Atlantic Climate Adaptation Solutions Association Solutions d'adaptation aux changements climatiques pour l'Atlantique
A case study of coastal aquifers near Richibucto, New Brunswick: 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, Fredericton, 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 Nova Scotia, New Brunswick, Prince Edward Island and Newfoundland and Labrador and supported through the Regional Adaptation Collaborative, a joint undertaking between the Atlantic provinces, Natural Resources Canada 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 Richibucto River/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 Northumberland Strait 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, Quebec 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 Moncton 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: