derived clasts and a siltyto-sandy matrix. Kames and associated ice-contact deposits are rare.

The study area is underlain primarily by the shallow to moderately north-dipping porous Horton Bluff and Wolfville formations. The precise location of the east-west-trending contact between these two units is uncertain due to a lack of surface outcrop and inconclusive well logs. To the south and up-slope of the study area, highly fractured Halifax Group with minimal primary porosity dominates (Fig. 1). These bedrock units are commonly overlain by <10 m of till, although greater thicknesses can occur at the crest of drumlins. Five significant hydrostratigraphic units are likely to affect groundwater in the Grand Pré study area.

The Halifax Group outcrops approximately 5 km south of the study area. These rocks have low primary porosity (<5%; Trescott 1968) and variable groundwater flow is dominated by secondary fracture systems controlled by the attitude of the rocks (Trescott 1968). Well yields range from 0.1–150 igpm (imperial gallons per minute) with a mean of 4.6 igpm and anomalous concentrations of Fe and Mn are common.

About 10 km south of the study area the Devonian South Mountain Batholith (SMB) forms a regional highland composed primarily of biotite granite. This highland is the primary source rock for HBF sediments (Martel and Gibling 1996). The amount of recharge through the SMB is probably not very large on a regional basis. Previous studies on recharge in this area suggest that the majority of recharge comes as downward flow through the Quaternary sediments throughout the region (Rivard et al. 2007).

Groundwater is stored in and transmitted through the SMB through surface-parallel joints that decrease in density with depth (Trescott 1968). Well yields vary from 0.2–220 igpm, with a mean of about 5.7 igpm. The pH of ground-water in the SMB granite is generally between 4.5–5.5. The SMB contains numerous uranium deposits and occurrences (Ryan and O’Beirne-Ryan 2006). These deposits are interpreted as having been formed due to fluid migration in the late stages of granitoid emplacement and occur within shear and fracture zones in altered rocks, either in the granitoids rocks or as peribatholithic occurrences in the metasedimentary country rocks of the Meguma Group (MacDonald 2001).

The most significant aquifers in the study region occur within porous sandstone and conglomerate of the Wolfville Formation and carbonaceous siltstone and sandstone of the HBF (Rivard et al. 2007). The HBF underlies the southern half of Grand Pré though its northern boundary has not been well established. These rocks characteristically have low permeability and transmit water primarily through joints, although intergranular porosity in coarser grained units accounts for some groundwater storage. Well yields from HBF sediments are highly variable and range from 0.5–1000 igpm, with a mean of 24.7 igpm. Water quality is generally moderate to poor with high concentrations of Fe, Mn, salinity, and hardness being common. Elevated concentrations of U and As have been noted in wells located near the contact with slate of the Halifax Group.

The Triassic Wolfville Formation is composed of highly permeable sandstone and conglomerate that transmit water primarily through intergranular pore space (Trescott 1968). Well yields vary from 1–450 igpm, with a mean of 77.8 igpm. Water quality is generally good (pH range from 6.0–6.5); where calcite is the dominant cement, elevated hardness has been noted. Elevated concentrations of U have been noted near the contact with slate of the Halifax Group (Ryan and O’Beirne-Ryan 2006).

Surficial deposits such as till, ice-contact driſt, and out-wash deposits are common in the Annapolis-Cornwallis Valley (Trescott 1968) but do not comprise significant aquifers in the study area. In Grand Pré a few dug wells access local or perched aquifers located in sandier facies of the Lawrencetown Till; well yields are generally less than 40 igpm. Both surface water and groundwater in the Annapolis-Cornwallis Valley ultimately discharge into the Bay of Fundy (Trescott 1968) and groundwater movement in the study area is generally northward into the Minas Basin.

Two suites of groundwater samples were collected from wells in Grand Pré. Samples were collected in acid-washed 250 ml polyethylene bottles from plumbing-access points located up-flow of any treatment or filtration devices. All well locations were recorded using a hand-held WAAS-enabled GPS receiver. Samples were stored on ice and in the dark during transportation to the laboratory.

Suite #1 includes seventeen samples gathered in August, 2006. Suite #2 was gathered in November, 2006, and contains nine samples from nine of the seventeen wells sampled initially. Re-sampled wells were chosen based on initial results from suite #1 and the requirement for further geochemical analysis. In sample suite #2, the concentration of ferrous iron (Fe²⁺) was measured immediately aſter sampling using a handheld Hach® (DR/870) colorimeter.

The pH of each sample was measured both in the field and at Acadia University immediately aſter returning from field sampling. Minimizing the time between sampling and measurement decreased the amount of gases lost and temperature changes that could alter the chemical equilibria of the water. Each sample was filtered through a 0.45 μm filter using a vacuum filtration unit. Filtering removed suspended particulate matter; most obvious of which was iron oxide (reddish-brown coating accumulated on filter surfaces). Samples that were to undergo cation analysis were acidified to a pH of 2 using concentrated nitric acid. Acidification and filtration minimized the dissolution of particulate matter and the formation of colloids in solution that could change the chemical equilibria of water samples between collection and analysis (Deutsch 1997).

