The stratigraphy of unconsolidated sediments beneath Fredericton Junction has been examined in several investigations of the subsurface that focused primarily on groundwater exploration and bridge construction (Subsurface Surveys 1976; ABP Consultants Limited 1979; Neill and Gunter Limited 2004, 2005; Giudice 2005). Locally, the surface is blanketed by extensive till cover, occasionally reaching depths up to 38 m. The till is sporadically overlain by Late Wisconsinan ice-contact and glacifluvial sand and gravel deposits of variable thickness. With the exception of the till unit, the unconsolidated sediments are present in greatest thickness underlying the bottom of the valleys and occur rarely at higher elevations (Giudice 2005). The stratigraphy of unconsolidated deposits (Fig. 3) represents a glacial retreat sequence (Broster 1994) consisting of, from bottom to top, till, glacifluvial sand and gravel, glacilacustrine-like clay-silt, overlain by a surface unit of alluvium (Gadd 1973; Rampton et al. 1984; Thibault 1981; Giudice 2005). The clayey-silt unit overlies bedrock and is rarely found along river exposures in the area.

Examination of the clay-silt at Fredericton (Daigle 2005) and Fredericton Junction (Giudice 2005) indicates that it is regionally extensive and was deposited during a marine transgression into isostatically-depressed areas and major river valleys, at the end of the last glaciation. The deposit varies in geochemical and geotechnical properties from location to location, likely in response to variation between freshwater from melting glacier ice-fronts and saline water from the marine incursion. At the city of Fredericton the sediment properties also vary between the two environments (Daigle 2005; Giudice 2005). However, at Fredericton Junction, Giudice (2005) concluded that the local deposits are primarily lacustrine-like because of their ‘ varved' appearance, low chloride content, mineralogy, and plasticity index and liquid limit values. The term glacilacustrine is used herein because it best reflects the character of the local deposits.

The bedrock underlying the study area is of sedimentary origin and comprises the Minto Formation that consists primarily of sandstone, conglomerate, and siltstone of Carboniferous age (or younger) (ABP Consultants 1979; Clark 1962; Gadd 1973; Thibault 1981). Weathered, reddish-grey sandstone interbedded with grey shale is exposed on the north side of the North Branch Oromocto River at the Highway 101 bridge crossing in the village (Fig. 4), whereas red pebble conglomerate bedrock is exposed on the south side. The bedrock strata dip shallowly to the north, with conglomerate underlying interbedded layers of sandstone and shale. Parting along bedding planes and a near-surface zone of intensive fracturing commonly is reported from excavations of the bedrock – till interface (Broster and Burke 1990; Broster and Park 1993a, b).

The glaciation and deglaciation of southwestern New Brunswick has been well studied (e.g., Lee 1957; Gadd 1973; Thibault 1981; Rampton et al. 1984; Seaman 1989, 2004; Seaman et al. 1993; Pronk et al. 2003). All studies agree that during the Late Wisconsinan the area was glaciated by southerly to southeasterly flowing glaciers. The land was isostatically depressed under the weight of the ice, enabling the lowland areas to be flooded as ice-masses retreated to higher elevations. Seaman (2004) suggested that deglaciation in this area began approximately 14 000 years ago with regional ice stagnation and down-wasting, which formed a large subglacial drainage network, likely following topographic depressions. The combination of isostatic depressed areas and eustatic change in sea-level resulted in the near-complete submersion of southern and eastern parts of the province under marine, brackish or freshwater (Gadd 1973; Thibault 1981; Rampton et al. 1984). Seaman (2004) suggested that ice retreating northward along the Saint John River Valley may have impeded drainage resulting in the creation of an ice-dammed lake in the Fredericton – Fredericton Junction area.

