A three-dimensional hydrostratigraphic model of the Waterloo Moraine area, southern Ontario, Canada

Abstract

Aquifers of the Waterloo Moraine play a key role as the main source of drinking water for the Region of Waterloo. For the effective management of this water source, a sound understanding of the aquifers contained within and below the Moraine is essential. Critical knowledge required for this understanding includes the definition of the sediment facies distribution, architectural elements and geological origin of the Quaternary-aged deposits. A basin analysis approach has been applied to geologic data collection and interpretation to unravel the paleogeographic history of the study area and to provide a predictive framework for understanding its geological variability. Coarse (sand and gravel) sediment within the Waterloo Moraine was deposited during a series of high-energy meltwater discharge events from several sediment input corridors (eskers), into a deep, large, ice-supported glacial lake. This depositional setting led to a complex three-dimensional architecture comprising sand-gravel and mud units that are increasingly interbedded away from the multi-directional influx sources around the perimeter of the Moraine. A recently completed digital, three-dimensional geologic model of the area provides details of the various geological units that help refine the understanding of the hydrostratigraphy. This information has improved the understanding of groundwater flow (including interaction between surface and groundwaters) and has provided valuable information critical for source water protection. Information on the distribution, thickness, geometry and properties of these units has resulted in a better understanding of the potential linkages between near-surface recharge areas and deep aquifers across the region. This geological information is important in developing predictive models, for example, determining the location of high transmissivity zones within the moraine. Derivative products such as aquifer vulnerability and recharge maps may help inform policy makers in developing land use and nutrient management plans in the vicinity of well fields and sensitive lands.

Les aquifères de la moraine de Waterloo jouent un rôle crucial dans l’approvisionnement en eau potable dans la région de Waterloo. Afin d’assurer la gestion efficace de cette ressource en eau, il est essentiel de bien comprendre les aquifères contenus dans la moraine et sous celle-ci. Parmi les connaissances essentielles à cette compréhension, mentionnons la définition de la distribution des faciès sédimentaires, les éléments architecturaux et l’origine géologique des dépôts datant du Quaternaire. Pour retracer l’histoire paléogéographique de la zone d’étude et établir un cadre prédictif afin d’en comprendre la variabilité géologique, une approche d’analyse de bassin a été appliquée à la collecte et à l’interprétation des données géologiques. Les sédiments grossiers (sable et gravier) dans la moraine de Waterloo se sont déposés pendant une série d’événements très énergétiques de décharge d’eau de fonte provenant de plusieurs corridors d’apports sédimentaires (eskers) dans un lac glaciaire profond et large, confiné par la glace. Ce milieu de dépôt a produit une architecture tridimensionnelle complexe comportant des unités de sable-gravier et de boue qui sont de plus en plus interstratifiées à mesure que l’on s’éloigne des sources d’afflux multidirectionnelles autour du périmètre de la moraine. Un modèle géologique numérique tridimensionnel, récemment terminé et couvrant la région, fournit des détails sur les diverses unités géologiques qui aident à mieux comprendre l’hydrostratigraphie. Cette information a permis de mieux comprendre l’écoulement des eaux souterraines (y compris les interactions entre les eaux de surface et les eaux souterraines) et a fourni des données précieuses pour la protection des eaux de source. L’information sur la distribution, l’épaisseur, la géométrie et les propriétés de ces unités a permis de mieux comprendre les liens possibles entre les zones de recharge près de la surface et les aquifères profonds dans la région. Ces informations géologiques sont importantes pour le développement de modèles prédictifs, comme par exemple pour déterminer l’emplacement des zones de grande transmissivité dans la moraine. Des produits dérivés, comme les cartes de recharge et de vulnérabilité des aquifères, peuvent aider les administrations publiques à développer des plans d’aménagement du territoire et de gestion des nutriments à proximité des champs de puits et des terres sensibles.

Introduction

Waterloo Region is one of the largest municipal users of groundwater in Canada. Approximately 80% of its water supply is derived from bedrock and sediment aquifers with about half that being pumped from aquifers contained within the Waterloo Moraine (Ministry of the Environment 2009; AquaResource Inc. 2009; Hodgins et al. 2012). The Moraine serves as a major recharge area for a portion of the municipal water supply and feeds numerous cold-water streams and rivers within the region. A population of just over 550,000 people is supported by 50 well fields comprising 126 wells. Collectively, these wells supply in excess of 269,000 cubic metres of water per day to the residents of the region (Lake Erie Region Source Protection Committee 2012). Of the 50 well fields, 25 are located in the Waterloo Moraine area and seven have wells screened in the shallow surface aquifers of the Waterloo Moraine. Ten well fields are screened in older deposits pre-dating the Waterloo Moraine, yet likely receive significant recharge from aquifers in the Waterloo Moraine.

Industrial development within this area over the past 100 years has resulted in more than 800 industrial sites being identified that contain potential contaminants (Sanderson et al. 1995; Hodgins et al. 2012). There are also concerns regarding aquifer contamination from modern land use, including the application of road salt (Bester et al. 2006) and agricultural pesticides and fertilizers.

Recommendations from the Walkerton Inquiry (O’Connor 2002) stressed the importance of understanding the geologic controls on surface and groundwater flows, and how geologic frameworks can be used to predict where significant recharge and discharge areas (source water) are located, as well as where aquifers are more susceptible to surface contamination. Vertical aquifer connectivity within sandy aquifer complexes (e.g. Fogg 1986) as well as hydrogeologic heterogeneity within them (e.g. Anderson 1989) have been shown previously to be important in understanding groundwater flow and contaminant transport, a prerequisite for the effective management of the water resource. In the Region, the role of vertical hydraulic connectivity has been demonstrated both at the stratigraphic (regional) scale (Martin and Frind 1998; Bester et al. 2006) and at the more local, depositional scale by the presence of low-permeability barriers (mud units) to groundwater flow within otherwise highly conductive bodies (Duckworth 1983).

The role of discrete features such as erosional channels through confining units or preferential flow paths has also been observed in the Region with the by-pass of monitoring wells by contaminant plumes (Hodgins et al. 2012). The distribution, scale and architecture of these features remain as yet poorly understood. Regional delineation of geologic and hydrostratigraphic units and a better understanding of the events and processes leading to their formation can provide insights into the features controlling groundwater flow. The stratigraphic model of the Oak Ridges Moraine has demonstrated how this understanding can be used to identify hydraulic connections across regional aquitards (Sharpe et al. 2002) and within depositional environments of hydrostratigraphic units (Russell, Arnott et al. 2003, Russell, Sharpe et al. 2003; Sharpe et al. 2003). At the scale of depositional building blocks (e.g. subaqueous fans), the identification of sedimentary architecture and depositional environments can advance our understanding of the distribution and arrangement of sediment facies and their link to the role and connectivity of hydrofacies (e.g. Weissmann et al. 1999). Important to this expert knowledge is not only the classification of sedimentary facies, but also the characterization of the physical properties of the sediments (Aigner et al. 1999).

The Waterloo Region aquifer system is complex due to the discontinuous nature of multi-level aquitard and aquifer units (Martin and Frind 1998). Each of the respective geologic units has different sedimentary hydrofacies characteristics and depositional environments that control aquifer and aquitard variability. Petroleum reservoir studies (e.g. Cross et al. 1993) illustrate that sedimentary processes exert a control on physical characteristics of strata and consequently reservoir properties are partially predictable from stratigraphic geometries. This control on aquifer/reservoir properties has been modelled by Edington and Poeter (2006) and related to hydrogeological implications. This approach has merit in the context of the Waterloo Moraine area as the glacial geology is complex and spatially variable on length scales of metres to kilometres related to lithofacies and depositional architecture and geometry.

In light of the complexity of processes responsible for aquifer-aquitard formation, it is necessary to employ a multi-disciplinary approach linked to robust geological process models to better characterize aquifer connectivity and heterogeneity. In this study, a basin analysis approach has been applied to data collection and interpretation to understand the paleogeographic history of the Waterloo Moraine area and to provide a predictive framework for understanding its geological variability. This paper sets out to provide a geological framework for groundwater studies within the Waterloo Moraine area, as follows:

• Present the key attributes of a recently developed three-dimensional (3D) geological model for the Waterloo Region;

• Characterize the geometry and facies characteristics of the principal geologic units to assist with regional and local subsurface correlations;

• Discuss how the information gained by the basin analysis and stratigraphic modelling can be used to explain geological controls on subsurface hydrogeology;

• Discuss areas of near-surface aquifer extent and recharge potential.

