Atlantic Climate Adaptation Solutions Association Solutions d'adaptation aux changements climatiques pour l'Atlantique

A case study of coastal aquifers near Richibucto, : Saline groundwater occurrence and potential impacts of climate change on seawater intrusion

By K. MacQuarriea, K. Butlerb, E. Mottb, and N. Greena

a Department of Civil Engineering b Department of Earth Sciences University of New Brunswick PO Box 4400, , NB E3B 5A3

October 2012

Report prepared by: K. MacQuarrie, K. Butler, E. Mott, and N. Green, and commissioned by the Atlantic Climate Solutions Association (ACASA), a non-profit organization formed to coordinate project management and planning for climate change adaptation initiatives in , New Brunswick, and Newfoundland and Labrador and supported through the Regional Adaptation Collaborative, a joint undertaking between the Atlantic provinces, Natural Resources and regional municipalities and other partners.

Project management: Climate Change Secretariat, New Brunswick Department of Environment. Report edited by the New Brunswick Department of Environment. P.O. Box 6000, Fredericton, NB, E3B 5H1. E-Mail: [email protected]

Disclaimer: This publication is not be used without permission, and any unauthorized use is strictly prohibited. ACASA, the authors, the provinces of Nova Scotia, New Brunswick, Prince Edward Island, Newfoundland and Labrador, and the Regional Adaptation Collaborative are not responsible for any unauthorized use that may be made of the information contained therein. The opinions expressed in this publication do not necessarily reflect those of ACASA, its associated provinces, or other partners of the Regional Adaptation Collaborative.

This report is also available for download from the ACASA website at: www.atlanticadaptation.ca

Cover photo: Drilling of borehole UNB2 near Richibucto in August, 2011. Photo by K. Butler.

SUMMARY

A case study was conducted between 2010 and 2012 to investigate saline groundwater occurrence and the potential impacts of future climate change and sea level rise on seawater intrusion into sandstone aquifers near the Town of Richibucto in eastern New Brunswick. The region investigated was the peninsula-like area of land bounded by the tidal Mill Creek, St. Charles River, and the /Harbour. The focus of the study was on water supplies obtained from municipal production wells. The primary field activities included geophysical surveys conducted at the ground surface, drilling of three boreholes, geophysical logging of boreholes, groundwater sampling at selected locations, and groundwater level monitoring. A three-dimensional numerical model of density-dependent groundwater flow coupled with solute transport was developed for the Richibucto region and this was then used to investigate the effect of future climate change and sea level on the distribution of groundwater salinity within the sandstone aquifers.

The information obtained from the geophysical investigations and borehole drilling generally supported preexisting hydrogeological information, suggesting that the main aquifers in the area are: a) a shallow unconfined sandstone unit that extends to depths of approximately 18 to 22 m, and b) a deeper sandstone aquifer that extends to depths of approximately 65 to 70 m. Electrical resistivity tomography survey results, constrained by geophysical logging and laboratory measurements on rock core, indicated that saline water was present in the shallow aquifer at locations near the Richibucto River/Harbour. Evidence for saline water occurrence was also obtained in at least three locations within the deeper sandstone units, including one location beneath a municipal well. Observed modes of saline water intrusion included both lateral intrusion and vertical movement though both sandstone and shale units under the influence of pumping (upconing).

The results from the numerical modelling investigation indicated that climate change (i.e. changes in groundwater recharge), sea level rise, and increased pumping were all significant to some extent in the context of lateral seawater intrusion in shallow to intermediate depth aquifers similar to those of the Richibucto region. However, sea level rise had the least significant effect of the three factors considered. The limitations imposed by sparse hydrogeological data, combined with the inherent uncertainties in predicting future climate conditions, mean that the simulation results are best suited for investigating the general sensitivity of the groundwater system to seawater intrusion, rather than predicting local conditions near individual pumping wells.

Seven recommendations are made based on this study; these deal with the methods used to investigate the occurrence of saline groundwater, locating and operating municipal wells, and general groundwater conservation.

TABLE OF CONTENTS

Page

1.0 Introduction 1

2.0 Study Area and Well Field Description 3

3.0 Investigation Methods 5

4.0 Key Results and Findings 8

5.0 Study Limitations 13

6.0 Recommendations 14

7.0 Acknowledgements 15

8.0 References 16

Appendix A: Salt water intrusion in a coastal sandstone aquifer at Richibucto, New Brunswick as revealed by geophysical and borehole surveys

Appendix B: Groundwater quality and hydraulic head data presentation and interpretation

Appendix C: An evaluation of the effects of climate change on seawater intrusion in coastal aquifers: A case study from Richibucto, New Brunswick

LIST OF FIGURES

Page

Figure 1. Regional map showing the location of the Town of Richibucto in eastern New Brunswick 2

Figure 2. Map of the study region, which was bounded by the tidal St. Charles River and the Richibucto River/Harbour 3

Figure 3. Aerial photograph of the northern part of the Town of Richibucto, showing the location of the seven ERT survey lines relative to the three municipal pumping wells, pre-existing monitoring well MW08-03, and boreholes UNB1 and UNB2 6

Figure 4. ERT section, 700 m long, showing inferred variations in electrical resistivity beneath Line 1 extending westward from the coast as shown in Figure 3 9

Figure 5. ERT section, 800 m long, showing inferred variations in electrical resistivity beneath Line 4 in Richibucto well field 10

Figure 6. Vertical section near the well field and Richibucto Harbour showing the simulated initial total dissolved solids concentration used for all odel simulations 11

LIST OF TABLES

Page

Table 1. Summary of primary field activities undertaken during the study 5 1

1.0 INTRODUCTION

Salt water, or saline water, is present in the ocean and in coastal bays, estuaries, and tidal rivers. This water is denser than fresh water because it contains relatively high concentrations of dissolved solids, especially sodium (Na) and chloride (Cl). Saline water naturally extends inland beneath the ground some distance from the coast; however, intrusion of saline water and mixing with fresh groundwater can be worsened by pumping large quantities of groundwater from water supply wells. Removing groundwater from coastal aquifers reduces the amount of fresh water that would otherwise discharge along the coastline, allowing the seawater to move further inland. In the future, changes in precipitation and temperature, and the rise in sea level that is forecast for Atlantic Canada, may also contribute to increased seawater intrusion.

The intrusion of seawater into water supply aquifers is a concern because of the associated deterioration in groundwater quality. It does not take much seawater to affect groundwater quality – for example, mixing about 2% seawater with fresh water would be sufficient to elevate dissolved chloride above the recommended maximum concentration (250 mg/L) for drinking water. Seawater also contains low concentrations of dissolved bromide, which may be a concern if the groundwater is disinfected using chlorination.

This report presents an overview of a project that was conducted to investigate saline groundwater occurrence and the potential impacts of future climate change and sea level rise on seawater intrusion into sandstone aquifers near the Town of Richibucto in eastern New Brunswick (Figure 1). The Richibucto region was selected for this study following consultation with the New Brunswick Department of Environment and Local Government, and the Town of Richibucto. In the past, several town wells have produced water with elevated chloride concentrations that were believed to be caused by saline water intrusion. The study area was also chosen because it was considered to be representative of hydrogeological conditions along much of the coastline of New Brunswick.

The project had two primary objectives: 1) to define the current hydrogeological conditions in the area, including the occurrence of saline groundwater in aquifers near the town water supply wells, and 2) to develop a conceptual hydrogeological model and related numerical model to assess the possible changes in lateral seawater intrusion caused by future climate change, sea level rise, and increased groundwater pumping. The project did not assess the potential for, or implications of, developing new pumping wells outside of the current well field area, nor did it address the potential for seawater intrusion in privately-owned wells within the study region. Other groundwater quality issues, such as contamination from human or natural sources, were not considered. Although the expected changes in groundwater recharge up to the year 2100 were evaluated, it was beyond the scope of this project to assess specific land use impacts (e.g. increased commercial development, peat harvesting) on groundwater recharge.

2

Figure 1. Regional map showing the location of the Town of Richibucto in eastern New Brunswick.

The geographic scope of the project was primarily limited to the peninsula-like area of land bounded by the tidal Mill Creek, St. Charles River, and the Richibucto River/Harbour (Figure 2). In general, the depth of subsurface investigations was limited to less than 100 m. The project began in May 2010 and was concluded in September 2012.

This report is intended to present a summary of the investigation methods, key results and findings, limitations, and recommendations of the study. Supporting technical information is presented in the three Appendices, which form essentially self-contained reports. In addition, several other technical reports and two theses have been produced during this project, including LeBlanc et al. (2012), Green (2012), and Mott (in preparation). Selected results have also been presented at regional and international conferences (e.g. Green et al. 2011; Green and MacQuarrie 2012; Mott et al. 2011a; Mott et al. 2011b; Mott and Butler 2012).

3

Figure 2. Map of the study region, which was bounded by the tidal St. Charles River and the Richibucto River/Harbour. The map shows the location of selected boreholes and monitoring wells (green dots) and the three active municipal pumping wells (blue symbols labeled PW1, PW3 and PW4). The red lines indicate three of the lateral boundaries for the numerical groundwater flow model that was developed for the area; in the northeast direction the model extended into the Northumberland Strait.

2.0 STUDY AREA AND WELL FIELD DESCRIPTION

A prominent land surface feature in the study region is the St. Charles Peat Bog/Plain (Figure 2). According to the New Brunswick Department of Natural Resources (NBDNR 1993), this deposit consists mainly of poorly humified (i.e. poorly decomposed) peat of Sphagnum genus. The total 4

surface area of the peat deposit is approximately is 2265 ha, with the thickness varying from 0 m on the fringes to about 7.6 m in the central portion of the deposit.

Covering much of the remainder of the study region, and presumably underlying the St. Charles Peat, is a poorly sorted glacial till up to 3 m in thickness. Despite its often sandy composition, the poorly sorted nature of this sediment excludes its use as an aquifer and there are no records of water wells being completed in the surficial sediments in the Richibucto region. However, rainfall or snow melt must infiltrate through this surficial till before recharging deeper aquifers.

The bedrock geology in the region is Richibucto Formation of the Pictou Group, late Carboniferous in age and part of the post-accretionary intermontane Maritimes Basin (St. Peter 1993). The Richibucto Formation is comprised of multilayered grey and lesser reddish brown sandstone, pebbly sandstones, and pebble conglomerates separated by reddish brown, very fine to fine grained sandstones and mudstones. It is the coarser-grained sandstone units that constitute the main aquifers in the region (Water Management Services 1986; Maritime Groundwater Inc. 1991, 1992, 2004; Stantec Limited 2009); these units may also contain fractures that contribute significantly to groundwater flow.

Historically, private wells were used to provide water for domestic, commercial, industrial and fire protection purposes in the vicinity of the Town of Richibucto. In 1986, a need was identified to provide the town with a higher quality, centralized water supply and the first groundwater exploration boreholes were drilled; two of these exploratory holes were later developed as production wells (PW1, Figure 2; and PW2). Two sandstone intervals were identified at these locations: an unconfined sandstone aquifer above a layer of mudstone/shale, and a deeper sandstone aquifer beneath the mudstone/shale. PW1 and PW2 were both completed in the deeper sandstone aquifer with total depths of 51 m and 67 m, respectively. Over time as demand on PW1 and PW2 increased, it became apparent that salinity was increasing in the water produced from the well field. In 2004, a new pumping well (PW3, Figure 2) was constructed near PW2; however, unlike previous wells, the new well was completed in the upper unconfined aquifer (Maritime Groundwater Inc. 2004). The shallower depth of PW3 (21 m) proved to be successful at avoiding elevated chloride levels, but the extracted water had to be monitored and treated for elevated levels of iron and manganese. PW2 was decommissioned after PW3 was brought into operation.

In 2008, work was performed to install another pumping well (PW4, Figure 2) approximately 300 m inland from PW1 and PW3, with the intent that distributing the groundwater extraction over a greater number of wells would minimize the potential for saline water intrusion. The same sequence of stratigraphy was intersected at this location (Stantec Limited 2009), and PW4 was completed to a depth of 49 m in the lower sandstone aquifer. To date, the location and operation of this pumping well has been effective at avoiding elevated chloride concentrations (Sean Sullivan, Town of Richibucto, pers. comm.). 5

3.0 INVESTIGATION METHODS

The study included field investigations, data analysis, and numerical simulations of future conditions. A review of existing information was also preformed; this information included groundwater exploration studies undertaken for the town’s water supply, geological studies, and aggregated information from the New Brunswick Department of Environment’s domestic water well database.

3.1 Field Activities The primary field activities included geophysical surveys conducted at the ground surface, drilling of three boreholes, geophysical logging of boreholes, groundwater sampling at selected locations, and groundwater level monitoring. Table 1 summarizes the main field activities.

Table 1. Summary of primary field activities undertaken during the study.

Activity Primary purpose Total amount

Ground-based Identification of the location(s) of 4.3 km of ERT data collected on geophysical saline groundwater seven lines surveys

Geophysical Geophysical properties of rock 370 m of geophysical logs obtained logging in units and pore fluids; in seven different boreholes (270 m boreholes identification of fractures in open holes; 100 m in cased holes)

Borehole drilling Confirmation of ERT survey Three boreholes drilled with a results; collection of rock core, combined depth of 176 m; acquired groundwater samples, and ~160 m of rock core groundwater level data

Groundwater Chemical analysis of groundwater One set of samples from 10 sampling and assessment of saline water boreholes and monitoring wells, occurrence some sampled at multiple depths

Groundwater level Determine hydraulic response of 10 boreholes and monitoring wells – and temperature shallow and deep aquifers to frequency and duration of monitoring monitoring pumping, tides, or recharge varied 6

The geophysical surveys consisted of electrical resistivity tomography (ERT) measurements made along seven lines as shown in Figure 3. The objective of the ERT surveys was to determine if the spatial distribution of saline groundwater could be identified from resistivity variations in the sandstone aquifer units and used to infer any preferential pathways or mechanisms of intrusion near the town’s well field. Ground penetrating radar (GPR) surveys were also tested but found to offer insufficient depth of investigation for aquifer characterization. Geophysical logging in open boreholes was used to determine resistivities of shale and sandstone layers, identify layers having saline water, and to confirm bedrock lithologies and identify fractures. Additional details regarding the geophysical investigation methods are provided in Appendix A.

Figure 3. Aerial photograph of the northern part of the Town of Richibucto, showing the location of the seven ERT survey lines relative to the three municipal pumping wells (black crosses with numbers indicating the production well ID), pre-existing monitoring well MW08-03, and boreholes UNB1 and UNB2.

Three boreholes (UNB1, UNB2 and UNB3; Figures 2 and 3) were drilled during the project to verify the ERT survey results, and to permit the collection of groundwater samples and water level data in regions that had not been previously investigated. Details related to the drilling and 7

hydrogeological data collected from these boreholes are provided by LeBlanc et al. (2012). Rock cores were also collected during drilling and these were logged in the field and stored at the University of New Brunswick (UNB) for later analysis; core from UNB1 and UNB2 were subsequently logged in more detail and tested (at 1 m depth intervals) to determine porosities and resistivity-salinity relationships.

Groundwater sampling and groundwater level and temperature monitoring were conducted at selected locations to determine the groundwater chemistry in the study area, including the identification of locations that were affected by mixing with seawater, and to assess the groundwater level response to municipal well pumping, precipitation (recharge), and tidal variations. All groundwater sampling and monitoring activities were undertaken in boreholes or monitoring wells; several samples were also collected from the Richibucto River/Harbour to serve as a reference for local surface waters. Appendix B provides more details on the locations and methods employed for obtaining water samples and groundwater level data.

3.2 Data Analysis and Numerical Modelling As discussed in Appendix A, the two-dimensional resistivity inversion program RES2DINV (Loke and Barker, 1995) was used to generate models or ‘sections’ of inferred resistivity variations in the subsurface beneath each ERT line to depths of up to 100 m. The methods used to relate the geophysical measurements of sandstone resistivity to the salinity of groundwater in the pore space of the rocks are also discussed in Appendix A and by Mott (in preparation).

The groundwater chemistry data were interpreted using Piper major-ion and chloride:bromide plots as discussed in Appendix B. Tritium results were used to infer the approximate dates of groundwater recharge.

The review of existing hydrogeological information and compilation of data obtained from the current project were used to develop a conceptual model of groundwater flow in the Richibucto area aquifers. This conceptual model is presented in Appendix C and by Green (2012), and provided the basis for formulating a three-dimensional numerical model of density-dependent groundwater flow coupled with solute transport. Although several numerical models are available to simulate density-dependent groundwater flow and solute transport, the SEAWAT code (Guo and Langevin 2002; Langevin et al. 2003; Langevin et al. 2008) was selected for this work because it is well tested and has been frequently applied to investigate seawater intrusion in coastal aquifers. The numerical model incorporated local topography, bathymetry of the surrounding tidal rivers and Northumberland Strait, stratigraphy from borehole and geophysical investigations, and well field characteristics. The spatial extent of the model is shown in Figure 2. Based on predictions of climate change for the Richibucto area (Environment and Sustainable Development Research Centre 2011) two scenarios for future variations in groundwater recharge were simulated. Two scenarios for future sea level rise and one scenario for future pumping were also simulated (see Appendix C). 8

4.0 KEY RESULTS AND FINDINGS

4.1 Existing Conditions The information obtained from the geophysical investigations and borehole drilling generally supported the preexisting hydrogeological information, suggesting that the main aquifers in the area are: a) a shallow unconfined sandstone unit that extends to depths of approximately 18 to 22 m, and b) a deeper sandstone aquifer, including some shale layers less readily correlated from hole to hole, that extends to depths of approximately 65 to 70 m. These two aquifers are separated by an intervening shale unit that has been identified in many of the boreholes drilled in the well field area and at the more distant UNB2 and UNB3 locations (Figure 2); however, as shown in Figure C2 (Appendix C), the lateral continuity of this shale unit over the entire study area is not well defined. The upper sandstone aquifer is unconfined, as indicated by rapid water table responses to precipitation (Figure B2, Appendix B), while the deeper sandstone aquifer does not display any notable response to precipitation events (LeBlanc et al. 2012). It therefore appears that, where present, the uppermost shale unit is effective at isolating the two aquifers. At a depth of approximately 68 m at borehole UNB1 (Figure 3), another relatively thick shale layer was also encountered (Appendix A).

Saline groundwater was not detected at the two available shallow aquifer monitoring well locations (Appendix C); however, the ERT survey results from Line 1 (Figure 4) show that saline water was present in the shallow aquifer near the shoreline of Richibucto Harbour. Given that this area is relatively low-lying (ground elevation less than approximately 1.5 m above CGVD28), it may be periodically inundated with salt water during storm events, allowing saline water to infiltrate into the shallow groundwater. The results from ERT Line 3 (L3 in Figure 3) also revealed a localized low resistivity zone in the shallow aquifer, indicating that saline water was present near monitoring well MW08-03. This finding suggests potential lateral migration of saline water from a tidal creek located approximately 200 m to the northwest of MW08-03 (Mott, in preparation). Low resistivities, indicating the presence of relatively saline water, were not detected at similar depths in any of the other ERT survey lines (Mott, in preparation). It appears that saline water intrusion in the upper aquifer was not extensive at the time of the ERT surveys, although it should be noted that no data were collected for the region to the southeast of the well field (i.e. towards the tidal Mooneys Creek, Figure 2).

9

Figure 4. ERT section, 700 m long, showing inferred variations in electrical resistivity beneath Line 1 extending westward from the coast as shown in Figure 3. Low resistivities (blue colours) at shallow depths near the coast are interpreted to represent salt water mixing in the upper sandstone aquifer as a consequence of overland inundation during storms. The low resistivity pattern at depth is suggestive of a salt water wedge intruding laterally into the lower sandstone aquifer. The interpretation is supported by hydrogeological, water salinity, and geophysical well logging data from borehole UNB2, shown as a superimposed geological log (yellow representing sandstone and brown representing shale).

At the time of the field investigations, several lines of evidence indicated that saline water was present in the deeper sandstone aquifer (i.e. between depths of approximately 20 to 70 m). First, as discussed in Appendix B, the groundwater chemistry results for major ions and bromide suggested that mixing of fresh groundwater and seawater had occurred in the deeper aquifer at the two locations investigated within 500 m of the shoreline of the Richibucto Harbour (i.e. MW08-03 and UNB2; Figure 3). These two monitoring locations also exhibited groundwater level and specific conductance responses that indicated a connection with tidal surface water (Appendix B). Furthermore, the ERT results presented in Figure 4 show a wedge-like zone of low resistivity extending inland from the coast, underlying more resistive formations above. The resistivity structure does not correlate with the lithological variations observed in borehole UNB2 particularly well, but is instead suggestive of a salt water wedge intruding laterally into the lower sandstones. The maximum measured chloride concentration in water collected from UNB2 was 1430 mg/L; however, specific conductance monitoring results from this borehole showed that specific conductance, and by inference chloride concentration, varied with tidal water levels (Appendix B).

Within the well field, results from geophysical logging of borehole UNB1 suggested the presence of elevated salinity in groundwater below the relatively thick layer of shale and sandy shale intersected between 67.5 and 77 m depth (Figure A8 of Appendix A). Two estimates of in-situ groundwater salinity below that layer, deduced from the resistivity log and laboratory measurements on core samples, ranged from 410 to 880 mg/L chloride assuming that NaCl was the dominant solute (Appendix A; Mott, in preparation). These salinities are an order of magnitude higher than estimates deduced from the resistivity log above the deep shale layer, 10

which were comparable to the 10 – 50 mg/L chloride found in most waters sampled directly from boreholes in the well field. Although water samples taken from the well bore of UNB1 did not indicate mixing with seawater, as discussed in Appendix B, it is quite likely that the water samples collected from this long open borehole were biased toward fresh groundwater present in the shallower sandstone aquifers.

Figure 5 shows the ERT section for Line 4, which passed within 20 m of borehole UNB1 and production well PW1 (Figure 3). This ERT section suggests that higher salinity groundwater was present beneath the production well. These results imply upward movement of saline water (i.e. upconing) through the lower shale unit, rather than lateral seawater migration through the sandstone aquifer in which PW1 is completed. Movement of water through the lower shale layer may occur along moderately dipping fractures such as those identified in various shale units by optical televiewer logging (Appendix A; Mott, in preparation). Mixing of saline waters originating below the deep shale layer with fresh waters in the aquifer above is a possible explanation for the elevated chloride contents of approximately 100 mg/L found in water produced from PW1 during the period 1999-2001, when this well was pumped at a higher rate (Figure A6, Appendix A).

Figure 5. ERT section, 800 m long, showing inferred variations in electrical resistivity beneath Line 4 in Richibucto well field (see Figure 3 for location). Low resistivities, indicated by cool colours reveal the locations of thicker layers of shale and/or sandstones saturated with more saline water. Geological logs for four boreholes, including UNB1 and PW1 are superimposed (yellow representing sandstone and brown representing shale). The resistivity structure is predominantly layered, in accordance with the flat- lying geology, except below PW1 where low resistivities are interpreted to represent upconing of saline water from depth.

The ERT section for Line 2, which passed within 90 m of PW3 and PW4 (Figure 3), showed slightly lower resistivities in the lower aquifer near PW3; although this could be indicative of 11

elevated salinity, the results were less definitive than for other locations (Mott, in preparation). As well, there was no evidence of the mixing of seawater with groundwater in the samples obtained from locations inland (southwest) of the current municipal well field (Appendix B). Samples from the deeper sandstone units at UNB3, TH3 and TH4 (Figure 2) had TDS concentrations generally around 400 mg L-1; however, they had low chloride (< 4 mg/L), an absence of bromide, and relatively low levels of tritium.

4.2 Numerical Simulation of Climate Change and Sea Level Rise The climate change, pumping, and sea level rise scenarios investigated with the numerical model are presented in detail in Appendix C, and in Green (2012). A two-dimensional depiction of the initial distribution of TDS in the model domain, which was used as the baseline condition to assess the changes to the groundwater system, is shown in Figure 6. Intrusion of seawater from the Northumberland Strait is simulated to produce a wedge of high TDS (i.e. saline) water that extends beneath Kouchibouguac National Park and the Richibucto Harbour (Figure 6). The Richibucto River/Harbour also contributes saline water to the aquifer, but the results presented in Figure 6 show that saline water from these sources is isolated to shallow zones directly beneath and adjacent to these bodies of water (Green 2012).

Richibucto Kouchibouguac National Park Harbour

Figure 6. Vertical section near the well field and Richibucto Harbour showing the simulated initial total dissolved solids concentration (TDS in mg L-1) used for all model simulations. Shale layers are shaded grey and labelled C2, C3 and C4. The open intervals of wells are shown by dotted lines. Note that UNB2 appears to be in the harbour because of the projection onto the cross-section. The source of the shallow zone of elevated TDS above shale layer C2 is the Richibucto Harbour, while the deeper salt water wedge represents intrusion from the Northumberland Strait. Vertical exaggeration is 20 times.

The data limitations in the current conceptual and numerical models, some of which are discussed in the following section, imply that the results obtained are unlikely to accurately 12

depict conditions at all locations in the aquifer. This means that the simulation results are best suited for investigating the general response of the aquifers.

The results from the numerical modelling investigation indicated that climate change (i.e. changes in groundwater recharge), sea level rise, and increased pumping were all significant to some extent in the context of seawater intrusion. However, the relative importance of these three factors changed depending on the specific location considered (e.g. locations “007C”, “mid”, “out” or “toe”, Figure 6). The effect of variations in groundwater recharge was greatest at shallow to intermediate depths (i.e. less than 60 m below sea level) in the aquifer, which are the depths typically targeted for groundwater extraction (Maritime Groundwater Inc. 1992; Stantec Limited 2009). The effect of the pumping rates was most significant relatively close (< ~ 2 km) to the well field (i.e. “007C”, Figure 6). Sea level rise appeared to be a significant factor only near the toe of the salt water wedge that is associated with intrusion from the Northumberland Strait. Given that a major objective is minimization of saline water intrusion into well fields, these findings suggest that pumping rates must be carefully managed in the future.

The numerical simulation results from this study are generally consistent with other recent studies. For example, Ferguson and Gleeson (2012) used an analytical model to predict the magnitude of seawater intrusion caused by future sea level rise and increased groundwater withdrawals (based on population density) for coastal aquifers throughout the United States. They found that coastal aquifers with large groundwater withdrawals (caused by higher population densities) would result in relatively large predicted increases in seawater intrusion. On the other hand, the effects of sea level rise were relatively insignificant when compared to the effects of increased groundwater withdrawal.

Although sea level rise is currently receiving significant attention in developed coastal areas because of the consequences for infrastructure and coastal erosion, this investigation suggests it has the least significant effect (of the three factors considered) on future seawater intrusion in shallow to intermediate aquifers similar to those of the Richibucto region. This finding should be encouraging to coastal water resource planners and operators of municipal well fields. Indeed, local and regional districts will not be able to control changing climate or rising sea levels; however, groundwater recharge is a process that is influenced by local land use, and pumping rates and locations can also be managed locally. Steps to control these two factors, such as the maintenance of groundwater recharge by limiting impervious surfaces in the watershed, and carefully considering the pumping rates and locations of municipal wells, should prove to be beneficial in protecting coastal fresh groundwater supplies despite an increase in sea level.

13

5.0 STUDY LIMITATIONS

The findings and recommendations from this study are subject to a number of limitations that arise from practical considerations, such as the inability to fully characterize subsurface conditions over the entire study area (Figure 2), and more theoretical aspects related to the use of numerical models.

The depth of investigations, whether based on geophysical methods or direct drilling, were limited to less than approximately 100 m, yet there are indications (Appendix A) that saline groundwater may be present at depths of greater than 80 m. The source and spatial extent of such water is currently unclear; it could be related to relatively modern seawater intrusion, or relict seawater that may have been trapped in sediments after the last ice age, as proposed by Webb (1982) and summarized in Figure A3 (Appendix A). A better understanding of groundwater salinity distributions at depth (> ~80 m), and the associated aquifer physical properties, is required to determine if such sources have/will impact municipal wells. This would require additional field investigations and the use of more advanced sampling and characterization methods. Such work could enable these deep saline water source(s) to be represented in future numerical models.

The spatial distribution of hydraulic conductivity is one of the key controls on groundwater flow and the position of the fresh water-seawater interface in a coastal aquifer (Green 2012). Although every attempt was made to accurately reproduce the location of low hydraulic conductivity layers (e.g. identified in borehole logs) in the numerical model, stratigraphic data were not available in large areas of the model domain including under the Richibucto Harbour and Northumberland Strait, which are two of the potential sources of saline water. As a result, there is uncertainty in how far offshore low hydraulic conductivity layers extend. Because seawater intrusion is especially sensitive to the sea-aquifer connections and the presence of preferential flow paths (Carrera et al. 2010), the simulated location and shape of the fresh water- seawater interface would be different if alternate shale configurations were used (Green 2012).

In addition, due to the regional extent of the numerical model, a relatively coarse discretization was used and hydrogeological units were represented as equivalent porous media. The limitations imposed by sparse hydrogeological data, combined with the inherent uncertainties in predicting future climate conditions, mean that the simulation results are best suited for investigating the general response of the aquifer to changing boundary conditions, rather than predicting local hydraulic heads and concentrations of TDS near individual pumping wells, where flow may be largely controlled by fractures in the sandstone bedrock.

14

6.0 RECOMMENDATIONS

The following recommendations are based on the findings of the work conducted in the Richibucto area. It is difficult to generalize these recommendations because no two well fields are identical in terms of aquifer properties, well depths, well spacing, pumping rates and schedules, etc.; however, many of these recommendations should be applicable in the broader context of coastal aquifers.

1. In this study, the spatial coverage and depth of investigation for saline waters were limited relative to the extent of the regional groundwater flow system. This was partly a consequence of the expense and time associated with accessing or renting specialized hydrogeological and geophysical equipment. Future studies would benefit from access to locally available equipment. For larger scale mapping of groundwater salinity, consideration should also be given to using other techniques, such as airborne electromagnetic surveying, which may prove cost-effective.

2. Municipal well fields are usually developed over a period of years, during which a number of exploratory boreholes are drilled. These boreholes provide critical hydrogeological information not only for developing production wells, but also for subsequent studies such as the one undertaken here. However, our investigations in the Richibucto area suggest that deep open boreholes, which penetrate low hydraulic conductivity shale/mudstone units, provide a hydraulic connection between aquifers that would not otherwise exist. Such holes do not provide groundwater quality or hydraulic data that can be attributed to a particular aquifer (e.g. shallow versus deep sandstone) and they may provide a pathway for saline water to move vertically between aquifers. To maximize the hydrogeological information obtained from future boreholes, it is recommended that open boreholes be geophysically logged and then completed as monitoring wells that isolate specific aquifer units.

3. For the Richibucto situation, it is recommended that new municipal production wells with open intervals in the depth range of approximately 10 to 20 m (shallow aquifer), or 25 to 50 m (deep aquifer), and planned pumping rates similar to those of the existing town wells, not be developed at distances of less than 500 m from saline coastal waters. Following such a recommendation will not guarantee that wells would never experience salinity problems but, based on the ERT and groundwater sampling results of this study, it appears the risk of lateral seawater intrusion can be reduced by maintaining an adequate set back from saline surface waters.

For other well fields, geophysical and hydrogeological investigations using some or all of the techniques employed here would be required to establish required distances from saline water sources (either surface or subsurface). 15

4. Based on ERT surveys, borehole logging, and core analyses conducted near PW1 in the Richibucto well field, it appears that saline waters have been drawn upward from beneath the shale/sandy shale layer that is present between 67.5 and 77 m depth. Although relatively thick, this shale layer has not prevented upconing of saline water in the vicinity of PW1, demonstrating that vertical migration of saline waters should be considered as a risk factor in the design (e.g. well depth) and operation of production wells.

5. Numerical modelling results and well operator experience (Sean Sullivan, Town of Richibucto, pers. comm.) both indicate that the rate of pumping of individual wells is an important factor influencing saline water ingress. Wells that are deemed to be at risk of saline water intrusion should be closely monitored for appropriate indicators; such monitoring is currently undertaken by the Town of Richibucto.

6. Because the amount of groundwater pumped from aquifers is an important factor in controlling lateral seawater instrusion, and potentially the upconing of deeper saline waters, municipal water suppliers should develop, or continue to promote, water conservation measures.

7. Maintaining groundwater recharge in inland (i.e. hydraulically upgradient) areas should be encouraged, as the numerical modelling undertaken here has shown that reductions in recharge may contribute to lateral seawater intrusion.

7.0 ACKNOWLEDGEMENTS

We appreciate the cooperation of Mr. Sean Sullivan (Town of Richibucto) in providing valuable logistical support during this study, for initiating contacts with local residents, assisting with borehole drilling arrangements, and for providing data related to the municipal pumping wells. We also appreciate the cooperation of the three private land owners who allowed borehole drilling and access to their properties.

Dennis Connor of the Department of Civil Engineering at UNB deserves recognition for his assistance in the field and lab. UNB students John Nichols and Shawn MacKenzie are acknowledged for their field assistance, as is summer-intern Guarav Lahoti for his technical and field support. UNB student Emily Jacobs provided the predictions of current and future groundwater recharge as part of an undergraduate project.

The authors would like to thank Dr. Larry Bentley of University of Calgary for the loan of his resistivity imaging system and advice on its use.

The New Brunswick Department of Environment and Local Government provided the water chemistry analyses as an in-kind contribution to the project, and Annie Daigle provided 16

information related to previous groundwater resource investigations and private water wells in the Richibucto area. We thank Robert Hughes, Mallory Gilliss, Darryl Pupek and Sherry McCoy of the New Brunswick Department of Environment and Local Government for their interest and support of the project. Sabine Dietz, Provincial Coordinator for the New Brunswick Regional Adaptation Collaborative, was instrumental in ensuring the project ran smoothly.

Primary funding for this study was provided by the Atlantic Climate Adaptation Solutions Association (ACASA; http://atlanticadaptation.ca/), a federal-provincial collaborative effort in Atlantic Canada involving Natural Resources Canada and coordinated provincially by the New Brunswick Department of the Environment and Local Government. The graduate students (E. Mott and N. Green) also received financial support from the Canada Research Chairs program, NSERC (Postgraduate Scholarship and Discovery Grants), the Department of Civil Engineering, and the Department of Earth Sciences at UNB.

8.0 REFERENCES

Carrera, J., J.J. Hidalgo, L.J. Slooten, and E. Vazquez-Sune, 2010, Computational and conceptual issues in the calibration of seawater intrusion models. Hydrogeology Journal, 18(1), 131-145.

Environment and Sustainable Development Research Centre (ESDRC), 2011, Climate Change Scenarios for New Brunswick Municipalities. ETF Project Number 09-0218, Prepared for New Brunswick Department of Environment, Fredericton, NB.

Ferguson, G., and T. Gleeson, 2012, Vulnerability of coastal aquifers to groundwater use and climate change. Nature Climate Change, 2(5), 342-345.

Green, N., 2012, A case study of the effects of climate change on seawater intrusion in coastal aquifers in New Brunswick, Canada. MScE thesis, Department of Civil Engineering, University of New Brunswick, Fredericton, NB, 165 p.

Green, N., and K. MacQuarrie, 2012, An evaluation of the importance of factors influencing seawater intrusion in coastal aquifers subject to climate change: A case study from Atlantic Canada. Poster presented at the 39th International Association of Hydrogeologists Congress, Niagara Falls, ON, September 2012.

Green, N.R., E.B. Mott, K.T.B. MacQuarrie, and K.E. Butler, 2011, Preliminary hydrogeological data and numerical modeling for a seawater intrusion study at Richibucto, New Brunswick. Presentation at the 37th Atlantic Geoscience Society Colloquium, Fredericton, NB, February 2011.

Guo, W., and C.D. Langevin, 2002, User's guide to SEAWAT; a computer program for simulation of three-dimensional variable-density ground-water flow. U.S. Geological Survey, Reston, VA. 17

Langevin, C. D., W.B. Shoemaker, and W. Guo, 2003, MODFLOW-2000, the U.S. Geological Survey Modular Ground-Water Model; documentation of the SEAWAT-2000 Version with the Variable-Density Flow Process (VDF) and the Integrated MT3DMS Transport Process (IMT). Open File Report 03-0426, U.S. Geological Survey, Reston, VA.

Langevin, C. D., D.T. Thorne, Jr, A.M. Dausman, M.C. Sukop, and W. Guo, 2008, SEAWAT Version 4; A computer program for simulation of multi-species solute and heat transport. U.S. Geological Survey, Reston, VA.

LeBlanc, N., K. MacQuarrie, K. Butler, E. Mott, N. Green, and D. Connor, 2012, Hydrogeological observations from boreholes installed to investigate seawater intrusion near Richibucto, New Brunswick. Unpublished report prepared for the Town of Richibucto, 8 p. plus appendices.

Loke, M.H. and R.D. Barker, 1995, Least-squares deconvolution of apparent resistivity pseudo-section. Geophysics, 60, 1682-1690.

Maritime Groundwater Inc., 1991, Drill Targets for the Town of Richibucto Water Supply. Unpublished report.

Maritime Groundwater Inc., 1992, Drilling and Yield Testing for the Town of Richibucto, NB; November - December 1991. Unpublished report.

Maritime Groundwater Inc., 2004, Hydrogeological Assessment - Water Source Supply Assessment, Town of Richibucto, N.B. Unpublished report.

Mott, E.B., N. Green, K.E. Butler, and K.T.B. MacQuarrie, 2011a, Preliminary interpretation of electrical resistivity tomography (ERT) surveys investigating seawater intrusion at Richibucto, eastern New Brunswick. Presentation at the 37th Atlantic Geoscience Society Colloquium, Fredericton, NB, February 2011.

Mott, E., K. Butler, N. Green, and K. MacQuarrie, 2011b, Application of ERT to the investigation of seawater intrusion at Richibucto, New Brunswick. Proc. of the Joint Mtg. of the Canadian Quaternary Association and the Canadian Chapter of the International Association of Hydrogeologists, City, August 2011, 6 p.

Mott, E.B., and K.E. Butler, 2012, Assessing saltwater intrusion using electrical resistivity tomography and borehole geophysics in a coastal sandstone aquifer, New Brunswick, Canada. Presentation at the 39th International Association of Hydrogeologists Congress, Niagara Falls, ON, September 2012.

New Brunswick Department of Natural Resources (NBDNR), 1993, Peatland Area 213 - Report, Isopach, Elevations and Profiles. New Brunswick Department of Natural Resources, Fredericton, NB.

St. Peter, C., 1993, Maritimes Basin evolution: key geologic and seismic evidence from the Subbasin of New Brunswick. Atlantic Geology, 29(3), 233-270.

Stantec Limited, 2009, Installation and Aquifer Test Analysis of Production Well PW08-01, Town of Richibucto, NB. Unpublished report. 18

Water Management Services Limited, 1986, Test Drilling and Pumping for the Village of Richibucto, 1986. Unpublished report.

Webb, T.C., 1982, Origin of brackish formational ground waters in the Pennsylvanian of New Brunswick. Open File Report 82-5, New Brunswick Department of Natural Resources, Mineral Development Branch, Fredericton, NB, 21 p.

APPENDIX A: SALT WATER INTRUSION IN A COASTAL SANDSTONE AQUIFER AT RICHIBUCTO, NEW BRUNSWICK AS REVEALED BY GEOPHYSICAL AND BOREHOLE SURVEYS Salt Water intrusion in a coastal sandstone aquifer at Richibucto, New Brunswick as revealed by geophysical and borehole surveys

Eric Mott1 and Karl E. Butler2 Department of Earth Sciences, University of New Brunswick, Fredericton, New Brunswick, Canada [email protected] [email protected]

ABSTRACT Hydrogeophysical surveys have been carried out in the vicinity of Richibucto, New Brunswick, as part of a project to assess the risk that salt water intrusion along the Northumberland Strait could increase over time as a consequence of climate change and rising sea level. The surveys involved electrical resistivity tomography (ERT) measurements along several lines extending inland from the coast in conjunction with geophysical logging of two new and several pre-existing boreholes. The investigation indicates that the elevated salinities that have intermittently affected Richibucto’s municipal well field are likely a consequence of salt water upconing beneath pumping wells. Evidence of a salt water wedge extending approximately 200 m inland was also observed beneath one ERT line acquired in a particularly low lying area adjacent to Richibucto Harbour. The results illustrate the ERT surveying can be an effective tool in identifying the extent and modes of salt water intrusion in the predominantly flat-lying, fractured, fluvial Carboniferous sandstone aquifers that supply water to many coastal communities in eastern New Brunswick. Data interpretation is, however, complicated by the presence of relatively thin, discontinuous layers of electrically conductive shale. The sensitivity of the ERT method to variations in groundwater salinity versus its sensitivity to variations in shale bed thickness are the subject of ongoing research employing ERT modelling constrained by geophysical well logs and laboratory measurements of drill core resistivity.

1 INTRODUCTION

The town of Richibucto, situated on the coast of the Northumberland Strait 65 km north of the city of Moncton (Figure A1), draws water from a municipal well field consisting of three production wells completed at depths between 23 and 51 m in a fractured sandstone aquifer. This aquifer is subdivided into upper and lower sandstone units separated by a possibly discontinuous layer of shale, with drilled thicknesses varying from 1.2 to 5 m. The production wells are located approximately 1 km from the coast and within 500 m of a tidally influenced brook (Figure A2). In recent years two production wells, PW1 and PW2, experienced elevated levels of groundwater salinity. The issue has been addressed by decommissioning PW2, restricting the pumping rate in PW1, and adding two new wells: PW3 was completed at a shallower depth, and most recently, in 2010, PW4 was commissioned farther inland (MGI 1991, Stantec 2009). However, the possibility of recurrence remains a concern.

In late July 2010 electrical resistivity tomography (ERT) surveys were conducted along several lines in and around the town of Richibucto in an effort to assess the current extent of seawater intrusion into the bedrock aquifer. The objective of this ERT survey was to determine if the spatial distribution of saline groundwater could be determined from resistivity variations in the sandstone

A1 aquifer units and used to infer any preferential pathways and mechanisms of intrusion in the Richibucto area. This information would aid a companion project (Green and MacQuarrie, 2012) in the development of a conceptual hydrogeological model for purposes of inferring the sensitivity of the area to increased salt water intrusion as a consequence of climate change

In August 2011, two boreholes, extending to depths of 56 m near the coast and 89 m within the well field, were drilled and cored to investigate signs of salt water intrusion evident in the ERT sections. These boreholes were logged in September 2011 with various geophysical tools. In this paper, we use the geological and geophysical logs, along with measurements of borehole water salinity to support interpretations of salt water intrusion evident in the two most informative ERT sections – one near the coast and one within the well field.

1.1 HYDROGEOLOGICAL SETTING

The surficial geology around Richibucto well field consists of poorly sorted glacial till up to 3 m thick. The area has very little relief, and slopes gently towards the Northumberland Strait. Bedrock geology in the region is Richibucto Formation of the Pictou Group, late Carboniferous in age and part of the post-accretionary intermontane Maritimes Basin (St. Peter, 1993) (see Figure A1). The Richibucto Formation is comprised of multilayered grey and lesser reddish brown sandstone, pebbly sandstones, and pebble conglomerates separated by reddish brown, very fine to fine grained sandstones and mudstones. The units are interpreted as fining up sequences deposited in a fluvial environment, possibly braided rivers (St. Peter and Johnson, 2009).

Figure A1. Regional map showing the extent of the Maritimes Basin in yellow and the location of Richibucto (modified from St. Peter, 1993).

A2

Figure A2. Aerial photo of the northern part of Richibucto, New Brunswick, showing the location of municipal pumping wells (black crosses with numbers indicating the production well ID) and ERT survey lines. Boreholes UNB1 and UNB2, drilled as part of this study, are shown as black circles.

The town of Richibucto has intermittently experienced elevated chloride levels within its well field with water samples from PW2 measuring as high as 322 mg/L (S. Sullivan, personal communication, 2011), in excess of the Canadian drinking water limit of 250 mg/L. These elevated chloride readings were associated with increased pumping rates to accommodate flushing of municipal water mains. However, naturally elevated levels of chloride in aquifers of eastern New Brunswick are not uncommon, as illustrated by water sampling that was carried out as part of the Carboniferous Drilling Project in late 1970’s (Ball et al., 1981). Figure A3 shows that chloride concentrations in excess of the drinking water guidelines (triangular symbols) were encountered at depths of 61 m (200 ft) and/or 121 m (400 ft) at four points along the Richibucto River extending southwest of Richibucto as well as to the east of Moncton and in several other localities (Webb, 1982). Many of these locations are far from the coast, which suggests that the saline waters may have been emplaced following the retreat of glaciers from this area when relative sea level was significantly higher. These observations seem to have been the basis for a recommendation (MGI 1991) that the town of Richibucto avoid drilling deeper than about 91 m (300 ft) to avoid intersecting saline water. Figure A3, shows that the validity of that recommendation is very site- dependent, as two wells drilled a few km to the northwest of Richibucto exhibited low salinities at both 61 and 121 m depth. Nonetheless, the data exhibit a general increase in salinity with depth which is a trend one would expect to be particularly evident in coastal areas, such as Richibucto, where modern sea water intrusion processes are occurring.

A3

Figure A3. Map showing the locations of boreholes drilled and sampled for water chemistry in central- eastern New Brunswick as part of the Carboniferous Drilling Project (Ball et al, 1981). The blue and red symbols represent chloride concentrations measured in groundwaters extracted at depths of 61 and 121 m (200 and 400 ft) respectively. The larger, triangular symbols represent chloride concentrations in excess of the Canadian Drinking Water Guideline of 250 mg/L.

2 METHODS

The two main hydrogeophysical methods used in the Richibucto area were surface ERT surveys and borehole geophysical logging. Figure A2 shows a map of the seven ERT survey lines, ranging from 400 to 800 m in length and oriented sub-perpendicular to the coast. The measurements presented in this paper were made using a 72-electrode Iris Syscal Pro resistivity imaging system using a Wenner array with unit electrode spacings (“a-spacings”) ranging from 6 to 10 m and up to 23 depth levels (i.e. “n-values” up to 23). The two-dimensional resistivity inversion program RES2DINV (Loke and Barker, 1995) was used to generate models or sections of inferred resistivity variations in the subsurface beneath each line to depths of up to 100m. The inverse modelling parameters includes smoothing constraints designed to favour the recovery of models having lateral continuity in agreement with the predominantly flat-lying geology. More details on the data acquisition, processing and inversion are presented by Mott et al. (2011).

A4

Areas of salt water intrusion were expected to be apparent as zones of anomalously low electrical resistivity (high electrical conductivity). However, it was recognized that shale layers would also be electrically conductive as a consequence of the high surface conductivity associated with clay particles. Two boreholes, UNB1 and UNB2 (Figure A2), were therefore drilled and cored with a geotechnical diamond drilling rig in order to investigate the geology and groundwater in two areas of particular interest identified by the ERT surveys. Borehole resistivity logging measurements were carried out in the new holes and in pre-existing holes for the purpose of determining typical resistivities of shale and sandstone layers. Additional geophysical logs were used to confirm lithologies and identify fractures as discussed in the examples presented below.

Relating geophysical measurements of sandstone resistivity to the salinity of its pore water requires knowledge of (i) how salinity affects pore fluid resistivity, and (ii) how pore fluid resistivity in turn affects rock resistivity. Relationship (i) was investigated by taking samples of water from boreholes in Richibucto and plotting their electrical conductivities against their salinities as expressed in mg/L chloride (sodium chloride being the dominant solute in seawater). These water samples were acquired using a down hole sampling pump or depth-specific sampling devices known as HydrasleevesTM after the wells had been purged. The purging process involved removing three well volumes of water from the borehole using a trash pump on surface connected to approximately 10 m of hose lowered down the open borehole. These samples give conductivities of the water within well bore and are subject to tidal effects near the coast (Leblanc et al., 2012) and vertical flow between fractures. They do not necessarily give salinities and conductivities of the pore water in the rock matrix at the depths from which they were taken but represent the best estimates available via open hole sampling.

Relating pore fluid resistivity to bulk rock resistivity is more complex, because the latter is also dependent on the porosity of the rock, the tortuosity of the pore space, and the possible presence of surface or grain conductivity. Visual inspection of sandstone cores from boreholes UNB1 and UNB2 suggested very little clay content so surface conductivity was not expected to be important. In that case, the relationship between fluid resistivity and bulk resistivity may be expected to take the form of Archie’s Law:

, [1] where ρr is the bulk resistivity of the rock, ρf is the resistivity of the pore fluid, F, the electrical formation factor,= aφ-m φ is the porosity m and a are empirically derived constants.

Drill core samples collected from UNB1 and UNB2 were analysed in the lab at UNB to determine porosities and formation factors, F, appropriate for the Richibucto formation sandstones making up the aquifer. The porosities of sandstone cores sampled at 1 m intervals were determined by hydrostatic weighing and range from 2.5% to 18.8% with an average of 14.2% and a standard deviation of 2.5% (Figure A4). Thus, variations in the bulk resistivities observed in the ERT and well logging data may be a consequence of variations in either sandstone porosity, pore water

A5 salinity, or sandstone vs shale lithology. The roles played by each of those factors are considered in interpretation of the ERT sections below.

40 35 30 25

Samples 20

of

15

No. 10 5 0

Porosity (%)

Figure A4. A histogram of porosities measured in 94 samples of sandstone core from boreholes UNB1 and UNB2. Labels on each bin refer to the porosity at the upper limit of the bin.

3 RESULTS AND DISCUSSION

We begin with a brief discussion of the subsurface geology and groundwater salinities encountered in boreholes UNB1 and UNB2, before moving on to a discussion of how spatial variations in groundwater salinity can be inferred from the ERT surveys.

3.1 Subsurface Geology in Boreholes UNB1 and UNB2

The recovered core (see example in Figure A5) revealed a subset of the lithologies previously reported for the Richibucto Formation. It was dominated by grey sandstone sequences generally fining upward into red shaley sands and grey shale. However, the two boreholes did not encounter any of the red sandstones (indicative of more oxidizing conditions) or pebble conglomerates that St. Peter and Johnson (2009) reported in other exposures of the Richibucto Formation. Using the core, it was difficult to identify natural fractures because core breaks occurred frequently during recovery and there was a lack of oxidation surrounding any of the breaks. To accurately identify the fractures an optical televiewer logging tool was used in the boreholes. The fractures are predominantly sub- horizontal and appear to lie mainly along bedding planes exhibiting a variety of low angle dips and dip directions as would be expected in a fluvial depositional environment. Just over 80% of the fractures in the two bore holes have dips below 20º. Some of the larger fractures appear at the top of fining up sequences where coarse grained sandstone overlies shale. Although not directly related to the issue of salt water intrusion, fracture characterization is relevant to understanding of the aquifer and water supplies as flow through fractures is expected to dominate over flow through the porous matrix. More details on fracture apertures and distributions can be found in the geophysical logs presented later in this section.

A6

Figure A5. Photograph of one end of a box of core spanning the contact between shale and overlying sandstone at 18.23 m depth in borehole UNB1. Lithologies from bottom (lower right) to top (upper left) include brown shale, grey shale, fine-grained grey sandstone, and coarse-grained grey sandstone. The fracture in the bottom right within the brown shale had a dip of approximately 16.5º. Core is type NQ, having a diameter of 47.6 mm (1.87 inches).

10000 1000 Well field 9000 900 UNB2 8000 800 PW1 1999-2001 Surface Water 7000 700

6000 600 (uS/cm) (uS/cm)

5000 500

4000 400 UNB1 Conductivity Conductivity 3000 300 MW08‐01 MW08‐02 2000 200 MW08‐04 1000 100 MW86‐2 PW1 0 0 0 1000 2000 3000 0 50 100 150 200 250 Chloride (mg/L) Chloride (mg/L)

Figure A6. Graph showing the relationship between chloride concentration and conductivity in water samples taken from wells within the well field, borehole UNB2 near the coast, and from the St Charles River (just northwest of the ERT surveys). The drinking water limit for chloride is shown with the dashed blue line. The graph on the right shows the lower salinities and conductivities, sampled from the well field, at a larger scale.

A7

3.2 Groundwater Salinities

Figure A6 shows two conductivity versus chloride concentration plots. The plot on the left shows the results of sampling the water column at various depths in monitoring wells within the well field, as well as in monitoring well UNB2 near the coast and at one surface site along the St. Charles River. The dashed blue line represents the Canadian Drinking Water Quality standard for chloride and its corresponding conductivity value (250 mg/L, ~1000 µS/cm). The second plot shows, at an expanded scale, results for only those waters sampled from monitoring wells within the well field. At the higher concentrations of chloride, a linear relationship can be established with conductivity. As chloride decreases well below the Canadian drinking water guidelines of 250 mg/L into the region of the samples from within the well field this linear relationship no longer applies as other ionic species (besides Cl- and Na+) become significant.

3.3 Electrical Resistivity Tomography Surveys

Of the seven ERT lines acquired, only the two most interesting and informative will be presented here. Line 1, acquired in a low-lying area on the north side of Richibucto Harbour (see Figure A2) shows clear evidence of seawater intruding laterally from the coast. Line 4, located within the well field, suggests that the elevated salinities in PW1 are a consequence of salt water upconing from depth.

3.3.1 Coastal ERT Survey

Figure A7 shows the ERT section for Line 1 which passes through borehole UNB2, drilled to a depth of 56.87 m, near the coast (see map in Figure A2). A generalized geological log for UNB2, overlain on the ERT section, shows that the geology is dominated by sandstone with three thin layers of shale/shaley sandstone ranging from 1.2 to 5.5 m thick. A more detailed geological log for UNB2 and the corresponding geophysical logs are shown in the lower part of the figure.

Apart from a low resistivity zone at surface below the eastern third of the line, which might represent infiltration of tidal surges, the ERT section is dominated by a wedge-like zone of low resistivity extending inland from the coast, underlying more resistive formations above. The resistivity structure does not correlate with the lithological variations observed in borehole UNB2, but is instead suggestive of a salt water wedge intruding laterally into the lower sandstones penetrated by the borehole. The rock resistivity measured within UNB2 by geophysical well logging is comparable to the ERT resistivity adjacent to the hole with the exception of the shale layer from 22 to 27.5 m which is not resolved in the ERT inversion. The ERT method is unable to resolve the low resistivity shale bed at this depth because of its small thickness to depth ratio. However, the transition from high to low resistivity at about 35 m depth in the resistivity log is reproduced in the ERT section.

Shale layers evident in core are also easily identified based on the geophysical logs. Higher gamma ray emission counts matched with low resistivities of 25 to 55 Ohm·m distinguish the shales from the sandstones. The sandstone resistivity varies widely, likely due to a combination of variations in porosity and pore water conductivity. Based on visual examination of the drill cores, it seems likely that short wavelength (metre-scale) variations in sandstone grain size and sorting may be accompanied by changes in porosity and/or permeability that in turn give rise to short

A8 wavelength changes in bulk resistivity. However, the gradual (long wavelength) decrease in resistivity with depth below the upper shale layer is more likely to be a consequence of increasing pore water salinity related to an increase in the chloride concentration as established earlier in the paper (see Figure A6) .

Note that the normal resistivity log contains four tracks with different electrode spacings (8, 16, 32, and 64 inches). The shorter spacings, which penetrate less deeply into the formation, tend to measure higher resistivities; this reflects the presence of fresh water introduced to the rock during drilling, forming a resistive flushed zone near the borehole wall. Such flushed zones are likely to be better developed in areas of higher sandstone porosity and permeability, thereby contributing to the short wavelength variations in resistivity that are most evident in the short electrode spacing data. The longer electrode spacings provide measurements closer to the true resistivity of the formation but due to the large sampling volume they are unable to resolve thin beds and become distorted when approaching shale layers.

Figure A7. Above: ERT section 1 - an estimate of the true resistivity structure beneath Line 1 suggesting the presence of a saltwater wedge intruding from the coast. Left: Geophysical and geological logs for borehole UNB2 intersected by the ERT survey. Grey shading is used to highlight the shale layers. Grey arrows indicate two fining upward sequences of grain size. Below: legend for the geological log.

A9

Two fluid conductivity logs, acquired on different dates, are shown side-by-side in Figure A7. The first, FCond 25°C (in red), was logged continuously with the Mount Sopris logging system at the same time as the other geophysical logs. The second, TLC-Cond25°C (in green), was logged about three weeks later with a Solinst Model 107 TLC meter at 1 m intervals. The logs have the same general shape with conductivity being nearly constant through the upper part of the borehole and increasing dramatically at 35 to 38 m depth. However, the conductivities were higher and the abrupt increase occurred at a shallower depth in the second log. Evidence from two water temperature-level-conductivity loggers emplaced in the borehole at depths of 8 and 28 m (LeBlanc et al., 2012) suggests that this variability is likely a consequence of tidal effects; it appears likely that saline water repetitively entered the borehole through fractures at depth during times of high tide and displaced some of the fresher water above until the tide receded again.

The geophysical logs in Figure A7 also include the locations of fractures, based on analysis of the optical televiewer data. The vast majority (80%) of fractures, such as those corresponding to bedding plane partings, are subhorizontal. The remainder of the fractures have dips as high as 67° with the steepest at 37.17 m depth. The fractures appear to have no preferred orientation. The geophysical log on the far right in Figure A7 includes estimates of the apertures of fractures that appeared to be 2 mm or greater in the optical televiewer data. Three of the 96 fractures in borehole UNB2 were found to have significant apertures, ranging from 2 to 9 mm. We note that fractures occur at or near the top of each of the three shale layers encountered in borehole UNB2. Given the abrupt increase in borehole water conductivity observed between 35 and 40 m depth, we suspect that the fractures occurring at the tops of the shale layers at 41 and 52 m depth could be the dominant sources of saline water in this borehole.

Combining the above observations from ERT surveying, geological and geophysical logging, and water conductivity logging, we can interpret the highly conductive area in the deeper part of the ERT section as a seawater intrusion wedge, mixed with fresh groundwater and extending inland beneath the coast.

3.3.2 Well Field ERT Survey

The ERT section for Line 4, located within the well field, is shown in Figure A8 with geological logs for four wells, including PW1 and UNB1, overlain on it. The geophysical logs for borehole UNB1, drilled to a depth of 88.54 m, are also shown. From these data a reasonable interpretation of the subsurface can be rendered. The ERT section reveals the upper and lower sandstone aquifer as two resistive layers with a more conductive unit between the depths of 20 and 35 m. Comparing the ERT section to the UNB1 geological and resistivity logs it appears that this conductive layer is an artifact of the ERT inversion code’s inability to resolve two shale layers separately. Another point of interest on the ERT section is the drop in resistivity observed at a depth of approximately 70 m. This drop can be explained, in part, by the presence of another shale to shaley sandstone layer at depth of 67.5 – 77 m, characterized by low resistivities of 25 – 55 Ohm m in the geophysical well logs. However, it is also a consequence of the fact that sandstones below the shale have an average resistivity that is at least a factor of two lower than the resistivity of the sandstones above. Comparing gamma emissions in the sandstones immediately above and below the lower shale shows that there is no reason to expect that they differ in clay content. Furthermore, lab measurements on core reveal that sandstone porosities above and below the shale are very

A10 comparable. As a result, it would appear that pore waters beneath the lower shale may be twice as saline as those above that layer. The fact that this is not observed in the borehole fluid conductivity log suggests that there is little or no flow of water from the formation into the borehole at depth.

Figure A8. Above: ERT section 4 - an estimate of the true resistivity structure beneath Line 4 revealing a predominantly layered subsurface modified by the possible upconing of saline water beneath PW1. The locations and geological logs for boreholes UNB1, PW1, MW86-1, and 04TW- 1 are also shown. Left: Geophysical and geological logs for borehole UNB1, annotated with measured sandstone porosities at two depths. Grey shading highlights the shale layers and three grey arrows indicating fining up sequences of grain size.

A11

The most prominent anomaly on the ERT section in Figure A8 is the low resistivity zone below PW1. This anomaly does not conform to the layered strata known to exist in the area. Forward modelling exercises (Mott et al., 2011) have demonstrated that the anomaly cannot be explained by variations in the shale thickness observed in the area or by the 26 m long steel well casing for PW1 located 17 m off-line. Instead, the most likely explanation is considered to be upconing of saline water from depth. The presence of more saline water at depth below the lower shale layer is supported by the low resistivities present at the base of the ERT section and by resistivity logs from borehole UNB1 as outlined above. While the shale layer would be expected to impede upward water movement, its vertical permeability may be increased by moderately to steeply dipping fractures. Examples of such fractures observed within the shale units include one at 23.05 m depth (dipping 53°) in borehole UNB2 (Figure A7) and two in borehole UNB1: one at 69.05 m depth (dipping 29) (Figure A8) and another at 19.80 m depth (dipping 16.5) (Figure A5).

3.4 Pore Fluid Conductivity and Salinity Estimates

To estimate the pore water conductivity from the borehole resistivity logs below the shale it was necessary to implement the relationship established in section 2 of this paper. The first step to accomplish this was to use ten 1inch diameter core samples chosen from various depths within the two boreholes and determine how their bulk conductivities varied with pore water conductivity. The samples were saturated in several NaCl solutions of different strength and allowed to equilibrate each time. The resistivities of the cores were measured with a two electrode apparatus (Merriam, 2009) after each equilibration period. Figure A9 shows a plot of the resulting bulk vs. pore water conductivity curves for each of the ten samples.

The curves from Figure A9 are expected to be linear, with slopes of 1/F according to Archie’s Law (equation 1). However the slopes depart from linearity towards the higher pore water conductivities. This can be attributed to non-fully equilibrated samples. Also none of the curves pass through the origin, suggesting that all samples have surface conductivity that cannot be ignored for low pore water conductivities. Equation [1] may be modified to incorporate the surface conductivity as follows:

, [2] where r is the bulk conductivity of the rock, f is the conductivity of the pore fluid, s is the surface conductivity, commonly attributed to clay content within the sample, F, electrical formation factor, = a/φ m φ is the porosity m and a are empirically derived constants.

For each sample we can now calculate 1/F and s by linear regression.

A12

200 R1‐50 R1‐65 160 R1‐83 R1‐87

S/cm) R2‐14 µ ( 120 R2‐16 R2‐32 R2‐36 80 R2‐39 Conductivity

R2‐43

Bulk 40

0 0 500 1000 1500 2000 2500 3000 Pore Water Conductivity (µS/cm)

Figure A9. Measured bulk conductivity vs. the pore water conductivity for the 10 cores used to establish electrical formation factors.

In order to estimate in-situ groundwater resistivities, we used Equation [2] in combination with the resistivities measured during borehole logging at the depths of the core samples taken from each hole. The long (64”) spacing normal resistivity measurement was used in order to minimize the influence of the zone close the borehole that would have been flushed by fresh water during drilling. The method assumes that the properties of the rock contributing to the borehole resistivity measurement are the same as the corresponding core sample. The error in this assumption is hard to quantify but can be minimized by selecting core samples from portions of the borehole over which the long normal resistivity readings are relatively stable. Table 1 shows the measured borehole log resistivities, and the equivalent bulk conductivities values after correction to a temperature of 25ºC using an estimated correction factor of 2.5% per degree Celsius (Keller and Frischknect, 1966). These temperature corrected bulk conductivities were substituted for r in Equation [2] in order to estimate the groundwater conductivities f (corrected to 25 deg C) presented in Table 1. The inferred chloride content (from the relationship established in Section 3.2) is also presented.

The pore water conductivity and salinity estimates obtained for a depth of 65 m in borehole UNB1 (above the lowermost shale) are 380 µS/cm and 64 mg/L respectively. This is within the range of expected values that are seen from water samples taken from the well field (Figure A6(b)). The inferred pore fluid conductivity at 50 m depth is even lower, at 96 µS/cm suggesting similarly fresh pore water (although the corresponding salinity is too low to be resolved by this method is therefore stated as 0 mg/L). Below the lowermost shale layer in UNB1 inferred pore water conductivity values are four to eight times higher at 1500 and 3000 µS/cm. This corresponds to chloride concentrations of 410 and 880 mg/L; well above the 250 mg/L drinking water guideline. Upconing

A13 of this deeper groundwater into the producing aquifer via the fractures seen in the shale layer could account for the anomalously high conductivites observed in water from PW1 during period 1999 to 2001 (see Figure A6(b)) when pumping rates were relatively high.

