Synthesis review on groundwater discharge to surface water in the Great Lakes Basin

JGLR-00708; No. of pages: 10; 4C: Journal of Great Lakes Research xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Journal of Great Lakes Research journal homepage: www.elsevier.com/locate/jglr

Review

Synthesis review on groundwater discharge to surface water in the Great Lakes Basin Kurt C. Kornelsen a,⁎, Paulin Coulibaly a,b,1 a b

School of Geography and Earth Science, McMaster University, Hamilton, Ontario L2S 4L8, Canada Department of Civil Engineering, McMaster University, Hamilton, Ontario L2S 4L8, Canada

a r t i c l e

i n f o

Article history: Received 17 August 2013 Accepted 28 February 2014 Available online xxxx Communicated by Harvey Thorleifson Index words: Groundwater discharge Groundwater–surface water interaction Hyporheic Scaling Riparian

a b s t r a c t Groundwater in the Great Lakes Basin (GLB) serves as a reservoir of approximately 4000 to 5500 km3 of water and is a significant source of water to the Great Lakes. Indirect groundwater inflow from tributaries of the Great Lakes may account for 5–25% of the total water inflow to the Great Lakes and in Lake Michigan it is estimated that groundwater directly contributes 2–2.5% of the total water inflow. Despite these estimates, there is great uncertainty with respect to the impact of groundwater on surface water in the GLB. In terms of water quantity, groundwater discharge is spatially and temporally variable from the reach to the basin scale. Reach scale preferential flow pathways in the sub-surface play an important role in delivering groundwater to surface water bodies, however their identification is difficult a priori with existing data and their impact at watershed to basin scale is unknown. This variability also results in difficulty determining the location and contribution of groundwater to both point and non-point source surface water contamination. With increasing human population in the GLB and the hydrological changes brought on by continued human development and climate change, sound management of water resources will require a better understanding of groundwater surface–water interactions as heterogeneous phenomena both spatially and temporally. This review provides a summary of the scientific knowledge and gaps on groundwater–surface water interactions in the GLB, along with a discussion on future research directions. © 2014 International Association for Great Lakes Research. Published by Elsevier B.V. All rights reserved.

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Overview of groundwater–surface water interactions . . . . . . . . . . . . . Quantity of groundwater discharge in the Great Lakes Basin . . . . . . . . . . Direct discharge to the great lakes . . . . . . . . . . . . . . . . . . . . Discharge to streams and tributaries in the Great Lakes Basin . . . . . . . Quality of groundwater discharge in the Great Lakes Basin . . . . . . . . . . . Great Lakes Basin scale water quality . . . . . . . . . . . . . . . . . . Hyporheic and riparian zones and residence time influences on water quality Groundwater–surface water contribution to non-point source pollution . . . Groundwater–surface water contribution to point source pollution . . . . . Impacts of environmental change on groundwater–surface water interactions . . Potential impacts of human development . . . . . . . . . . . . . . . . Potential impacts of climate change . . . . . . . . . . . . . . . . . . . Temporal variability of groundwater–surface water interactions . . . . . . . . Discussion and conclusions . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

⁎ Corresponding author. Tel.: +1 905 525 9140x27875. E-mail addresses: [email protected] (K.C. Kornelsen), [email protected] (P. Coulibaly). 1 Tel.: +1 905 525 9140x23354.

http://dx.doi.org/10.1016/j.jglr.2014.03.006 0380-1330/© 2014 International Association for Great Lakes Research. Published by Elsevier B.V. All rights reserved.

Please cite this article as: Kornelsen, K.C., Coulibaly, P., Synthesis review on groundwater discharge to surface water in the Great Lakes Basin, J Great Lakes Res (2014), http://dx.doi.org/10.1016/j.jglr.2014.03.006

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

2

K.C. Kornelsen, P. Coulibaly / Journal of Great Lakes Research xxx (2014) xxx–xxx

Introduction The Great Lakes Basin contains approximately 20% of the world's supply of fresh surface water and serves as a groundwater reservoir which Granneman et al. (2000) estimated to be 4000 km3 of groundwater for the entire GLB and Coon and Sheets (2006) estimated to be 5500 km3 of groundwater for the U.S. portion of the GLB only. Traditionally, groundwater and surface water stores have been considered as independent resources from the perspective of both water quality and quantity. However, the inter-connections between groundwater and surface water create a requirement to consider both as a single water resource (Winter et al., 1998) and there is a need to develop an integrated understanding of groundwater–surface water interactions (GWSWI), including riparian, hyporheic and biogeochemical processes that impact groundwater. Groundwater discharge is estimated to directly contribute to approximately 2–2.5% of the annual water budget for Lake Michigan (Feinstein et al., 2010; Granneman et al., 2000) while discharge from groundwater to tributaries is responsible for an estimated 40–75% of the tributary inflow to the Great Lakes and is therefore important for the overall health of the Great Lakes ecosystem (Neff et al., 2005). However, determining the correct impact of the role of groundwater in the Great Lakes Basin (GLB) is difficult as direct groundwater contribution is highly uncertain and poorly estimated (Neff and Nicholas, 2004) and baseflow separation estimates also vary widely as seen in Fig. 2 (Neff et al., 2005). In light of the significant role of groundwater as a source to surface waters in the Great Lakes, there is a need for a thorough understanding of GWSWI in terms of water quantity as well as an understanding of the role played by GWSWI in the management of surface water quality in the GLB. Despite early recommendations for such an integrated understanding of GWSWI (IJC, 2010) the heterogeneity throughout the basin, lack of available data and limitations of research methods have hampered efforts to consider groundwater and surface water as a single resource. With increased pressures on water resources as a result of continued human development and the potential impacts of climate variability, a comprehensive baseline understanding of the role of GWSWI is critical for the forecasting, management and preservation of this important resource. To promote a comprehensive understanding of the impact of groundwater on surface water in the GLB a review has been conducted which encompasses peer-reviewed scientific literature, reports and other studies specific to the Great Lakes or which are considered relevant for extension to the Great Lakes. The approach selected for the literature review was to search based on combinations of keywords ‘groundwater’, ‘surface water’ or ‘groundwater surface water interactions’ in combination with ‘Great Lakes’ and each of the Great Lake names individually. The material presented herein is derived from an unpublished report produced for Environment Canada (Coulibaly and Kornelsen, 2013) where specific emphasis was placed on the role of scale and spatial–temporal heterogeneity of groundwater discharge to surface waters in the Ontario portion of the GLB. Herein, the review covers the entire GLB.

floods (Herman et al., 2001). A brief overview of select groundwater– surface water interactions (GWSWI) is provided, however the reader is referred to Winter et al. (1998), Granneman et al. (2000) and Sophocleous (2002) for a more comprehensive description of the processes. Groundwater, as a mobile part of the hydrosphere, can be loosely grouped into two categories, deep or regional aquifers, which underlay relatively large geographic areas and shallow aquifers which are local in scope. While both are sources of groundwater discharge, shallow aquifers represent shorter flow pathways, are more susceptible to contamination and provide a significantly larger contribution to surface waters (Granneman et al., 2000; Sophocleous, 2002; Winter et al., 1998). Infiltration of precipitation and surface water causes a rise in the water table, resulting in increased flow toward surface water bodies and producing baseflow in many streams. This process is reversed during dry periods when the water table lowers, decreasing or reversing groundwater flow. Lateral groundwater flow also occurs in the unsaturated zone, particularly following precipitation, where a rapid increase in the soil water content results in interflow (shallow subsurface flow) (Sophocleous, 2002). The build-up of pressure displaces ‘old water’, stored in the soil, into surface water bodies, producing a considerable

Overview of groundwater–surface water interactions The discharge of groundwater to surface water bodies occurs through a variety of flow pathways, the interaction of which is governed by relative potential of the groundwater table, the surface water level and the potential rate of flow (hydraulic conductivity) of the system. While there are many definitions of an aquifer, it is considered here in a loose sense to refer to subsurface areas (or geologic units) that can store and transmit water. It is also noteworthy that not all groundwater is stored in ‘aquifers’ as saturated zones occasionally form above what would traditionally be considered an aquifer. Generally, groundwater can provide a range of ecosystem goods and services such as water purification, maintaining ecosystem health and mitigation of erosion and

Fig. 1. Mean BFI (a) and the range of BFI values (b) produced by different hydrograph separation methods for each watershed using data from Neff et al. (2005).