Full inundation of the Saint John River Valley is believed to have occurred at about 12 000 BP (Rampton et al. 1984) with incursion of marine water into central New Brunswick forming the so-called inland Acadia Sea (Fig. 5). The maximum height of the marine incursion is not known and estimates of this level may lack precision due to the occurrence of raised deltas and strandlines that represent proglacial or glacilacustrine lakes deposited prior to marine inundation. Blankets of marine sediments are common at elevations below 15 m (Gadd 1973) along the valleys of major rivers emptying into coastal waters. However, Thibault in Seaman et al. (1993) suggested that the incursion may have reached 89 m in elevation at its maximum. Nonetheless, raised deltas and clay/silt deposits believed to be of marine origin are present up to elevations of 60 m along Kennebecasis Bay (Rampton et al. 1984). Flooding at that level would have inundated the Oromocto River Valley as well, regardless of the possibility of small differences in isostatic rebound between areas. Although the exact elevation of the marine incursion in the study area remains unknown, Seaman et al. (1993) suggested that at about 10 500 BP sea level was near the present level due to isostatic rebound of the interior. Thus, samples of relict saline water from this time could be expected to return radiogenic dates near this age.

Three aquifers have been interpreted (Giudice 2005) to underlie the Fredericton Junction study area, informally named the "Upper", "Middle" and "Lower" Aquifers (Fig. 3). The Lower (supply) Aquifer is fractured bedrock consisting of interbedded sandstone and shale of the Minto Formation. It is separated from the Middle Aquifer by a till unit, herein referred to as the lower aquitard. The lower aquitard is a reddish-brown lodgement till layer that tends to be dominantly clayey or silty in composition. The till forms a dense layer at depth and has been found to occur up to 38 m in thickness (Giudice 2005). It is consistently present in the study area and is relatively impermeable, allowing it to hydraulically and stratigraphically confine the lower aquifer and provide protection from surface contamination. The Middle Aquifer consists of a confined, glacifluvial, sand and gravel unit that has been previously reported to occur in thicknesses of up to 12 m in the study area (Giudice 2005). If these deposits can be found in sufficient quantities, such as a buried bedrock valley, they would provide a suitable target for the extraction of potable water.

The Middle Aquifer is separated from the Upper Aquifer by a lacustrine clay-silt unit, herein referred to as the upper aquitard. This unit is not consistently present in the area and has been found in the study region only at elevations below 20 m above sea level (Rampton et al. 1984; Daigle 2005; Giudice 2005). It therefore acts as a semi-confining layer to the underlying Middle Aquifer and has been documented by Giudice (2005) to occur locally at thicknesses of up to 27 m. The Upper Aquifer consists of unconfined, unconsolidated post-glacial sands and gravels. Given the range of thickness of this unit in the study area (1 to 10 m) (Giudice 2005), it is conceivable that shallow wells could draw water from this aquifer. However, the Upper Aquifer is the unit most prone to contamination from surface sources. It should be noted that this unit has been excavated extensively in the area for use as aggregate (Hamilton and Carroll 1975; Thibault 1981).

Several municipal groundwater exploration programs have occurred and are in progress in Fredericton Junction (ABP Consultants Limited 1979; Godfrey and Associates Limited 1996; Neill and Gunter Limited 2004, 2005). Two of the municipal production wells, Well 1 and Well 2 (Fig. 2), were drilled as part of the 1979 groundwater exploration program. The wells are approximately 30 m apart, the aquifer characteristics are similar in each well, and they are hydraulically connected to each other (Jacques Whitford Environment Limited 2003). Freeze and Cherry (1979) considered transmissivity (ability of the aquifer to transmit water) of 0.015 m²/s or greater as being representative of a good aquifer for water well exploitation. By using theoretical values for hydraulic conductivity for a sandstone (range of 10^-10 to 10^-6 m/s) presented by Freeze and Cherry (1979; Table 2.2), theoretical transmissivity values for Well 1 and Well 2 with comparable aquifer thicknesses of 30 m (Well 1) and 40 m (Well 2), range from 3 x 10^-8 to 0.03 m²/s for Well 1 and from 4 x 10^-8 to 0.04 m²/s for Well 2. Measured transmissivities for Well 1 and Well 2 varied from 0.00037 to 0.00043 m²/s (32 to 37 m²/d), which are within the above theoretical ranges (ABP Consultants Limited 1979; Jacques Whitford Environment Limited 2003). Specific capacity (an expression of the productivity of a well), was reported for both wells as 42.9 m³/d per meter of water level drawdown (ABP Consultants Limited 1979). LeBlanc and Lloy (2003) noted that Well 1 operates as stand-by well at a pumping rate of 40 igpm while Well 2, the main well, operates at 45 igpm and is pumped for 7 to 15 hours per day (average of 11 hours per day).

