STRATIGRAPHIC ANALYSIS OF LATE WISCONSIN AND HOLOCENE GLACIOLACUSTRINE DEPOSITS EXPOSED ALONG THE NOTTAWASAGA RIVER, SOUTHERN ONTARIO, CANADA

Analysis of 56 outcrop exposures in cut banks along the Nottawasaga River in southern Simcoe County, Ontario, Canada, has led to the identification of eight stratigraphic units (SU1–SU8) that represent a record of changing environmental conditions during deglaciation and exhibit strong controls on shallow groundwater flow in the region. The stratigraphic succession is floored by the Late Wisconsin Newmarket Till (SU1), which is locally overlain by ice-proximal debris flow deposits (SU2). These glacial sediments are overlain by glaciolacustrine silt rhythmites (SU3) that pass upwards into deltaic sand (SU4) and channelized fluviodeltaic sand and gravel (SU5). Lying above the fluvial deposits are widespread interbedded glaciolacustrine sands and silt (SU6), which coarsen up-section toward the ground surface. The succession is locally capped by fluviodeltaic (SU7) and younger fluvial (SU8) deposits. These SUs record sedimentary environments that existed during deglaciation of the region and provide insight into the evolution of glacial lakes Schomberg and Algonquin and the Nipissing phase of the upper Great Lakes. The environmental changes described from sediments along the Nottawasaga River provide insights into basin-scale events that occurred throughout the upper Great Lakes during deglaciation. Qualitative observations of groundwater discharge from sediments at outcrop faces are used to characterize the hydraulic function of the stratigraphic units as well as possible preferential groundwater flow pathways in the shallow subsurface.

D r a f t  81 Large areas of Canada and the northern United States were covered by glacial lakes at the 82 end of the Late Wisconsin glaciation, as meltwaters draining from retreating ice margins were 83 ponded against the ice front in isostatically-depressed basins (Teller 2001;Fig. 1a). In southern 84 Simcoe County, Ontario (Fig. 1b), thick successions of glaciolacustrine sediment found at 85 surface, particularly within low-lying areas, record the evolution and drainage of a series of Late 86 Wisconsin to Holocene deglacial and postglacial lakes (Chapman and Putnam 1984;Bajc et al. 87 2014). Glacial lakes Schomberg (>13 14 C kyr BP (15,700 cal yr BP)), Algonquin (13-10.5 14 C 88 kyr BP (15,700-12,400 cal yr BP)), and the Nipissing phase of the upper Great Lakes (6-4 14 C 89 kyr BP (6,900-4,400 cal yr BP)) inundated parts of the study area and deposited sediments that 90 contain a record of environmental changes as ice margins withdrew from the region (Fig. 1b;91 Deane 1950;Karrow et al. 1975;Lewis et al. 2008) and currently host significant regional and 92 local aquifers (Sibul and Choo-Ying 1971). Detailed sedimentological analysis of these 93 glaciolacustrine deposits will enhance understanding of the three-dimensional (3D) distribution 94 and character of sediments within the shallow subsurface (Anderson 1989), information that is 95 essential for future hydrogeological investigations in the region. Southern Simcoe County has 96 been identified as a region where extensive urban and industrial development is predicted to 97 occur (Ministry of Public Infrastructure and Renewal 2006) and performing this type of 98 investigation is essential for the evaluation, management and protection of the region's potable 99 groundwater resources. 100 This paper presents the first sedimentological description and analysis of sediments 101 exposed in cut banks along the Nottawasaga River in southern Simcoe County (Fig. 1b). 102 Outcrop data are used to develop an improved stratigraphic framework and paleoenvironmental 103 reconstruction of deglacial and postglacial events for lowland areas in the region. These data 104 D r a f t help to better constrain information obtained from continuously cored boreholes drilled as part of 105 ongoing regional-scale 3D mapping efforts conducted by the Ontario Geological Survey (OGS; 106 Bajc and Rainsford 2011;Bajc et al. 2012Bajc et al. , 2014Fig. 3a), and to further refine and test previous 107 interpretations of glacial and postglacial lake evolution in the region that were based primarily on 108 geomorphological observations (Deane 1950;Chapman and Putnam 1984;Finamore 1985).