Please cite this article as: Kornelsen, K.C., Coulibaly, P., Synthesis review on groundwater discharge to surface water in the Great Lakes Basin, J Great Lakes Res (2014), http://dx.doi.org/10.1016/j.jglr.2014.03.006

K.C. Kornelsen, P. Coulibaly / Journal of Great Lakes Research xxx (2014) xxx–xxx

portion of the storm hydrograph in many watersheds. Also, during large storm events GWSWI plays an important role in mitigating floods, as excess water in streams can infiltrate the stream bank and be retained as bank-storage, which is subsequently released over time (Winter et al., 1998). Due to stable temperatures, groundwater discharge zones provide regions of high biodiversity and productivity which are critical to the maintenance of ecosystem health in streams, lakes and wetlands (Sophocleous, 2002). However, since shallow subsurface flow is a major pathway for water, it is also an important pathway for contaminants. Nutrients, agri-chemicals and other non-point source contaminants infiltrate the shallow aquifer with precipitation and are subsequently transferred into surface waters. Surface water entering groundwater is usually most significantly chemically altered within the first few meters (Sophocleous, 2002) although aquifers in general provide an ecosystem service and potentially remove contamination. The rate of groundwater discharge to surface waters varies through time, but transitions are generally gradual. Changes in the rate of groundwater discharge occur due to both seasonality of recharge, as well as the seasonality of groundwater removal by evapotranspiration, particularly in the mid-latitudes, demonstrating the important relationship between vegetation and GWSWI (Sophocleous, 2002; Winter et al., 1998). Also important is the spatial variability of groundwater discharge which varies regionally with aquifer properties (Feinstein et al., 2010; Meriano and Eyles, 2002) as well as at the catchment to reach scales with streambed properties (Conant, 2004; Drake et al., 2010). Quantity of groundwater discharge in the Great Lakes Basin The amount of groundwater stored in the U.S. portion of the Great Lakes Basin has been estimated at 5500 km3, with uncertainty in this estimate due to assumptions about representative aquifer properties, resulting in a possible range of values between 1900 km3 and 9200 km3 (Coon and Sheets, 2006). The amount of groundwater stored in the GLB has increased by 10–20 mm/yr over the past decade, with the greatest increases along the Michigan Peninsula (Wang et al., 2012). Direct groundwater contributions to Lake Michigan were estimated to only account for approximately 2% of the annual water budget for the basin (Granneman et al., 2000) whereas indirect groundwater discharge to the entire GLB from tributaries has been estimated between 22 and 42% of the water budget for the Great Lakes (Holtschlag and Nicholas, 1998). Differences in aquifer properties, recharge rates and water sources for the lakes result in discharge estimates that vary both between and within each of the Great Lakes. Model results in the Oak Ridges Moraine and Rouge River Watersheds are highly spatially variable and suggest that 83% of groundwater recharge in this area is eventually released to streams as baseflow, 64% of which is from the upper aquifer, and discharge to springs and Lake Ontario accounts for 12% of the total outflow (Meriano and Eyles, 2002). Granneman et al. (2000) point out that most studies of groundwater focus on regional scale aquifer systems, whereas the majority of groundwater discharge to surface waters occur as a result of the shallow groundwater flow systems. A visible example occurs in Lake Ontario's Rouge River and Highland Creek Watersheds, where 92% of direct groundwater discharge seeps through the exposed upper and middle aquifers in the Scarborough Bluffs (Meriano and Eyles, 2002). Direct discharge to the great lakes

With respect to direct groundwater discharge to a Great Lake, the greatest concentration of research has been on Lake Michigan as well as the Michigan Peninsula and the lakes to which it contributes. A survey of results from Michigan provides an example of the heterogeneity which must be considered at the scale of a Great Lake. Direct discharge to Lake Michigan accounts for 2.6% (76.5 m3/s) of the total basin inputs, which, in contrast, is much smaller than the groundwater discharge to streams which account for 31.2% (906 m3/s) of water inputs (Granneman et al., 2000). In comparison to the 2.6% basin wide estimate, there are regional differences in groundwater discharge. Along the Door Peninsula 43% of water that enters the upper aquifer is quickly discharged to the lake (Cherkauer et al., 1992). In the Grand Traverse Bay, direct groundwater discharge accounts for approximately 7% (0.23 m3/s) of the total water budget (Boutt et al., 2001), while only accounting for 5% (1.8 m3/s) of the total water budget of the lower Michigan Peninsula (Hoaglund et al., 2002). In Lake Huron's Saginaw Bay, direct groundwater discharge is estimated at 1.13 m3/s (Hoaglund et al., 2002, 2004), although chloride transport times to the bay were found to be shorter than suggested by transport modeling, suggesting a mechanism of preferential flow through fractures (Kolak et al., 1999). During the International Field Year for the Great Lakes (IFYGL), watersheds that were considered to be hydro-geologically representative were selected and monitored. Table 1 shows the estimates of groundwater discharge for the individual watersheds. When findings were assumed representative of the shoreline region it was calculated that between the Niagara and Humber Rivers 0.2 m3/s was discharged (Ostry, 1979a). Between the Humber River and Oshawa 0.14 m3/s was discharged with the majority of that discharge concentrated around the Scarborough Bluffs (Ostry, 1979b). Along the shoreline representing the Trent to St. Lawrence Rivers a total of 1.2 m3/s was estimated to be discharged, with approximately half of that value originating from Prince Edward County (Ostry and Singer, 1981). In Hamilton Harbour, Harvey et al. (2000) estimated the direct groundwater contribution to the harbor at 0.7 m3/s and Meriano and Eyles (2002) estimated that 17% (0.3 m3/s) of the recharge in the Rouge River and Highland Creek Watersheds was directly released to Lake Ontario. The variability of IFYGL and subsequent measurements demonstrates the impact of spatial variability on direct groundwater discharge. Given the logistical limitations of representative and continuous in situ measurements, modeling work at the basin scale such as was conducted by Feinstein et al. (2010) could provide scientists and decision makers much needed quantitative information about groundwater discharge in the GLB. From a water management perspective, direct discharge only contributes a small proportion of water to the Great Lakes however, its impact on water quality and human populations may be relatively disproportionate to its volume. Processes which result in greater discharge in the near-shore zone (Feinstein et al., 2010) and concentration of groundwater discharge in bays (Cherkauer et al., 1992) may result in greater concentrations of groundwater in populated regions (i.e. Traverse City; Feinstein et al., 2010), although more work is required before definitive conclusions can be drawn.

Table 1 Estimated groundwater discharge to Lake Ontario during the IFYGL. Source: Ostry (1979a,b) and Ostry and Singer (1981). Watershed

Direct groundwater discharge to large water bodies, such as the Great Lakes, is generally highest in the near shore areas and decays offshore (Cherkauer and Taylor, 1990; Feinstein et al., 2010; Harvey et al., 2000). Despite this general trend, unique systems such as the Lake Huron sinkhole (Ruberg et al., 2008) and heterogeneous near shore features such as sand lenses (Harvey et al., 2000) provide pathways of preferential discharge to the lakes.

3

Discharge rate

Total discharge m3/s

L/s/km Forty Mile Creek Oakville Creek Duffins Creek Moira River Wilton Creek Thousand Islands

8.1 2.6 4.2 7.1 4.2 11.3

Scarborough Bluffs Upper Aquifer Lower Aquifer Prince Edward County Shoreline

L/s

0.05 0.09

50 90

0.63

630

Please cite this article as: Kornelsen, K.C., Coulibaly, P., Synthesis review on groundwater discharge to surface water in the Great Lakes Basin, J Great Lakes Res (2014), http://dx.doi.org/10.1016/j.jglr.2014.03.006

4

K.C. Kornelsen, P. Coulibaly / Journal of Great Lakes Research xxx (2014) xxx–xxx

Table 2 Basin Averaged Baseflow Index (%). Source: Neff et al. (2005). Basin

Min BFI

Max BFI

Mean BFI

Superior Michigan Huron Erie Ontario

44.5 47.2 46.1 34.2 38.7

81.0 78.5 77.1 61.9 67.8

64.7 63.0 63.7 50.7 55.3

Discharge to streams and tributaries in the Great Lakes Basin Where direct groundwater discharge to the Great Lakes is relatively minor, indirect groundwater discharge to contributing streams and tributaries is a major source of water within the basin. In streams that have not been impacted by human development, groundwater discharge provides the primary source of baseflow, with some baseflow originating from wetland discharge and anthropogenic sources. Granneman et al. (2000) suggest that most groundwater flow in the Great Lakes Region occurs in shallow flow paths. This was demonstrated to be the case for the Rouge River and Highland Creek where 64% of baseflow originates in the upper aquifer and only 12% and 13% originate in the middle and lower aquifers respectively (Meriano and Eyles, 2002). During a storm event in the Kintore Creek catchment, Cey et al. (1998) observed that rainfall infiltrated the soil and forced groundwater into the creek. The result was that even during storms, pre-event water contributed greater than 60% of the total flow. This finding reveals that GWSWI is also important during periods which are not traditionally considered as baseflow events. At the scale of the entire GLB, Holtschlag and Nicholas (1998) regionalized baseflow in selected tributaries from the eight Great Lakes States to estimate the indirect groundwater component of the water budget. Neff et al. (2005) extended the baseflow approach by including a greater number of gaged basins, including many in Canada, and by using six hydrograph separation and two regionalization methods to provide an estimate of uncertainty. Fig. 2 shows a synthesis of the results of Neff et al. (2005) which shows the range in baseflow index (BFI) from six hydrograph separation and methods with the two regionalization models. Table 2 shows a basin averaged synthesis of the results of Neff et al. (2005) showing the mean of the minimum, maximum and average BFI estimates from the 12 hydrograph separation/ regionalization methods. It is obvious that different portions of the GLB

have different stream baseflow contributions, but it is also evident that there is still significant uncertainty when it comes to estimating groundwater discharge using baseflow due to the sometimes large discrepancies in BFI estimates. The groundwater contribution to total stream flow is highest in the Lake Superior, Michigan and Upper Huron basins and is lowest in Lakes Erie, Ontario and lower Lake Huron. The connections of the lower lakes to the upper lakes also result in a lower portion of the total water budget in Lakes Erie and Ontario coming directly from tributary inflow, resulting in less relative importance of groundwater discharge in these lakes. At the reach scale various studies have found that the amount and direction of hydrologic exchange between ground and surface water are highly heterogeneous. Bustros-Lussier (2008) found that coarse sediment channels such as eskers act as preferential flow paths. Similarly, variations in bed topography and stream flow can produce regions with a buildup of fine (coarse) pore sediments and organic matter which inhibit (enhance) GWSWI (Ashworth, 2012; Boulton et al., 1998; Conant, 2004). At the regional/watershed scale, the study of GWSWI in tributaries of the Great Lakes was most extensively analyzed by Neff et al. (2005) who only considered flow in major streams as has been previously discussed. In terms of modeling, quantitative flow estimates in a few watersheds in Ontario, such as the Oak Ridges Moraine (Meriano and Eyles, 2002), have been well studied due to special interest in protection. Other watersheds have been comprehensively modeled as part of Tier 1 and 2 Source Water Protection Assessments. Fig. 1 shows the extent of GWSWI presented in the assessment reports and it can be seen that the Essex, Long Point, Toronto and Simcoe areas have had the most thorough analysis of groundwater discharge in their water budget studies. Due to the large population the Lake Michigan Basin has also been extensively modeled for groundwater discharge in the entire basin by Feinstein et al. (2010) and in regional ‘watersheds’ by Boutt et al. (2001), Cherkauer et al. (1992), Hoaglund et al. (2002) and others. Cherkauer and Taylor (1990) sampled a transect along the Detroit and St. Clair Rivers, joining Lake Huron to Lake Erie, to quantify the amount of flow added by direct groundwater discharge between the two lakes. They found that hydraulic conductivity varied along the channel, fluctuating between 4.6 × 10−6 and 1.5 × 10−3 cm/s, where the hydraulic conductivity generally decreased between Lake Huron and Lake St. Clair, increased significantly in the first 20 km of the Detroit River and decreased again thereafter. Except for Cherkauer and Taylor (1990), none of the studies found at this scale account for small scale variability,

Fig. 2. Groundwater discharge considerations in Ontario Source Water Protection Assessments which (a) do not consider GW discharge, (b) qualitatively identify GW discharge areas, (c) quantify GW discharge in the watershed and (d) identify and quantify GW discharge at reach scales.

Please cite this article as: Kornelsen, K.C., Coulibaly, P., Synthesis review on groundwater discharge to surface water in the Great Lakes Basin, J Great Lakes Res (2014), http://dx.doi.org/10.1016/j.jglr.2014.03.006

K.C. Kornelsen, P. Coulibaly / Journal of Great Lakes Research xxx (2014) xxx–xxx

but presume uniform hydraulic properties throughout a large hydrological response unit. At the reach scale, GWSWI studies have been carried out in many small streams by Bustros-Lussier (2008), Cey et al. (1998), Conant et al. (2004), Kasahara and Hill (2006), and others and demonstrate the importance of methods that account for streambed variability when quantifying groundwater discharge. The catchment to sub-watershed scale represents a knowledge gap where few studies could be found relating to the role of groundwater discharge. It is also unclear from the literature what, if any, scaling relationships can be used to up/down scale measurements or model estimates of groundwater discharge from the reach to watershed scales. Similarly there is no indication if the variability identified at the reach scale has important consequences at the watershed scale. Quality of groundwater discharge in the Great Lakes Basin The quality of groundwater is directly related to ecosystem health and diversity, as groundwater provides a source of nutrients, stable temperatures, and acts as a buffer to stabilize wetland pH (Granneman et al., 2000). In many watersheds, degradation of the water in the shallow aquifer has resulted in a sharp chemical interface between the shallow and deep groundwater (Haack et al., 2005; Hill, 1990). As with the quantity of groundwater discharge, the quality of exchanged groundwater varies in response to variations in discharge, hydraulic gradient, bed topography, porosity and ecological interactions (Boulton et al., 1998). Therefore, when considering water quality sampling it is important that sampling accounts for the variability in flow pathways (Conant et al., 2004; Hill, 1990). In Ontario, groundwater quality is considered good in most aquifers but has traditionally been poorly protected (MacRitchie et al., 1994). Following the tragedy in Walkerton and the introduction of the Clean Water Act (2006) sources of groundwater have received greater levels of protection through Source Water Protection Plans, which were submitted to the Ontario Ministry of the Environment in 2012 with implementation steps following. As was found with water quantity, much of the water quality studies in the Great Lakes Basin is focused on study areas in the Michigan Peninsula. Great Lakes Basin scale water quality The impact of groundwater discharge on water quality at the scale of the GLB is difficult to determine due to the effects of dilution and the mixing with contamination from surface waters. However, a conceptual understanding of the primary pathways by which certain contaminants enter the Great Lakes allows for inferences to be made relating the overall quality of water in the lakes to GWSWI. For example, nutrients are major contaminants of concern. Nitrogen is highly mobile and follows a largely subsurface flow path to surface water bodies (Hill, 1996), whereas phosphorus is easily bound to soil particles and is predominantly transported by suspended sediments, but also has dissolved sub-surface flow paths (Boulton et al., 1998; Great Lakes Commission, 2012). Chloride is relatively non-reactive in groundwater and occurs naturally at high concentrations in formation brines and deep aquifers (Kolak et al., 1999). In shallow aquifers, it is often from anthropogenic sources such as septic systems, oil brines and primarily the dissolution of road salt (Boutt et al., 2001). Chloride is therefore an indicator of the impact of human development on shallow groundwater (Boutt et al., 2001). In 2012, a comprehensive examination of long term monthly and annual surface water quality samples from the Great Lakes by Chapra et al. (2012), Dolan and Chapra (2012) and Chapra and Dolan (2012) has identified current trends in the basin. In general, the concentration of ions and nutrients in the Great Lakes is related to the flow of water through the lakes. Lake Superior was found to have the lowest concentrations of contaminants and urban-industrial development in Lake Michigan has resulted in a high concentration of many contaminants.

5

The mixing of Lake Superior and Michigan waters results in Lake Huron having intermediate water quality and Lakes Erie and Ontario having the lowest water quality as they accumulate contaminants from the upstream lakes. In terms of Cl−, Na+ and SO24 − ions, which are largely indicative of human development, the long residence time in the upper Great Lakes has resulted in persistent increases in the concentration of these ions (Chapra et al., 2012). The lower lakes had decreasing ion concentrations between the 1970s and mid-1990s, but recent data show that Na+ and Cl− ions are again rising in the lower basin, resulting partially from previous loading in the upper basin and partially from increasing human population (Chapra et al., 2012). Chapra et al. (2012) attribute the decreases in ion concentrations between the 1970s and 1990s to the reduction in industrial discharges (point sources), whereas the current increasing trends suggest nonpoint sources of ions in the basin require more attention. The greatest impact will be to near-shore waters which are influenced by regional groundwater chemistry (Haack et al., 2005). Hyporheic and riparian zones and residence time influences on water quality The hyporheic and riparian zones are important biogeochemical interfaces, which despite their small size (centimeters to meters) play an important role in mitigating groundwater contamination, particularly with respect to nutrients, prior to discharge. The primary mechanism by which the hyporheic and riparian zones improve water quality results from the introduction of dissolved oxygen (DO), which changes the reduction–oxidation (redox) status of the water (Boulton et al., 1998) and in combination with nutrients from the groundwater allows for highly productive microbial and benthic communities which can remove or alter nutrients, ions and other water chemistry constituents (Hayashi and Rosenberry, 2001). In the Oak Ridges and Waterloo Moraines, shallow groundwater chemistry was significantly impacted by nitrate and chloride from anthropogenic sources, whereas deep groundwater chemistry is dominated by ions from the dissolution of bedrock (Hill, 1990; Stotler et al., 2011). At one site in the Oak Ridges Moraine nutrients in the groundwater flowed toward streams where the riparian zone reduced ammonium loads by microbial immobilization in aerobic rivulets, but had little effect on nitrate (Hill, 1990). The improvement of groundwater quality in the riparian zone was seasonal and therefore nutrient concentrations in the shallow groundwater were greatest in the winter, but declined steadily during the growing season as vegetation arose from dormancy (Hill, 1990). The depletion of oxygen due to the mineralization of organic matter in the outer hyporheic zone can result in the release of P from soil particles when water is flowing into the sub-surface (Boulton et al., 1998). The hyporheic zone may also play a role in regenerating inorganic nitrogen as oxygen becomes available from the mixing of surface waters as groundwater flows out of the subsurface (Boulton et al., 1998). Besides oxygenation, the residence time in the streambed also partially determines the potential for (de)nitrification, where longer residence times are directly related to the amount of (de)nitrification that can occur (Puckett et al., 2008). The result is that the response of nutrients to hyporheic and riparian zone processes varies. In the Blue Springs and Hillman Creeks, hyporheic and riparian processes resulted in a decrease in dissolved nitrogen in the groundwater (Gillham et al., 1978a,b). In the Oak Ridges Moraine, the riparian and hyporheic zones had little impact on nitrates (Hill, 1990) and in the Emmons Creek Watershed of Wisconsin, the nitrate concentration in the hyporheic zone was only 10% to 60% of that found in deeper groundwater (Stelzer et al., 2011). In the northeastern United States, Puckett et al. (2008) found that the hyporheic zone could act as either a source or a sink of nitrates. The variability of ecological and microbial communities provides different responses to the chemical constituents of shallow groundwater. Since these processes vary between catchments, the best approach in

Please cite this article as: Kornelsen, K.C., Coulibaly, P., Synthesis review on groundwater discharge to surface water in the Great Lakes Basin, J Great Lakes Res (2014), http://dx.doi.org/10.1016/j.jglr.2014.03.006

6

K.C. Kornelsen, P. Coulibaly / Journal of Great Lakes Research xxx (2014) xxx–xxx

determining the role of the riparian and hyporheic zones in terms of GWSWI is likely by identifying processes at this scale. The difficulty of a catchment based effort at the scale of the GLB necessitates the identification of broad predictor variables applicable within the basin to identify areas where denitrification, phosphorus release and ammonium immobilization are likely to occur. Groundwater–surface water contribution to non-point source pollution A primary flow pathway for common non-point source contaminants, such as chloride and nitrate, is through shallow groundwater (Boutt et al., 2001; Hill, 1990; Meriano, 2007). As a result of agricultural and urban land-uses these two contaminants have become a major chemical constituent in shallow aquifers in both the Oak Ridges and Waterloo Moraines (Hill, 1990; Stotler et al., 2011) in Canada and in many aquifers in urban areas of the Great Lakes States (Boutt et al., 2001; Cherkauer et al., 1992; Konrad et al., 1979). Groundwater contamination by non-point source pollution is particularly problematic in urban areas where there are often few wells to monitor subsurface water quality and environmental protection against point-source contamination is easier. ‘It is significant, for example, that water quality in the Don River remains severely degraded despite the recent elimination of combined sewer overflows and the closure of several sewage treatment plants in the catchment,’ (from Howard and Livingstone, 2000). At the regional scale, flow paths of non-point source pollution have been studied in the Door Peninsula of Wisconsin (Cherkauer et al., 1992) and along the Grand Traverse Bay of Michigan (Boutt et al., 2001) using MODFLOW, as well as in the Waterloo Moraine (Stotler et al., 2011) and Lake Simcoe Watersheds (Winter et al., 2007) using water sampling in wells. In most watersheds road salt was the primary contributor of chloride in groundwater and surface water (Boutt et al., 2001; Chapra et al., 2012; Howard and Maier, 2007; Meriano, 2007; Stotler et al., 2011; and others). Agriculture and urban land uses around the city of Green Bay were responsible for 58% of chloride in the shallow groundwater while only receiving 38% of recharge water volume (Cherkauer et al., 1992). Of the total chloride that enters the upper aquifer of the Green Bay basin, 33% was transported into Lake Michigan, where the rest was diverted into wells and other sinks (Cherkauer et al., 1992). A possible explanation for this lower than expected amount of discharge to Lake Michigan may be related to changes in the water table resulting from concentrated pumping in the urban areas, which also contributes heavily to chloride. In the Seaton Lands north of Pickering, Howard and Maier (2007) modeled the impact of urban expansion on the underlying aquifer. The results of the model forecasted that development will cause significant degradation of both upper and lower aquifers with chloride and that the aquitards between aquifers provided little barrier to contaminant transport (Howard and Maier, 2007). The presence of nitrates in groundwater of the GLB has been long linked to agricultural land use and lawn care (Gillham et al., 1978a,b; Konrad et al., 1979), whereas mature forested catchments tend to have low concentrations of nitrogen (Detenbeck et al., 2003). Cherkauer et al. (1992) found that along Green Bay 38% of nitrate that entered the upper aquifer subsequently entered Lake Michigan and 50% of the nitrate contribution was focused within the periphery of the urban area. Similarly, the concentrations of nitrates were generally higher in shallow groundwater closer to their source (Gillham et al., 1978a,b; Hill, 1990; Stelzer et al., 2011) and the discharge of nitrates to surface waters was highest from preferential flow pathways (Gillham et al., 1978a,b). These complicated flow patterns require careful sampling that accounts for flow pathways to ensure non-point source estimates are accurate (Hill, 1990) and that remediation strategies are tailored to the processes operating within the watershed (Passeport et al., 2013). Phosphorus is a major contaminant of concern in the GLB as it has been found to contribute to algal blooms particularly in Lake Erie (Great Lakes Commission, 2012). In Ontario, MacRitchie et al. (1994)

reported that between 60 and 70% of phosphorus and pesticides entering the Great Lakes originated from tributary flow, which suggests groundwater discharge as a potential contributing pathway. Recently, Winter et al. (2007) found that total phosphorus in Lake Simcoe has decreased, primarily due to a reduction in tributary discharge of phosphorus. The reason for this decrease was attributed to better handling of nutrients from agricultural sources. In a report for the Ohio Environmental Protection Agency the OLEP Task Force (2010) found that mitigation measures in the 1970s to 1980s reduced point-source loading of phosphorus to the lake, while non-point sources of phosphorus have remained relatively constant. Traditionally phosphorus has been considered inactive in groundwater as it is adsorbed onto soil particles and is transported with sediment. However, particulate phosphorus has been on the decline with better soil management, but dissolved reactive phosphorus has increased to levels measured in the early 1970s (Great Lakes Commission, 2012). The increase in dissolved phosphorus occurs because soil has a finite phosphorus sorption capacity, which when exceeded allows phosphorus to become mobile in groundwater in a dissolved form (Domagalski and Johnson, 2011; Tesoriero et al., 2009). Groundwater–surface water contribution to point source pollution Many common supplies of point source pollution, such as wastewater treatment plants and storm sewers, do not involve GWSWI by design and most known waste disposal sites are monitored for groundwater contamination (MacRitchie et al., 1994). However, a recent report on the “State of the Science of Groundwater Contribution to the Great Lakes–St. Lawrence River Basin Watershed and Ecosystem” prepared for the Ontario Ministry of the Environment stated that the influence of point sources on Great Lakes water quality is a knowledge gap to be addressed because of the presence of unknown contaminant sources (Byerley and Velderman, 2012). The Science Advisory Board to the International Joint Commission reported that 90% of water-borne pathogenic disease is from groundwater contamination (IJC, 2010), the primary source of which is from human fecal waste leaking from septic systems. They similarly identify leaky underground storage tanks as a threat to ground and surface water quality. In both cases, the greatest threat for point-source contamination results from the fact that the number and location of potential sources are largely unknown (IJC, 2010). In the Niagara River Area of Concern, the Science Advisory Panel reports that the contaminants from Lake Erie are decreasing but the groundwater contamination along the river has not, despite the cessation of many legacy point source industrial operations (IJC, 2010). As a result of the difficulty in identifying point sources of groundwater contamination the association between GWSWI and point sources of pollution remains a significant knowledge gap. Despite this, a small number of studies can be presented that identify interactions of point sources of pollution and groundwater discharge. In the sandy barrier bar aquifer of the Point Pelee Marsh, Ptacek (1998) studied the transport of septic leachate plumes. As with non-point sources of pollution, oxygen plays an important role in determining the redox status of groundwater contaminants, resulting in distinct redox zones. Ptacek (1998) also found that the relative mobility of PO3− varied as the flow 4 of PO3− in the oxidized zone was retarded, whereas in the reduced 4 areas there were indications that PO3− may be relatively mobile. In 4 Muskoka, nitrate was found to be released from septic systems and was present in the groundwater; however, the concentration of nitrate was significantly reduced prior to discharge by passage through the hyporheic zone (Robertson et al., 1991). When studying a tetrachloroethene (PCE) plume released from a drying–cleaning facility, Conant et al. (2004) found that the riparian and hyporheic zones resulted in anaerobic biodegradation of the plume in the top 2.5 m of the streambed. Dilution of the point-source plume made it difficult to detect in surface water and differences in streambed hydraulic conductivity resulted in order of magnitude difference in contaminants over short distances (Conant et al., 2004).