Well 3 is an emergency stand-by well operated at 12 igpm and is situated adjacent to the Village's water tower, approximately 1 km northwest of Wells 1 and 2. A pump test report was not located for this production well. However, given that the operational yield of the well is significantly lower than the 4 to 45 igpm pumping rates of the other wells and that Well 3 is used only in emergencies, it can be inferred from this low yield that the Well 3 is located in a less transmissive area of the Lower Aquifer. This low transmissivity is likely due to fewer fractures at that location and demonstrates the importance of identifying areas of dense fracturing prior to conducting exploratory water resource drilling in bedrock.

Within the village proper, approximately 40% of the population obtains its water supply from private wells. A search of the New Brunswick Department of Environment's Domestic Water Well Database (2003) found two well logs within the village proper, which have been named DW1 and DW47 for the purpose of this study (Fig. 2). Both wells draw their water from fracture zones in the Lower Aquifer at depths of 25.6 m (DW1; 6.6 m below mean sea level) and 56.1 m (DW47; 41 m below mean sea level). Estimated well yields, as indicated on the well driller logs, are 30 igpm and 12 igpm in DW1 and DW47, respectively. This lower well yield with greater depth is likely due to the decreasing frequency of fractures, which may reflect the decrease in effect of surface processes (i.e., weathering and glaciation) with depth.

Occurrences of brackish water at depth have been documented for many areas of central New Brunswick. For example, several testholes drilled for various exploration purposes within the Carboniferous strata encountered brackish water at depths (Table 1) ranging from 58.97 m to 256.5 m below sea level at sites within 75 km of the coast (ABP Consultants Limited 1979; Neill and Gunter Limited 2005; Webb 1981). A review of borehole logs, geophysical logs, and water analysis data in the area of Fredericton Junction indicates that deeper aquifers are commonly found to have a high saline content. This situation was previously examined by Roy and Elliott (1980) who used combined resistivity and induced polarization soundings to delineate saline water and fresh water zones in the municipalities of Fredericton Junction, Tracy, and Rusagonis (Fig.1). Their studies confirmed the existence of underlying pockets of saline water containing from 2000 to 4000 ppm NaCl and demonstrated the application of this geophysical technique as a tool for the identification of saline groundwater, but offered no explanation for the origin for the saline water (Roy and Elliott 1980).

Four possible explanations for the origin of this brackish water: (1) road salt contamination, (2) dissolution of salt-rich bedrock, (3) saltwater intrusion from underlying saline ground-water linked to coastal marine source, and (4) relict seawater trapped from the late glacial marine incursion. The occurrence of bromide normally precludes the possibility that the brackish water results from road salt contamination as bromide is not typically part of the chemical composition for road salt (Snow et al. 1990). In those studies that were carried out (Table 1), Cl/Br ratios are closer to that typically found for seawater (in the vicinity of 289) than that commonly found for rock salt (>1200), as suggested by Snow et al. (1990). While road salt run-off could migrate into the North Branch Oromocto River, it would be further diluted by river water and infiltration before arriving at the supply wells. Local river levels, approximately 10 m above sea level, are well above the depths of saltwater occurrences at Fredericton Junction, indicating that migration of road salt to the aquifer from river recharge is an unlikely source of brackish water at present.

Webb (1981) suggested that pockets of saltwater were retained in the formation upon isostatic uplift following the last deglaciation. This conclusion was reached after examining and discounting, (a) possible sources of chloride from subsurface evaporates in Pennsylvanian strata, (b) subsurface evaporates in Mississipian strata, and (c) complex weathering reactions within Pennsylvanian strata.

Table 1 shows chemical and geophysical parameters that are typically associated with saltwater occurrences. The more elevated the parameter, the higher the influence of saltwater in that particular area of the formation. Examination of test borehole sampling depths and related end of borehole depths are quite variable. The groundwater sample depths range from 58.97 m to 256.5 m below sea level. If the source of the brackish water is from the underlying freshwater/saltwater interface, one would expect a much closer agreement in depth with the occurrence of saltwater (assuming a gradual landward-dip in the ‘ saltwater table'). In addition, the concentrations in Table 1 are more representative of brackish water than seawater and are probably derived from an interface zone of mixing. The wide variance in depths of brackish water would indicate either a freshwater/saltwater interface mixing zone of variable thickness, or the occurrence of discontinuous ‘ pockets' or ‘ islands' of saline-rich groundwater. Such saline-rich pockets could be the result of trapped accumulations from the Acadia Sea incursion. Mixing could also occur at the boundaries of saline pockets, thus both situations are feasible and could likely be distinguished by hydrogeological mapping and finite dating.