Table 1. Estimates of pore water conductivity based on long-spacing normal resistivity measurements in boreholes UNB1 and UNB2 and core measurements. Fluid conductivities are reported for a temperature of 25° C as is standard. Inferred chloride concentrations are also shown. R64 R64 Formation Surface Inferred Inferred Bore- Resistivity Conductivity Factor Conductivity Pore water Chloride hole Depth Temp (Measured) (25 deg C) F s Conductivity Concentration m oC Ohm m µS/cm µS/cm µS/cm mg/L UNB1 50 7.2 219 45 18.0 40.3 96 0 (-23) UNB1 65 7.2 192 52 19.4 32.4 380 64 UNB1 83 7.4 76 131 14.7 29.1 1500 410 UNB1 87 7.4 92 108 32.6 15.3 3000 880 UNB2 14 9.3 61 170 19.2 24.6 2800 800 UNB2 16 9.2 58 178 55.2 15.9 9000 2700 UNB2 32 8 106 96 17.2 28.8 1100 300 UNB2 36 8.2 44 233 23.2 37.8 4500 1300 UNB2 39 7.8 37 272 21.6 34.4 5100 1500 UNB2 43 7.4 26 380 18.7 44.6 6200 1900

4 CONCLUSIONS

ERT surveys results suggest that elevated groundwater salinities in the Richibucto well field are likely a consequence of salt water upconing beneath pumping wells. The presence of increased salinities at depth is confirmed by the observation that sandstones underlying a shale layer, approximately 80 m deep in borehole UNB1, have relatively low resistivities that cannot be readily explained by an increase in clay content or porosity. ERT surveys and borehole geophysics also reveal a salt water wedge extending inland from the coast in a very low-lying area. These results show that ERT surveys can be an effective tool in identifying the extent and modes of salt water intrusion into the similar Carboniferous sandstone aquifers that underlie many coastal communities in eastern New Brunswick. Repeat surveys may be used to assess changes in the extent of salt water intrusion that may accompany climate change and sea level rise or other stressors such as increased pumping. Ongoing work, including (i) an examination of the depth of investigation of ERT surveys in this environment and (ii) forward modeling to determine the method’s relative sensitivity to variations in groundwater salinity versus variations in shale layer thickness will serve to refine and strengthen these assertions.

ACKNOWLEDGEMENTS

Funding for this research was provided by the Atlantic Regional Adaptation Collaborative (RAC) Climate Change Program initiated by NRCan and the Departments of Environment of the four Atlantic provinces. The program is being managed by the Atlantic Climate Adaptation Solutions

A14

Association (ACASA). The authors would like to thank Larry Bentley of University of Calgary for the loan of his resistivity imaging system and advice on its use. We thank the Town of Richibucto, especially Supervisor of Public Works and Utilities Sean Sullivan, for help with access and logistics. We thank Annie Daigle, Mallory Gillis, Darryl Pupek and Sherry McCoy of the New Brunswick Department of Environment for their support of the project. We also thank UNB colleagues Nathan Green, Kerry MacQuarrie and Dennis Connor for their help in the field, for water quality data, and for their insights and cooperation in the project as a whole.

REFERENCES

Ball,F. D., Sullivan, R. M., and Peach, A. R., 1981, Carboniferous Drilling Project, New Brunswick Department of Natural Resources, Fredericton, New Brunswick, Canada, 109 pp. Green, N., and MacQuarrie, K., 2012, An evaluation of the effects of climate change on seawater intrusion in coastal aquifers: A case study from Richibucto, New Brunswick. Report submitted to the Atlantic Climate Adaptation Solution Association (this volume), Keller, G. V. and Frischknecht, F. C., 1966, Electrical Methods in Geophysical Prospecting, Pargamon Press, Oxford Loke, M. H. and Barker, R. D., 1995. Least-squares deconvolution of apparent resistivity pseudo-section. Geophysics, 60, 1682-1690. Maritime Groundwater Inc. (MGI), 1991, Test Drilling and Pump Testing for the town of Richibucto, N.B., Report prepared for Boissonnault McGraw, November. 1991, 30 pp. LeBlanc, N., MacQuarrie, K., Butler, K., Mott, E., Green, N., and Connor, D., 2012, Hydrogeological Observations From Boreholes Installed to Investigate Seawater Intrusion Near Richibucto, New Brunswick, Unpublished UNB report prepared for the Town of Richibucto, May, 2012, 25 pp. Merriam, J.B. 2009. A two electrode apparatus for electrical impedance measurements. EOS Transactions of the Amer. Geophysical Union, 90(52), Fall meeting Supplement, Abstract # NS31B-1165. Mott, E., Butler, K., Green, N. and MacQuarrie, K., 2011, Application of ERT to the investigation of seawater intrusion at Richibucto, New Brunswick. Proceedings of GeoHydro 2011: Joint Mtg. of the Canadian Quaternary Association and the Canadian Chapter of the International Association of Hydrogeologists, Aug. 28 – 31, Quebec, Canada, 6 pp. Stantec, 2009. Installation and Aquifer Test Analysis of Production Well PW08-1, Town of Richibucto, NB, Report prepared for Crandell Engineering Ltd., March 2009, 94 pp. St. Peter, C., 1993, Maritimes Basin evolution: key geologic and seismic evidence from the Moncton Subbasin of New Brunswick. Atlantic Geology, 29(3), 233-270. St. Peter, C and Johnson, S.C., 2009, Stratigraphy and structural history of the late Paleozoic Maritimes Basin in southeastern New Brunswick, Canada. Memoir 2008-8, New Brunswick Department of Natural Resources; Minerals, Policy and Planning Division., 348 pp. Reynolds, J. M., 2011, An Introduction to Applied and Environmental Geophysics 2nd Ed, John Wiley & Sons Ltd, Baffins Lane, Chichester, West Sussex, England. Webb, T. C., 1982, Origin of Brackish Formational Ground Waters in the Pennsylvania of New Brunswick, New Brunswick Department of Natural Resources, Fredericton, New Brunswick, Canada, 21 pp.

A15

APPENDIX B: GROUNDWATER QUALITY AND HYDRAULIC HEAD DATA PRESENTATION AND INTERPRETATION

Groundwater quality and hydraulic head data presentation and interpretation Nathan Green1, Kerry MacQuarrie2, and Dennis Connor3

Department of Civil Engineering, University of New Brunswick, P.O. Box 4400, Fredericton, NB, E3B 5A3 1 email: [email protected] 2 email: [email protected] 3 email: [email protected]

1. INTRODUCTION A field campaign was undertaken between June 2010 and November 2011 to collect hydrogeological data from the Richibucto area aquifers. Water level and specific conductance data were recorded and groundwater quality samples were collected from selected boreholes and monitoring wells. These data, along with previously published information, were used to help develop the conceptual and numerical hydrogeological models used for this project.

This appendix provides details of the field methods used, the results obtained, and a discussion of how these findings apply to the current study.

2. DATA COLLECTION METHODS All the monitoring data reported here were collected from boreholes or wells that were installed prior to this investigation, with the exception of boreholes UNB1 and UNB2, and monitoring well UNB3 (Figure B1), which were drilled between August 17, 2011 and August 25, 2011. UNB1 and UNB2 were subsequently decommissioned in November 2011. Decommissioning involved full borehole grouting and casing removal. The locations of all boreholes and wells for which data are reported are shown in Figure B1.

2.1 WATER LEVEL AND SPECIFIC CONDUCTANCE Water level and specific conductance were recorded in the boreholes and monitoring wells shown in Table B1. The majority of the holes had steel casing that extended below the ground surface, with the entire interval below the bottom of the casing being open to the formation rocks. Two wells (UNB3 and MW08-02, Table B1) were completed with PVC well screens.

To record water level and specific conductance, Model 3001 LTC Levelogger Juniors (Solinst Canada Ltd., Georgetown ON) were deployed on stainless steel cables attached to the well cap (Table B1), and changes in barometric pressure were compensated for by using data collected from two Barologgers (Solinst Canada Ltd., Georgetown, ON) that were deployed approximately a metre below ground surface in the MW08-01 and UNB3 monitoring wells. The loggers were programmed to record data every half-hour, which was considered sufficient to capture changes in water levels and specific conductance caused by pumping in the well field or tides. Water B1

levels and specific conductance were also recorded in the St. Charles and Richibucto Rivers at locations S1 and S4 (Figure B1-A). Data loggers were not installed at TH3 and TH4 due to a lack of secure deployment options (i.e. no locking well caps).

Water level and specific conductance were periodically measured manually to verify the operation of the data loggers. Specific conductance was measured manually using a Model 107 TLC meter (Solinst Canada Ltd., Georgetown, ON). Values of specific conductance collected by the data loggers were corrected to the more frequently calibrated TLC meter readings.

2.2 GROUNDWATER SAMPLE COLLECTION A set of groundwater samples was collected from existing monitoring wells MW08-01, MW08- 02, MW08-03, 04TW-1, TH3, TH4 and MW86-2, and surface water locations S1, S2, S3 and S4 (Figure B1-A), between June 20, 2011 and June 22, 2011. A set of samples was also collected from locations UNB1, UNB2, and UNB3 (Figure B1) on October 6 and October 13, 2011. Because most of the monitoring wells were completed as open boreholes below the surficial casing, it is possible that the water chemistry within the open intervals may be influenced by hydraulic conditions arising from the intersection of the hole with multiple fractures, or fracture zones, with different hydraulic heads. This can lead to ambient flow of water within the borehole and ambiguous results for water quality (e.g. Shapiro 2002). Straddle-packer equipment was not available for the sampling, so in an effort to address the possible vertical variations of water chemistry within the open boreholes two methods were used to collect water samples: depth specific grab sampling and pumping.

B2

Figure B1: A: Regional map of the Richibucto study area showing selected groundwater monitoring locations (green dots) and surface water sampling sites (S1 to S4). B: Close-up of well field area showing pumping wells (blue dots) and monitoring wells/boreholes (green dots).

B3

Table B1: Information for the monitoring wells used in this study. Elevation is reported in m above Canadian Geodetic Vertical Datum of 1928 (CGVD28). Data logger depth is reported in metres below top of casing (m btoc). NB Double Elevation Stereographic (m above Current Data Diameter Casing depth Monitoring Year Projection CGVD28) depth logger (cm) and (m below Comment well drilled (m below depth Current Current type ground) Easting Northing top of ground ground) (m btoc) casing surface 25.7 (slotted UNB3 2011 2618638.2 7519120.1 24.87 24.02 30.3 5 PVC PVC below 14.4 this) TH3 1991 2621896.9 7521381.1 11.85 11.60 49.1 12.5 steel 24.4 N/A open to deeper TH4 1991 2622217.0 7521634.6 11.02 10.72 86.9 12.5 steel 51.2 N/A interval than UNB3 and TH3 MW08-01 2008 2624123.3 7522656.8 6.46 6.01 48.8 15 steel 23.9 31.6 N/A (all open to surficial MW08-02 2008 2624134.9 7522670.2 6.38 6.08 8.8 5 PVC 5.5 slotted PVC) aquifer only open to surficial 04TW-1 2004 2624162.4 7522915.1 6.98 6.48 36.4 15 steel 7.6 11.3 aquifer and confined aquifer open to surficial UNB1 2011 2624281.0 7523075.9 5.26 4.91 88.5 7.5 steel 4.4 9.5 aquifer and confined aquifer MW86-2 1986 2624379.6 7522955.2 7.13 7.13 37.4 10 steel 30.0 32.2 MW08-03 2008 2624666.6 7523253.9 5.38 4.91 48.8 15 steel 27.1 31.3 open to surficial 8.4 and UNB2 2011 2625889.4 7523192.7 2.04 1.48 56.9 7.5 steel 4.2 aquifer and 28.4 confined aquifer

B4

HydraSleeve (GeoInsight, Las Cruces, NM) discrete interval, no purge, samplers were used to collect water samples from multiple depths within the uncased interval of selected boreholes. Sampling depths were selected to correspond with the depth of fractures indicated on borehole logs (when available). Shapiro (2002) has suggested that grab samples taken from the water column near transmissive fractures may be representative of aquifer water in fractures provided the borehole has been open for a sufficient time to establish ambient flow. Where no fractures were documented in a borehole, a minimum of two samples were collected.

The single-use HydraSleeve sampler captures water from a discrete interval of the borehole. While being lowered into position, external water pressure keeps the sample bag collapsed and check valve closed to prevent water from entering. Samplers were left in place at the desired depth for approximately 15 minutes to allow the water column to stabilize, and sample collection was initiated by raising the sampler faster than approximately 0.3 m s-1. Sterile, high-density polyethylene bottles (supplied by the New Brunswick Department of Environment) were filled by puncturing the HydraSleeve bag with the supplied straw, and allowing water to pour through the straw into the bottle. Once samples had been decanted, the specific conductance of the remaining water in the sampler was measured using the Model 107 TLC meter. Multiple samples from the same well were collected from increasingly greater depths to minimize disturbance of the water column at the point of collection. The deepest HydraSleeve sample collected from each well was also used for the analysis of tritium.

A Model 140-6 submersible air operated piston-sampling pump (Bennett Sample Pumps, Inc., Amarillo, TX) with 41 m of polypropylene tubing was also used to collect a composite sample from each well after discrete interval sampling was completed. During sampling the pumping rate varied from approximately 2.5 to 4.8 L min-1 and was dependent on pump depth in the well, and occasional clogging of the pump intake screen by fine sediments. While pumping, sample water was directed through two flow-through cells where several water quality parameters were measured. In the first cell dissolved oxygen (D.O.) was monitored using a YSI 85 meter (YSI Inc., Yellow Springs, OH). Outflow from this cell was directed into a second flow-through cell where specific conductance, temperature, pH and redox potential were measured using a YSI 3500 meter (YSI Inc., Yellow Springs, OH). Prior to sample collection, pumping was carried out until at least one well volume was purged, and the water quality parameters measured in the flow-through cells were sufficiently stable. The pump tubing was purged after sampling each well using distilled water.

Water samples were stored for a maximum of three days at 5°C until analyzed by the New Brunswick Department of Environment Laboratory in Fredericton, New Brunswick. Minerals were analyzed by inductively coupled plasma optical emission spectrometry (ICP-OES) using EPA method 200.7 (with modifications), metals were analyzed by inductively coupled plasma mass spectrometry (ICP-MS) using EPA method 200.8 (with modifications), and chloride, sulfate and bromide ions were analyzed by ion chromatography (IC) using method 4110B (Standard Methods 21st ed.).

B5

Tritium samples were analyzed at Isotope Tracer Technologies, Inc. in Waterloo, , by counting beta (β-) decay events using a liquid scintillation counter after the samples were enriched fifteen times by electrolysis. The detection limit for this method is 0.8 TU (tritium units) and accuracy is ± 1 standard deviation (typically around 0.8 TU) of the count of tritium units.

Selected groundwater quality results from water well exploration studies conducted near the municipal well field (Stantec 2009), and from domestic water well testing in the area, were used to compare and contrast to the results obtained from the 2011 sampling. Results were included from domestic water wells (Annie Daigle, New Brunswick Department of Environment, pers. comm.) that had concentrations of chloride greater than 250 mg L-1 (the aesthetic limit set by Health Canada (2008)) and measurable levels of bromide. The purpose of including these results was to provide examples of local water chemistry for wells impacted by seawater intrusion.

3. RESULTS

3.1 WATER LEVEL AND SPECIFIC CONDUCTANCE Although water level and specific conductance were recorded in many of the monitoring wells for the full duration of field work (i.e. June 2010 to November 2011), only selected data sets are shown for brevity. Water level, temperature and specific conductance results from UNB1, UNB2 and UNB3 are also reported by LeBlanc et al. (2012).

Figure B2 shows a sample of data collected from MW08-02 (Figure B1), and daily precipitation recorded by Environment Canada (2012) at Kouchibouguac National Park. The small daily oscillations in water level that were recorded at this location were likely caused by groundwater withdrawals from municipal well PW3, located approximately 450 m to the northeast. The larger fluctuations in water level correlate well with precipitation. During dry periods, the groundwater level slowly declined, while significant rain fall events (e.g. September 29, 2011 to October 5, 2011, or November 10, 2011) caused the groundwater levels to increase by 0.1 to 0.5 m. Although there appeared to be some correlation between the specific conductance recorded at this location and precipitation (i.e. increase of approximately 60 μS cm-1 during the September 29, 2011 to October 5, 2011 precipitation event) other shorter-duration fluctuations are unexplained.

B6

Figure B2: Water level and specific conductance recorded at MW08-02. Also shown is precipitation measured at Kouchibouguac National Park (Environment Canada 2012).

Figure B3 shows four weeks of data collected at MW08-01 (Figure B1). During the first two weeks, the regular pattern of water level fluctuations of approximately 0.2 to 0.5 m were caused primarily by withdrawals at Pumping Well 1 (PW1), located approximately 300 m away. The two short duration, large amplitude, declines in water levels during this period were due to equipment tests that were conducted at PW4, located approximately 9 m away. On September 21, 2011, PW4 was returned to regular operation and the water level fluctuations in MW08-01 increased significantly. When PW4 was in operation, the water level in the aquifer at MW08-01 dropped by approximately 2 m, but showed quick recovery when the pumping ceased. Also shown in Figure B3 is the specific conductance of the water at a depth of 31.6 m below ground surface. An inverse correlation between the water level and the specific conductance was noted, and results from groundwater samples taken from MW08-01 (Table B2) suggest that the fluctuations in specific - conductance were primarily caused by changes in bicarbonate (HCO3 ) concentration. The chloride (Cl-) concentrations remained stable and low (< 7.79 mg L-1) in MW08-01 during the same sampling period (Table B2).

B7

Figure B3: Hydraulic head and specific conductance measured at MW08-01.

Table B2: Major ion chemistry at MW08-01. Data logger for specific conductance was 31.6 m btoc. Specific Depth Alkalinity HCO - Cl- TDS Date/Time 3 Conductance (m btoc) (mg L-1) (mg L-1) (mg L-1) (mg L-1) (μS cm-1) MW08- June 21, 25.2 118 115.2 7.79 247 205.8 01-H1 2011 8:57 MW08- June 21, 31.3 128 125 6.14 256 207.7 02-H2 2011 9:35 MW08- June 21, 34.6 159 155.7 5.2 310 253.0 01-H3 2011 10:03 MW08- June 21, 42.2 191 188.6 4.3 345 304.7 01-H4 2011 10:35 MW08- June 21, 25.0 180 174.2 6.85 350 298.1 01-B1 2011 13:25 MW08- July 12, 31.7 149 146 6.33 318 260.7 01-HL1 2011 12:23 MW08- July 12, 32.0 158 154.7 6.23 342 271.9 01-HL2 2011 14:40 MW08- July 12, 32.3 186 181.8 4.61 387 316.6 01-HL3 2011 17:12

B8

Figure B4 shows water levels and specific conductance at MW08-03 (Figure B1), located approximately 800 m from MW08-01 towards the shore. Also shown in Figure B4 is the predicted tidal water level in Richibucto Harbour (Canadian Hydrographic Service 2010) during the same time period (June 21, 2010 to July 31, 2010). It should be noted that tidal predictions may not reflect actual water level fluctuations in the harbour because of local weather conditions. Comparisons between the pumping records from the municipal well field and the water levels at MW08-03 did not show strong correlations; however, the water levels at MW08-03 appeared to follow the fluctuating tidal water levels, as shown by the close alignment of these two time series. Figure B4 also shows very large fluctuations (up to 800 μS cm-1) in specific conductance, which unlike MW08-01, had no apparent correlation with tidal variations or municipal well pumping. The fluctuations in specific conductance began on approximately July 3, 2010 and stopped by August 11, 2010 and fluctuations of this magnitude were not observed again during the study. One possible explanation for these fluctuations might be the presence of private pumping wells in the area of MW08-03, which if operated intermittently may have temporarily drawn saline water inland through fractures that intersect the open interval of MW08-03. When pumping ceased, a reversal in hydraulic gradient may have allowed fresh water to re-enter the monitoring well.

Figure B4: Hydraulic head and specific conductance at MW08-03. Also shown is the predicted tidal water level for Richibucto Harbour.

B9

Water level and specific conductance data for UNB2 (Figure B1) are shown in Figure B5. The similarity between the temporal variations in the water levels in the well and those of the predicted tides suggest that groundwater levels at this location were strongly controlled by tides. It also appeared that the specific conductance at this location was related to tidal fluctuations. These results are not surprising given the close proximity (100 m) of this monitoring well to the shoreline of Richibucto Harbour. The deeper specific conductance logger (Figure B5, UNB2_28 located 28.4 m below the surface) recorded more frequent, longer duration, and larger variations than the shallower logger (Figure B5, UNB2_8 located 8.4 m below the surface). These differences may have been caused by having more saline, and thus denser, water entering near the bottom of the open borehole. This dense water may then have been pushed upward by increasing tidal water pressure within the open hole, first contacting the deeper specific conductance logger. On several occasions (e.g. Sept. 25, approximately 18:00, Figure B5) it appears that the driving head (i.e. tidal amplitude) was not large enough to cause the saline water to move far enough upward to contact the shallower logger.

Figure B5: Hydraulic head and specific conductance from two depths within UNB2. Also shown are predicted tidal water levels in Richibucto Harbour.

Average groundwater levels for the month of September 2011 for all monitoring locations are reported in Table B3. Values range from an average of 17.52 m at UNB3, to 0.8 m at UNB2. The decrease in hydraulic head towards the shore indicates a general direction of groundwater flow from the St. Charles Peat Bog towards Richibucto B10

Harbour. TH3 and TH4 are open to different intervals of sandstone separated by what is assumed to be a low hydraulic conductivity shale layer, which may explain their contrasting hydraulic head values considering their close proximity to one-another. The higher head in TH3 indicates a downward gradient across the shale layer. A downward gradient is also inferred across the shale layer that vertically separates MW08-02 (shallow sandstone aquifer) and MW08-01 (deeper sandstone).