Geologic setting

The Waterloo Moraine is centrally located within the inter-lake region of southwestern Ontario and lies almost entirely within the limits of the Regional Municipality of Waterloo (RMOW, or the Region). It is an irregular tract of gently rolling to hummocky topography occupying an area of approximately 600 km². The central part of the Waterloo Moraine is situated due west and southwest of the urban centres of Kitchener and Waterloo and consists of a broad, ridge-like upland measuring 10 to 20 km wide and 50 km long (Figure 1). Elevations range between 330 and 400 m above sea level (asl) with a local relief of 20 to 25 m. The moraine surface occurs > 155 m above Lake Erie and up to 90 m above the Grand River, a key local drain. Closed depressions or areas of internal drainage occur frequently within the hummocky terrain of the Moraine and occur as irregular hollows 10–50 m across and 5–10 m in depth. Ridges of sand and gravel converge from the north, west, south and east towards the centre of the moraine, forming a radial topographic pattern (Figure 1). These hummocks and ridges are identified as probable fans and eskers – building blocks of the moraine complex which have been subsequently dissected by fluvial processes.

Figure 1. Digital elevation model of the Waterloo Moraine area showing the important landscape elements discussed in the text (after Chapman and Putnam 1984). The Waterloo Moraine is outlined by the thick red line and Waterloo Region by the dashed black and white line. The yellow arrows delineate the major corridors along which sediment was delivered into the moraine. asl: above sea level.

Waterloo Moraine footprint

The understanding of the geographic extent of the Waterloo Moraine continues to evolve as new topographic and geological data become available. The recent availability of high-resolution imagery and topographic information as well as improved 3D Quaternary geology mapping has led to significant new understanding. Taylor’s (1913, 70) morphological definition of the Waterloo Moraine as a “finely formed moraine ridge running south from Waterloo to Ayr and west to Bamberg” is hardly recognizable in relation to current understanding of the moraine. Ensuing revisions, including those of Taylor (1939) and Chapman and Putnam (1951) as well as 1:50,000 scale surficial geological mapping (e.g. Karrow 1993) resulted in a refined definition. This mapping depicted areas of hummocky terrain, including genetically related radiating spurs of sand and gravel attributed to the Waterloo Moraine. This information served as the basis for the delineation of the planning boundary of the Waterloo Moraine used in the Region’s Official Plan (Blackport and Dorfman 2014, this issue). Subsequent subsurface geologic and hydrogeologic modelling (Martin and Frind 1998; Bajc and Shirota 2007; Bajc and Dodge 2011) has further improved the definition of the Moraine’s extent.

A revised boundary for the Waterloo Moraine is proposed which integrates surficial geology and geomorphology, the extent of modelled subsurface units and sedimentological understanding of important hydrofacies. Figure 1 shows this current interpretation of the limits of the Waterloo Moraine and highlights some of the important physiographic features mentioned in the discussion of the origin of the Waterloo Moraine. Significant additions to the Moraine include large tracts of sand and gravel in the southeastern corner of the Region that have similar sedimentologic and stratigraphic attributes to subsurface units identified in the main mass of the moraine. The Dryden Tract esker complex and its eastern extension beneath the Paris and Galt Moraines are also currently recognized as important elements of the Moraine. In addition, sand and gravel sediment bodies overlain by extensive till units in the peripheral areas are included as part of the Moraine. The recognition of lateral sedimentary facies changes within the Moraine has also resulted in the inclusion of fine-grained sediments (i.e. Maryhill muds), which are often interbedded with sand and gravel. These advancements in the understanding of sediment facies and depositional processes, which have led to an improved understanding of the origin of the Waterloo Moraine, provide a framework to better understand geological controls on the spatial distribution of hydrofacies and both regional and local groundwater flow. Thus, the common occurrence of interbedded mud, sand and gravel sequences within the moraine may explain complex aquifer-aquitard relationships.

Regional stratigraphy

Waterloo Region is underlain by west-dipping, Silurian-aged dolostone, shale and evaporite (mainly gypsum) belonging to the Guelph, Salina and Bass Islands formations (Armstrong and Carter 2010). These are overlain to the west by Devonian-aged cherty limestones of the Bois Blanc Formation. The contact between Devonian and Silurian bedrock units is marked by the Onondaga Escarpment extending along the west edge of Waterloo Region. The Guelph Formation dolostones outcrop along valleys of the Grand and Speed Rivers in the vicinity of Breslau and Cambridge. The depth to bedrock within the Waterloo Moraine area ranges between 70 and 115 m with depths > 140 m being documented in the central part of the Region.

Glacial and nonglacial sediments predating the last major advance of ice into southern Ontario are informally referred to as Pre-Catfish Creek sediment. This part of the sedimentary succession can be subdivided into: (1) a lower unit consisting of stony, silty to sandy diamicton (poorly sorted sediment, likely till) and associated/interbedded deposits of stratified sand and gravel, and (2) an upper unit of clast-poor, variably-textured diamicton with a red to mauve-grey colour (Canning Till), commonly interbedded with fine-textured glaciolacustrine sediment (Karrow 1963). Rare occurrences of organic-bearing sediments, deposited in alluvial, lacustrine, pond and bog environments during a non-glacial interval have been reported throughout the region (Bajc et al. 2009) and, in most cases, unconformably overlie Canning Till. Radiocarbon ages on wood recovered from these organic-bearing sediments range between 23,500 and > 50,500 radiocarbon (C¹⁴) years before present (BP) (Bajc et al. 2009).

Catfish Creek Till is present in the subsurface over much of southwestern Ontario and is mapped from exposures in deeply incised river valleys. This till is typically an over-consolidated and stony, silty to sandy diamicton containing frequent cobbles and boulders from Precambrian-aged bedrock outcropping several hundred kilometres to the north of Waterloo Region. Due to its reported consistent properties and regional distribution, this till is considered an important marker horizon for regional stratigraphic studies (Karrow 1988). Water well drillers often report this unit as “hardpan” in their drilling logs because of its stony and very dense character. Catfish Creek Till is locally absent where meltwater has down cut through and into the underlying sediments (Bajc and Shirota 2007).

Deposits forming the Waterloo Moraine both conformably and unconformably overlie Catfish Creek Till. These deposits include distinct units of stratified sand and gravel grading both upward and laterally into fine-grained sand that is locally interbedded with laminated mud (silt and clay) and silty to clayey diamicton. These gradations can occur over very short distances (tens of metres) (Bajc and Karrow 2004). Discontinuous sheets of clayey to sandy diamicton (Maryhill, Mornington, Port Stanley and Tavistock tills) and associated fine-grained glaciolacustrine sediments cap the moraine sequence.

Geological history

The oldest Quaternary-aged deposits preserved in the Waterloo Moraine area are considered to be older than 50,000 C¹⁴ years BP (Bajc et al. 2009). Diamicton and lacustrine sediments contained within these older sequences have been locally incised by subaerial fluvial processes that down cut and partially filled eroded channels with organic-bearing sediments of variable texture and permeability (Bajc and Shirota 2007; Bajc et al. 2009; Bajc and Dodge 2011). The last major ice advance across southern Ontario deposited the stony, silty to sandy Catfish Creek Till and associated waterlain sediments approximately 25,000 to 15,000 years BP.

The deglaciation of southern Ontario was topographically controlled; inland uplands became ice free first as thin ice lobes occupied and flowed out of the lake basins (Chapman and Putnam 1984). During this period, meltwater floods carved a system of subglacial channels across a broad region centred on the Oak Ridges Moraine (Sharpe et al. 2004). These floodwaters likely traversed the Niagara Escarpment and possibly flowed as far south and west as the Waterloo Moraine area (Kor and Cowell 2008). Subsequently, a thinner, reorganized ice sheet continued to disintegrate, expanding the interlobate zones. The ice-free area expanded progressively towards the northwest and southeast into the lake basins with large, ice-walled lakes occupying low-lying areas between the ice lobes (Bajc and Karrow 2004). The Waterloo Moraine was constructed into an ice-walled lake within this interlobate zone. Ridges composed of sand and gravel (eskers) extending radially out from the Moraine (Figure 1) represent the corridors along which subglacial meltwaters flowed into the interlobate zone (Bajc and Karrow 2004). The Waterloo Moraine is one of several moraines (Dorchester, Orangeville) that were constructed in this interlobate zone. Channels within and under the ice delivered meltwater and sediment into the ice-walled lake (possible sub-glacial lake), to form coalescing subaquatic fans composed of stratified deposits of sand and gravel (Bajc and Karrow 2004; Russell et al. 2007).

Later in the deglaciation, the margins of Lake Erie-Ontario and Huron-Georgian Bay ice lobes retreated from the Waterloo Moraine area towards the east-southeast and west-northwest, respectively (Bajc and Karrow 2004). Diamictons were deposited either during slight oscillations in the overall retreat of the ice margins or by debris flows along the margins of ice-contact lakes (Bajc and Karrow 2004). Glaciofluvial outwash and deltaic deposits graded to higher-level lakes in the Erie basin were formed in ice-marginal channels following the recession of the Huron and Erie-Ontario ice lobes (Thames, Grand and Speed Rivers). Some of the hummocky terrain along the southern and southeastern margins of the Waterloo Moraine was formed by this late-stage glaciofluvial activity, which altered the previous depositional landscape. All but the highest features have been truncated by glaciofluvial meltwater discharge with the intervening lows filled with outwash sand and gravel.

Development of 3D geological model

The 3D geological model presented in this paper is restricted to the Waterloo Region in spite of the fact that morainal sediment and associated landscape morphology extends beyond the municipal boundary. The model domain chosen includes the area of Waterloo Region plus a 1-km buffer strip around its periphery (Figures 3, 4, 6 and 11). Subsequent modelling efforts by the Ontario Geological Survey (OGS) south of the Region have covered off small portions of the Moraine not included in the original study (Bajc and Dodge 2011). The Easthope Moraine, located west of the Region, is the last remaining segment of the Waterloo Moraine that requires investigation. Development of the 3D model of the bedrock surface and overlying glacial sediments involved four principal stages: (1) data compilation and standardization, (2) collection of new geoscience information and the development of a conceptual geological model, (3) data analysis and model creation, and (4) data synthesis and interpretation. Details of this process are contained in Bajc and Newton (2007) and Bajc and Shirota (2007).

The subsurface database for Waterloo Region contains approximately 17,000 records having a total of 73,000 descriptive entries or layers of varying data quality (Table 1). Nearly 60% of the records are classified as “definitive”, which means that they were compiled by a geoscientist. These definitive records include information from boreholes drilled to install monitoring wells, as engineering test holes, or where continuous cores were collected. These records also include descriptions of natural and man-made exposures. A critical dataset for the geological modelling included a set of 110 records for boreholes with continuous cores (“golden spikes”) that provide the highest-quality stratigraphic information and sedimentary detail to develop and substantiate the conceptual geological model. In addition, 344 boreholes with downhole geophysical logs, 17.5 km of seismic reflection profiling and 16 km of ground penetrating radar data were used to further constrain the 3D geological model (Bajc and Hunter 2006; Endres et al. 2006).

Table 1. Subsurface information used to create the three-dimensional (3D) hydrogeological model. WAGAIS: Waterloo Geotechnical Automated Information System; OGS: Ontario Geological Survey; University: Department of Earth Sciences, University of Waterloo; RMOW: Regional Municipality of Waterloo; MOE: Ministry of the Environment.

Source	# Records	% of total	Depth (m) min–max (mean)	Details
RMOW monitoring wells	1084	6.4	0.5–244.5 (36.2)	Consultants’ boreholes
Urban geology database	3918	23.1	0.2–109.7 (7.9)	WAGAIS geotechnical data
Field mapping sites	4687	27.6	0.2–35.4 (2.2)	Shallow exposures
Cored boreholes	110	0.6	8.8–112.8 (60.0)	OGS, University, RMOW
MOE water well records	6832	40.3	1.3–179.8 (37.6)	Moderate-low quality
Geophysical database	344	2.0	5.8–157.8 (54.3)	Gamma ray logs, etc.
Total	16975	100	0.2–244.5 (21.5)

The subsurface geological information, in conjunction with the surficial geology maps (Karrow 1968, 1993; Cowan 1975), was the basis for developing an initial conceptual geological model for the Waterloo Region that consisted of 18 stratigraphic units and bedrock (Bajc and Shirota 2007). Many of these units, however, have a limited spatial extent, their boundaries pinching out at inferred ice margins, meltwater channels and limits of glaciolacustrine inundation. Consequently, the model layers were amalgamated and only seven lithostratigraphic units are discussed for the purposes of this study (Figure 2, Table 2). The Grand River outwash and Wentworth Till are not included in the model because their deposition post-dates the formation of the Waterloo Moraine and they are of limited extent within the Waterloo Region.

Figure 2. Idealized conceptual geological model for the Waterloo Moraine area. Aquifer units are shaded in dark grey, aquitards in white and bedrock in light grey.

Table 2. Summary characteristics of the important hydrostratigraphic units modelled in the Waterloo Moraine area. OGS: Ontario Geological Survey. OGS hydrostratigraphic classes are defined in Figure 2.

Stratigraphic units	OGS hydrostratigraphic class	Thickness (m)	Sediment character	Geophysical signature (terminology after Cant 1984)	Description/interpretation	Wellfields (from Martin and Frind 1998)
Mornington Till	ATB1	< 30	Pebbly, sandy silt to stone-poor, silty clay and clayey silt diamictons, massive to laminated	Shapes: irregular, bell, cylindrical, funnel	Subglacial lodgement till, subaquatic and ice-marginal debris flows, lacustrine rainout
Port Stanley Till
Tavistock Till
Upper Maryhill Till
Waterloo Moraine aquifer	AFB1	< 80	Bedded to massive silt, sand and gravel, silt- to clay-rich diamicton and glaciolacustrine deposits with granules and isolated pebbles, sharp-based fining upward successions	Shapes: irregular, bell, funnel	Coalescing subaquatic fans; paleoflows towards the west-northwest; some input from the west, northwest and north; discontinuous lenses/sheets of fine-grained waterlain deposits	Wilmot Center, Erb St., Strange St., Baden, Waterloo North, New Dundee, St. Clements, St. Agatha, Mannheim
Middle Maryhill Till	ATB2
Waterloo Moraine aquifer	AFB2					Strange St.
Lower Maryhill Till	ATB3
Waterloo Moraine aquifer	AFB3					William St., Greenbrook
Catfish Creek Till	ATC1	< 30	Massive to laminated, pebbly to cobbly sandy silt and sand mud diamicton with minor silt, sand and gravel interbeds	Shapes: irregular, bell, P-wave > 2600 m/s provides a nearly unique characterization, S-wave above 1200m/s are unique to this unit	Subglacial lodgement till; ice-marginal debris flows; minor waterlain deposits
Pre-Catfish glaciofluvial deposits	AFD1	< 20	Bedded sand and gravel	Shapes: irregular, bell; gamma P-wave 2600 m/s	Proglacial/subglacial fluvial deposits	Parkway, Greenbrook, Forwell, Heidelberg, Wellesley, Roseville
Canning sediment	ATE1	< 15	Massive to laminated silt and clay and stone-poor silty to clayey diamicton, minor sand	Shapes: cylindrical, symmetric, P-wave < 2700 m/s	Basinal glaciolacustrine deposits and glacial till
Pre-Canning glaciofluvial deposits	AFF1	< 30	Bedded sand and gravel	MS 2, P-wave 2600 m/s	Proglacial gravel bed rivers, braided fluvial, constrained by bedrock topography
Silurian and Devonian bedrock	Bedrock		Dolostone, shale, gypsum, cherty limestone			Forwell, William St.

Modelling of stratigraphic layers involved the manual interpretation of borehole records in Datamine Studio^® version 2 (software developed by CAE Mining, Montreal) along east-to-west and north-to-south sections spaced at 100 m. These interpretations were guided by preliminary (training) surfaces generated from the continuous core records and other datasets with high quality. The upper surface of the geological model was derived from the seamless surficial geology map of southern Ontario (Ontario Geological Survey 2003) projected over a 10-m digital elevation model (DEM, sourced from the Ontario Ministry of Natural Resources) as an elevation datum. The model thus honoured the field information with the greatest data support and highest level of confidence. The other surfaces were interpolated by applying user-defined search radii requiring a minimum number of geological picks for elevation values to be assigned to individual strata. The search radii ranged between 500 and 6000 m with the largest radii being applied to units which are presumed to be more continuous in the subsurface (i.e. the bedrock surface and aquitard units). A set of Boolean operations was subsequently applied to resolve surface crossovers and to allow for the creation of continuous wireframes of the surfaces for each stratum. A 3D block model representing all strata was created by filling the gaps between wireframes with sub-cells measuring 100 m by 100 m and of variable thickness, resulting in a model with approximately 560,000 cells (Bajc and Shirota 2007). This model serves as a basis for the current understanding of aerial extent, thickness and continuity of important lithostratigraphic units present within the Region of Waterloo.

Modelled stratigraphic units

The geometry and physical characteristics of the seven regionally identified lithostratigraphic units that were modelled as part of this study are reviewed in this section. These units include: (1) Paleozoic bedrock, (2) Pre-Canning glaciofluvial deposits (unit AFF1), (3) Canning sediment (unit ATE1), (4) Pre-Catfish glaciofluvial deposits (unit AFD1), (5) Catfish Creek Till (unit ATC1), (6) Waterloo Moraine deposits (units AFB1, ATB2, AFB2, ATB3 and AFB3) and (7) Upper Tills (unit ATB1) (Figure 2). Summary information on the sediment character, borehole geophysical signature and depositional environment of these seven regionally identified units is presented in Table 2.

Paleozoic bedrock surface

The bedrock surface in Waterloo Region is at an elevation of between 220 and 368 m asl and slopes gently from north to south (Figure 3), reflecting the regional bedrock structural control and topographic gradient. A large bedrock depression/trough, open to the south, occurs within the central part of the region and is coincident with the sub-crop belt of soft erosive shale of the Salina Formation. A number of distinct buried bedrock valleys up to several kilometres wide, > 30 km long and up to 60 m deep are incised into the bedrock. The largest recognized feature on the bedrock surface is the trunk valley of a southeast-trending buried valley system which enters the region at its west-central edge and is likely connected to a deeply entrenched re-entrant along the Niagara Escarpment at Dundas, approximately 35 km to the east (Figure 3). Northwest of the Waterloo Moraine, this bedrock valley is a few kilometres wide, 30 m deep, with local over-deepening to 60 m, and buried by up to 70 m of Quaternary sediment (Figures 3 and 4; Bajc and Karrow 2004). Locally, the sediment cover is much thicker. For example, to the east of the study area at Copetown along the Niagara Escarpment (Figure 1), an OGS borehole was drilled to a depth of nearly 200 m without encountering bedrock (Marich et al. 2011). The main thalweg of the Dundas/Wellesley buried bedrock valley is poorly constrained beneath the Waterloo Moraine as few wells are drilled to the bedrock within this region. A number of valleys, particularly in the western part of the region, are not well constrained, and their location is inferred from isolated deep boreholes and linear patterns of gravity lows (Figure 3; Marich et al. 2011).

Figure 3. Topography of the bedrock surface in the Waterloo Region. The red dashed line delineates the trunk axis of the Dundas/Wellesley buried bedrock valley. asl, above sea level.

Figure 4. Thickness of Quaternary sediments within the Waterloo Region.

Glacial stratigraphic succession

The sediment in the Waterloo Region is up to ~140 m thick with a mean value of 54 m (Figure 4). The thickest sediment is found over the geographic footprint of the Waterloo Moraine as well as along thalwegs of buried bedrock valleys (Figures 3 and 4). Sediment thins rapidly over the eastern third of the region as well as in valleys of the Nith, Conestogo, Grand and Speed Rivers. Locally, in the eastern part of the region, the sediment thickens along the crests of the Paris, Galt and Breslau Moraines. Two cross-sections (Figure 5) provide an overview of the stratigraphy and stratigraphic architecture that is supplemented in plan view by the isopach maps of the units (Figure 6). The salient stratigraphic information highlighted in boreholes of the two cross-sections are: (1) coarse gravel (unit OD) lying on the bedrock surface in apparent bedrock lows, (2) variable and diverse sediment facies beneath the Catfish Creek Till (units OD and AFD1), (3) the variable surface elevation and thickness of the Catfish Creek Till (unit ATC1), (4) highly variable unit thickness, bed thickness and composition of the Lower Maryhill Till (unit ATB3), (5) lateral facies variability and thickness changes of the sediments forming the Waterloo Moraine (units AFB1, ATB2 and AFB2) and (6) the discontinuous nature of the surface aquitard unit (unit ATB1).

Figure 5. Two geologic cross-sections along transects transverse and parallel to the long axis of the Waterloo Moraine. The cross-sections were constructed using information from cored boreholes. Refer to Figure 2 and Table 2 for keys to abbreviations of main lithologic units. OD refers to Older Drift and includes the sediments assigned to units ATE1, AFF1 and ATG1.

Figure 6. Isopachs of the major aquifer and aquitard units in the Waterloo Moraine area. White areas indicate locations where unit is not present.

Pre-Canning glaciofluvial deposits

Sand and gravel deposits pre-dating Canning sediment are generally discontinuous and have a sporadic distribution (Figure 6f). These deposits are most commonly intercepted in buried bedrock valleys, for example the Dundas/Wellesley bedrock valley, and in a bedrock valley partially exhumed by the Grand River near Breslau (Figure 1). These deposits may either directly overlie the bedrock surface or older deposits of till. The portion of boreholes that penetrate this sand and gravel is generally < 30 m long. In outcrop, the unit was observed to consist of cobble to small boulder gravel, well-bedded pebble to cobble gravel with lesser sand, and pebbly sand forming gently-dipping sheets and broad shallow channels filled with tabular bar forms (Bajc and Karrow 2004). The lower gravel beds contain boulders up to 1 m in diameter. The occurrence of Gowganda Formation and Huronian quartz arenite boulders indicates a source area to the north. Imbricated clasts and steeply-dipping bar-form foresets indicate a southerly or southwesterly paleoflow.

On the basis of sedimentological evidence and stratigraphic position, deposits of pre-Canning sand and gravel are interpreted to be part of a proglacial (beyond the ice margin) or potentially subglacial (beneath the glacier) glaciofluvial system. The large clast size (up to 1 m) would require high-velocity water flow to transport this material. The limited information on the stratigraphic architecture of this unit makes it difficult to interpret the exact depositional setting. This unit provides an important deeper aquifer target in the Waterloo area and its stratigraphic position beneath several regional aquitards would suggest a higher degree of protection from surface contaminants.

Canning sediment

Canning sediment is mapped across about half of Waterloo Region (Figure 6e) and includes the Canning Till and associated stratified, fine-grained glaciolacustrine sediment. It extends beyond the study area to the west and pinches out to the east of the Grand River valley. Modelling the sediment distribution suggests there are numerous places where the unit has been eroded (hydraulic windows), with a large linear northwest-southeast incision in the southwest part of the Region that is > 3 km wide and 15 km long. The Canning sediment is locally up to 45 m thick, although a thickness of 10 m is more representative. The Canning Till is typically clast-poor and has a silty mud matrix, although silty to sandy facies have been encountered. The till often has a red to grey-mauve colour suggesting a possible Erie-Ontario lobe source, possibly from shale below the Niagara Escarpment. The fine-grained texture of this till is attributed to the incorporation of clast-poor silty to clayey glaciolacustrine sediment (Karrow 1993). Interbeds of rhythmically laminated silt and clay associated with subglacial and proglacial facies of the unit suggest a predominant glaciolacustrine depositional environment.

Hydrogeologically, the unit is important as a regional aquitard where present over large areas. Hydraulic windows through the unit occur in NW-SE oriented erosional corridors that may provide hydrogeologic linkages between Waterloo Moraine and pre-Canning aquifers.

Pre-Catfish glaciofluvial deposits

Sand and gravel units situated stratigraphically between Canning sediment and Catfish Creek Till are referred to as Pre-Catfish glaciofluvial deposits. Sediments comprising this unit are found throughout the Waterloo Region and have a thickness of up to 30 m within the Dundas/Wellesley buried bedrock valley and are thinner in secondary bedrock valleys (Figure 6d). These deposits are exposed locally along rivers and in aggregate operations along the Grand and Nith River valleys. The coarsest sediment facies contain 0.5-m-thick beds of open-work cobble gravel (Figure 7a) that are laterally continuous over tens of metres. These beds are commonly associated with extensive deposits of cross-stratified medium sandy gravel. Cobble and boulder horizons occur infrequently. Discontinuous sand beds, up to a few metres thick and tens of metres in lateral extent, are interbedded with the gravel.

Figure 7. Photographs of sand and gravel units that are from the pre-Catfish Creek Till and Waterloo Moraine aquifers. (a) Open-framework cobble gravel with manganese oxide staining. (b) Isolated, discontinuous deposits of open-framework gravel (denoted by the arrows) within matrix-supported medium sand and gravel. (c) Contact between cobble gravel and silty Catfish Creek Till. (d) Stratigraphic architecture in the Waterloo Moraine which includes successions containing (i) gravel overlain by (ii) lacustrine mud, (iii) gravel, (iv) sand, and (v) silt. (e) Cobble gravel shown in Figure 6d with an open framework and basal medium sand matrix. (f) Bedded, well-sorted medium sand fining upwards to fine sand.

Pre-Catfish aquifer materials are interpreted to have been deposited in proglacial rivers with gravel beds, in subglacial conduits and in subaquatic fan settings. Their apparent spatial association with buried bedrock valleys may simply reflect either selective preservation of a previously more extensive unit or their original depositional location within the valley systems. Deposition in subglacial (tunnel) channels (channels carved by meltwater into sediments and/or bedrock beneath the glacier) is also possible and may help explain the association with valley settings. Many of these valleys have a north-south preferred orientation.

Catfish Creek Till

The Catfish Creek Till occurs as a nearly continuous lithostratigraphic unit within the region and is, for the most part, considered to be a regional till sheet and marker bed (Figure 6c). The unit consists primarily of dense, massive sandy diamicton; however, stratified deposits of both fine- and coarse-textured sediment occur as lenses and interbeds within the unit. Catfish Creek Till has also been observed to be interbedded with stratified to laminated glaciolacustrine sediment.

A number of valleys appear to be cut into the upper surface of this till, the most notable of which extends south-southeastward (Bajc and Shirota 2007) in the vicinity of the Dundas/Wellesley bedrock valley (Figure 3). This valley is 1.5 to 2.0 km wide, up to 35 m deep and, locally, is completely eroded through the till, forming potential hydraulic windows that are up to 1000 m across. There is also a poorly developed or diffuse pattern of east- to west-trending areas of thin Catfish Creek Till that may represent areas of incomplete erosion beneath the northeast side of the Waterloo Moraine. No significant erosional valleys have been identified or modelled beneath the Moraine. The Catfish Creek Till is modelled to be up to 65 m thick, with a mean value of 15 m (Figure 6c); intercepted thickness in boreholes is up to 45 m. The unit pinches out to the southeast where younger Port Stanley and Wentworth tills are present. The Catfish Creek Till is an overconsolidated stony, silty to sandy diamicton with local finer-grained facies. In continuous cores, this till is often dry, unoxidized and displays very little evidence of water saturation (water content is < 10%). Locally, isolated lenses of sand and gravel are interbedded within the Catfish Creek Till; these small bodies of sediment are only mapped over tens of metres at most. Both subglacial and proglacial facies (flow till) of the Catfish Creek Till have been observed, the latter often interbedded with stratified glaciolacustrine sediment.

This readily identified till sheet is considered to be a regional aquitard (Figure 6c). Isolated lenses of sand and gravel interbedded within Catfish Creek Till may transmit water but will not produce much flow unless connected to overlying moraine sand and gravel, or underlying pre-Catfish Creek sand and gravel.

Waterloo Moraine deposits

Waterloo Moraine sediment abruptly overlies Catfish Creek Till and is generally > 45 m thick with local accumulations of 100 m (Figure 6b). Linear, northwest-trending accumulations of thicker sediment are present within the core of the Waterloo Moraine and may represent corridors or conduits along which sediment was transported into the moraine sediment wedge. An abrupt decrease in sediment thickness occurs along the eastern edge of the moraine, beneath the urbanized parts of the cities of Kitchener and Waterloo. This may represent an important influx point along which coarse sediment accumulated (of note, a number of important well fields including the Waterloo North, Strange St., Greenbrook, Parkway and Strasburg well fields appear to lie along this NW-SE trend). Meltwater flowing to the ice margins in several subglacial conduits carried large volumes of sediment into a deep water body (proglacial or subglacial lake) from the southeast (Dryden Tract Esker), east (Sportsworld Complex Esker), north (Hawkesville Spur), northwest (Crosshill Spur) and possibly the west (Baden Hills). Moraine sediment sequences comprise three sharply bounded, fining-upward successions that contain sediments grading upward from gravel to mud (silt and clay). The lower, fine-grained mud units are commonly referred to as the lower (ATB3) and middle Maryhill (ATB2) Tills. In places, however, lower Maryhill sediment rests directly on Catfish Creek Till. The units of stratified sediment range in texture from gravel with an open-work structure to fine sand with small-scale cross-stratification. Moraine sediment is coarsest in the southeast, where it is predominantly sand and gravel, and finest to the west, although this is modified along the sediment influx corridors (eskers) located around the moraine (Figure 1). The stratal architecture of the middle and upper moraine packages is well understood and mapped from exposures at aggregate sites (e.g. Russell et al. 2007). The lower stratified package is described primarily from borehole data (Bajc and Hunter 2006). The lower and middle fine-grained units consist of laminated to massive, occasionally pebbly, clayey silt and silty clay. The thickest deposits of the lower Maryhill sediment appear to be coincident with the trend of the valley carved into Catfish Creek Till in the south-central part of the region (Bajc and Shirota 2007).

A sedimentological study of these moraine deposits by Russell et al. (2007) indicates that deposition occurred in a series of high-magnitude meltwater discharge events which arranged the sediment facies in predictable patterns along the path of water flow (Figure 8). Coarse sediment was deposited in subglacial tunnels, and finer sediment was deposited downflow (possibly beyond the ice-margin) as subaquatic fans. Sheets of diamicton and glaciolacustrine sediment, up to approximately 3–4 km wide and 10–20 m thick, occur within or between sediment of the subaquatic fans (Unit ATB2 in Figure 5). These muds were deposited by suspension settling in low-energy, ponded environments; diamicton was deposited as debris flows from melting ice or as resedimented mud. In a subaquatic fan depositional model, the transition from coarse, gravelly sediment to silty fine and very fine sand can occur over variable distances in the subsurface (Figure 8). For example, parallel to meltwater inflow, gravel can grade to sand over tens to hundreds of metres, whereas sand facies can extend downflow for hundreds to thousands of metres, or more, before grading to mud (silt and clay) in the basin. However, perpendicular to meltwater influx, gravel can grade to sand and mud in approximately one to tens of metres close to the influx point and in hundreds of metres in mid-fan settings. Where fans do not interfere with one another, this results in a predictable arrangement of sediment facies (Figure 8).

Figure 8. Subaquatic fan sediment facies model highlighting the proximal to distal hydraulic conductivity (in m/s) associated with distinct facies associations.

The above subaquatic fan sediment facies model can be associated with a hydrofacies model based on available knowledge of the Waterloo Moraine sediments. A predictable regional model would include knowledge of sediment input locations and paleoflow directions to assist in tracing highly permeable aquifer units downflow. As a first pass, groundwater exploration projects should target the sediment input corridors as highlighted in Figure 1.

Upper tills

Tills (Port Stanley, Stratford, Upper Maryhill, Tavistock and Mornington) form a mappable, nearly continuous surficial unit in areas peripheral to the Waterloo Moraine and become discontinuous over and adjacent to the core of the moraine (Figure 6a). This till package is generally < 20 m thick, with local modelled thicknesses of up to 45 m. The thickest deposits are situated in the western part of the Region and over the drumlinized area east of the cities of Kitchener-Waterloo (Guelph drumlin field on Figure 1). The Tavistock, Mornington and Upper Maryhill Tills are fine-textured diamictons (silty) whereas the Stratford and Port Stanley Tills are silty to sandy in texture and slightly more permeable. The diamictons were deposited as either subglacial till or as subaquatic debris flows and rainout deposits in a large glacial lake and they form a discontinuous, interfingered cover over the Waterloo Moraine and associated radial esker system peripheral to the moraine.

This upper discontinuous till package forms a variable, protective aquitard cover over the Waterloo Moraine and associated radial esker system. Where absent over the core Moraine area, both rainfall and snow melt recharge directly into the Moraine aquifers, whereas less recharge occurs towards the Moraine fringes where the confining unit is generally thicker and more continuous.

Physical properties

To improve the delineation and correlation of the key stratigraphic units, a variety of physical property data including grain size, carbonate content and geophysical response was compiled to develop a high-quality dataset for regional stratigraphic modelling.

Grain size

Particle size data reported by Karrow (1993) and compiled by Gautrey (1996) have been assessed according to the reported stratigraphic units summarized in Table 2. Clay was reported as the < 2 micron fraction in all analyses; however, no control or comparison samples exist across the datasets. The results from Karrow (1993) (Figure 9a) show that the surface, Upper Maryhill and Mornington Till (aquitard) units have a strong clustering in the mud and sandy mud textural fields, whereas the Tavistock and Port Stanley Tills group along the silt and sand boundary. For the lower aquitard units (Catfish Creek and Canning Tills), the texture is predominantly sandy silt, silty sand and sandy mud (Figure 9a). The clay content of most samples is generally low with an average reported value of 28%, a maximum value of 70% and a standard deviation of 15%. These grain size data provide a first-order constraint on estimating hydraulic conductivity values across the Waterloo Moraine area where other measured or derived hydrogeological values are not available.

Figure 9. Overview of physical properties of lithostratigraphic units discussed in this study: (a) plot of grain-size data on Folk’s (1954) ternary diagram, (b) cumulative curves of calcite/dolomite ratio and total carbonates and (c) plot of S- and P- wave vertical seismic velocities.

Carbonate mineral data

The carbonate mineral data derived from Chittick analyses of till units are summarized in the calcite/dolomite (Ca/Do) ratio and total carbonate cumulative frequency plots (Figure 9b). The percentage of total calcite in the respective units does not vary significantly (9.5 to 29.8%). However, both sediments of the surface tills and Canning Till have generally higher calcite concentrations than diamictons of the Waterloo Moraine and Catfish Creek Till. Percent dolomite plots in the range of 7.9 to 58.6%, with sediments of the surface tills having much lower concentrations than the Catfish Creek Till (Figure 9b). The dolomite concentration for the Canning Till plots in the middle of the graph and has a tighter range than either of the other two units. This variation may simply reflect differences in the grain-size characteristics of the respective units and possible partitioning of carbonates in the silt fraction. The Ca/Do ratios have good separation for two units. Catfish Creek Till has a low Ca/Do ratio, with the 75th percentile value being lower than the 10th percentile of materials comprising the surface till/mud. Canning Till plots in the middle of the graph and contains nearly equal concentrations of calcite and dolomite. Total carbonate ranges between 24 and 86% with the surface till/mud generally having the lowest and Catfish Creek Till having the highest values. The 10th percentile for the Catfish Creek Till is approximately equal to the 90th percentile of materials composing the surface till/mud.

Variable carbonate concentrations in the sedimentary sequence may allow for hydrostratigraphic markers to be estimated where other geological data are not available. Variable carbonate concentrations may also show up as differences in the hydrochemistry of natural groundwater from these geological units.

Geophysics

To support the stratigraphic correlations and development of the conceptual and 3D geological models, a suite of nine downhole geophysical logs was collected in OGS boreholes (e.g. Figure 10). The geophysical methods used included natural gamma, conductivity, magnetic susceptibility, neutron, relative density, spectral density, S- and P-wave seismic velocity and temperature (Bajc and Hunter 2006). The signal range of each tool for lithostratigraphic units has considerable overlap with very limited discrimination possible from absolute values (Table 2). The exception is the record for S- and P-wave seismic velocities of the Catfish Creek Till. The measured P-wave velocities provide a near unique characterization for the Catfish Creek Till. They are typically > 2600 m/s and the 10th percentile value of Catfish Creek Till is greater than the 90th percentile values for all the other units (Figure 9c). It should be noted that high P-wave velocities can occur in some sand and gravel units (potential aquifers), but these coarse-grained facies show chaotic reflectors in seismic profile lines (e.g. Pugin et al. 1999). Similarly, the S-wave velocities in the Catfish Creek Till are > 1200 m/s, which is unique to this unit.

Figure 10. Graphical presentation of the sedimentary and geophysical logs of continuous core from borehole OGS-03-04, west-central Waterloo Moraine (from Bajc and Hunter 2006).

The suite of data from natural gamma, conductivity and magnetic susceptibility have a large overlap (Table 2), but logs of these data show discrete changes or patterns that correspond to stratigraphic breaks and transitions in grain size (Figure 10). Specifically, the log shapes can be identified as irregular, symmetrical, bell-shaped and funnel-shaped responses (Table 2; terminology after Cant 1984) with sharp terminations at both the base and top depending upon the shape. These signatures may be identified in one or more of the logs and at scales ranging from several metres to tens of metres (e.g. Figure 10).

Discussion

Physical properties

The grain-size and carbonate datasets were integrated to better understand the aquitard matrix properties, controls on over-consolidation, and the properties controlling the geophysical log response. Besides the spatially distributed grain-size data for near surface units reported by Karrow (1993), the grain-size data from Bajc and Hunter (2006) include the downhole analyses of samples from continuous core that characterizes the sedimentary stratigraphy in the subsurface. Overall, the grain size clusters in a relatively tight range dominated by silt and sand. Clay-size particles comprise a minor component of the sediment. Using the Folk (1974) classification for fine-grained sediment, < 1% of the samples analyzed by Karrow (1993) have concentrations that plot in the clay group (i.e. containing > 67% clay; Figure 9a). This compositional distribution corresponds to previous work on the terminal grain size in glaciated carbonate basins and the terminal grain size of mineral grains (Folk 1954). In a study of glacial sediment (tills) of southern Ontario, Dreimanis and Vagners (1971) report that the terminal grain size of calcite is 62 to 4 microns and for dolomite it is coarser, between 62 to 16 microns. Karrow (1993) only reports the sand-silt-clay data; consequently, the distribution relative to reported grain-size range for dolomite and calcite cannot be assessed. Nevertheless, the Catfish Creek Till has a higher dolomite content than sediments of the surface tills and is correspondingly coarser (Figure 10). The results of the carbonate analyses indicate that Catfish Creek Till has the highest carbonate content, with an average concentration of 50% and a maximum of 85%. The grain-size data also suggest that the matrix of the surface till units is coarser than commonly described in the field by visual inspection. Muddy diamictons, such as the Maryhill, Mornington and Tavistock Tills are predominantly composed of silt, and no samples had more than 67% clay, the concentration used to differentiate a clay texture.

Sediments that lie at or near the bedrock surface have compositions that more closely reflect the mineralogy of the bedrock – sediments higher in the stratigraphic sequence contain compositions that are indicative of more mixing with a variety of farther-travelled geochemical constituents. This relative position of the sediments with respect to bedrock source rock is important geologically as compositional variations must be taken into consideration when drawing geological and perhaps hydrogeological correlations. Hydrogeological implications may involve the evolution of carbonate-rich groundwater. For example, when assessing whether elevated sodium and chloride concentrations in deep groundwaters are due to surface road salt or upwelling of saline waters, it may be useful to consider whether saline surface water has reacted with carbonate-rich sediment during infiltration to deeper aquifers to produce more sodium-bicarbonate water.

Both the grain size and carbonate contents have important implications for the interpretation of the downhole geophysical logs and resulting geophysical response of various sediments. The overall response of gamma radiation for the hydrostratigraphic units is relatively low with maximum values (> 150 counts per second or cps) only recorded < 2% of the time reflecting the lack of clay mineralogy and predominance of glacial flour of carbonate and quartz composing the mud fraction. The natural gamma, conductivity and magnetic susceptibility measurements are in similar ranges as reported in studies of the Oak Ridges Moraine area (Pullan et al. 2002), where higher magnetic susceptibility indicates a greater influx of Shield sediment during meltwater floods.

For the till (aquitard) units, the sediments display a variety of geophysical signals at both the stratigraphic and sedimentary facies scales. Cylindrical and bell-shaped geophysical logs are characteristic of some tills included in the surface units. For the Waterloo Moraine, the muddy Maryhill Till displays sharp-based, elongate, and bell-shaped curves for natural gamma and conductivity data that are responses associated with coarsening upward successions interpreted to be glaciolacustrine sediment (Bajc and Shirota 2007). Smaller, 3–5-m-long funnel-shaped curves in the natural gamma and conductivity logs represent fining-upward successions in glaciolacustrine sediment of the surface units. More symmetric log responses for natural gamma and conductivity characterize units of silty glaciolacustrine facies of the Canning sediment.

The Pre-Canning, Pre-Catfish and Waterloo Moraine stratified sediment (aquifers) all have low gamma and conductivity responses with irregular flat log curves. These measurements are similar to log responses from the sediments in the Oak Ridges Moraine area (Pullan et al. 2002). The most diagnostic characteristic of the geophysical response from the Waterloo sediment units is the abrupt, high amplitude changes in magnetic susceptibility that correlate with pebble to cobble gravel in sediment cores from shield-sourced areas (Bajc and Shirota 2007).

The most characteristic geophysical property of the Catfish Creek Till that is significant to the hydrogeology is the distinctive vertical seismic velocity profiles. The P-wave and S-wave velocities for this till are the maximum recorded in the glacial succession, > 2000 m/s and >1100 m/s respectively, and are unique to this unit. Hence, the physical properties of Catfish Creek Till make it an ideal geological and hydro-stratigraphic marker bed in the Waterloo Moraine area (and across southern Ontario). These velocities are also similar to, though slightly lower than, measurements for the Newmarket Till in the Oak Ridges Moraine (ORM) area (Pullan et al. 2002). This similarity supports a regional correlation that Newmarket Till is the Catfish Creek Till equivalent in the ORM area; significantly, the channel dissection of Newmarket Till with overlying ORM sand and gravel moraine sediments (e.g. Sharpe et al. 2002) is analogous to dissected Catfish Creek Till beneath Waterloo Moraine sand and gravel and supports the comparison between the two geological and hydrogeological settings, as discussed below.

Significance of aquitard continuity

Three regional aquitards (two occurring at depth) are generally considered to provide protection for the lowest aquifer units. The continuity of these aquitards, however, has been questioned following the geochemical detection of surface contaminants in deep well field groundwaters previously believed to be isolated from the surface environment (Bester et al. 2006). Structural and isopach contour maps for the Canning and Catfish Creek Till aquitards both show areas where the aquitard has been eroded or is thin. Apparent discontinuities in the Canning aquitard are up to 1 km wide and tens of kilometres in length. The breaches in the Catfish Creek Till aquitard are significantly smaller, generally measuring hundreds of metres in width and potentially thousands of metres in length. The presence of these hydraulic windows is very important in developing the 3D hydrogeological model. This information provides a predictive geological model and can be used to guide follow-up hydrogeological investigations/tests to verify whether in fact these hypotheses are correct. At the same time, follow-up investigations/tests provide an opportunity to assess the reliability of hydrogeological models constructed mainly from water well records. Breaches in aquitards are only hydrogeologically significant if there are hydraulic connections between shallow aquifers and a drain beneath the aquitard, or if pumping is occurring, as has been shown in two-dimensional numerical models of channel-breached aquitards in the Oak Ridges Moraine (Sharpe et al. 2002). Under non-pumping conditions, breaches across the Catfish Creek Till aquitard would only prove to be hydraulically significant if the underlying aquifers were hydraulically connected to the ground surface (i.e. groundwater discharge points along the Nith and Grand Rivers along the south flanks of the Moraine). If deeper aquifers below the breached aquitard are pumped, as is the case around some municipal water-supply wells, then the vertical connection through the aquitard breaches could allow groundwater to be drawn downward from the shallower aquifer systems at an increased rate. In this scenario, municipal wells could be susceptible to potential contamination within the capture zone above breached aquitards.

Near-surface aquifer distribution

On the basis of the new structural contour and isopach maps (Bajc and Shirota 2007), it is possible to map areas of near surface aquifer and recharge potential. In this study, a mappable unit of sand and gravel having a thickness > 3 m is considered to have aquifer potential, where it intersects the water table. The Grand River Conservation Authority (GRCA) used similar criteria for its assessment of aquifer vulnerability (Holysh et al. 2001). An assessment was also undertaken using a minimum thickness of 5 m with no significant change in the observed patterns of aquifer potential. By mapping the depth to the first aquifer with a thickness > 3 m as well as areas where sediment cover over bedrock is < 1 m (Figure 11), it is possible to determine where aquifers are easily recharged with surface precipitation as well as where they may be more susceptible to surface contamination by chemicals for agriculture, livestock wastes, road salt or industrial contaminants.

Figure 11. Depth to first aquifer of thickness greater than 3 m in the Waterloo Region. The areas in white are where aquifers of thickness greater than 3 m have not been modelled or where the depth to bedrock is greater than 1 m.

Areas shaded in red, orange and purple (Figure 11) are where aquifers occur at shallow depths. Note that in much of the Waterloo Moraine area, including the areas with sediment influx ridges radiating toward the moraine, the shallowest aquifers are located < 1 m below the ground surface. A notable exception to this distribution is the large area where the Upper Maryhill Till outcrops immediately west of the cities of Kitchener and Waterloo. In this area, the surface sediment is composed of clayey diamictons > 10 m thick. The Waterloo landfill site is situated within this area of thick clayey sediment.

The distribution of shallow deposits of sand and gravel in the Waterloo Moraine area and local topographic highs as identified in the new geological model show similarities to a groundwater recharge layer developed by the GRCA for the watershed (Bellamy et al. 2002). The groundwater recharge area identified by the GRCA was delineated using the Guelph All-Weather Sequential-Events Runoff (GAWSER) hydrologic model (Schroeter and Associates 1996) that includes precipitation, temperature, surficial geology and land cover as the main inputs. The GAWSER model results identified three main categories, namely: (1) areas that produced 40% of the total recharge and constituted 10% of the total watershed area, (2) areas that produced 30% of the total recharge and constituted 20% of the total watershed area, and (3) areas that produced 30% of the total recharge and constituted 70% of the total watershed area. These GAWSER model results indicate that much of the rural area over the Waterloo, Orangeville, Paris, and Galt Moraines accounted for a significant proportion (70%) of the total recharge within the Grand River watershed. The undeveloped lands on the Waterloo Moraine in Wilmot and North Dumfries townships (western, southwestern and southeastern parts of the region) show strong downward hydraulic gradients (Holysh et al. 2001) and clearly stand out as important areas of recharge that should be protected from future development. The frequent occurrence of closed depressions, in the topographic undulations of the hummocky portions of the Waterloo Moraine, described earlier as subaquatic fans, means that these areas of permeable surface sediments facilitate groundwater recharge following spring snow melt and storm events. The recharge areas of the Waterloo Moraine have also been mapped by the Region of Waterloo and included in the Regional Official Plan (Blackport et al. 2014, this issue; Blackport and Dorfman 2014, this issue).

Event stratigraphy and depositional models for hydrogeological studies

A sound understanding of the event stratigraphy, including allostratigraphy (a sedimentary unit identified and defined by its bounding discontinuities) (e.g. Zuchiewicz, 1988) or glacial sequence stratigraphy (Powell and Cooper 2002; Cummings et al. 2011), complements the lithostratigraphy described above. This allows for the identification of a stratigraphic architecture and depositional sedimentary facies that can improve understanding of aquifer distribution and geometry, aquifer heterogeneity and delineation of preferred paths and directions for groundwater flow. The event stratigraphy for the Waterloo Moraine area has been developed previously from the geomorphology (Chapman and Putnam 1984), till stratigraphy (Karrow 1974), post-glacial shorelines (e.g. Karrow and Calkin 1984) and, more recently, recognition of regional meltwater events (Shaw and Gilbert 1990; Sharpe et al. 2004; Kor and Cowell 2008). There is limited information on the older aquifer and aquitard units below the Catfish Creek Till aquitard, although the modelling of Bajc and Shirota (2007) and the studies of Marich et al. (2011) provide the most complete delineation of Pre-Canning and Pre-Catfish aquifers in the Waterloo Region. These units most commonly include deposits of proglacial and ice-marginal settings; however, deposition within bedrock-controlled subglacial tunnel valleys and subaquatic fans cannot be ruled out. As a regionally extensive unit in southwestern Ontario, the Catfish Creek Till is correlated with the Newmarket Till (northern till) in the Oak Ridges Moraine area. This diamicton is interpreted as subglacial till deposited during the main, Late Wisconsinan ice advance of the Laurentide Ice Sheet into southern Ontario (e.g. de Vries and Dreimanis 1960; Karrow 1988; Boyce and Eyles 2000). Extrapolation from the Oak Ridges Moraine area and work along the Niagara Escarpment by Kor and Cowell (2008) indicates the next major glacial event prior to deglaciation was the occurrence of subglacial meltwater floods that in places eroded large areas of the Newmarket Till and possibly the Catfish Creek Till (Sharpe et al. 2004). This flooding is believed to have occurred in tunnel valleys that breached the regional till/aquitard sediments. The interpreted hydraulic windows in the Waterloo Moraine area appear to have significantly different characteristics (length, width and depth) from the same features mapped in the Oak Ridges Moraine area (Sharpe et al. 2002; Russell, Arnott et al. 2003, Russell, Sharpe et al. 2003; Brennand et al. 2006; Sharpe et al. 2013). These differences are thought to be due to variations in the basin topography and sediment thickness (thick lower sand) that may have resulted in enhanced downcutting in the Oak Ridges Moraine area. Hydraulic windows through Catfish Creek Till appear to be present; however, it is likely that they are smaller and have different infills than similar features in the Oak Ridges Moraine area. Mapping of these features in the Waterloo Moraine area, however, has been principally delineated by borehole interpolation in contrast to high-resolution reflection seismic data in the Oak Ridges Moraine area. Consequently, the availability of continuous transects of high resolution seismic data could provide valuable additional data support for mapping hydraulic windows in the Waterloo Region.

Aquifer connectivity across regional aquitards may be controlled by the location of hydraulic windows resulting from downcutting and incision during regional meltwater events and localized erosion during nonglacial periods. Locally, the degree of aquifer connectivity may be reduced where subsequent infill of these breaches by fine-grained diamictons and glaciolacustrine sediment has occurred, as described for the lower Maryhill Till. This sediment predominantly has a silt to sandy mud texture and hydraulic conductivities ranging from 10⁻⁵ to 10⁻¹⁰ cm/sec based on its grain-size distribution. Based on current sedimentological understanding and data support, the fills are finer-grained and less permeable than documented from the Oak Ridges Moraine area (Russell, Arnott et al. 2003; Sharpe et al. 2003; Sharpe et al. 2013).

Deposition of the Waterloo Moraine is inferred to have occurred as deglaciation proceeded and an ice-walled glaciolacustrine environment developed (Bajc and Karrow 2004). Influx of sediment into this ice-walled lake basin occurred via a network of radially arranged subglacial channels or conduits. The largest meltwater influx came from the east based on paleocurrent data, the location and orientation of the thickest sediment bodies in the moraine, the moraine morphology and the proximal-to-distal sediment fining from east to west. The multiple, sharp-based fining-upward successions suggest the moraine was constructed by a series of large meltwater pulses. Studies conducted along the axis of the moraine (Russell et al. 2007) suggest this area was a focal point of coarse-grained sediment delivered through meltwater conduits and deposition of ice-marginal subaqueous fans with fine-grained sediment accumulating in the basin (see Figure 8). The architectural elements of the coarse-grained sediment remain poorly documented across the region, but appropriate analogues for their formation are provided in studies of the flood-sourced origin of eskers by Brennand (1994) and Burke et al. (2008). Analogues for the origin of subaquatic fans and their rapid gravel-sand-mud transitions are provided by Russell and Arnott (2003) from the Oak Ridges Moraine.

Two related interpretations are proposed for the deposition of mud in the Waterloo Moraine and lower and middle Maryhill Tills. The gravel and sand components of sediments composing the Waterloo Moraine represent the process of rapid deposition by high-energy meltwater flow, whereas the fine-textured glaciolacustrine sediment and Maryhill diamictons were formed in adjacent ice-marginal or downflow suspension and slumping environments. Locally, Maryhill Tills and related fine-grained sediment represent longer, quiescent periods of low meltwater inputs into the basin formed by suspension settling, ice-marginal debris flow activity and occasional ice grounding in basin muds. Similar fine-grained sediment associations have been described, and similar process models have been inferred, for the Halton Till sediments which are interbedded with and overlie Oak Ridges Moraine sandy, gravelly sediments (Sharpe and Russell 2013). Consequently, units of fine-grained sediment may be expected to be both lateral to and interbedded with the coarse-grained moraine sediment on length-scales of 10, 100 or 1000 metres depending on paleoflow direction. This sediment association (and the underlying process models) explains why deposits of fine-grained sediment have been intercepted within the stratigraphy, and have confounded our understanding of variable permeability, connectivity and preferential pathways for groundwater flow and recharge (e.g. Duckworth 1983). The unexpected discovery of an intermediate-level aquitard at the Region’s Mannheim Aquifer Storage and Recovery (ASR) system demonstrates this complexity (Hodgins et al. 2012).

Regional context

The Waterloo, Orangeville and Oak Ridges Moraines all share a number of common geological elements that relate to similar stratigraphic sequences, sediment facies and a common depositional setting (Barnett et al. 1998; Bajc and Karrow 2004; Russell et al. 2008; Burt 2011). All three moraines were constructed during glacial times punctuated by major meltwater drainage events during down wasting of the main Late Wisconsinan ice sheet approximately 13,000–18,000 years BP. The moraines also overlie a regional till sheet (aquitard) and marker bed, Catfish Creek and Newmarket Tills, that have been dissected by meltwater channels. Large, ice-controlled glacial lakes (possibly subglacial initially) formed depositional basins where meltwater flow was focused and sediments trapped, generally grading from gravel to sand to mud. The moraines consist primarily of sandy to silty sediment, with localized beds of gravel that was deposited by high-energy water flows in esker ridges and subglacial and ice-proximal channels. Discontinuous layers of muddy diamicton and glaciolacustrine sediments are interfingered with and overlie coarse-grained sediment and are significant controls on the pattern of groundwater flow. In the case of the Waterloo and Oak Ridges Moraines, two or three fining-upward cycles have been recognized in the coarse-grained sedimentary sequences that possibly represent deposits formed by the catastrophic release of meltwater into an ice-supported body of water. Coarse-grained sediment zones should be targeted for future exploration of water resources. The onlapping and interfingering of muddy diamictons and glaciolacustrine sediment with underlying sand and gravel is a result of synchronous pulses of sedimentation during waning stages of meltwater flow within ice-marginal lakes. Final lake drainage locally eroded the moraine sequences.

These regional geological/sedimentological environments of stratified moraines produce complex hydrogeological conditions including variable hydraulic conductivities, transmissivities and semi-confined and confined aquifer conditions. Distinct aquifer recharge zones as delineated in the Waterloo Moraine relate to areas where fine-grained sediment was either not deposited or eroded during final drainage of the lake basin during ice margin retreat. The dissected till aquitards that underlie stratified moraines allow for hydraulic connection from moraine aquifers to lower, apparently better protected, aquifers.

Conclusions

As the need to protect and preserve the quality and sustainability of southern Ontario’s groundwater resources intensifies, the importance of 3D geologic mapping, such as highlighted in this paper, will become more apparent. Having a thorough understanding of the distribution and character of aquifers and aquitards in the subsurface is essential for developing and implementing future source water protection plans in the province. This is best accomplished using a basin analysis approach to 3D geological mapping which involves the development of robust conceptual models tied to key hydrostratigraphic units. A comprehensive understanding of the depositional origin and facies characteristics of the modeled units is critical for the successful construction of 3D geological models.

Protocols for 3D geological modelling presented in this paper are designed to improve our understanding of the stratigraphy in the Waterloo Region and other areas of the province. A key aspect of the model development is maximizing the integration of high-quality data. This approach to 3D mapping (e.g. Bajc and Shirota 2007; Logan et al. 2006) blends expert geological knowledge with interpretation guided by a testable conceptual geological model. These elements are integrated by the powerful capabilities of 3D modelling software capable of performing complex tasks quickly and repeatedly as new data are obtained. It is essential that interpretations be evaluated with spatial geostatistics, as performed in this study, to establish a procedure for assessing the reliability and probability estimates of the interpolations because datasets have disparate qualities and regional data are scarce.

It is clear from these experiences that the resulting geological and hydrogeological information must be delivered in a manner that can be understood by both technical experts and non-technical users. Derivative or value-added maps, such as those that estimate aquifer recharge or vulnerability, require sufficient knowledge about the distribution and character of the subsurface sediments. Much of this information is easily extracted from the detailed 3D geological model developed in this study.

Acknowledgements

The authors wish to thank M. Hinton and R. Knight for internal Geological Survey of Canada review and three anonymous journal reviewers of this manuscript for their constructive comments and suggestions. This project was funded under the Groundwater Programs of both the Ontario Geological Survey and the Geological Survey of Canada. The manuscript is published with the permission of the Director of the Ontario Geological Survey. Earth Science Sector contribution number 20130195 (Natural Resources Canada).