Gemtec Limited recently obtained a sample of saline water from a depth of approximately 45 m in fractured sandstone (Gemtec Limited, personal communication, 2006). The sample was recovered during drilling at Youngs Cove located approximately 65 km northeast of Fredericton Junction and 80 km north of the Bay of Fundy (see insert Fig. 1). Radiocarbon dating of the sample by the Isotrace Radiocarbon Laboratory at the University of Toronto returned a date of 7400 BP ± 70 years (Gemtec Limited, personal communication, 2006). The date is less than the expected age of 10 000 – 12 000 BP and is likely the result of mixing with younger formational water.

Two of the test holes from the 1979 drill program (TH4-79 and TH8-79; Fig. 2) as well as the 2005 test well (TH05-1; Fig. 2) drilled as part of the current groundwater exploration program intersected saline groundwater at depths (Jacques Whitford Environment Limited 2003). Brackish water in these wells occurs at depths of 58.97 m (TH4-79), 67.56 m (TH8-79) and 63.34 m (TH05-1) below sea level. The approximate level of the North Branch Oromocto River in the Fredericton Junction area is 10 m above sea level, or about 69 m higher in elevation than the reported findings of brackish water. Surface water quality monitoring from 1969 to 2002 at a sampling site on the North Branch of the Oromocto River at Tracy (Fig. 1) shows conductivity measurements in the order of 50 µSIE/cm (New Brunswick Department of Environment, unpublished data). Based on the low values of measured conductivity, it is reasonable to state that the North Branch Oromocto River is representative of a freshwater river system.

The occurrence of brackish groundwater at locations within a few 10's of kilometers from a sea coast as a result of saltwater intrusion is not unexpected. However, occurrences of saline groundwater at locations greater than 50 km from a sea coast, while common in the Carboniferous bedrock of New Brunswick (Webb 1981), may be due in part to pockets of remnant late glacial seawater. After consideration of the previous studies of glacial history and surrounding geology, we are inclined to agree with the conclusion of Webb (1981) that the saline water is a relict of glacial marine incursion. Although the maximum elevation remains unknown, Rampton et al. (1984) suggest that the area could have been inundated during the DeGeer Sea (about 12 400 BP) and Acadia Sea (about 12 000 BP), and submerged under both fresh and marine waters (Fig. 5). From examination of the mineralogy and geochemistry, Giudice (2005) concludeed that much of the sediment was deposited under fresh water conditions, possibly due to melting of surrounding ice masses at higher elevations.

Grant (in Seaman et al. 1993) suggested that after 14 000 BP isostatic uplift began to exceed the rate of eustatic sea level rise. By 10 500 BP relative sea level may have been close to the present mean sea level and it continued to fall until about 7000 BP when it reached an elevation approximately 37 m below present mean sea level (Seaman et al. 1993). The inland sea may have existed for a period of less than 2500 years and possibly for less time at locations experiencing rapid isostatic rebound. Yet this time would have been likely sufficient to enable sea water to displace any fresh water in the underlying bedrock. Webb (1981) suggested that the significant density contrast between seawater and freshwater, as well as the greater volume and greater hydrostatic saline groundwater head would have contributed to the freshwater displacement. Webb (1981) also noted that most of the brackish wells in New Brunswick are located in topographic depressions and are particularly common along river valleys overlying Carboniferous bedrock and emptying into coastal waters. These locations would form natural topographic traps for relict seawater.

It is likely that some expansion of the sedimentary bedrock porosity occurred during deglaciation. With removal of the glacier load and isostatic rebound, the underlying strata would fracture, opening new voids (cf. Broster and Burke 1990; Park and Broster 1996). This development of cavities would have increased the permeability of the bedrock and may have facilitated absorbency. Marine water entering these depressed basins, likely saturated the underlying rock at that time and/or created brackish water by mixing with freshwater from surrounding melting glaciers.