D r a f t study area, but are commonly deeply buried in the subsurface (Eyles et al. 1985;Bajc et al. 2012, D r a f t the Lake Huron basin forming glacial Lake Algonquin (12.5-10.5 14 C kyr BP (15,000-12,400 cal 151 yr BP); Fig. 2c; Eschman and Karrow 1985). Several discrete phases of this lake have been 152 identified from shoreline studies across the region and record progressive retreat of ice from the 153 basin and isostatic uplift of outlets (Figs. 1,2). Early Lake Algonquin (Karrow et al. 1975) 154 developed after the final phase of drainage of glacial Lake Schomberg, as water levels fell to 155 incrementally lower levels (from 300 m to approximately 250 m asl; Fig. 2b,c). A low-water 156 phase (Kirkfield Low; 12-11.4 14 C kyr BP (14,000-13,300 cal yr BP)) marks the opening of a 157 new outlet to the east near Fenelon Falls (Finamore 1985;Fig.,2c). Isostatic rebound of this lake 158 outlet allowed water levels to rise towards southern outlets and form main Lake Algonquin 159 Finally, uplift of the North Bay outlet following deglaciation triggered water level rise in 165 Georgian Bay; water levels south and west of the outlet rose towards southern outlets, up to 15 m 166 above modern Lake Huron (192 m asl) within the study area, during the establishment of the 167 Nipissing phase of the upper Great Lakes (6-4 14 C kyr BP (6,900-4,400 cal yr BP); Eschman and 168 Karrow 1985). 169 170

171
Fifty six exposures were logged as part of this study (Fig. 3a). Sections were cleaned and logged 172 using standard sedimentological logging techniques, recording grain size, bedding, sedimentary 173 D r a f t structures, bed contacts, clast lithologies, unit geometry and continuity of contacts between 174 sediment groups, and paleocurrent directional indicators (Fig. 3b). Ice flow directions were 175 interpreted from analysis of the long axes of regional subglacial landforms, as well as the 176 orientations of striae on clasts and large boulders with locally well-developed stoss-lee features, 177 Paleocurrent measurements in stratified sediments were obtained by cutting horizontal sections 178 into bedforms and measuring the direction normal to foresets. Qualitative descriptions of 179 groundwater seepage zones were also made. Fossils were isolated from organic bearing 180 sediments by wet sieving to allow for preliminary paleoecological identification of specimens. 181 Material for dating was picked under a binocular microscope to isolate wood, needles, leaves, 182 seeds, molluscs, insect fragments. In rarer cases, larger sticks and logs were sawed and removed 183 from an outcrop, cleaned, then cut into smaller pieces suitable for submission. Wood and leaves 184 of terrestrial plants were sent out to the Illinois State Geological Survey Radiocarbon Dating 185 Laboratory and the University of Ottawa André E. Lalonde AMS Laboratory for radiocarbon age 186 determination by accelerator mass spectrometry (Crann et al. 2016). Wood and leaf samples 187 were subjected to a standard acid-alkali-acid pretreatment prior to combustion to remove any 188 modern contaminants (humic and fulvic acids). Radiocarbon dates are reported as both 189 uncalibrated and calibrated ages (Table 1). Digital data used to create maps include a 5 m digital 190 elevation and hillshade model (DEM; OMNRF 2010) and surficial geology data (OGS 2010). 191

RESULTS 192
Exposures along the Nottawasaga River are composed of eight distinct stratigraphic units 193 (SU1-8) that can be traced consistently along the valley (Figs. 3b,4,5). SUs are distinguished 194 based on an integrated analysis of individual sediment types and genetically-related sediment 195 associations, stratigraphic context, geomorphologic setting and sediment-landform associations 196 D r a f t identified through regional surficial mapping, and chronostratigraphic information from local 197 and regional radiocarbon age determinations.  Silt to sand-rich diamict (SU1) is observed at many locations at the base of cut bank 200 exposures along the Nottawasaga, River (Fig. 3a). The observed thickness of SU1 in outcrops 201 ranges from 1 to 15 m (Fig. 4) and its occurrence is discontinuous along the river due to a highly 202 undulating upper surface. The lower bounding surface of this unit is not exposed in the study 203 area. The diamict is poorly sorted and appears massive at most locations ( Fig. 6a,b), but crude 204 interstratification with sandy and gravelly interbeds (up to 1 m thick) and stacked alternating 205 beds of massive sand-and silt-rich matrix textures (up to 4 m thick) is observed at some outcrops 206 ( Fig. 6c,d). The diamict matrix is moderately to highly consolidated, and displays well-207 developed fissility at several sites (Fig. 6c). Clasts comprise less than 5-10% of the diamict and 208 are dominantly sub-angular to sub-rounded, composed primarily of locally-derived carbonate 209 lithologies. Clast size ranges from granules to boulders over 3 m in diameter. Carbonate clasts 210 are often striated and faceted with flattened tops or bullet shapes (Fig. 6b). Long axes of the 211 clasts tend to be sub-horizontal and show a preferred northeast-southwest orientation (Fig. 6b). A 212 few horizontal to undulating silt to gravel lenses and interbeds, ranging from 10 cm to 2 m thick, 213 are locally observed (Fig. 6c). Groundwater is commonly observed discharging from the outcrop 214 faces at the upper surface of the diamict unit or along the base of interbeds composed of sand or 215 gravel (Fig. 6d). 216 The characteristics of SU1 suggest it is a subglacial till. The presence of striated clasts 217 with a preferred parallel and sub-horizontal orientation, poor textural sorting, and high 218 consolidation and fissility of the diamict matrix, are all consistent with subglacial transport and 219 D r a f t deposition (e.g. Boulton and Deynoux 1981). Although no detailed clast fabric analyses were 220 performed, the crude preferential long-axis alignment, development of stoss-lee features on 221 boulders, and strong parallelism of striae indicate widespread deposition by lodgement 222 processes, consistent with a well-drained and strongly coupled ice-bed interface (Evans et al., 223 2006). Coarse-grained interbeds within the diamict probably record meltwater flow that created 224 local bed separation events during emplacement of the till (Boyce and Eyles 2000). The physical 225 characteristics of SU1 are similar to those reported for the Newmarket Till (Gwyn 1972;Russell 226 and Dumas 1997;Mulligan 2017), as are its surface topography, and stratigraphic position (Todd 227 et al. 2008;Bajc et al. 2012; (Fig. 13b). Outcrops of SU1 are typically found close to 228 areas where Newmarket Till has been mapped at surface near the margins of streamlined, Interbedded, crudely stratified sand-rich diamict and variably textured sorted beds 237 directly overlie SU1 at two locations along the Nottawasaga River (SU2; Fig. 7). Interbedded 238 planar-laminated silt, asymmetrical ripple-scale cross-laminated very fine-to fine-grained sand, 239 planar-laminated fine-to coarse-grained sand, and massive to crudely-stratified gravel with till 240 intraclasts (derived from SU1) comprise the sorted beds, which are up to 0.75 m thick. 241 Sedimentary structures within sorted beds are often highly deformed by shearing, folding and 242 D r a f t faulting (with common low-angle reverse faults) when overlain by diamict beds (Fig. 7b,c). 243 Diamict beds vary from stratified, stone-poor, silt-rich diamict to structureless, well-244 consolidated, clast-and sand-rich. Individual diamict beds in SU2 are up to 1 m thick. Clasts 245 range up to 30 cm in diameter, and lack the preferential long-axis orientation observed in SU1. 246 Contacts between individual beds comprising SU2 are typically sharp, highly deformed, and 247 irregular (Fig. 7). SU2 is 1.2 -3 m thick and is only observed locally, directly overlying SU1, in 248 areas where the surface of SU1 is highly irregular, commonly near the flanks of isolated ridges of 249 Newmarket Till (drumlins) mapped at surface (Figs. 1b, 4;Mulligan and Bajc 2012). 250 Sharp bed contacts, rapid transitions in sediment texture, poor sorting, and pervasive 251 deformation features throughout SU2 suggest a chaotic depositional environment, characterized 252 by rapid fluctuations in sediment deposition rates and processes (Evenson et al. 1977;Glasser et 253 al. 2009). The localized distribution of SU2 on the flanks of highs on the upper surface of SU1 254 suggests that its formation has been controlled in part by the topography of the underlying 255 sediment. Interbedding of well-sorted sand and silt in SU2 records low-energy sedimentation into 256 an ice-marginal lake by density underflows (Smith and Ashley 1985). Gravelly beds with till 257 intraclasts record higher-energy meltwater activity and erosion of the underlying SU1. SU2 is 258 interpreted to record slumping from the ice front and/or local topographic highs during the 259 earliest phases of deglaciation (Fig. 13d-f;Mills 1983;Benn 1989;Mulligan 2013). Some 260 diamict beds may be attributed to minor readvances of the ice front, which could promote 261 subglacial till deposition and lead to glaciotectonization (shearing and faulting) of the bed (White