Please cite this article as: Kornelsen, K.C., Coulibaly, P., Synthesis review on groundwater discharge to surface water in the Great Lakes Basin, J Great Lakes Res (2014), http://dx.doi.org/10.1016/j.jglr.2014.03.006

K.C. Kornelsen, P. Coulibaly / Journal of Great Lakes Research xxx (2014) xxx–xxx

Therefore, to best detect the PCE plume in surface water, areas of high hydraulic exchange needed to be identified as they are responsible for a disproportionate amount of discharge to surface water (Conant, 2004). Since dilution impacts the concentration of contaminants, Conant et al. (2004) recommend sampling interstitial water in the shallow streambed where contaminant plumes will be most prevalent. Impacts of environmental change on groundwater–surface water interactions The interactions between groundwater and surface water are dynamic at all scales and vary temporally (Boulton et al., 1998). While the influence exerted on groundwater discharge by the hydrogeologic cycle occurs very slowly, the hydrological cycle is spatially and temporally variable. Therefore, management of water resources and the mitigation of point and non-point source pollution require an appreciation of the conceivable impacts of potentially uncertain changes in the environment. Potential impacts of human development The impact of human development on water in the GLB has long been recognized as an area of concern and continues to require attention (IJC, 2010). The links between surface water quality and the application of road salt (Boutt et al., 2001; Meriano, 2007), use of fertilizers (IJC, 2010; Kidmose, 2010; SOGL, 2009), leakage from septic beds (IJC, 2010) and other point sources are well established. The correlation between land-use and contamination has been verified from catchment scale (Gillham et al., 1978a,b; Meriano, 2007) to lake scale, where contrast between Lakes Michigan and Superior demonstrates the distinct impacts of human development (Chapra et al., 2012). With increased population, land use in the Great Lakes Basin is changing as urban development is occurring at a high rate and forest is converted to agriculture lands to sustain the increasing human population (Létourneau, 2010). Concerns about potential changes to GWSWI due to human development are related to the impact of groundwater residence time on groundwater contaminants, the alteration of flow paths due to continued urbanization and the role of pumping in altering hydrological regimes. Howard and Livingstone (2000) also noted a particular problem of groundwater contamination in urban areas that rely on surface water. They suggest that these communities may be at a greater risk because groundwater is not monitored and contamination is rarely noticed until it becomes manifest in surface waters (Howard and Livingstone, 2000). A source of concern, in terms of water quality, is the residence time of groundwater. In urban areas of Toronto, Meriano (2007) found that 50% of road salt is stored in the shallow sub-surface. Howard and Maier (2007) found that road salt application results in a continued increase in the concentration of chlorides in groundwater and that the intensity of urban development is related to chloride contamination. In Michigan, Boutt et al. (2001) found that road salt was the primary source of chloride in shallow groundwater, but also pointed out that the slow transport and long residence time of groundwater resulted in the accumulation of chloride over time. These findings in Michigan are also valid for most Toronto Watersheds (Howard and Livingstone, 2000; Howard and Maier, 2007). The lower portion of the GLB in is highly impacted by human development. According to Neff et al. (2005), baseflow accounts for greater than 50% of water to contributing streams, which, results suggest, will become increasingly contaminated with non-point source pollution in developed areas. Despite improvements in urban design, agricultural practices and water treatment technology the concentrations of some contaminants appears to be on the rise (Chapra et al., 2012; Eimers and Winter, 2005). This trend is accelerated by the fact that urban land use change is exceeding what would be predicted by population growth (SOGL, 2009).

7

The relationship between land-use change and groundwater contamination is well established, however the relationship between land-use change and changes to hydrologic flow pathways, or the connection between recharge and discharge areas, is difficult to conceptualize. In the Oak Ridges Moraine, Howard et al. (1995) conducted a modeling study to gage potential responses of the Oak Ridges Aquifer to different urbanization scenarios. In some simulations urbanization resulted in a loss in groundwater recharge as high as 50%. Even in scenarios which included soakways and recharge ponds, urbanization was still found to result in a 5% decrease in recharge (Howard et al., 1995). This finding contrasts with findings in the urbanized Frenchman's Bay catchment in Toronto. Due to the infiltration from roof runoff and losses from water mains, Meriano (2007) found that groundwater recharge in the catchment was maintained at pre-urban levels. This discrepancy likely results from differences in the sources of urban water, whether pumping from an aquifer, which may decrease recharge over time (i.e. Howard et al., 1995), or pumping from a lake, which may allow maintenance of aquifer properties with urban development (i.e. Meriano, 2007). Changes in land-use and land-cover have also been found to alter the flow paths to tributaries (SOGL, 2009), which is an important and unknown consideration given the variability of GWSWI in streams. Increasing populations require larger amounts of water for agricultural, urban and industrial uses. In areas where surface water sources are not sufficient, groundwater is withdrawn from aquifers, causing depressions of hydraulic head in the aquifer around the supply wells and altering local hydraulic gradients (Winter et al., 1998). Feinstein et al. (2010) concluded that groundwater systems in the Lake Michigan Basin were predominantly unaltered from their pre-development state, but that pumping from wells had created strong local impacts. In their model, throughout the entire basin, pumping did not cause regional modifications, but diversion of groundwater from streams to wells did result in decreased baseflows. The overall result was a decrease in the groundwater discharge to Lake Michigan by approximately 2% (Feinstein et al., 2010). The largest impact was centered around the cities of Chicago, Milwaukee and Green Bay where well withdrawals have caused major reversals in the direction of groundwater flow and pumping wells have replaced Lake Michigan as the regional sink (Feinstein et al., 2010; Granneman et al., 2000; Haack et al., 2005). In Chicago and Milwaukee, the drawdown of the water table has been observed to expose aquifer sediments to oxygen, altering biogeochemical cycles (Granneman et al., 2000) and in Milwaukee pumping has long been identified as a cause factor for a loss of water in the Menomonee River Watershed (Konrad et al., 1979). If the rate of pumping remains constant, the local water table impacted by a well will eventually reach a new equilibrium by diverting surface water into groundwater, a process which occurs more rapidly if the well is located near the surface water body (Sophocleous, 2002). Potential impacts of climate change The science of climate change is continually evolving, but remains inherently characterized by a high level of uncertainty because the future impacts of human decisions and complex climate system feedbacks make accurate prediction of future conditions difficult. For this reason, few studies have been conducted with respect to the impact of changing climate on the integrated hydrological system that links groundwater, surface water and wetlands in the Great Lakes region. In the Grand River Watershed, Jyrkama and Sykes (2007) estimate that climate change will result in higher intensity and higher frequency precipitation as well as increased temperatures, which will increase runoff and evapotranspiration. They conclude that the impact of these climate influences is likely to be an increase in groundwater recharge, but the impact will vary based on soil and vegetation as well as to the nature of changes in winter infiltration (Jyrkama and Sykes, 2007). In des Anglais, PQ, along the St. Lawrence River, Sulis et al. (2011) found that

Please cite this article as: Kornelsen, K.C., Coulibaly, P., Synthesis review on groundwater discharge to surface water in the Great Lakes Basin, J Great Lakes Res (2014), http://dx.doi.org/10.1016/j.jglr.2014.03.006

8

K.C. Kornelsen, P. Coulibaly / Journal of Great Lakes Research xxx (2014) xxx–xxx

groundwater recharge is expected to increase during the winter and decrease in the spring due to the changing snow–rain ratio and diminished spring melt. The result of these changes is significant spatiotemporal variations in river discharge response as a consequence of a different partition between overland and baseflow (Sulis et al., 2011). At the basin scale climate change is expected to decrease the snowpack, soil moisture and therefore direct runoff to the Great Lakes (Croley, 2003). In terms of groundwater, most scenarios evaluated resulted in lowered water tables (Croley, 2003) which would presumably result in decreased discharge to surface water bodies and results in lowered GLB water levels (Croley, 2003). In order to account for the many interactions and feedbacks in the climate system, Wiley et al. (2010) linked land-cover, climate, hydrologic, hydraulic, thermal, loading and biological response models to assess the impact of climate and land use change in the Muskegon River Watershed. Application of the business as usual scenario in the multi-model resulted in hydrological changes of increased recharge, storm runoff, baseflow and median discharge. In terms of land-cover the model indicated that a future reduction in agriculture and forested land combined with an increase in urban areas is likely (Wiley et al., 2010). A resulting increase in groundwater discharge was modeled to produce channel destabilization, and increased urban cover is likely to result in increases in nutrient loadings. Under the same climate scenario, if urban sprawl is reduced, the model indicated that forest cover may increase and stabilize stream and base flows (Wiley et al., 2010) demonstrating the variability of response to human decisions. Temporal variability of groundwater–surface water interactions Much of the research relating to GWSWI emphasizes the spatial variability relating to groundwater discharge, while few studies provide explicit consideration of temporal variability. Over long time periods Coulibaly and Burn (2005) and Coulibaly (2006) have shown that both rainfall and streamflow patterns not only are seasonal, but also vary with longer cycles such as the El Niño Southern Oscillation and the North Atlantic Oscillation, which inherently has implications for groundwater recharge and water table elevation. Inter-annual variation in rainfall also directly impacts water quality, where the inter-annual variability in total phosphorus in Lakes Michigan and Huron was mainly due to differences in rainfall (Dolan and Chapra, 2012), resulting in greater tributary and groundwater flow. Long term shifts in climate, soil/hydrogeology or ecology can also result in changes to streambed morphology, where stream incision, redistribution of organic matter and sediments or changes in ecological composition can lead to changes in the spatial distribution and magnitude of GWSWI (Sophocleous, 2002). The most prominent temporal cycle in the GLB is the seasonal cycle which affects recharge (MacRitchie et al., 1994), water table elevation (Crowe et al., 2004; Haack et al., 2005) and groundwater discharge (Ashworth, 2012). Throughout the basin, water table recharge peaks during the spring snow melt and declines throughout the growing season. The dormancy of vegetation in the fall reduces evapotranspiration and recharge again increases until winter frost. This creates a seasonal shift in the rate and sometimes direction of groundwater flow. In barrier wetlands such as the Point Pelee and Deer Marshes, the difference in relative elevation between the marsh and the lake varies seasonally, resulting in a reversal of groundwater and contaminant flow (Bailey and Bedford, 2003; Crowe et al., 2004). In the western Lake Erie basin, Haack et al. (2005) observed that under natural conditions, the water table was drawn flat during the summer by evapotranspiration, temporarily decreasing groundwater discharge. Similar shifts in groundwater discharge were observed at the reach scale in Clythe Creek (Ashworth, 2012) and resulted in seasonal changes in nitrate concentrations in the Oak Ridges Moraine (Hill, 1990). While seasonal shifts impact natural processes such as evapotranspiration they also impact human activity. Agricultural and urban usage of water increases during the summer

months and in regions where the source of water is groundwater, the increased water demand has the potential to lower the water table locally and thus decrease the groundwater contribution to baseflow. It is expected that this consideration would be important in some aquifers with high rates of withdrawal or in areas where wells are located near surface water bodies, however the seasonal impact of pumping was not addressed in the reviewed literature. The cyclical changes in the water table also alter the catchment response to periodic storm events. The partitioning of infiltration and runoff and the groundwater response to a precipitation event are well known to vary with antecedent conditions (Sklash et al., 1978), where a high water table results in greater streamflow and a greater proportion of event water in stream runoff (Cey et al., 1998). Stoor et al. (2006) note that the flushing of groundwater and wetlands by both spring melt and large storms is an important process that increases the concentration of methyl mercury in surface waters. Similarly, Winter et al. (2007) point out that high discharge events are important for the transport of phosphorus, which must be accounted for to determine accurate loadings. Both examples demonstrate the importance of understanding the temporal variability of (shallow) groundwater discharge as well as spatial variability. Groundwater flow paths also vary greatly in terms of residence time. The shallow flow pathways responsible for the majority of groundwater discharge generally have a residence time on the order of days to years depending on the specific hydraulic conductivity and distance to a surface water body (Winter et al., 1998). Along the Grand Traverse Bay, Boutt et al. (2001) found that the transport of chloride after 50 years of application was approximately 10 km. In the narrow and sandy Point Pelee aquifer, contaminants introduced to the groundwater are expected to move 50 m in 7.5 years (Crowe et al., 2004). In the unconfined portion of the upper aquifer in the Waterloo Moraine, Stotler et al. (2011) identified aerobic groundwater with elevated Cl− and NO3 − concentrations with origins dating back to the mid-1960s precipitation. In Toronto, Howard and Livingstone (2000) modeled conservative contaminant transport in the aquifer and found that contaminants released within a few kilometers of rivers and Lake Ontario will discharge within a 50 year period. The result is that 80% of conservative contaminants released over the last 30 years will be flushed from the aquifer over the next 100 years, although some contamination may linger for as many as 500 years (Howard and Livingstone, 2000). While these examples of long residence times have the potential for groundwater remediation, they also indicate that groundwater is a potential reservoir of legacy contamination. In some cases, persistent toxins and conservative ions, such as chloride, from both point and non-point sources build up in groundwater with continual application and are being transported toward surface waters (Howard and Livingstone, 2000). If contaminants do not biodegrade, concentrations in aquifers will increase until a steady-state condition is reached which may take thousands of years (Howard and Maier, 2007).

Discussion and conclusions Surface water and groundwater have traditionally been considered as independent water resources, however a process based understanding determines that groundwater directly interacts with surface water and therefore, the two should be considered as a single resource (Granneman et al., 2000; Winter et al., 1998). In Lake Michigan, it is estimated that 2% of the total water budget comes from direct groundwater contribution to the lake, approximately 50% from direct precipitation and the remainder from tributaries (Granneman et al., 2000). Of that part of the Great Lakes water budget that is derived from tributaries between 40 and 75% of water is the result of baseflow or groundwater discharge (Neff et al., 2005), primarily through shallow subsurface flow pathways. Therefore, both directly and indirectly, groundwater is a major source of water to the Great Lakes, accounting

Please cite this article as: Kornelsen, K.C., Coulibaly, P., Synthesis review on groundwater discharge to surface water in the Great Lakes Basin, J Great Lakes Res (2014), http://dx.doi.org/10.1016/j.jglr.2014.03.006

K.C. Kornelsen, P. Coulibaly / Journal of Great Lakes Research xxx (2014) xxx–xxx

for approximately 20–40% of the total water budget, and is expected to significantly impact both quantity and quality of water in the GLB. The management of this resource is plagued by both spatial and temporal variability of processes which result in the discharge of groundwater and the great disparity of scales that must be considered in the basin. The approach required for managing the impact of groundwater discharge in the GLB requires a process-based approach that is synthesized into a holistic understanding of the GLB system (Harte, 2002; Montari et al., 2013). In practical terms, a greater assessment of both spatial and temporal scaling relationships in the GLB is required. For instance, at the reach scale, the groundwater contribution to surface water is highly variable in space and time (Bustros-Lussier, 2008; Conant et al., 2004; Drake et al., 2010). Preferential flow paths, such as sand lenses, provide a major transport mechanism resulting in point scale groundwater discharge varying by orders of magnitude at this scale (Conant, 2004; Drake et al., 2010). Similarly, the hyporheic and riparian zones near streams have been found to significantly alter water chemistry over short distances, but their impact remains difficult to predict a priori (Bonta and Rodgers, 2005; Crowley, 2012). This spatial variability is integrated to the watershed and basin scales, where it is difficult to ascertain the relative importance of smaller scale processes that result in the response at the basin level. This spatial variability is further complicated by the temporal variability of groundwater discharge and disparate timescales of groundwater processes, where the water contribution to a surface water body is the amalgamation of groundwater stored for time periods of days to centuries. Similarly, the residence time of groundwater has both positive and negative implications where natural remediation of some contaminants occurs simultaneously with the accumulation and slow discharge of other legacy contamination sources (Conant et al., 2004; Hill, 1990; Howard and Maier, 2007; Ptacek, 1998; Stelzer et al., 2011). Therefore, significant knowledge gaps remain the influence of scale on groundwater discharge processes and the identification of both spatial and temporal scaling relationships. Such information would help determine the impact of reach to catchment scale heterogeneity on water quantity and quality at the watershed and basin scale. Similarly, groundwater discharge is temporally variable and is influenced by event, seasonal and long term climatic oscillations as well as hydrogeologic and watershed characteristics. Complicating natural variability, anthropogenic influences directly impact not only water quality but also the processes by which groundwater and surface water interact. The limited extent of human development, when compared to the entire GLB, likely has a minor impact on the quantity of groundwater discharge (Feinstein et al., 2010), but a disproportionate impact on the quality of groundwater discharge, an influence which requires better understanding as a multi-scale phenomenon. In the future, a nested research framework would allow for the identification and monitoring of high groundwater discharge zones and lead to a better understanding of the hydrological cycle in the GLB. Extension of the basin scale modeling approach of Feinstein et al. (2010) to the other Great Lakes could provide much needed water balance information and serve as a foundation for higher resolution modeling studies. At the basin/watershed scales, areas of high discharge at the catchment to reach scale can be investigated through velocity–area methods (Becker et al., 2004; Cey et al., 1998), GIS (Bonta and Rodgers, 2005; Crowley, 2012), temperature sensing (Becker et al., 2004; Conant, 2004; Drake et al., 2010; and others), ground penetrating radar (Kidmose, 2010) and other methods. Such a multi-scale approach will also provide important information to connect groundwater recharge zones with the areas in which the groundwater will eventually be discharged (Conant et al., 2004; Hill, 1990; Sophocleous, 2002). The variability of baseflow estimates found by Neff et al. (2005), and the lack of available information on groundwater discharge for source water protection (Fig. 2), and the insufficient quantification of direct ground water discharge to the GLB (arguably except for Lake Michigan), indicate a poor baseline understanding of the water budget in the GLB which has