Table B3: Water level results for boreholes and monitoring wells for September 2011. Water level at TH3 and TH4 measured on August 12, 2011. Groundwater level Monitoring (m above CGVD28) well Minimum Average Maximum UNB3 17.44 17.52 17.62 TH3 9.73 TH4 3.97 MW08-01 -0.26 1.45 2.15 MW08-02 4.81 5.11 5.39 04TW-1 2.78 3.06 3.39 UNB1 2.17 2.70 3.52 MW86-2 0.51 0.99 1.47 MW08-03 0.58 0.95 1.26 UNB2 0.44 0.80 1.14

3.2 GROUNDWATER CHEMISTRY Chloride (Cl-) and bromide (Br-) ions generally behave conservatively in groundwater flow systems and the ratio between these two anions can be used to help distinguish the source of salinity in a water sample. Alcalá and Custodio (2008) provide a review and discussion of how different sources of salinity, and geochemical reactions that may occur during groundwater flow, affect the ratio of chloride to bromide in groundwater. In seawater the mass ratio of Cl- to Br- is 284 (Hem 1985); if seawater is diluted with fresh water, the Cl-:Br- mass ratio will remain close to this value. Significant deviations away from this ratio generally indicate sources of bromide and/or chloride other than seawater. For example, increased quantities of Cl- from wastewater and other anthropogenic sources can cause the Cl-:Br- mass ratio to be greater than 650, while the use of bromide based halocarbons in pesticides may reduce this ratio to below 130 (Alcalá and Custodio 2008). Snow et al. (1990) also used Cl-:Br mass ratios to differentiate between water samples containing relict seawater, contamination from road de-icing salt, and modern seawater intrusion. Samples contaminated by road salt had Cl-:Br- ratios greater than 1200 and significant levels of tritium, those contaminated by relict seawater had ratios similar to modern seawater, but contained no tritium, while those contaminated by modern seawater had Cl-:Br ratios similar to seawater and contained tritium (Snow et al. 1990).

The mass ratios of chloride to bromide in groundwater samples collected from the Richibucto area are displayed in Figure B6. Also included are samples from the St. Charles River and Richibucto River and Harbour, selected historic samples from the B11

municipal well field, and samples from domestic wells in the region where Cl- concentrations have been recorded at levels greater than 250 mg L-1. At concentrations of chloride greater than approximately 30 mg L-1, both the historic and current data fall close to the seawater dilution line (i.e. mass ratio of 284). Elevated Br- concentrations for three samples from UNB2 cause those samples to deviate from the seawater dilution line; however, other hydraulic and specific conductance data (see results above) have confirmed the presence of seawater in this monitoring well. The deviation of these samples from the ideal dilution line may reflect small concentration differences that arise due to sampling or analytical methods. Overall, the data confirm seawater as the source of salinity in the aquifer and do not suggest the presence of other sources of salinity such as road salt.

Figure B6: Scatter plot of the concentrations of bromide (Br-) and chloride (Cl-) from historic samples (obtained from NB Department of Environment, Town of Richibucto, and Stantec (2009)) and samples collected during the current study. The theoretical seawater dilution line indicates a constant mass ratio (Cl-:Br-) of 284.

The major ion results for water samples from the Richibucto area are presented in the format of a Piper plot (Fetter 2001), which consists of two triangular ternary plots and one central diamond-shaped figure (Figure B7). Because there were generally only small differences in the major ion results for samples collected by pumping and the HydraSleeve™ sampler, Figure B7 presents the results obtained from samples collected

B12

by pumping (with the exception of MW86-2). An additional Piper plot comparing the major ion results for samples collected using the two sampling methods is provided by Green (2012). All groundwater chemistry results are also tabulated at the end of this Appendix.

On a Piper plot, each axis of the cation ternary plot represents one of the major cations (i.e. Ca2+, Mg2+, or Na+ + K+) as a percentage of the sum of these cations. Similarly, the - 2- - 2- anion ternary plot represents the major anions (i.e. Cl , SO4 and HCO3 + CO3 ) as a percentage of the sum of these anions. Points on these two ternary plots are projected to the central diamond region to obtain a single point that represents the bulk chemical composition of the water sample. Water samples with distinct chemistries (e.g. fresh groundwater versus saline estuary water) will be located in different regions on a Piper plot, and thus it is a useful method to visually distinguish different water types.

Figure B7 shows two distinct “end-member” water types, including one that is typical of seawater with relatively high proportions of Cl- and Na+ + K+ (e.g. Hem 1985). In this seawater region of the Piper plot (i.e. upper right region of central diamond plot) are samples from saline surface waters (St. Charles River and Richibucto Harbour), and UNB2. Once again, samples from domestic wells with Cl- greater than 250 mg L-1 are also included to illustrate the effect of seawater intrusion. In contrast with these samples are those from MW08-02 and 04TW-1, which are screened partially or entirely in the upper (i.e. surficial) aquifer and thus have major ion chemistry (relatively high Ca2+ and - HCO3 ) that is believed to represent relatively fresh groundwater that has been recharged by precipitation or snow melt. Several of the remaining samples (e.g. MW08-03) plot between these end members of fresh and saline water and likely represent various degrees of seawater mixed with fresh groundwater. At least one water sample appears to be distinct from either fresh shallow groundwater or seawater; TH4 plots to the lower left region on the central diamond (Figure B7) which suggests that the water chemistry may have been modified by geochemical reactions (e.g. cation exchange, or reaction with carbonate minerals; Appelo and Postma 2005).

B13

Figure B7: Piper plot of water chemistry results. The cation ternary plot is the lower left triangle, the anion ternary plot is the lower right triangle. Locations of water samples within the ternary plots are projected into the center diamond. Unless otherwise noted in the legend, all results are for samples collected by pumping.

3.3 TRITIUM Tritium is a radioactive isotope of hydrogen with a half-life of 12.43 years. It decays by the emission of a low energy beta particle with no associated gamma rays or x-rays, meaning that the amount present in natural waters is harmless. Tritium is measured in 3 18 tritium units (TU) where 1 TU is equivalent to one tritium atom ( H) per 10 hydrogen atoms or 3.221 pCi L⁻¹ (picocurries per litre) (Clark and Fritz 1997). From 1951 to 1976, thermonuclear bomb testing produced large quantities of tritium that were incorporated into precipitation that subsequently recharged groundwater. Decay of tritium from natural “pre-bomb” levels is such that it cannot normally be detected in groundwater recharged before 1951 (Clark and Fritz 1997).

A number of complicating factors make it difficult to use tritium results to determine an exact groundwater age (Lindsey et al. 2003). For example, mixing of pre- and post “bomb” waters, smoothing of the highly variable input signal, and the decay of the input signal at the same rate as that of the isotope itself all generate a degree of ambiguity. A graphical technique (Lindsey et al. 2003; Clark and Fritz 1997) is often used to determine

B14 an approximate age of a groundwater sample based on tritium results. First, a prediction of tritium in local precipitation is developed using correlations between short-term local records and the long-term records that are available for Ottawa, Canada (Water Isotope System for Data Analysis, Visualization, and Electronic Retrieval, 2011). This predicted time-series of tritium in local precipitation defines the amount of tritium that would have entered the aquifer during groundwater recharge (Figure B8). Tritium results for groundwater samples are then plotted at the time of sample collection (e.g. 2011), and lines are drawn backward in time to represent the natural decay curve of tritium. The time(s) at which these decay curves intersect the precipitation time series is used to assign an approximate age of the groundwater sample.

The tritium results for selected groundwater samples and the local (Truro, NS) predicted tritium in precipitation are displayed in Figure B8. As explained above, the measured values have been extrapolated backward in time to obtain the approximate times associated with groundwater recharge and these are tabulated in Table B4.

Levels of tritium range from 7.07 TU in UNB2, to less than 0.8 TU in TH3 and appear to be separated into two groups (Figure B8 and Table B4). The first group with low values includes samples from MW08-01, TH3, TH4, UNB3 and MW08-03. The prediction of the recharge time for these samples all fall in the 1950’s (Figure B8). The remaining samples have higher levels of tritium, which considered in isolation, would suggest more recent recharge.

Figure B8: Time series of tritium in precipitation at Ottawa, ON, predicted time series of tritium in precipitation at Truro, NS, and radioactive decay curves of tritium for selected groundwater samples.

B15

Table B4: Tritium results for selected groundwater samples and estimated water recharge dates. Sample Name Result (TU) Estimate of recharge date MW08-01-H4 1.12 1950’s TH3-H2 <0.8 1950’s or earlier TH4-H1 1.82 1950’s TH4-H4 1.36 1950’s UNB3-B1 1.57 1950’s MW08-03-H4 2.10 1950’s MW86-2-1H 6.26 undetermined: late 1950s, or 1970’s to 2000’s UNB1-H3 5.24 undetermined: 1950’s, or 1970’s to 1990’s UNB1-H5 6.16 undetermined: 1960’s, or 1970’s to 2000’s UNB2-H1 6.39 undetermined: 1960’s, or 1970’s to 2000’s UNB2-H3 7.07 undetermined: 1950’s, or 1970’s to 2000’s MW08-02-B1 6.07 undetermined: 1950’s, or 1970’s to 2000’s

4. DISCUSSION The groundwater monitoring results presented above were assessed individually by location in an attempt to determine if the various lines of evidence indicated the influence of groundwater mixing with seawater. It should be noted that the discussion is based solely on the monitoring results obtained in 2010/2011 and, as such, gradual long-term trends that may be indicative of seawater intrusion cannot be determined.

04TW-1 and MW08-02: Both of these wells are open, partially (04TW-1) or entirely (MW08-02), to the upper unconfined sandstone aquifer. Water level and specific conductance data recorded in these wells (e.g. MW08-02 in Figure B2) indicate that hydraulic head and TDS at these locations respond quickly to precipitation. The position of the major ion results for both wells on the left side of the Piper plot (Figure B7) indicates relatively fresh groundwater that has not been influenced by mixing with seawater. Low concentrations of chloride and a lack of bromide in the water samples also support this. Considering the shallow depth of MW08-02 (i.e. 8.8 m) and that it is screened only in the upper sandstone, the high tritium result for this location (i.e. 6.07 TU) can be interpreted as indicating relatively recently recharged water (e.g. perhaps as recently as the mid-2000’s; Figure B8).

Because the major ion results from 04TW-1 are very similar to those of MW08-02 (Figure B7), it is believed that the water samples collected from 04TW-1 were probably dominated by groundwater from the shallow sandstone, even though the well extends to a depth of 36.4 m. This is supported by a comparison of the hydraulic heads in monitoring wells isolated within the upper (i.e. MW08-02) and lower aquifers (e.g. MW08-01), which indicates that there is a strong downward hydraulic gradient that may promote downward flow in a conduit such as a borehole open to both aquifers. B16

MW08-01: Since MW08-01 is open only to the lower, semi-confined, aquifer it was expected that water sampled from this location would have different characteristics than water from wells in the shallower unconfined aquifer (i.e. 04TW-1, MW08-02). Water sampled from this well had lower TDS than water from the upper aquifer, with the differences being accounted for by lower concentrations of bicarbonate and calcium (by approximately 20 mg L-1 each). Based on the position of these water samples on the left side of the Piper plot (Figure B7), it appears that the groundwater chemistry at MW08-01 was also unaffected by seawater. This is confirmed by the relatively low levels of bromide and also by the distal position of the Cl-:Br- mass ratio from the theoretical seawater dilution line (Figure B6). The relatively low quantity of tritium from this monitoring well suggests a date of recharge in the early 1950’s.

UNB1: Similar to the well completion at 04TW-1, UNB1 was open to both the upper, unconfined and lower, semi-confined sandstone aquifers. Based on the results discussed above, the water samples collected from UNB1 may have been more indicative of groundwater from the upper unconfined aquifer. The position of the major ion results on the left side of the Piper plot (Figure B7), as well as the absence of Br-, indicate that the groundwater sample from this location has not experienced mixing with seawater. The relatively high levels of tritium (> 5 TU) detected in two groundwater samples from this well are also probably reflective of a shallow aquifer origin.

MW86-2: MW86-2 is located adjacent to the now abandoned production well PW2 (lower aquifer) and still producing PW3 (upper aquifer), and is between MW08-01 and MW08-03 (Figure B1-B). Based on the fact that the well is open exclusively to the lower confined aquifer, it was expected that groundwater sampled from this well would be similar to that of MW08-01. However, calcium and bicarbonate concentrations were noticeably lower, while chloride levels were higher. The position of the major ion results on the Piper plot (Figure B7) indicates that the water sampled from this well may have been affected by mixing with seawater. This is consistent with the history of this portion of the well field. During the late 1990’s production well PW2, located only 5 m away and open to the same interval as MW86-2, started to experience increased levels of chloride and by 2004 it was abandoned. The presence of bromide in MW86-2, and a Cl-:Br- ratio close to the theoretical seawater dilution line (Figure B6), corroborates this evidence.

The high level of tritium (6.26 TU) in the groundwater sample from MW86-2 was somewhat unexpected. Given the open interval in the semi-confined lower aquifer, lower tritium levels, as obtained for MW08-01 and MW08-03, were anticipated. However, if the intrusion of seawater through the aquifer toward MW86-2 was a relatively recent process (e.g. caused by increased pumping in the 1990’s) then this may explain the relatively high tritium level. Also, it should be noted that the tritium result for this location yields ambiguous recharge dates because it can be interpreted to be related to

B17

recharge that occurred after the 1970’s, or during the late 1950’s (Figure B8 and Table B4).

MW08-03: MW08-03 is located approximately 550 m from the shore, approximately halfway between the St. Charles River/Richibucto Harbour and the municipal well field (Figure B1-B). It was constructed with an open interval in the lower, semi-confined aquifer. The major ion results for MW08-03 are noticeably different from those above, and lie in a region on the Piper plot that indicates mixing of fresh water and seawater (Figure B7). The concentrations of Cl- were higher than in the municipal well field (approximately 20 to 60 mg L-1), and the Cl-:Br- ratios fall on the theoretical seawater dilution line (Figure B6). The higher concentrations of chloride and the location of the observation well may indicate the close proximity of the freshwater-seawater transition zone. The occurrence of tidal fluctuations in the groundwater level records and sporadic high specific conductance (~ 800 μS cm-1; Figure B4) in this well also provide evidence that seawater was impacting the groundwater at this location. However, the relatively low value of tritium obtained for water from this well (2.1 TU) indicates that some portion of the water was sourced inland and was recharged in the 1950’s.

UNB2: UNB2 was located approximately 100 m from the Richibucto Harbour, and was open to both unconfined and semi-confined aquifers. This was the only monitoring well that definitively showed evidence that it had intersected the freshwater-seawater interface. The position of the major ions to the upper right on the Piper plot (Figure B7), the proximity of the Cl-:Br- ratio to that of the seawater dilution line (Figure B6), and the relatively high tritium content (> 6 TU) all point towards high proportions of seawater being sampled from this well. The high value of TDS in the water samples (1200 mg L-1 to 2500 mg L-1) is also indicative of mixing with seawater at this location in the aquifer. Water level and specific conductance recorded here (Figure B5) suggest that the salinity of the water at this location varied with the tides, and that higher salinities were present in the lower semi-confined sandstone aquifer at depths greater than 28 m.

TH3, TH4 and UNB3: These wells are located in the centre of the study region and are the farthest away from the tidal rivers and harbour (Figure B1-A). The interpretation of the stratigraphy between these wells implies that UNB3 and TH3 are open to the same interval of semi-confined aquifer, while TH4 is open to a deeper sandstone layer, isolated by a layer of shale. The position of the major ion results on the Piper plot indicates that this inland portion of the aquifer is unaffected by seawater. Three other factors support this: low concentrations of Cl-; the absence of bromide (see “Additional Data” for concentrations of Cl- and Br-) and similar levels of tritium to MW08-01 (Table B4). However, these wells had the highest levels of TDS of any of the wells not suspected of being influenced by seawater, with a range of 145 mg L-1 to 429 mg L-1 (see “Additional Data” for tabulated TDS). The high levels of TDS are likely indicative of a longer groundwater residence time, allowing for

B18

increased dissolution of aquifer solids. This is consistent with the location and depth of these wells, and also with the low levels of tritium (< 2 TU, Table B4).

5. CONCLUSIONS Fieldwork was conducted between June 2010 and November 2011 in an attempt to improve the understanding of groundwater flow and seawater intrusion in the Richibucto area. Hydraulic head and specific conductance data were recorded in 10 monitoring wells in the area, and groundwater samples were collected and analyzed for major cations, anions and tritium.

The data suggest that the shallow unconfined sandstone aquifer contains relatively fresh groundwater that has been recently recharged. There was no evidence of the presence of seawater in the unconfined aquifer at the locations investigated (e.g. MW08-02, 04TW- 1); however, no shallow monitoring wells (i.e. isolated in the upper sandstone) were available within 1200 m of the shoreline. Groundwater samples from monitoring locations that had open intervals that intersected both the shallow unconfined aquifer and the deeper semi-confined aquifer(s) (e.g. 04TW-1, UNB1) appeared to have chemical compositions reflecting shallow groundwater (e.g. MW08-02). Such holes may be providing a cross-connection between the shallow and deeper aquifers.

Groundwater withdrawals from the unconfined portion of the aquifer are minimal: few private wells are completed this shallow, and there is only one municipal pumping well (PW3) currently operating in this interval. With little pumping from this interval, the potential for seawater intrusion in the vicinity of the well field appears to be minimal.

Mixing of fresh groundwater and seawater appears to have occurred in the deeper semi- confined aquifer at the two locations investigated within 500 m of the shoreline of the Richibucto harbour (i.e. MW08-03 and UNB2). This finding likely indicates close proximity to the freshwater-seawater transition zone.

In the past, high pumping rates in the vicinity of MW86-2 caused elevated levels of chloride within the well field (i.e. PW2). Although chloride concentrations in groundwater samples from MW86-2 were less than 25 mg L-1, it is believed that the effects of past seawater intrusion were detected in the major ion chemistry results based on their position on the Piper plot (Figure B7).

There was no evidence of the presence of seawater in groundwater samples from locations inland (southwest) of the current municipal well field. Samples from the deeper sandstone units at UNB3, TH3 and TH4 had TDS concentrations generally around 400 mg L-1, with relatively low levels of tritium, indicating an older source of freshwater.

It is concluded that at least three distinct types of water occur within the Richibucto area aquifers. Freshwater, recharged recently from precipitation and snow melt, is present in the unconfined sandstone aquifer. Low tritium values and relatively high TDS for deeper

B19 water samples from inland wells (TH3, TH4 and UNB3) suggest the presence of older groundwater recharged in the 1950’s, possibly indicating a regional source of freshwater sourced farther to the west. In the vicinity of the municipal well field, these two sources of freshwater may experience some degree of mixing, explaining the combination of lower TDS and low tritium levels. The third type of water in the Richibucto aquifer is groundwater mixed with seawater, as revealed by the higher TDS and distinctive major ion chemistry.

ACKNOWLEDGEMENTS

Funding for this research was provided by the Atlantic Regional Adaptation Collaborative (RAC) Climate Change Program initiated by NRCan and the Departments of Environment of the four Atlantic Provinces. The program is being managed by the Atlantic Climate Adaptation Solutions Association (ACASA). We thank the Town of Richibucto, especially Supervisor of Public Works and Utilities, Sean Sullivan, for help with access and logistics. We thank Robert Hughes, Annie Daigle, Mallory Gilliss, Darryl Pupek and Sherry McCoy of the New Brunswick Department of Environment and Local Government for their interest and in-kind support of the project. We also thank UNB colleagues Eric Mott and Karl Butler for their help in the field, and for their insights and cooperation in the project as a whole.

B20

REFERENCES

Alcalá, F.J., and Custodio, E. (2008). Using the Cl/Br ratio as a tracer to identify the origin of salinity in aquifers in Spain and Portugal. J Hydrol, 359, 189-207. Appelo, C. A. J., and Postma, D. (2005). Geochemistry, groundwater and pollution. Balkema, Leiden, New York. 649 pp. Canadian Hydrographic Service. (2010). Tides, Currents and Water Levels. http://www.lau.chs- shc.gc.ca/english/Canada.shtml. Fisheries and Oceans Canada. (January 15, 2010). Clark, I.D. and Fritz, P. (1997). Environmental Isotopes in Hydrogeology. Lewis Publishers, Boca Raton, Florida. 328 pp. Environment Canada. (2012). National Climate Data and Information Archive. http://climate.weatheroffice.gc.ca (May 9, 2012). Fetter, C. W. (2001). Applied Hydrogeology. Prentice Hall, Inc., Upper Saddle River, NJ. Green, N. (2012). A case study of the effects of climate change on seawater intrusion in coastal aquifers in New Brunswick, Canada. MScE thesis, Department of Civil Engineering, University of New Brunswick, 165 p. Health Canada. (2011)."Environmental and Workplace Health - Reports and Publications - Water Quality. http://www.hc-sc.gc.ca/ewh-semt/pubs/water-eau/index-eng.php (November 9, 2010). Hem, J.D. (1985). Study and Interpretation of the Chemical Characteristics of Natural Water. United States Geological Survey Water Supply Paper, 2254. LeBlanc, N., K. MacQuarrie, K. Butler, E. Mott, N. Green, and D. Connor. (2012) Hydrogeological observations from boreholes installed to investigate seawater intrusion near Richibucto, New Brunswick. Unpublished report prepared for the Town of Richibucto, 8 p. plus appendices. Lindsey, B.D., Phillips, S.W., Donnelly, C.A., Speiran, G.K., Plummer, L.N., Böhlke, J-K., Focazio, M.J., Burton, W.C., and Busenburg, E. (2003). Residence times and nitrate transport in ground water discharging to streams in the Chesapeake Bay watershed. Water Resources Investigations Report, 03-4035. Shapiro, A. M. (2002). Cautions and suggestions for geochemical sampling in fractured rock. Ground Water Monit. Remediat., 22(3), 151-164. Snow, M. S., Kahl, J. S., Norton, S. A., and Olson, C. (1990). Geochemical determination of salinity sources in groundwater wells in Maine. National Water Well Association, Groundwater Management Book 3: FOCUS Conference on Eastern Regional Groundwater Issues (Proceedings). Springfield, MA, 313-327. Stantec Limited. (2009). Installation and Aquifer Test Analysis of Production Well PW08-01, Town of Richibucto, NB. (Unpublished report). Water Isotope System for Data Analysis, Visualization, and Electronic Retrieval (WISER). (2011). Monitoring Programmes Data http://www-naweb.iaea.org/napc/ih/IHS_resources_isohis.html. International Atomic Energy Association. (September 26, 2011).

B21

Additional Data

Complete results for water samples

Sample label key:

Groundwater samples: the last two digits in the sample label indicate the type of sampling method (H: HydraSleeve™; B: Bennett™ pump) and the sample number from that location (1-4). The set of digits before the last two indicate the sample location (i.e. monitoring well name: 04TW-1, MW08-01, MW08-02, MW08-03, MW86-2, TH3, TH4)

Surface water samples: The first digit (S) indicates a surface water sample and the next digit (1-4) indicates the location (Figure B1).

B22

Units Detection limit 04TW-1-H1 04TW-1-H2 04TW-1-H3 04TW-1-B1 MW08-02-B1 MW08-01-H1 MW08-01-H2 MW08-01-H3 Date 2011-06-20 2011-06-20 2011-06-20 2011-06-20 2011-06-21 2011-06-21 2011-06-21 2011-06-21 Time 15:05 15:34 16:08 20:00 09:12 08:57 09:35 10:03 Sample depth (m btoc) 13.96 19.08 33.47 19.05 7.62 25.18 31.27 34.56 Alkalinity mg/L 141 140 138 137 122 118 128 159 Aluminum mg/L <0.025 <0.025 <0.025 <0.025 <0.025 0.054 <0.025 <0.025 <0.025 Antimony μg/L <1 <1 <1 <1 <1 <1 <1 <1 <1 Arsenic μg/L <1.5 <1.5 <1.5 <1.5 <1.5 <1.5 <1.5 <1.5 <1.5 Barium mg/L <0.01 0.346 0.352 0.363 0.364 0.308 0.729 0.317 0.432 Boron mg/L <0.01 0.01 <0.01 <0.01 <0.01 <0.01 0.018 0.017 0.023 Bromide mg/L <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 0.31 <0.1 <0.1 Cadmium μg/L <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 Calcium mg/L <0.1 47.3 46.6 46.8 45.5 41 24.3 18.6 23 Chloride mg/L <0.05 6.73 7.06 7.05 6.58 7.65 7.79 6.14 5.2 Chromium mg/L <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 Specific conductance μS/cm 293 292 290 288 261 247 256 310 Copper mg/L <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 Fluoride mg/L <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 Iron mg/L <0.01 0.247 0.705 0.681 0.171 4.6 0.53 1.23 0.932

B23 Lead μg/L <1 <1 <1 <1 <1 <1 <1 <1 <1 Magnesium mg/L <0.1 6.44 6.41 6.44 6.14 5.2 10.6 9.32 10.5 Manganese mg/L <0.005 0.15 0.16 0.17 0.16 0.53 0.26 0.13 0.17 pH 8.11 8.16 8.19 8.19 7.39 8.4 8.39 8.34 Potassium mg/L <0.1 0.8 0.8 0.8 0.8 0.7 7.5 6 6.9 Selenium μg/L <1.5 <1.5 <1.5 <1.5 <1.5 <1.5 <1.5 <1.5 <1.5 Sodium mg/L <0.1 6.79 6.74 6.7 6.74 6.88 9.4 11.8 13.2 Sulfate mg/L <0.05 6.32 6.46 6.61 6.29 3.46 2.24 <0.05 0.88 Thallium μg/L <1 <1 <1 <1 <1 <1 <1 <1 <1 Total Hardness mg/L <0.67 145 143 143 139 124 104 84.8 101 Turbidity NTU <0.2 Uranium μg/L <0.5 0.5 0.7 0.6 <0.5 <0.5 <0.5 <0.5 <0.5 Zinc mg/L <0.005 <0.005 <0.005 <0.005 <0.005 <0.005 <0.005 <0.005 <0.005 2- CO3 mg/L as CaCO3 1.7 1.9 2 2 0.3 2.7 2.9 3.2 - HCO3 mg/L as CaCO3 139.2 138.1 135.9 135 121.7 115.2 125 155.7 Total Dissolved Solids mg/L as ion 245.9 244.8 242.5 238.5 218.9 205.8 207.7 253.0 Field spc. cond. (TLC 107) μS/cm 260 263 261 300 198 202 Field spc. cond. (YSI 3500) μS/cm 288 275 Field pH (YSI 3500) 7.67 6.43 Field D.O. (YSI 85) % 0.5 1.1 Field ORP (YSI 3500) mV -118 -164 Field temp (YSI 3500) oC 7 7.3 Units Detection limit MW08-01-H4 MW08-01-B1 MW86-2-H1 MW86-2-H2 MW08-03-H1 MW08-03-H2 MW08-03-H3 MW08-03-H4 Date 2011-06-21 2011-06-21 2011-06-21 2011-06-21 2011-06-21 2011-06-21 2011-06-21 2011-06-21 Time 10:35 13:25 12:03 12:37 15:37 16:03 16:30 16:55 Sample depth (m btoc) 42.15 24.99 33.01 36.00 28.47 33.07 39.47 42.85 Alkalinity mg/L 191 180 88.8 84.5 83.4 78.7 93.9 102 Aluminum mg/L <0.025 <0.025 <0.025 0.028 0.07 <0.025 <0.025 <0.025 <0.025 Antimony μg/L <1 <1 <1 <1 <1 <1 <1 <1 <1 Arsenic μg/L <1.5 <1.5 <1.5 <1.5 <1.5 <1.5 <1.5 <1.5 <1.5 Barium mg/L <0.01 0.626 0.865 0.458 0.206 0.925 0.637 1 0.78 Boron mg/L <0.01 0.033 0.023 0.051 0.031 0.015 0.021 0.04 0.041 Bromide mg/L <0.1 <0.1 0.114 <0.1 0.122 0.137 <0.1 0.122 <0.1 Cadmium μg/L <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 Calcium mg/L <0.1 30.7 37.1 9.1 9.34 17.3 12 20.3 16.1 Chloride mg/L <0.05 4.3 6.85 12 24.6 26.4 26.9 23.4 33.5 Chromium mg/L <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 Specific conductance μS/cm 345 350 214 248 266 253 273 328 Copper mg/L <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 Fluoride mg/L <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 0.113 <0.1 Iron mg/L <0.01 0.863 0.126 1.57 4 0.859 1.18 0.55 1.15

B24 Lead μg/L <1 <1 <1 <1 1 <1 <1 <1 <1 Magnesium mg/L <0.1 12.5 14.3 2.89 2.98 4.42 3.89 3.94 4.13 Manganese mg/L <0.005 0.23 0.29 0.076 0.19 0.12 0.08 0.18 0.14 pH 8.12 8.54 8.4 8.34 8.34 8.39 8.43 8.46 Potassium mg/L <0.1 7.2 8.9 5.9 4.4 2.5 2.5 2.2 2.6 Selenium μg/L <1.5 <1.5 <1.5 <1.5 <1.5 <1.5 <1.5 <1.5 <1.5 Sodium mg/L <0.1 15.9 10.8 30 30.6 24.1 29.2 31.1 38.4 Sulfate mg/L <0.05 1.06 2.89 4.54 4.36 11.6 7.67 10 9.91 Thallium μg/L <1 <1 <1 <1 <1 <1 <1 <1 <1 Total Hardness mg/L <0.67 128 152 34.6 35.6 61.5 46 66.9 57.1 Turbidity NTU <0.2 Uranium μg/L <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 Zinc mg/L <0.005 <0.005 <0.005 <0.005 0.018 <0.005 <0.005 <0.005 0.007 2- CO3 mg/L as CaCO3 2.3 5.7 2 1.7 1.7 1.8 2.3 2.7 - HCO3 mg/L as CaCO3 188.6 174.2 86.6 82.7 81.6 76.8 91.5 99.2 Total Dissolved Solids mg/L as ion 304.7 298.1 173.4 182.8 188.9 178.8 205.9 229.3 Field spc. cond. (TLC 107) μS/cm 265 198 185 246 213 253 262 Field spc. cond. (YSI 3500) μS/cm 346 Field pH (YSI 3500) 7.61 Field D.O. (YSI 85) % 0.2 Field ORP (YSI 3500) mV -160 Field temp (YSI 3500) oC 7.4 Units Detection limit MW08-03-B1 TH3-H1 TH3-H2 TH3-B1 TH4-H1 TH4-H2 TH4-H3 TH4-H4 TH4-B1 Date 2011-06-21 2011-06-22 2011-06-22 2011-06-22 2011-06-22 2011-06-22 2011-06-22 2011-06-22 2011-06-22 Time 20:17 08:43 09:10 12:40 12:20 15:50 13:25 14:05 17:20 Sample depth (m btoc) 31.09 30.30 45.29 33.83 55.93 63.92 79.28 86.93 37.19 Alkalinity mg/L 117 249 253 256 237 85.7 121 210 261 Aluminum mg/L <0.025 0.037 <0.025 <0.025 <0.025 <0.025 <0.025 <0.025 <0.025 <0.025 Antimony μg/L <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 Arsenic μg/L <1.5 <1.5 <1.5 <1.5 <1.5 <1.5 <1.5 <1.5 <1.5 <1.5 Barium mg/L <0.01 0.765 3.8 2.3 3.8 0.483 0.143 0.235 0.492 0.665 Boron mg/L <0.01 0.07 0.037 0.048 0.041 0.119 <0.01 0.027 0.068 0.116 Bromide mg/L <0.1 0.195 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 Cadmium μg/L <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 Calcium mg/L <0.1 20.2 52.8 39.2 52 19.6 7.34 11.1 22.2 26.1 Chloride mg/L <0.05 61.8 3.12 3.04 3.12 3.18 4 3.94 3.25 2.88 Chromium mg/L <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 Specific conductance μS/cm 447 450 458 462 440 173 236 391 474 Copper mg/L <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 Fluoride mg/L <0.1 <0.1 0.167 0.199 0.184 <0.1 <0.1 <0.1 <0.1 <0.1 Iron mg/L <0.01 0.641 1.92 2.6 1.23 3 9.4 5.7 4.9 0.559

B25 Lead μg/L <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 Magnesium mg/L <0.1 5.08 15.7 13.6 15.4 5.07 1.9 3.15 6.84 8.46 Manganese mg/L <0.005 0.15 0.83 0.58 0.87 0.17 0.41 0.39 0.29 0.17 pH 8.48 8 8.05 7.96 8.24 7.37 7.62 8.1 8.36 Potassium mg/L <0.1 3.1 8 6.5 8 3.1 0.8 1.3 2.3 2.9 Selenium μg/L <1.5 <1.5 <1.5 <1.5 <1.5 <1.5 <1.5 <1.5 <1.5 <1.5 Sodium mg/L <0.1 62.4 20.1 34.9 19.7 74.5 15.4 25.9 52 63.9 Sulfate mg/L <0.05 11.6 0.071 0.956 <0.05 5.21 0.935 1.38 2.35 3.14 Thallium μg/L <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 Total Hardness mg/L <0.67 71.3 196 154 193 69.9 26.2 40.8 83.6 100 Turbidity NTU <0.2 Uranium μg/L <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 Zinc mg/L <0.005 0.007 <0.005 <0.005 <0.005 <0.005 <0.005 <0.005 <0.005 <0.005 2- CO3 mg/L as CaCO3 3.2 2.3 2.6 2.2 3.8 0.2 0.5 2.5 5.5 - HCO3 mg/L as CaCO3 113.6 246.6 250.3 253.8 233.1 85.5 120.5 207.5 255.4 Total Dissolved Solids mg/L as ion 306.5 408.6 410.7 415.1 400.9 144.7 200.3 349.2 423.6 Field spc. cond. (TLC 107) μS/cm 431 418 435 166 196 230 Field spc. cond. (YSI 3500) μS/cm 446 472 483 Field pH (YSI 3500) 8.22 7.45 7.9 Field D.O. (YSI 85) % 0.6 0.3 0.3 Field ORP (YSI 3500) mV -188 -145 -239 Field temp (YSI 3500) oC 7.1 7.3 6.9 Units Detection limit S1 S2 S3 S4 Seawater MW08-01-HL1 MW08-01-HL2 MW08-01-HL3 Date 2011-06-22 2011-06-22 2011-06-22 2011-06-22 (Hem 1985) 2011-07-12 2011-07-12 2011-07-12 Time 16:15 16:50 17:10 17:27 12:23 14:40 14:12 Sample depth (m btoc) ---- 31.70 32.00 32.31 Alkalinity mg/L 70.2 81.9 30.5 18.7 149 158 186 Aluminum mg/L <0.025 <0.5 <0.5 <0.5 0.16 0.001 <0.025 <0.025 <0.025 Antimony μg/L <1 <20 <20 <20 <1 0.0003 <1 <1 <1 Arsenic μg/L <1.5 <30 <30 <30 <1.5 0.003 <1.5 <1.5 <1.5 Barium mg/L <0.01 <0.2 <0.2 <0.2 0.034 0.02 0.811 0.87 0.933 Boron mg/L <0.01 1.9 2.5 0.83 <0.01 4.5 0.028 0.023 0.045 Bromide mg/L <0.1 41.4 60.3 8.97 <0.1 67 <0.1 <0.1 <0.1 Cadmium μg/L <0.5 <10 <10 <10 <0.5 0.00011 <0.5 <0.5 <0.5 Calcium mg/L <0.1 183 230 72.9 6.03 410 36.4 36.8 42.3 Chloride mg/L <0.05 9670 11900 2460 3.72 19000 6.33 6.23 4.61 Chromium mg/L <0.01 <0.2 <0.2 <0.2 <0.01 0.00005 <0.01 <0.01 <0.01 Specific conductance μS/cm 26600 32600 8350 57.1 318 342 387 Copper mg/L <0.01 <0.2 <0.2 <0.2 <0.01 0.003 <0.01 <0.01 <0.01 Fluoride mg/L <0.1 0.407 0.687 0.245 <0.1 1.3 <0.1 <0.1 0.107 Iron mg/L <0.01 0.056 0.042 0.471 0.676 0.003 0.307 0.272 0.257

B26 Lead μg/L <1 <20 <20 <20 <1 0.00003 <1 <1 <1 Magnesium mg/L <0.1 559 710 235 1.12 1350 14 14.7 15.9 Manganese mg/L <0.005 <0.1 <0.1 0.14 0.056 0.002 0.3 0.31 0.34 pH 7.84 7.94 7.43 7.4 8.32 8.34 8.38 Potassium mg/L <0.1 170 220 72 0.4 390 7.8 8.8 7.8 Selenium μg/L <1.5 <1.5 <1.5 0.00009 <1.5 <1.5 <1.5 Sodium mg/L <0.1 4474 5800 1882 2.17 10500 13.1 11.1 18.8 Sulfate mg/L <0.05 1140 1510 351 1.56 2700 1.86 2.25 1.38 Thallium μg/L <1 <20 <20 <20 <1 <1 <1 <1 Total Hardness mg/L <0.67 2760 3500 1150 19.7 149 152 171 Turbidity NTU <0.2 Uranium μg/L <0.5 <10 <10 <10 <0.5 0.003 <0.5 <0.5 <0.5 Zinc mg/L <0.005 <0.1 <0.1 <0.1 <0.005 0.01 <0.005 <0.005 <0.005 2- CO3 mg/L as CaCO3 0.5 0.7 0.1 0 0 2.9 3.2 4.1 - HCO3 mg/L as CaCO3 69.7 81.2 30.4 18.6 116.5 146 154.7 181.8 Total Dissolved Solids mg/L as ion 16324.7 20532.5 5120.6 38.6 34564.8 260.7 271.9 316.6 Field spc. cond. (TLC 107) μS/cm 22000 27000 11800 51 53000 318 319 375 Field spc. cond. (YSI 3500) μS/cm Field pH (YSI 3500) Field D.O. (YSI 85) % Field ORP (YSI 3500) mV Field temp (YSI 3500) oC Units Detection limit UNB1-H1 UNB1-H2 UNB1-H3 UNB1-H5 UNB1-B1 UNB2-H1 UNB2-H2 UNB2-H3 UNB2-B1 Date 2011-10-06 2011-10-06 2011-10-06 2011-10-13 2011-10-06 2011-10-13 2011-10-13 2011-10-13 2011-10-13 Time 13:30 14:00 14:30 17:08 20:30 ~10:30 ~11:00 ~11:30 16:29 Sample depth (m btoc) 12.8 39.75 52.2 84.3 40 16.5 34.9 40.6 40 Alkalinity mg/L 102 96.7 97.1 97 96 13.5 12.9 16.2 48.4 Aluminum mg/L <0.025 <0.025 <0.025 <0.025 0.062 <0.025 0.31 0.28 0.4 0.033 Antimony μg/L <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 Arsenic μg/L <1.5 14 5.9 6 5.9 4.4 <15 <15 <15 3.3 Barium mg/L <0.01 0.452 0.339 0.324 0.39 0.335 0.114 0.12 0.14 0.117 Boron mg/L <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 0.103 0.106 0.104 0.137 Bromide mg/L <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 6.6 6.46 6.41 5.89 Cadmium μg/L <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 Calcium mg/L <0.1 35.8 31.9 29.8 36.9 33.1 39.2 39.7 43.2 156 Chloride mg/L <0.05 9.5 7.77 7.86 7.9 7.75 729 711 731 1430 Chromium mg/L <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.1 <0.1 <0.1 <0.01 Specific conductance μS/cm 254 236 239 243 226 2410 2410 2460 4720 Copper mg/L <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.1 <0.1 <0.1 0.018 Fluoride mg/L <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 Iron mg/L <0.01 4.3 0.78 0.511 0.771 0.428 7.6 7 7.5 1.23

B27 Lead μg/L <1 <1 <1 <1 <1 <1 2.5 1.6 1.6 <1 Magnesium mg/L <0.1 2.62 3.17 3.01 3.72 3.44 33.5 33.6 33.9 65.7 Manganese mg/L <0.005 0.99 0.49 0.47 0.53 0.41 0.44 0.42 0.47 1.1 pH 7.46 7.72 7.76 7.76 7.79 6.17 6.16 6.28 6.84 Potassium mg/L <0.1 0.5 0.9 0.8 1.1 1 8.8 8.8 8.9 11 Selenium μg/L <1.5 <1.5 <1.5 <1.5 <1.5 <1.5 <15 <15 <15 7.8 Sodium mg/L <0.1 6.62 6.38 5.95 7.39 6.43 341 351 344 660 Sulfate mg/L <0.05 7.84 5.36 5.39 5.74 5.35 42.2 42.1 42.8 119 Thallium μg/L <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 Total Hardness mg/L <0.67 100 92.6 86.8 107 96.8 236 238 247 660 Turbidity NTU <0.2 4.4 2.4 2.3 6.7 0.53 2.1 2 4.3 6.6 Uranium μg/L <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 Zinc mg/L <0.005 <0.005 <0.005 <0.005 <0.005 <0.005 <0.050 <0.050 <0.050 <0.005 2- CO3 mg/L as CaCO3 0.30.50.50.50.60000 - HCO3 mg/L as CaCO3 101.7 96.2 96.5 96.4 95.4 13.5 12.9 16.2 48.4 Total Dissolved Solids mg/L as ion 192.8 174.7 172.1 182.3 174.9 1225.3 1216.3 1238.6 2509.2 Field spc. cond. (TLC 107) μS/cm Field spc. cond. (YSI 3500) μS/cm 237 3950 Field pH (YSI 3500) 7.69 6.3 Field D.O. (YSI 85) % 3.1 1.2 Field ORP (YSI 3500) mV -178 83 Field temp (YSI 3500) oC 7.5 8.7 Units Detection limit UNB3-B1 Date 2011-10-13 Time 20:41 Sample depth (m btoc) 28.5 Alkalinity mg/L 257 Aluminum mg/L <0.025 0.041 Antimony μg/L <1 <1 Arsenic μg/L <1.5 <1.5 Barium mg/L <0.01 4 Boron mg/L <0.01 0.032 Bromide mg/L <0.1 <0.1 Cadmium μg/L <0.5 <0.5 Calcium mg/L <0.1 65.6 Chloride mg/L <0.05 2.74 Chromium mg/L <0.01 <0.01 Specific conductance μS/cm 522 Copper mg/L <0.01 <0.01 Fluoride mg/L <0.1 <0.1 Iron mg/L <0.01 1.16

B28 Lead μg/L <1 <1 Magnesium mg/L <0.1 16.3 Manganese mg/L <0.005 0.99 pH 7.9 Potassium mg/L <0.1 5.4 Selenium μg/L <1.5 <1.5 Sodium mg/L <0.1 20.6 Sulfate mg/L <0.05 0.112 Thallium μg/L <1 <1 Total Hardness mg/L <0.67 231 Turbidity NTU <0.2 6.7 Uranium μg/L <0.5 <0.5 Zinc mg/L <0.005 <0.005 2- CO3 mg/L as CaCO3 1.9 - HCO3 mg/L as CaCO3 255.1 Total Dissolved Solids mg/L as ion 429.1 Field spc. cond. (TLC 107) μS/cm Field spc. cond. (YSI 3500) μS/cm 503.0 Field pH (YSI 3500) 7.37 Field D.O. (YSI 85) % 1.4 Field ORP (YSI 3500) mV -133 Field temp (YSI 3500) oC 6.7

APPENDIX C: AN EVALUATION OF THE EFFECTS OF CLIMATE CHANGE ON SEAWATER INTRUSION IN COASTAL AQUIFERS: A CASE STUDY FROM RICHIBUCTO, NEW BRUNSWICK

An evaluation of the effects of climate change on seawater intrusion in coastal aquifers: A case study from Richibucto, New Brunswick Nathan Green1 and Kerry T.B. MacQuarrie2 Department of Civil Engineering, University of New Brunswick, P.O. Box 4400, Fredericton, NB, E3B 5A3 1 email: [email protected] 2 email: [email protected]

ABSTRACT The position of the freshwater-seawater interface in a coastal aquifer is ultimately controlled by the amount of freshwater flowing towards the coastal boundaries. In turn, this is controlled by the hydraulic conductivity distribution, groundwater recharge rates, hydraulic gradients and withdrawals from the aquifer. A three-dimensional numerical model of density dependent groundwater flow coupled with solute transport was developed with the SEAWAT code for the Richibucto region of New Brunswick, Canada. This was then used to investigate the effect of climate change on the distribution of groundwater salinity within the sandstone aquifer. The model incorporated local topography, bathymetry of the surrounding tidal rivers and Northumberland Strait, stratigraphy from borehole and geophysical investigations, and well field characteristics. Based on predictions of climate change for the Richibucto area, two scenarios for variations in each of groundwater recharge and sea level rise and one scenario for pumping were developed. For the period of 2011 to 2100, the maximum cumulative sea level rise that was simulated was 1.86 m, the maximum net decline in recharge was approximately 85 mm year- 1, and the maximum net increase in municipal well field pumping was 775 m3 day-1. Simulations were performed using various combinations of the scenarios to quantify the magnitude of the effects of these three factors on the change in total dissolved solids at selected locations during the 2011-2100 simulation period. Results indicated that the relative importance of the three factors (i.e. groundwater recharge, sea level rise and pumping) varied depending on the specific location considered. The effect of generally declining recharge was the most significant at shallow to intermediate depths (i.e. less than 60 m below sea level) along the transition zone, while the effect of increasing pumping rates was most important for a location relatively close to the well field and at a depth similar to the depth of groundwater extraction. The effect of sea level rise was found to be significant only at the much deeper inland toe of the transition zone. Although sea level rise is currently receiving significant attention in developed coastal areas because of the consequences for infrastructure and coastal erosion, this investigation suggests it has the least significant effect (of the three factors considered) on future seawater intrusion in shallow to intermediate aquifers similar to those of the Richibucto region.