Cation analysis was performed at ACME Analytical Laboratories Ltd. in British Columbia. Sample duplicates were included in both suites to monitor data quality. All samples were analyzed for the concentrations of 72 cations using an Inductively Coupled Plasma Mass Spectrometer (ICP-MS). Major anion concentrations were measured in suite #2 using ICP-MS to provide data for a more detailed geochemical analysis. Alkalinity was measured in non-acid-ified samples from suite #2 by titration using 0.02 N H₂SO₄ (with an end-point pH of 4.5).

Five of the 72 cations in the Grand Pré samples were present in concentrations that exceeded a Health Canada (2011) maximum allowable concentration (MAC) or aesthetic objective (AO) in suite #1 and again in suite #2 (excluding Pb; Table 2). These cations are U (exceeding MAC = 0.02 mg/L in 4 wells), Se (exceeding MAC = 0.01 mg/L in 4 wells), Fe (exceeding AO = 0.3 mg/L in 3 wells), Mn (exceeding AO = 0.05 mg/L in 12 wells), and Pb (exceeding MAC = 0.01 mg/L in 3 wells).

The concentrations of chloride (Cl^-), bromide (Br^-), sulphate (SO₄^2-), nitrate (NO₃^-), nitrite (NO₂^-), and phosphate (PO43-) were below the limits listed in the Guidelines for Canadian Drinking Water Quality (Health Canada 2011). All concentrations of nitrite, phosphate, and bromide (excluding sample 5D) were below the detection limits (Table 3). Alkalinity ranged from 57 to 157 mg/L CaCO₃; the majority of the samples fell within the range 130 to160 mg/L CaCO₃ (Table 3). The pH values of groundwater in Grand Pré ranged from 6.45 to 7.7.

Weathered granite from the South Mountain Batholith and arkosic arenite and shale from the Horton Group have been associated with the release of U into groundwater and may be a regional source of U in the Grand Pré area (O’Beirne-Ryan and Zentilli 2003; Kronfeld et al. 2004; Ryan and O’Beirne-Ryan 2006). Ryan and O’Beirne-Ryan (2006) noted that in the Windsor region, U enrichment in the HBF is related to reduction-oxidation fronts in the sandstone. The U is associated with carbon-rich horizons and disseminated pyrite in the sandstone and shale, the oxidation of which reduces the valence state of U to an insoluble U⁴⁺ state, causing its deposition at or near pyrite and/or organic carbon debris.

The contact between the Horton Group and the Wolfville Formation lies north of Highway 101 but is poorly constrained by both outcrop and subsurface data (Moore et al. 2000; White 2010). The erosional contact marks a significant change in stratigraphy and rock composition, and is likely to affect the chemistry and composition of the groundwater that flows through it. Interbedded carbon-rich shale and siltstone of the HBF likely produce a reduced groundwater environment, and the clean, porous sandstone (mature quartz arenite) in the Wolfville Formation produce an oxidized groundwater environment. This stratigraphic dichotomy may explain the different groundwater environments (oxidized and reduced) observed in the Grand Pré area.

The higher concentrations of U are located in the north-ernmost wells in the sample area (Fig. 5), the area most likely underlain by the Wolfville Formation ( Fig. 1). Groundwater flowing through reducing conditions in the Horton Group is characterized by low U concentrations and high Fe and Mn concentrations (e.g., samples 15 and 16). When water flows into the Wolfville Formation, oxidized conditions are encountered (Fig. 8). This change results in the precipitation of Fe and Mn and the dissolution of U (e.g., samples 11 and 12).

This model presents a paradox that is supported by the data, and is of importance to well owners. Well water that is aesthetically undesirable due to high concentrations of Fe and Mn may pose a lower health risk than water with low Fe and Mn concentrations that is more aesthetically desirable but likely has higher U concentrations.

Significant scope exists for continued investigation of groundwater quality and the occurrence of U in the Grand Pré region. The effect of well depth and seasonal variations on groundwater chemistry should be examined. Also, determination of the composition of the lithologies from which samples were derived and locating the contact between the Horton Group and the Wolfville Formation would provide further insight into the dominant influences on groundwater in the study area.

The health risks associated with the consumption of Grand Pré groundwater were not within the scope of this study. Unpredictable and poorly understood environmental influences on groundwater make predictions unrealistic. Importantly, this study has determined that residents and industrial users in the Grand Pré area will likely have to treat water with high Fe and Mn concentrations in order to obtain water that poses the lower health risks (i.e., low U concentrations).