SU 3: Silt and clay rhythmites with diamict horizons 267
Silt and clay rhythmites abruptly overlie and drape the upper surfaces of both SU1 and 268 SU2 and typically grade upward into sand-and silt-rich sediments of SU4 (Figs. 3b,4,8,9). SU3 269 is observed in all sections logged along the Nottawasaga River (Fig. 4)  underlying fine-grained units as well as at small circular groundwater piping conduits. 296 The overall fine-grained texture of SU3 suggests a low-energy, subaqueous 297 (glaciolacustrine) depositional environment, consistent with an ice-marginal lake. Clay and silt 298 beds were likely deposited from suspension settling whereas coarser-grained sediments were 299 delivered via density underflows entering the lake from subglacial meltwater streams (Van Der 300 Meer and Warren 1997). Seasonal ice cover and precipitation patterns associated with summer 301 melting were probably important controls on sedimentation (e.g. Gilbert 1975). Rhythmites may 302 record a combination of annual (Breckenridge et al. 2004) or diurnal (Schneider andBronge, 303 1996) cycles of sedimentation, storms, or rapid snow melt events (Lamoureux 2000), sediment 304 gravity flows in a prodeltaic environment (Harrison 1975), or distal subaquatic fan sedimentation 305 (Gravenor and Coyle 1985). Small deformation structures are likely caused by loading and 306 dewatering of sediments during discrete sedimentation events. Interbedded diamict horizons may 307 record deposition of coarse-grained debris by rainout of ice-rafted material or by slumping along 308 oversteepened slopes of SU1 and SU2 (e.g. Bennett et al. 2002). Isolated scattered clasts higher 309 up-section within rhythmites of SU3 are interpreted as more distal ice-rafted debris from icebergs 310 calved off a retreating ice margin (Dowdeswell and Dowdeswell 1989;Condon et al. 2002), 311 D r a f t possibly recording reduced glacial influence on the succession as ice retreated northward. 312 Alternatively they may record the break-up of seasonal shore ice (Martini et al. 1993). 313 Gradational up-section textural changes combined with the consistency of northward 314 paleocurrent directional indicators in sands within SU3 suggests continuous sedimentation 315 patterns, consistent with gradual ice retreat from the study area. There is a gradual transition from the silt and clay rhythmites with sand interbeds of SU3 319 into a thick unit of rippled, cross-stratified very fine-to fine-grained sand with silt interbeds 320 comprising SU4 (Figs. 3b,9a). SU4 is between 4 and 10 m thick, generally thins northward, and 321 is not observed north of log 39 9). Both stoss-erosional (Type A) and stoss- The lower part of the coarsening-upward succession that comprises SU6 is interpreted to 397 record inundation of the area by rising lake levels (Karrow et al., 1975) and deposition of 398 Near Angus (Fig. 3a), a unit of very fine-to medium-grained sand (SU7) is observed in 418 several outcrops (Figs. 12,13). The unit has a gradational lower contact with underlying 419 laminated silts which coarsen upward into rippled and planar laminated fine-grained sand (Figs. D r a f t A unit of highly fossiliferous fine-to medium-grained sand with minor silt (SU8; Fig.  449 13c,d) overlies the subaerial weathering horizon at the top of SU7. SU8 is predominantly planar 450 laminated and ripple-scale cross-laminated sand with rare, thin (<2 cm), semi-discontinuous silt 451 drapes and contains an abundance of detrital wood and molluscs, including large unionid clams 452 and some bones (logs 51,52,53;Figs. 4,13). Radiocarbon age determination of wood recovered 453 from a massive peat bed underlying SU8 (Fig. 13e) suggests it was deposited after 5,880±25 14 C 454 yr BP (6,650 -6,750 cal yr BP; Table 1). In this area, SU8 is found at low elevations (192 m asl;455 Figs. 4,13a), overlying accumulations of peat and thin paleosols that overlie SU7 (Fig. 13c-e). 456 The weathering profile above SU7 records a hiatus caused by lowered base levels prior to 457 deposition of SU8, which is interpreted to record rising water levels in the area (Karrow et al., 458 1975;Chapman and Putnam, 1984;Fitzgerald, 1985). The sediments of SU8 record deposition in 459 either a low-energy fluviodeltaic or shoreface setting in a lake ponded within the Minesing basin D r a f t SU3 contains no visible organic material for radiocarbon dating, or evidence of 494 unconformities that could be attributed to rapid changes in water level; hence, it is not possible to 495 differentiate between the deposits of glacial Lake Schomberg and early Lake Algonquin within 496

SU3.