9

important implications for water quality and management of this important resource, especially in light of established environmental change. Acknowledgments This study was partially supported by Environment Canada (KW405-12-1762-0). The authors are grateful to Nancy Stadler-Salt and Dale VanStempvoort (Environment Canada), and Scott MacRitchie (Ministry of Environment of Ontario) for their review and edits of the initial work report. We are also grateful for the comments and suggestions of two anonymous reviewers who have helped to improve this manuscript. References Ashworth, H.E., 2012. Groundwater–Surface Water Interactions and Thermal Regime in Clythe Creek, Guelph Ontario: Threats and Opportunities for Restoration. (MSc Thesis) University of Guelph, Guelph, ON. Bailey, K.M., Bedford, B.L., 2003. Transient geomorphic control of water table and hydraulic head reversals in a coastal freshwater peatland. Wetlands 23 (4), 969–978. Becker, M.W., Georgian, T., Ambrose, H., Siniscalchi, J., Fredrick, K., 2004. Estimating flow and flux of ground water discharge using water temperature and velocity. J. Hydrol. 296, 221–233. Bonta, C., Rodgers, G., 2005. Wetland–groundwater linkage study for the Lower Trent Conservation watershed. Final Report. Lower Trent Conservation Authority, Trenton, ON. Boulton, A.J., Findlay, S., Marmonier, P., Stanley, E.H., Valett, M., 1998. The functional significance of the hyporheic zone in streams and rivers. Annu. Rev. Ecol. Syst. 29, 59–81. Boutt, D.F., Hyndman, D.W., Pikanowski, B.C., Long, D.T., 2001. Identifying potential land use-derived solute sources to stream baseflow using ground water models and GIS. Ground Water 39 (1), 24–34. Bustros-Lussier, E., 2008. Identifying Deep-Groundwater Discharge in Rivers of Eastern Canada. (MSc Thesis) University of Ottawa, Ottawa, ON. Byerley, B., Velderman, B., 2012. State of the science review: groundwater contribution to the Great Lakes–St. Lawrence River Basin watershed and ecosystem. Report No 09-1127-0060-1Ministry of the Environment, Toronto, ON. Cey, E.E., Rudolph, D.L., Parkin, G.W., Aravena, R., 1998. Quantifying groundwater discharge to a small perennial stream in southern Ontario Canada. J. Hydrol. 210, 21–37. Chapra, S.C., Dolan, D.M., 2012. Great lakes total phosphorus revisited: 2. Mass balance modeling. J. Great Lakes Res. 38, 711–754. Chapra, S.C., Dove, A., Warren, G.J., 2012. Long-term trends of Great Lakes major ion chemistry. J. Great Lakes Res. 28, 550–560. Cherkauer, D.S., Taylor, R.W., 1990. The spatially continuous determination of groundwater flow to surface water bodies: application to the connecting channels between Lakes Huron and Erie. J. Hydrol. 114, 349–369. Cherkauer, D.S., McKereghan, P.F., Schalch, L.H., 1992. Delivery of chloride and nitrate by ground water to the Great Lakes: case study for the Door Peninsula, Wisconsin. Ground Water 30 (6), 885–894. Conant, B., 2004. Delineating and quantifying ground water discharge zones using streambed temperatures. Ground Water 42 (2), 243–257. Conant, B., Cherry, J.A., Gillham, R.W., 2004. A PCE groundwater plume discharging to a river: influence on the streambed and near-river zone on contaminant distributions. J. Contam. Hydrol. 73, 249–279. Coon, W.F., Sheets, R.A., 2006. Estimate of groundwater in storage in the Great Lakes Basin, United States, 2006. U.S. Geological Survey Scientific Investigations Report 2006-5180U.S. Geological Survey, Reston, VA. Coulibaly, P., 2006. Spatial and temporal variability of Canadian seasonal precipitation (1900–2000). Adv. Water Resour. 29, 1846–1865. Coulibaly, P., Burn, D., 2005. Spatial and temporal variability of Canadian seasonal streamflows. J. Clim. 18 (1), 191–210. Coulibaly, P., Kornelsen, K.C., 2013. Groundwater discharge to surface water in the Great Lakes Basin. Report 03-1312‐17621Environment Canada, Hamilton, ON. Croley, T.E., 2003. Great Lakes Climate Change Hydrologic Impact Assessment I.J.C. Lake Ontario–St. Lawrence River Regulation Study. NOAA Great Lakes Environmental Research Laboratory, Ann Arbor, MI. Crowe, A.S., Shikaze, S.G., Ptacek, C.J., 2004. Numerical modelling of groundwater flow and contaminant transport to Point Pelee march, Ontario, Canada. Hydrol. Process. 18, 293–314. Crowley, J., 2012. Determining the Spatial Distribution of Groundwater and Surface Water Exchange using Heat as a Tracer. (MSc Thesis) University of Buffalo, Buffalo, NY. Detenbeck, N.E., Elonen, C.M., Taylor, D.L., Anderson, L.E., Jicha, T.M., Batterman, S.L., 2003. Effects of hydrogeomorphic region, catchment storage and mature forest on baseflow and snowmelt stream water quality in second-order Lake Superior Basin tributaries. Freshw. Biol. 48, 912–927. Dolan, D.M., Chapra, S.C., 2012. Great Lakes total phosphorus revisited: 1. Loading analysis and update (1994–2008). J. Great Lakes Res. 38, 730–740. Domagalski, J.L., Johnson, H.M., 2011. Subsurface transport of orthophosphate in five agricultural watersheds, USA. J. Hydrol. 409, 157–171. Drake, J., Bradford, A., Joy, D., 2010. A multi-scale method for identifying groundwater exchanges sustaining critical thermal regimes in streams. Int. J. River Basin Manag. 8 (2), 173–184.

Please cite this article as: Kornelsen, K.C., Coulibaly, P., Synthesis review on groundwater discharge to surface water in the Great Lakes Basin, J Great Lakes Res (2014), http://dx.doi.org/10.1016/j.jglr.2014.03.006

10

K.C. Kornelsen, P. Coulibaly / Journal of Great Lakes Research xxx (2014) xxx–xxx

Eimers, M.C., Winter, J.G., 2005. Lake Simcoe water quality update 2000–2003. Lake Simcoe Environmental Management Strategy Implementation Technical Report Imp. B.20. Lake Simcoe Region Conservation Authority, New Market, ON. Feinstein, D.T., Hunt, R.J., Reeves, H.W., 2010. Regional Groundwater-Flow Model of the Lake Michigan Basin in Support of Great Lakes Basin Water Availability and Use Studies: USGS Scientific Investigations Report 2010-5109. U.S. Geological Survey, Reston, VA. Gillham, R.W., Cherry, J.A., Hendry, M.J., Sklash, M.G., Frind, E.O., Blackport, R.J., Pucovsky, M.G., 1978a. Studies of the Agricultural Contribution to Nitrate Enrichment of Groundwater and the Subsequent Loading to Surface Waters: Part 1. International Reference Group on Great Lakes Pollution from Land Use Activities, Waterloo, ON. Gillham, R.W., Blackport, R.J., Cherry, J.A., 1978b. Studies of the Agricultural Contribution to Nitrate Enrichment of Groundwater and the Subsequent Loading to Surface Waters: Part 2. International Reference Group on Great Lakes Pollution from Land Use Activities, Waterloo, ON. Grannemann, N.G., Hunt, R.J., Nicholas, J.R., Reilly, T.E., Winter, T.C., 2000. The Importance of Ground Water in the Great Lakes Region: USGS Water Resources Investigations Report 00-4008. U.S. Geological Survey, Lansing, MI. Great Lakes Commission, 2012. Priorities for Reducing Phosphorus Loadings and Abating Algal Blooms in the Great Lakes–St. Lawrence River Basin: Opportunities and Challenges for Improving Great Lakes Aquatic Ecosystem. Phosphorus Reduction Task Force, Great Lakes Commission, Ann Arbor, MI. Haack, S.K., Neff, B.P., Rosenberry, O., Savino, J.F., Lundstrom, S.C., 2005. An evaluation of effects of groundwater exchange on nearshore habitats and water quality of Western Lake Erie. J. Great Lakes Res. 31 (Suppl. no. 1), 45–63. Harte, J., 2002. Towards a synthesis of Newtonian and Darwinian worldviews. Phys. Today 29–34 (October). Harvey, F.E., Rudolph, D.L., Frape, S.K., 2000. Estimating ground water flux into large lakes: application in the Hamilton Harbour, Western Lake Ontario. Ground Water 38 (4), 550–565. Hayashi, M., Rosenberry, D.O., 2001. Effects of groundwater exchange on the hydrology and ecology of surface water. Ground Water 40 (3), 309–316. Herman, J.S., Culver, D.C., Salzman, J., 2001. Groundwater ecosystems and the service of water purification. Stanford Environ. Law J. 20, 479–495. Hill, A.R., 1990. Ground water flow paths in relation to nitrogen chemistry in the nearstream zone. Hydrobiologia 206, 39–52. Hill, A.R., 1996. Nitrate removal in stream riparian zones. J. Environ. Qual. 25, 743–755. Hoaglund, J.R., Huffman, G.C., Grannemann, N.G., 2002. Michigan basin regional ground water flow discharge to three Great Lakes. Ground Water 40 (4), 390–406. Hoaglund, J.R., Kolak, J.J., Long, D.T., Larson, G.J., 2004. Analysis of modern and Pleistocene hydrologic exchange between Saginaw Bay (Lake Huron) and the Saginaw Lowlands area. Geol. Soc. Am. Bull. 116 (1–2), 3–15. Holtschlag, D.J., Nicholas, J.R., 1998. Indirect ground-water discharge to the Great Lakes. Open File Report 98-579. U.S. Geological Survey, Lansing, MI. Howard, K.W., Livingstone, S., 2000. Transport of urban contaminants into Lake Ontario via sub-surface flow. Urban Water 2, 183–195. Howard, K.W., Maier, H., 2007. Road de-icing salt as a potential constraint on urban growth in the Greater Toronto Area, Canada. J. Contam. Hydrol. 99, 146–170. Howard, K.W.F., Eyles, N., Smart, P.J., Boyce, J.I., Gerber, R.E., Salvatori, S.L., Doughty, M., 1995. The Oak Ridges Moraine of southern Ontario: a ground-water resource at risk. Geosci. Can. 22 (3), 101–121. IJC, 2010. Groundwater in the Great Lakes Basin: A Report of the Great Lakes Science Advisory Board to the International Joint Commission. International Joint Commission, Windsor, ON. Jyrkama, M.I., Sykes, J.F., 2007. The impact of climate change on spatially varying groundwater recharge in the Grand River watershed. J. Hydrol. 338, 237–250. Kasahara, T., Hill, A.R., 2006. Hyporheic exchange flows induced by constructed riffles and steps in lowland streams in southern Ontario, Canada. Hydrol. Process. 20, 4287–4305. Kidmose, J., 2010. Groundwater–Surface Water Interaction: From Catchments to Interfaces at Lakes and Streams. (PhD diss) University of Copenhagen, Copenhagen, Denmark. Kolak, J.J., Long, D.T., Matty, J.M., Larson, G.J., Sibley, D.F., Councell, T.B., 1999. Groundwater, large lake interactions in Saginaw Bay, Lake Huron: a geochemical and isotopic approach. Geol. Soc. Am. Bull. 111 (2), 177–188. Konrad, J.G., Chesters, G., Bauer, K.W., 1979. The IJC Menomonee River Watershed Study: Groundwater Hydrology, vol. 7. U.S. Environmental Protection Agency, Chicago, IL. Létourneau, G., 2010. Land Cover Along the Great Lakes and the St. Lawrence River. Environment Canada, Gatineau, QC. MacRitchie, S.M., Pupp, C., Grove, G., Howard, K.W.F., Lapcevic, P., 1994. Groundwater in Ontario: hydrogeology, quality concerns and management. NHRI Contribution No. CS-94011Environment Canada, Toronto, ON. Meriano, M., 2007. Groundwater–Surface Water Interactions in Frenchmans' Bay Watershed, Ontario: Implications for Urban Recharge, Contaminant Storage and Migration. (PhD diss.) University of Toronto, Toronto, ON.