C1

1. INTRODUCTION Lower population densities along Canadian coastlines, compared to those of many other countries, means that the problem of seawater intrusion is seldom exacerbated to the extent commonly reported in the scientific literature. However, the history of local well fields such as the one in Richibucto, New Brunswick, demonstrate that this problem does exist in the Atlantic region. Municipal wells have been used to supply water for the town (population of approximately 1300) since 1991. During the past 10 years, some of these wells have experienced increases in chloride concentrations suggestive of seawater intrusion (e.g. Maritime Groundwater Inc. 2004; Stantec Limited 2009). The town currently supplies about 600 m³ day-1 from three wells; one located in an unconfined sandstone aquifer that is about 20 m in thickness, and two located in a deeper (~ 40 m below sea level) confined or semi-confined sandstone unit. The objective of this study was to assess the effects of climate change on the coastal aquifer utilized by the Town of Richibucto. A conceptual model was developed based on detailed topography and bathymetry, stratigraphic information from borehole and geophysical investigations, and hydrogeologic properties from previous investigations (e.g. Maritime Groundwater Inc. 2004; Stantec Limited 2009). To simulate the possible future evolution of conditions in the aquifer, a three-dimensional variable-density groundwater flow model coupled with solute transport was developed. Climate change was considered by investigating the effect of changes in recharge and sea level boundary conditions on the groundwater salinity through the year 2100. Potential increases in the pumping rates from the municipal well field were also investigated. Although there have been several recent studies dealing with the effect of climate change on coastal aquifers (e.g. Feseker et al. 2007; Carneiro et al. 2010; Oude Essink et al. 2010; Werner and Simmons 2009), this is one of the first case studies to be conducted in Atlantic Canada. The southern Gulf of St. Lawrence coastal area is generally characterized as being highly sensitive to sea level rise (Shaw et al. 1998), and the adjoining Maritime provinces have a large percentage of the population that is reliant on groundwater. Because many similar groundwater supplies exist in the Carboniferous sedimentary rocks along the shoreline of the Northumberland Strait in New Brunswick, Prince Edward Island, and Nova Scotia, it is expected that this study will provide useful insight for other locations in the region with similar hydrogeological conditions.

2. STUDY AREA DESCRIPTION The conceptual hydrogeological model for the Richibucto area includes a number of stratigraphic units. The Saint Charles peat deposit located in the centre of the study area (Figure C1) consists mainly of poorly humified (i.e. poorly decomposed) peat of Sphagnum genus, with well-humified peats making up less than 10 to 15% of the deposit (New Brunswick Department of Natural Resources 1993). The thickness of this surficial layer varies from 0 m to 7.6 m with an average of approximately 2 m. Underlying the peat deposit, and covering much of the rest of the region, is a layer of poorly sorted Wisconsinan aged sandy till with an average thickness of 3 m. The sole source of groundwater in the study area is the Carboniferous-aged Richibucto Formation that underlies the surficial materials. This predominately grey and minor brownish red, trough cross-bedded medium- to coarse-grained sandstone includes discontinuous inter- stratified red and grey very fine-grained sandstone, siltstone and mudstone layers (St. Peter and C2

Johnson 2009). The channel fill sandstones can be upwards of 20 m thick, and the overbank finer sandstone, siltstone and mudstone are typically only a few metres thick (van de Poll 1973). This model of aquifer architecture is supported by pumping tests that reveal evidence of discontinuous layers of lower permeability sediments (e.g. Rivard et al. 2008). Recharge to the groundwater system is provided by rainfall and snowmelt infiltration. Jacobs (2011) modeled the recharge to the Richibucto aquifer using past climate data and the characteristics of the surficial sandy loam and peat materials with the United States Environmental Protection Agency code HELP (Schroeder et al. 1994). Based on climate parameters averaged from 1971 to 2000, values of recharge were predicted to be, on average, 322 mm year-1 and 331 mm year-1 through the peat and sandy loam materials, respectively (Jacobs 2011).

Figure C1: Regional map of the study area bounded by the tidal St. Charles and Richibucto Rivers, showing model boundaries (red lines), location of monitoring (green symbols) and pumping wells (blue symbols), and the location of cross-section (Figure C2).

C3

3. NUMERICAL MODEL Although analytical solutions exist for approximating the location of the freshwater- seawater interface (FSI) (e.g. Bobba 1993; Mahesha and Nagaraja 1996; Werner and Simmons 2009), they are limited to conditions of homogeneous hydraulic conductivity, simple domain shapes and boundary conditions, and are often two-dimensional. Numerical models are better suited for predicting the salinity distribution in more complicated hydrogeological settings such as encountered in the current study. Although several numerical models are available to simulate density-dependent groundwater flow and solute transport, the SEAWAT code (Guo and Langevin 2002; Langevin et al. 2003; Langevin et al. 2008) was selected for this investigation because it is well established and has been frequently applied to coastal aquifers to predict the position of the FSI (e.g. Scheider and Kruse 2005; Lin et al. 2009; Chang et al. 2011; Masterson and Garabedian 2007; Praveena and Aris 2010; Rozell and Wong 2010; Webb and Howard 2010; Sanford and Pope 2010). SEAWAT iteratively couples a variable density form of the groundwater flow model MODFLOW (McDonald and Harbaugh 1988; Harbaugh et al. 2000) and the solute transport application MT3DMS (Zheng and Wang 1998).

3.1 GRID PROPERTIES Since it was assumed that local groundwater discharge occurs to the tidal rivers that effectively surround the study area (Figure C1), the boundaries of the model were chosen to include the natural constant head and salinity conditions imposed by the Northumberland Strait, the Richibucto River and the St. Charles River. The model domain was 15000 m long and 9500 m wide. Horizontal cell dimensions varied from 300 m by 300 m on the outside edges of the model, to 100 m by 100 m near the well field (Figure C2). This discretization scheme provided the resolution needed for capturing major surface features and for simulation of the coastal freshwater-seawater transition zone, while still providing efficient simulation times. The model cells represented the hydrogeological units as equivalent porous media and, as such, the results should not be used to infer local conditions near individual pumping wells, which may be largely controlled by fractures in the sandstone bedrock. The topography of the surface layer of the model was created from digital elevation data (Service New Brunswick 2010; Service New Brunswick 2001) and bathymetric data for the Richibucto harbour and the Northumberland Strait (Canadian Hydrographic Service 1988a, 1988b). Deformed grid layers were discretized initially to match local stratigraphy from records of the town well field (Water Management Services 1986; Maritime Groundwater Inc. 1991, 1992, 2004; Stantec Limited 2009) and 218 private water wells (Annie Daigle, New Brunswick Department of Environment, pers. comm.) and refined vertically to improve the resolution of the FSI (Figure C2). The model domain extended to an elevation of approximately -125 m, which is thought to be the base of the Richibucto Formation (Ball et al. 1981).

C4

Figure C2: Cross-section of model domain showing grid discretization, shale layers (shaded grey and labelled C1-C4), well locations and surface waters (constant head above cells). Open intervals of wells are shown by dotted lines. Vertical exaggeration is 20. Note that UNB2 appears to be in the harbour because of the projection onto the cross-section.

The hydraulic conductivity and effective porosity of the model layers are summarized in Table C1. Where stratigraphic layers pinched out or were not present, the model layer was set at 1 m thick and assigned the properties of the layer above it. Because little is known about the hydraulic properties of the sandstone at elevations other than those encountered by the well field, the hydraulic conductivity was decreased with depth due to increasing lithification and decreasing fracture frequency (Rivard et al. 2008). Due to poor knowledge of the aquifer properties in areas away from the well field, the model parameters were propagated laterally through the entire model domain (i.e. aquifer properties are constant throughout the same model layer). Other model properties and SEAWAT solution parameters are listed in Table C2.

3.2 BOUNDARY AND INITIAL CONDITIONS A constant head boundary was applied to the southwestern, or inland, boundary of the study area (e.g. left side of Figure C2). It was assigned a value so that the piezometric surface of the boundary cells represented a subdued replica of the land surface at approximately 3 m depth. In an attempt to match measured groundwater chemistry throughout freshwater portions of the domain, constant total dissolved solids (TDS) concentrations specified at this boundary varied from 200 mg L-1 at the surface to 400 mg L-1 at the base of the domain. At this location, hydraulic head and salinity boundary conditions remained unchanged throughout the climate change sensitivity simulations. Freshwater was also supplied to the model from recharge (mentioned previously).

C5

Table C1: Key model hydrogeological properties.

MODFLOW Material Kx = Ky Kz ne Source layer # and label (m day⁻¹) (m day⁻¹) based on Mesri et al. 1997; Letts et al. 1 Peat 1 0.5 0.9 2000; Carrier et al. 2002; Radforth 1977 2 Surficial till 0.2 0.1 0.1 based on Rivard et al. 2008 3 Sandstone 1 20 5 0.2 4 Shale (C1) 1 x 10-6 1 x 10-7 0.35 5 Sandstone 2 15 3 0.2 6 Shale (C2) 1 x 10-6 1 x 10-7 0.35 Water Management Services 1986; 7-9 Sandstone 3 10 1 0.15 Maritime Groundwater Inc. 1991, 1992, 2004; Stantec Limited 2009 10 Shale (C3) 1 x 10-6 1 x 10-7 0.35 11-13 Sandstone 4 5 0.5 0.15 14 Shale (C4) 1 x 10-6 1 x 10-7 0.35 15-17 Sandstone 5 2 0.2 0.15 18-21 Sandstone 6 1 0.1 0.1

Table C2: Key model transport and numerical parameters. Parameter Name (units) Value Longitudinal dispersivity (m) 1 Horizontal transverse dispersivity (m) 0.1 Vertical transverse dispersivity (m) 0.01 Molecular diffusion (m² day-1) 0 Specific Storage (m-1) 0.00016 Freshwater density (kg m-3) 1000 Seawater density (kg m-3) 1025 Seawater salinity (mg L-1) 35,000 Change in density over salinity (-) 0.7143 Flow Solution Preconditioned Conjugate Gradient Head change criterion (m) 1 x 10-5 Transport Solution Advection Explicit Total Variation Diminishing (TVD) solver Dispersion and source/sink terms Implicit Generalized Conjugate Gradient solver Initial time step Automatically calculated Automatically calculated based on stability criteria Maximum time step for TVD scheme

Constant head values of 0.22 m and 0.01 m were also applied to selected cells in layer 1 to represent rivers and the Northumberland Strait, respectively (Figure C1). These elevations reflect the average tidal water levels (above mean sea level) as predicted by Fisheries and Oceans Canada (2010) for Richibucto Harbour and Richibucto Bar, respectively, and were confirmed by

C6

comparison with continuously monitored water levels in each river. The constant head boundaries in the rivers and harbour were also set as constant concentration boundaries. For example, salinity at the mouth of the Mill Creek estuary (Figure C1) was set to 21 000 mg L-1 TDS, and this value was gradually increased to that of seawater (35 000 mg L-1) offshore of Kouchibouguac National Park (Koutitonsky et al. 2004). The entire depth of the model in the eastern-most column (e.g. right side of Figure C2) was also set to a constant head and concentration equal to that of seawater. All constant concentration boundaries remained unchanged throughout the climate change sensitivity simulations. The northwestern and southeastern sides of the model domain (Figure C1) were assigned as zero flux boundaries. This was deemed appropriate since they are aligned roughly parallel to that of regional groundwater flow (e.g. Rivard et al. 2008), and their distance from the well field meant they should not have a large effect on this area. Short term (i.e. daily) tidal water level fluctuations were not included in the simulations. Studies in which the influence of tidal fluctuations have been shown to be important are often performed at much finer temporal and spatial resolutions with numerical codes that can simulate unsaturated conditions (e.g. Robinson et al. 2006; Mulligan et al. 2011), or are concerned with tidal amplitudes larger than those observed at Richibucto (e.g. an amplitude of 6.4 m in Carey et al. 2009; Werner and Gallagher 2006). Because the tidal range of the Richibucto area is approximately 1.15 m, the exclusion of daily tidal water level fluctuations was not expected to have a significant influence on the long-term results that are of primary interest in this study.

3.3 DATA LIMITATIONS The spatial distribution of hydraulic conductivity is one of the key controls on aquifer flow and the position of the FSI in a coastal aquifer (e.g. Ranjan et al. 2006; Ranjan et al. 2009; Narayan et al 2007; Rozell and Wong 2010; van der Kamp 1981). Although every attempt was made to accurately reproduce the location of low hydraulic conductivity layers (e.g. identified in borehole logs) in the numerical model, stratigraphic data were not available in large areas of the model domain including under the Richibucto harbour and Northumberland Strait, which are two of the potential sources of saline water. As a result, there is uncertainty in how far offshore low hydraulic conductivity layers extend. Because salt water intrusion (SWI) is especially sensitive to the sea-aquifer connections, usually associated with the presence of preferential flow paths (Carrera et al. 2010), the simulated location and shape of the FSI may be different if different shale unit extents had been used. To illustrate this effect, Green (2012) presents results for a simulation in which the C2 confining layer (Figure C2) did not extend as far under the harbour. Although some spatially-limited groundwater salinity information was collected in the Richibucto area (Appendix A, this report), we were unable to confidently extrapolate field results to assign initial TDS concentrations to the entire domain, which is a requirement of the numerical simulations. The potential occurrence of remnant seawater, emplaced following the retreat of glaciers from this region approximately 13,000 years ago when relative sea level was significantly higher (Appendix A, this report; Webb 1982), has likewise not been considered in the current model for similar reasons. Instead, initial conditions were established by allowing the model to reach a quasi-steady state (or dynamic equilibrium) with prescribed present-day boundary conditions. This was achieved by running the model for very long periods of simulated time (e.g. Schneider and Kruse 2005; Bobba 2002; Carneiro et al. 2010; Comte and Banton 2006; Feseker et al. 2007; Kopsiaftis et al. 2009; Masterson and Garabedian 2007;

C7

Mulligan et al. 2011; Narayan et al. 2007; Park et al. in press; Watson et al. 2010; Yechieli et al. 2010). The initial conditions, along with simplifications required by the conceptual and numerical model, imply that the results obtained are unlikely to accurately depict conditions at all locations in the aquifer. This uncertainty may in part contribute to the inability of the model to reproduce current groundwater TDS concentrations at several monitoring locations near the well field. The limitations imposed by sparse hydrogeological data, combined with the inherent uncertainties in predicting future climate conditions, mean that the simulation results are best suited for investigating the general response of the aquifer to changing boundary conditions, rather than predicting local hydraulic heads or TDS concentrations.

4. CLIMATE CHANGE SCENARIOS

4.1 RECHARGE, SEA LEVEL RISE, AND PUMPING Estimates of changes in precipitation and air temperature have been determined for the Richibucto area by the Environment and Sustainable Development Research Centre et al. (ESDRC 2010). Using data from the Canadian Climate Change Scenarios Network (www.cccsn.ca), ESDRC (2010) selected the results from 10 global climate models and two emission scenarios (A1B and A2) to calculate mean deviations for climate indices for the 2020’s (2011-2040), 2050’s (2041-2070), and the 2080’s (2071-2100). For example, a mean increase in annual air temperature of 3.7°C (standard deviation of 1.2°C) was determined for the 2071 to 2100 period. Conversely, ESDRC (2010) estimated minimal and fluctuating changes to seasonal precipitation, with mean ratios consistently between 0.9 and 1.1 (i.e. ratio of future seasonal precipitation to normal precipitation for the 1971 to 2000 period). Jacobs (2011) used the predictions of ESDRC (2010) to estimate recharge through the peat and sandy loam surficial materials in the Richibucto area for the 2020’s, 2050’s, and the 2080’s. For the current study two scenarios were defined for recharge based on the results of Jacobs (2011): one reflecting the predicted changes developed by Jacobs (2011), and a more pessimistic one that doubled the percent change of those predictions relative to 1971 to 2000 conditions (Figure C3). It should be noted that the precipitation and temperature changes estimated by ESDRC (2010) result in an increasing recharge trend through the 2050’s; however, by 2100 there is a net decline in recharge for both scenarios relative to current conditions. For example, for the more pessimistic scenario the annual recharge rate for the peat at 2100 is 87 mm yr-1 less than estimated for 2011. All recharge was assigned a TDS concentration of 0 mg L-1.

C8

Figure C3: Groundwater recharge scenarios for different soil types. Values for current and predicted recharge presented by Jacobs (2011) are indicated by the data points at 2011, 2025, 2055 and 2085. A cubic spline function was used to interpolate.

R.J. Daigle Enviro (2011) provided predictions of sea level rise for Richibucto based on global sea level rise trends and estimates of local vertical motion (crustal subsidence). The anticipated change in relative sea level for the Richibucto area is a rise of 0.93 m ± 0.38 m by the year 2100. As shown in Figure C4, one model scenario conformed to these predictions and one doubled these predictions for sea level in the Richibucto harbour and Northumberland Strait. Changes in sea level were implemented consistently across all surface water boundaries. Since no information was available on how increases in sea level would be propagated inland into estuaries or harbours, it was assumed that an increase of 0.93 m by 2100 in the Northumberland Strait would result in an equivalent increase of 0.93 m in the tidal portions of the Richibucto and St. Charles Rivers. It should be noted that sea level rise was simulated by increasing the hydraulic head of spatially-constant surface water boundaries; the current study did not consider the effects of land inundation by seawater. In the current study, the municipal wells were represented by point locations of specified flux from cells intersected by the open intervals shown in Figure C2. Future pumping scenarios were applied as shown in Figure C5. One model scenario assumed that the current pumping rates for the three municipal wells remained constant until 2100, while the other considered increased pumping based on the assumption that a portion of the neighbouring community would also be supplied from the Richibucto well field, in addition to a 0.5% per year increase in water consumption. Possible alternative well locations outside of the area of the current well field were not considered.

C9

Figure C4: Sea level rise scenarios. Data points reflect increases reported by R.J. Daigle Enviro (2011) at times of 2025, 2055, 2085 and 2100 over current conditions.

Figure C5: Pumping scenarios for municipal wells PW1, PW3 and PW4. The scenario with increased pumping rates applied the changes stepwise, so that rates during 2085 to 2100 were 2.3x the 2011 values.

C10

4.2 SIMULATION DESIGN AND ANALYSIS The relative importance of changing groundwater recharge, sea level and municipal well field pumping was assessed using a 23 factorial analysis (e.g. Levine et al. 2001; Green 2012). This analysis evaluates the results from a series of experiments (i.e. simulations) to determine the effect of increasing individual factors, or combinations of factors (interaction effects), from a “low” setting to a “high” setting. In this case, the factors considered in the numerical simulations were the transient boundary conditions of recharge, sea level, and pumping rates. The “low” setting was chosen to be the current conditions, while the “high” setting was the predicted future conditions (Table C3). Since two future scenarios for recharge and sea level were tested, this required two separate factorial analyses: Group I investigated the increase of factors from current to less pessimistic future conditions, and Group II investigated the increase of factors from current to more pessimistic future conditions (Table C3 and Table C4). The result, or response variable, used to assess the effects of the changing boundary conditions was the maximum increase in TDS at specific monitoring locations (i.e. “toe”, “mid”, “out”, “007C” in Figure C2) during the simulation period of 2011 to 2100. Location “007C” was chosen to act as a sentinel location for seawater intrusion into the well field, while the “toe”, “mid” and “out” locations were chosen to monitor the effects along the FSI. In all, two factorial analyses were performed for each location, for a total of eight analyses.