Meriano, M., Eyles, N., 2002. Groundwater flow through Pleistocene glacial deposits in the rapidly urbanizing Rouge River-Highland Creek watershed, City of Scarborough, southern Ontario, Canada. Hydrogeol. J. 11, 288–303. Montari, A., Young, G., Savenije, H.H.G., Hughes, D., Wagener, T., Ren, L.L., Koutsoyiannis, D., Cudennec, C., Toth, E., Grimaldi, S., Blöschl, G., Sivapalan, M., Beven, K., Gupta, H., Hipsey, M., Schaefli, B., Arheimer, B., Boegh, E., Schymanski, S.J., Di Baldassarre, G., Yu, B., Hubert, P., Huang, Y., Schumann, A., Post, D.A., Srinivasan, V., Harman, C., Thompson, S., Rogger, M., Viglione, A., McMillan, H., Characklis, G., Pang, Z., Belyaev, V., 2013. ‘Panta Rhei — everything flows’, change in hydrology and society — the IAHS scientific decade 2013–2022. Hydrol. Sci. J. 58 (6), 1–20. Neff, B.P., Nicholas, J.R., 2004. Uncertainty in the Great Lakes Water Balance: U.S. Geological Survey Scientific Investigations Report 2004-5100. U.S. Geological Survey, Reston, VA. Neff, B.P., Day, S.M., Piggott, A.R., Fuller, L.M., 2005. Base Flow in the Great Lakes Basin: USGS Scientific Investigations Report 2005-5217. U.S. Geological Survey, Reston, VA. Ostry, R.C., 1979a. The Hydrogeology of the IFYGL Forty Mile and Oakville Creeks Study Area. Ministry of the Environment, Toronto, ON. Ostry, R.C., 1979b. The Hydrogeology of the IFYGL Duffins Creek Study Area. Ministry of the Environment, Toronto, ON. Ostry, R.C., Singer, S.N., 1981. The Hydrogeology of the IFYGL Moira River, Wilton Creek and Thousand Islands Study Areas. Ministry of the Environment, Toronto, ON. Passeport, E., Vidon, P., Forshay, K.J., Harris, L., Kaushal, S.S., Kellogg, D.Q., Lazar, J., Mayer, P., Stander, E.K., 2013. Ecological engineering practices for the reduction of excess nitrogen in human-influenced landscapes: a guide for watershed managers. Environ. Manage. 51, 392–413. Ptacek, C.J., 1998. Geochemistry of a septic-system plume in a coastal barrier bar, Point Pelee, Ontario Canada. J. Contam. Hydrol. 33, 293–312. Puckett, L.J., Zamora, C., Essaid, H., Wilson, J.T., Johnson, H.M., Brayton, M.J., Vogel, J.R., 2008. Transport and fate of nitrate at the ground-water/surface water interface. J. Environ. Qual. 37, 1034–1050. Robertson, W.D., Cherry, J.A., Sudicky, E.A., 1991. Ground-water contamination from two small septic systems on sand aquifers. Ground Water 29 (1), 82–92. Ruberg, S.A., Kendall, S.T., Biddanda, B.A., Black, T., Nold, S.C., Lusardi, W.R., Green, R., Casserley, T., Smith, E., Sanders, T.G., Lang, G.A., Constant, S.A., 2008. Observations of the Middle Island sinkhold in Lake Huron— a unique hydrogeologic and glacial creation of 400 million years. Mar. Technol. Soc. J. 42 (4), 12–22. Sklash, M.G., Gillham, R.W., Cherry, J.A., 1978. Studies of the Agricultural Contribution to Nitrate Enrichment of Groundwater and the Subsequent Nitrate Loading to Surface Waters: Part 3. International Reference Group on Great Lakes Pollution from Land Use Activities, Waterloo, ON. SOGL, 2009. Nearshore areas of the Great Lakes 2009. State of the Lakes Ecosystem Conference 2008 Background Paper. Environment Canada and U.S. Environmental Protection Agency. Sophocleous, M., 2002. Interactions between groundwater and surface water: the state of the science. Hydrogeol. J. 10, 52–67. Stelzer, R.S., Bartsch, L.A., Richardson, W.B., Strauss, E.A., 2011. The dark side of the hyporheic zone: depth profiles of nitrogen and its processing in stream sediments. Freshw. Biol. 56 (8), 2021–2033. Stoor, R.W., Hurley, J.P., Babiarz, C.L., Armstrong, D.E., 2006. Subsurface sources of methyl mercury to Lake Superior from a wetland-forested watershed. Sci. Total Environ. 368, 99–110. Stotler, R.L., Frape, S.K., El Mugammar, H.T., Johnston, C., Judd-Henrey, I., Harvey, F.E., Dribbie, R., Jones, J.P., 2011. Geochemical heterogeneity in a small, stratigraphically complex moraine aquifer system (Ontario, Canada): interpretation of flow and recharge using multiple geochemical parameters. Hydrogeol. J. 19, 101–115. Sulis, M., Paniconi, C., Rivard, C., Harvey, R., Chaumont, D., 2011. Assessment of climate change impacts at the catchment scale with a detailed hydrological model of surface–subsurface interactions and comparison with a land surface model. Water Resour. Res. 47, W01513. Task Force, O.L.E.P., 2010. Ohio Lake Erie Phosphorus Task Force Final Report. Ohio Environmental Protection Agency, Columbus, OH. Tesoriero, A.J., Duff, J.H., Wolock, D.M., Spahr, N.E., 2009. Identifying pathways and processes affecting nitrate and orthophosphate inputs to streams in agricultural watersheds. J. Environ. Qual. 38, 1892–1900. Wang, H., Jia, L., Steffen, H., Wu, P., Jiang, L., Hsu, H., Xiang, L., Wang, Z., Hu, B., 2012. Increased water storage in North America and Scandinavia from GRACE gravity data. Nat. Geosci. 6, 38–42 (NGEO1652). Wiley, M.J., Hyndman, D.W., Pijanowski, B.C., Kendall, A.D., Riseng, C., Rutherford, E.S., Cheng, S.T., Carlson, M.L., Tyler, J.A., Stevenson, R.J., Steen, P.J., Richards, P.L., Seelbach, P.W., Koches, J.M., Rediske, R.R., 2010. A multi-modeling approach to evaluating climate and land use change impacts in a Great Lakes river basin. Hydrobiology 657, 243–262. Winter, T.C., Harvey, J.W., Franke, O.L., Alley, W.M., 1998. Ground Water and Surface Water: A Single Resource: USGS Survey Circular 1139. U.S. Geological Survey, Denver, CO. Winter, J.G., Eimers, M.C., Dillon, P.J., Scott, L.D., Scheider, W.A., Willcox, C.C., 2007. Phosphorus inputs to Lake Simcoe from 1990 to 2003: declines in tributary loads and observations on lake water quality. J. Great Lakes Res. 33, 381–396.

Please cite this article as: Kornelsen, K.C., Coulibaly, P., Synthesis review on groundwater discharge to surface water in the Great Lakes Basin, J Great Lakes Res (2014), http://dx.doi.org/10.1016/j.jglr.2014.03.006