Table C3: Coding used for 23 factorial experimental design (see Table C4). Group I Simulations Group II Simulations Factor "Low" "High" level "Low" "High" level level (-1) (+1) level (-1) (+1) A: Recharge Predicted 2x predicted Current Current (Figure C3) change change B: Sea level Predicted 2x predicted Current Current (Figure C4) increase increase C: Pumping Predicted Predicted Current Current (Figure C5) increase increase

Table C4: 23 scenarios for studying the effects of recharge, sea level and pumping on TDS. “+1” denotes the high level of the factor, while “-1” denotes the low level. Note that Simulations 1 and 9, and 5 and 13 are identical. Simulation Number Factors and Levels Group I Group II A: Recharge B: Sea level C: Pumping 1 9 -1 -1 -1 2 10 +1 -1 -1 3 11 -1 +1 -1 4 12 +1 +1 -1 5 13 -1 -1 +1 6 14 +1 -1 +1 7 15 -1 +1 +1 8 16 +1 +1 +1

C11

5. RESULTS AND DISCUSSION

5.1 DEVELOPMENT OF INITIAL CONDITIONS FOR CLIMATE CHANGE SIMULATIONS As previously mentioned, initial conditions for the climate change simulations were developed using present-day boundary conditions, and running the model to quasi-steady state. This was accomplished by assigning initial conditions of freshwater in the entire model domain and, using present-day recharge rates and sea level, running a transient flow and transport simulation for 5000 years until a stable TDS distribution and mass were obtained. To obtain quasi-steady state conditions under the influence of pumping, an additional 120 year simulation including pumping from the municipal wells was then preformed. A generally favourable comparison between simulated hydraulic heads and available field data was found (Green 2012). The spatial distribution of TDS at the end of the 5120 year simulation is shown in Figure C6. A base case simulation (Simulation 1) was then run to produce results for the period of 2011 to 2100 using the Figure C6 results as initial conditions and with the present-day boundary conditions unchanged. When the initial and final conditions for this simulation were compared, it was found that there were minor TDS changes occurring within the area of the FSI. To remove the influence of these changes from the simulations that included climate change scenarios, the results for Simulation 1 were subtracted from the results obtained for each of the Simulations 2 to 8, and 10 to 16. Thus, TDS changes (TDS) were obtained relative to those assuming present-day boundary conditions during the period 2011 to 2100 (i.e. Simulation 1).

Figure C6: Initial TDS (mg L-1) conditions for all simulations. Note that the results are for an enlarged portion of the model cross section shown in Figure C2. The source of the shallow zone of elevated TDS is the Richibucto harbour, while the deeper salt water wedge represents intrusion from the Northumberland Strait. Vertical exaggeration is 20.

C12

5.2 CLIMATE CHANGE SCENARIOS Several key locations (“007C”, “mid”, “out” and “toe”; Figure C2) were chosen to illustrate the temporal variations in TDS through each simulation (e.g. Figure C7); spatial changes in TDS were also visualized along cross sections (e.g. Figure C8). It can be noted, by comparing Figures C7 and C8, that the temporal trends presented for the four monitoring locations do not necessarily represent the locations of maximum TDS change. For example, the maximum TDS change of approximately 13,000 mg L-1 for Simulation 16 occurs within the FSI at an elevation of about -60 m (Figure C8). Table C5 presents the maximum TDS for each simulation at the four monitoring locations. The results show that for the “007C” location the maximum changes in TDS were less than 100 mg L-1, while changes as large as 5600 mg L-1 were obtained at the “out” location. As expected, the Group II simulations produced larger TDS changes because larger changes in sea level and groundwater recharge were imposed.

Figure C7: Example of the temporal variations of ΔTDS (mg L-1) at monitoring locations during Simulation 16 (Table C4). This particular simulation had the combination of the highest variations in recharge, the highest sea level rise, and increased pumping rates. Note the different scales for ΔTDS.

C13

Figure C8: Example of the spatial changes in TDS for Simulation 16 (Table C4). The colour scale shows the TDS (mg L-1) at the end of the simulation (2100) compared to the initial conditions while disregarding base case changes. The dashed concentration contours represent the absolute TDS for initial conditions (Figure C6), while the solid concentration contours represent the absolute TDS at end of the simulation. Vertical exaggeration is 20.

Table C5: Maximum change in TDS (mg L-1) for the four selected monitoring locations through the 89 year simulation period. Note that Simulations 1 and 9, and 5 and 13 are identical and that the results are normalized to those for Simulation 1.

Simulation "007C" "toe" "out" "mid" 1 0 0 0 0 2 14 223 2031 846 3 -3 1051 8 -48 Group I maximum 4 13 1307 2025 794 TDS 5 27 -178 18 227 results 6 49 48 2056 1093 7 22 869 25 170 8 44 1122 2052 1037 9 0 0 0 0 10 38 463 5631 1858 11 -6 2179 15 -97 Group II maximum 12 35 3116 5622 1766 TDS 13 27 -178 18 227 results 14 76 287 5664 2173 15 18 1989 31 111 16 67 2857 5654 2067

C14

The results for the factorial analyses, including normalization of the results presented in Table C5 and probability plots of the effects, are presented by Green (2012). The factorial analyses indicated that all three of the factors investigated were significant to some extent, but their relative importance changed with location (i.e. “007C”, “mid”, “out” or “toe”). The effect of variations in groundwater recharge was greatest at shallow to intermediate depths (i.e. less than 60 m below sea level) in the aquifer, which are the depths typically targeted for groundwater extraction (Maritime Groundwater Inc. 1992; Stantec Limited 2009). Sea level rise appeared to be a significant factor only near the toe of the salt water wedge that is associated with intrusion from the Northumberland Strait. The effect of the pumping rates was most significant relatively close (< ~ 2 km) to the well field (i.e. “007C”, Figure C2). At this location the maximum increase in TDS was 76 mg L-1 for the parameter combination of Simulation 14 (Table C5). This represents a relatively small increase in TDS compared to the TDS range of 21,000 to 35,000 mg L-1 in the FSI; however, increases of this magnitude may be associated with increased production of brominated disinfection by-products in treated drinking water (e.g. Zhai et al. 2010; Sohn et al. 2006). No significant seawater intrusion into the well field was found for the current set of simulations. The spatial variability concerning the relative importance of groundwater recharge, sea level rise, and pumping is likely related to the type of hydraulic control on the freshwater flow. Head controlled freshwater systems are dominated by specified head, or Dirichlet, boundaries (e.g. non-tidal surface water features). Flux controlled systems are dominated by specified flux, or Neuman, conditions (e.g. recharge). Numerical models of confined and unconfined aquifers dominated by flux controlled freshwater boundaries have been found to be relatively insensitive to the effects of sea level rise when compared to those where freshwater flow is controlled by specified head boundaries (Werner and Simmons 2009; Chang et al. 2011). Because the tidal rivers in the current numerical simulations are connected to the Northumberland Strait, their elevations were assumed to increase at the same rate as the sea level rise predictions. This resulted in similar increases to freshwater heads in the interior of the model domain, indicating that the model tends to be flux controlled. This may explain the relative insensitivity of the TDS results to sea level rise. For the majority of the modeled aquifer thickness, it appears that sea level rise by itself is not causing a decrease in the freshwater flow. This factor only becomes significant at a depth where the influence of recharge flux diminishes and constant head boundaries (simulating regional inflow of freshwater on one side and sea level on the other side) become more important. A number of recent studies have revealed that freshwater boundary conditions can control how an aquifer will respond to climate change, and that sea level rise may have a relatively insignificant effect on the position of the FSI. For example, Loáiciga et al. (2012) used a numerical model, in which the freshwater boundaries were controlled by constant flux conditions, to model seawater intrusion near Monterey, California. They found that groundwater extraction was the predominate cause of seawater intrusion in their study area, and that sea level rise had little effect. Although Loáiciga et al. (2012) did not test the effect of declining recharge, their results support the current finding that pumping may be more significant to SWI than sea level rise. Similarly, Carneiro et al. (2010) investigated the effects of decreasing recharge and rising sea level on a surficial aquifer in Morocco. They simulated three future climate and sea level rise scenarios and found that the reduction in freshwater flow volume responsible for the

C15

increased salinity was attributed mainly to a combination of declining recharge and declining contributions from an adjacent aquifer, rather than to the predicted increase in sea level.

6. CONCLUSIONS A conceptual and numerical model of variable density groundwater flow coupled with solute transport was developed using available hydrogeological data for Richibucto, NB. Although there are limitations with the model, notably a lack of hydrogeological properties and groundwater TDS data in regions outside the municipal well field, it is considered a suitable tool for assessing the general response of the sandstone aquifers to climate change and sea level rise. We considered two scenarios each for future changes in recharge and sea level and one scenario for pumping for the period from 2011 to 2100. Scenarios of groundwater recharge were based on predictions made by Jacobs (2011); one scenario had a net decline in groundwater recharge of 40 mm year-1 by 2100, while the other had a net decline of approximately 85 mm year-1 by 2100. One sea level scenario considered a net rise of 0.93 m, while the other considered a net rise of 1.86 m during the simulation period. Pumping rates from the municipal well field increased by a factor of 2.3 for the future scenario. Various combinations of these scenarios were simulated in two groups and the maximum change in total dissolved solids at selected locations within the model during the 2011-2100 simulation period was used as the model response. The results indicate that the relative importance of the three factors (i.e. groundwater recharge, sea level rise and pumping) changes depending on the specific location considered. The effect of generally declining recharge is the most significant at shallow to intermediate depths (i.e. less than 60 m below sea level) along the transition zone, while the effect of increasing pumping rates was most important for a location relatively close to the well field and at a depth similar to the depth of groundwater extraction. The effect of sea level rise was found to be significant only at the much deeper inland toe of the transition zone. Not surprisingly, the more pessimistic changes to boundary conditions resulted in larger effects on the seawater distribution, but the relative significance of the factors was similar to that obtained for the less pessimistic scenarios. The variability of the most significant factors with position can at least in part be explained by the spatial variability of the controls on freshwater flow. For example, the importance of recharge diminishes with depth, while the overall hydraulic gradient imposed by the boundary conditions becomes more significant (i.e. increasing the significance sea level rise). Although sea level rise is currently receiving significant attention in developed coastal areas because of the consequences for infrastructure and coastal erosion, this investigation suggests it has the least significant effect (of the three factors considered) on future seawater intrusion in shallow to intermediate aquifers similar to those of the Richibucto region. This finding may be encouraging to coastal water resource planners. Indeed, local and regional districts will not be able to directly control changes in climate and sea level. However, because groundwater recharge is a process that is influenced by land use, and pumping rates and locations can be managed, steps to control these two factors may prove to be beneficial in protecting coastal fresh groundwater supplies.

C16

ACKNOWLEDGEMENTS Funding for this research was provided by the Atlantic Regional Adaptation Collaborative (RAC) Climate Change Program initiated by NRCan and the Departments of Environment of the four Atlantic provinces. The program is being managed by the Atlantic Climate Adaptation Solutions Association (ACASA). We thank the Town of Richibucto, especially Supervisor of Public Works and Utilities, Sean Sullivan, for help with access and logistics. We thank Robert Hughes, Annie Daigle, Mallory Gilliss, Darryl Pupek and Sherry McCoy of the New Brunswick Department of Environment and Local Government for their interest and support of the project. We also thank UNB colleagues Eric Mott, Karl Butler and Dennis Connor for their help in the field, and for their insights and cooperation in the project as a whole.

REFERENCES Ball, F. D., Sullivan, R. M., and Peach, A. R. (1981). "Carboniferous Drilling Project." Report of Investigation 18, New Brunswick Department of Natural Resources, Mineral Development Branch, Fredericton, New Brunswick. Bobba, A. G. (1993). "Mathematical models for saltwater intrusion in coastal aquifers." Water Resour. Manag., 7(1), 3-37. Bobba, A. G. (2002). "Numerical modelling of salt-water intrusion due to human activities and sea-level change in the Godavari Delta, India." Hydrolog. Sci. J., 47, 67-80. Canadian Hydrographic Service. (1988a). "Chart 4909 - Détroit de Northumberland/Northumberland Strait - Partie Ouest/Western Portion - Ports / Harbours." 30,000. Canadian Hydrographic Service. (1988b). "Chart 4906 - West Point à/to Baie de Tracadie." 100,000. Carneiro, J. F., Boughriba, M., Correia, A., Zarhloule, Y., Rimi, A., and Houadi, B. E. (2010). "Evaluation of climate change effects in a coastal aquifer in Morocco using a density-dependent numerical model." Environ. Earth Sci., 61(2), 241-252. Carrera, J., Hidalgo, J.J., Slooten, L.J., and Vazquez-Sune, E. (2010). “Computational and conceptual issues in the calibration of seawater intrusion models.” Hydrogeology Journal, 18(1), 131-145. Carrier, C., Michaud, Y., Crowe, A. S., and Allard, M. (2002). "Caracterisation hydrogeologique d'une tourbiere ombrotrophe dans le sud-est du Nouveau-Brunswick: resultats preliminaries." Proceedings of the 3rd Joint IAH-CNC/CGS Conference, Ground and Water: Theory to Practice, 227-234. Comte, J., and Banton, O. (2006). "Modelling of Seawater Intrusion in the Magdalen Islands (Quebec, Canada)." Proceedings 1st SWIM-SWICA Joint Saltwater Intrusion Conference, 303-310. Datta, B., Vennalakanti, H., and Dhar, A. (2009). "Modeling and control of saltwater intrusion in a coastal aquifer of Andhra Pradesh, India." J. Hydro-Environ. Res., 3(3), 148-159. Environment and Sustainable Development Research Centre (ESDRC). (2011). "Climate Change Scenarios New Brunswick Municipalities." ETF Project Number 09-0218, Prepared for New Brunswick Department of Environment, Fredericton, New Brunswick. Feseker, T. (2007). "Numerical studies on saltwater intrusion in a coastal aquifer in northwestern Germany." Hydrogeol. J., 15(2), 267-279. Fisheries and Oceans Canada. (2010). "Tides, Currents and Water Levels." http://www.lau.chs- shc.gc.ca/english/Canada.shtml (September 29, 2011).

C17

Green, N. (2012). “A case study of the effects of climate change on seawater intrusion in coastal aquifers in New Brunswick, Canada”. Master of Science in Engineering thesis, University of New Brunswick, Fredericton. Guo, W., and Langevin, C. D. (2002). User's guide to SEAWAT; a computer program for simulation of three-dimensional variable-density ground-water flow. U. S. Geological Survey, Reston, VA. Harbaugh, A. W., Banta, E. R., Hill, M. C., and McDonald, M. G. (2000). "MODFLOW-2000, The U.S. Geological Survey Modular Groundwater Model - User Guide to Modularization Concepts and the Groundwater Flow Process." Open File Report 00-92, United States Geological Survey, Reston, VA. Jacobs, E. E. (2011). "The Application of the HELP Model to Examine the Potential Impact of Climate Change on Recharge Rates in Coastal Regions of New Brunswick." CE5943: Senior Report. Unpublished. University of New Brunswick, Fredericton. Koutitonsky, V. G., Guyondet, T., St-Hilaire, A., Courtenay, S. C., and Bohgen, A. (2004). "Water renewal estimates for aquaculture developments in the Richibucto estuary, Canada." Estuaries, 27(5), 839-850. Langevin, C. D., Shoemaker, W. B., and Guo, W. (2003). "MODFLOW-2000, the U.S. Geological Survey Modular Ground-Water Model; documentation of the SEAWAT-2000 Version with the Variable-Density Flow Process (VDF) and the Integrated MT3DMS Transport Process (IMT)." Open File Report 03-0426, U. S. Geological Survey, Reston, VA. Langevin, C. D., Thorne, D. T., Jr, Dausman, A. M., Sukop, M. C., and Guo, W. (2008). SEAWAT Version 4; A computer program for simulation of multi-species solute and heat transport. U. S. Geological Survey, Reston, VA. Letts, M. G., Roulet, N. T., Comer, N. T., Skarupa, M. R., and Verseghy, D. L. (2000). "Parametrization of peatland hydraulic properties for the Canadian Land Surface Scheme." Atmos. Ocean, 38(1), 141- 160. Levine, D. M., Ramsey, P. P., and Smidt, R. K. (2001). "Factorial designs involving three or more factors." Applied statistics for engineers and scientists: using Microsoft Excel and MINITAB, Prentice Hall, New Jersey, 525-541. Mahesha, A., and Nagaraja, S. H. (1996). "Effect of natural recharge on sea water intrusion in coastal aquifers." J. Hydrol., 174(3-4), 211-220. Maritime Groundwater Inc. (1991). "Drill Targets for the Town of Richibucto Water Supply." Unpublished report. Maritime Groundwater Inc. (1992). "Drilling and Yield Testing for the Town of Richibucto, NB; November - December 1991." Unpublished report. Maritime Groundwater Inc. (2004). "Hydrogeological Assessment - Water Source Supply Assessment, Town of Richibucto, N.B." Unpublished report. McDonald, M. G., and Harbaugh, A. W. (1988). A modular three-dimensional finite-difference groundwater flow model. United States Geological Survey, Reston, VA. Mesri, G., Stark, T. D., Ajlouni, M. A., and Chen, C. S. (1997). "Secondary compression of peat with or without surcharging." J. Geotech. Geoenviron. Eng., 123(5), 411-420. Mott, E., Butler, K., Green, N., and MacQuarrie, K. (2011). “Application of ERT to the investigation of seawater intrusion at Richibucto, New Brunswick.” Proceedings of GeoHydro 2011, Water and Earth: The Junction of Quaternary Geoscience and Hydrogeology.

C18

Narayan, K. A., Schleeberger, C., and Bristow, K. L. (2007). "Modelling seawater intrusion in the Burdekin Delta Irrigation Area, North Queensland, Australia." Agr. Water Manage., 89(3), 217-228. New Brunswick Department of Natural Resources (NBDNR). (1993). "Peatland Area 213 - Report, Isopach, Elevations and Profiles." New Brunswick Department of Natural Resources. Fredericton, New Brunswick. Oude Essink, G. H. P., Van Baaren, E. S., and De Louw, P. G. B. (2010). "Effects of climate change on coastal groundwater systems: A modeling study in the Netherlands." Water Resour. Res., 46(10), W00F04. Park, S., Kim, J., Yum, B., and Yeh, G. (in press). "Three-dimensional numerical simulation of salt water extraction schemes to mitigate seawater intrusion due to groundwater pumping in a coastal aquifer system." J. Hydrologic Eng. R. J. Daigle Enviro. (2011). "Sea-Level Rise Estimates for New Brunswick Municipalities." Report prepared for the Atlantic Climate Adaptation Solutions Association. Radforth, N. W. (1977). "Muskeg hydrology." Muskeg and the Northern Environment in Canada, N. W. Radforth, and C. O. Brawner, eds., University of Toronto Press, Toronto, Ontario, 82-147. Ranjan, P. S., Kazama, S., and Sawamoto, M. (2006). "Effects of climate and land use changes on groundwater resources in coastal aquifers." J. Environ. Manage., 80(1), 25-35. Ranjan, P. S., Kazama, S., Sawamoto, M., and Sana, A. (2009). "Global scale evaluation of coastal fresh groundwater resources." Ocean Coast. Manage., 52(3-4), 197-206. Rivard, C., Michaud, Y., Lefebvre, R., Deblonde, C., and Rivera, A. (2008). "Characterization of a regional aquifer system in the Maritimes Basin, Eastern Canada." Water Resour. Manag., 22(11), 1649-1675. Sanford, W. E., and Pope, J. P. (2010). "Current challenges using models to forecast seawater intrusion: lessons from the Eastern Shore of Virginia, USA." Hydrogeol. J., 18(1), 73-93. Service New Brunswick. (2010). "Geographic Data and Maps." http://www.snb.ca/gdam-igec/e/2900e.asp (February 8, 2010). Service New Brunswick. (2001). "User Guide to the Digital Topographic Data Base 1998 (DTDB98) of New Brunswick." Service New Brunswick. Schroeder, P. R., Aziz, N. M., Lloyd, C. M. and Zappi, P. A. (1994). "The Hydrologic Evaluation of Landfill Performance (HELP) Model: User’s Guide for Version 3." Rep. No. EPA/600/R-94/168a. U.S. Environmental Protection Agency Office of Research and Development, Washington, DC. Shaw, J., Taylor, R. B., Solomon, S., Christian, H. A. and Forbes, D. A. (1998) “Potential impacts of global sea-level rise on Canadian coasts.” The Canadian Geographer, 42(4), 365-379. Sohn, J., Amy, G., and Yoon, Y. (2006). "Bromide ion incorporation into brominated disinfection by- products." Water Air Soil Pollut., 174(1-4), 265-277. St. Peter, C. J., and Johnson, S. C. (2009). " New Brunswick Bedrock Lexicon: Richibucto Formation." http://dnr-mrn.gnb.ca/Lexicon/Lexicon/Lexicon_View.aspx (June 3, 2011). Stantec Limited. (2009). "Installation and Aquifer Test Analysis of Production Well PW08-01, Town of Richibucto, NB." Unpublished report. van der Kamp, G. (1981). "Salt-water intrusion in a layered coastal aquifer at York Point, Prince Edward Island." NHRI Paper No. 14; IWD Scientific Series No. 121. National Hydrology Research Institute, Inland Waters Directorate, Ottawa, Canada.

C19 van de Poll, H. W. (1973). "Stratigraphy, Sediment Dispersal and Facies Analysis of the Pennsylvanian Pictou Group in New Brunswick." Marit. Sediments, 9(3), 72-77. Water Management Services Limited. (1986). "Test Drilling and Pumping for the Village of Richibucto, 1986." Unpublished report. Webb, T.C. (1982) “Origin of brackish formational ground waters in the Pennsylvanian of New Brunswick.” Open File Report 82-5, New Brunswick Department of Natural Resources, Mineral Development Branch, Fredericton, New Brunswick, 21 p. Werner, A. D., and Gallagher, M. R. (2006). "Characterisation of sea-water intrusion in the Pioneer Valley, Australia using hydrochemistry and three-dimensional numerical modelling." Hydrogeol. J., 14(8), 1452-1469. Werner, A. D., and Simmons, C. T. (2009). "Impact of sea-level rise on sea water intrusion in coastal aquifers." Ground Water, 47(2), 197-204. Zhai, X., Liu, J., Liu, F., and Guo, Y. (2010). "Trihalomethane formation of source water contaminated by seawater intrusion." 2nd Conference on Environmental Science and Information Application Technology, ESIAT 2010, 565-568. Zheng, C., and Wang, P. (1999). "MT3DMS: A modular three-dimensional multispecies transport model for simulation of advection, dispersion, and chemical reactions of contaminants in groundwater systems; documentation and user’s guide." Contract Report SERDP-99-1, U.S. Army Engineer Research and Development Center, Vicksburg, MS.

C20