However, the up-section increase in sand within SU3 may record decreasing 497 accommodation space and/or increasing influence of fluviodeltaic systems prograding northward 498 during gradual or stepwise water level fall from 300 m asl (glacial Lake Schomberg) to below 499 250 m asl (early Lake Algonquin) as ice withdrew from the study area (Figs. 14b-d, 15;Mulligan 500 et al., 2015). Raised shorelines, which are particularly well-developed west of Lake Simcoe (  Fig. 14c,d). 504 The gradational transition and consistent northward paleoflow indicators from silt-clay 505 rhythmites (SU3) into ripple-scale cross-stratified and deformed fine-grained sands with 506 interbedded silt of SU4 (Figs. 4,5,9) are interpreted to record the transition from prodelta to 507 delta front sedimentation in response to decreasing accommodation space in early Lake 508 Algonquin due to a combination of sediment aggradation and lowering of water levels following 509 exposure of a new, lower outlet for the Huron basin at Fenelon Falls (Figs. 1, 14c,d, 15). The 510 thick packages of rippled and deformed sands of SU4 suggest relatively high rates of sediment 511 supply, delivered by sediment-laden streams re-working glacial deposits along elevated ground 512 to the south and west (Figs. 1b, 14c,d;Mulligan and Bajc 2012). Unlike many deltas deposited at 513 the margins of ice-marginal lakes during deglaciation of the Great Lakes region, which formed 514 thick (>30 m), well-defined delta bodies (Vader et al., 2010;Arbogast et al., 2017;Connallon 515 and Schaetzl, 2017; Barnett and Karrow this volume), fluviodeltaic sediments exposed along the 516 D r a f t Nottawasaga River are comparatively thinner and no distinct delta morphologies are identified. 517 Large deltas were constructed into stable lake bodies that pre-date SU4 in the area (Glacial Lake 518 Schomberg and equivalents; Chapman and Putnam 1984;Mulligan and Bajc, 2012;Mulligan et 519 al., 2015). These represent the coarse-grained proximal equivalents to the silt and clay rhythmites 520 (SU3) that forms the bulk of the sedimentary infill within the valleys in the study area. However, 521 during deposition of SU4 and SU5, drastically-lowered water levels (Fig. 15)  infill within the valleys in the study area. However, during deposition of SU4 and SU5, 533 drastically-lowered water levels (Fig. 15) and significantly reduced accommodation space 534 prevented the construction of classic Gilbert-type deltaic sediment within the Nottawasaga River 535 cut bank exposures. 536 The delta front deposits of SU4 are overlain by sand and gravel of SU5, which also infill 537 incisions into the underlying units (Figs. 4, 5). This erosional phase likely records a drop in 538 regional base level during the Kirkfield low water phase, following the ice retreat beyond an 539  Organic-rich sediments found in several outcrops elsewhere in the paleolake basin have 551 been used to date the Kirkfield lowstand to between 12 and 11.2 14 C kyr BP (approximately 552 14,615 to 12,973 cal yr BP; Table 1; Fig. 15). Accelerator mass spectrometer radiocarbon dating 553 of Dryas integrifolia leaves found in sand near the top of a coarsening-upward succession 554 interpreted to be correlative with SU4-5 in boreholes SS-11-06 and SS-13-04 ( Fig. 3a; Bajc et al., 555 2014), yielded ages of 12,410±25 14 C yr BP (14,320-14,640 cal yr BP) and 12,810±40 14 C yr BP 556 (15,170-15,320 cal yr BP), respectively (Table 1). These dates provide a maximum age for the 557 timing of this low-water event in Simcoe County. 558 At Angus, a younger package of fluviodeltaic sediment (SU7) is observed (Figs. 4, 12, 582 13), recording declining water plane elevations of post-Algonquin lakes (Fitzgerald 1985 ; Fig.  583 15). The lower, sand-rich part of SU7 is consistent with high sedimentation rates (possibly a 584 single depositional event), while interbedded sand and silt in the upper part of the unit (Fig. 12a) 585 D r a f t likely represent more sluggish accumulation rates, with fewer and smaller sediment pulses. The 586 upper surface of SU7 grades to the Wyebridge phase shoreline mapped in the vicinity (Fitzgerald 587 1985;Mulligan et al., 2015;Mulligan 2017 Fig. 14f). Together, these form a sediment-landform 588 assemblage consistent with a rapid drop in water levels from the upper Orillia to Wyebridge 589 phases (likely recording the opening of a new low outlet in the north), followed by a period of 590 stable water levels. Organic material (plant debris) sampled from SU7 provides a maximum age 591 of 9,950±490 14 C yr BP (10,800-12,200 cal yr BP) for the Wyebridge phase (Fitzgerald, 1985; 592 Table 1). 593 Truncation of SU7 by an unconformity surface marked by a paleosol and associated 594 terrestrial organic deposits records lowered water levels in the region following construction of 595 the delta (Figs. 13, 14g, 15). A similar succession of sediments was exposed in a borehole drilled 596 3.5 km north of Angus within the Minesing basin ( Fig. 1b; Fitzgerald 1982). A likely correlative 597 paleosol is identified in fine-to medium-grained sand beneath the surficial lacustrine silty clay 598 and peat deposits at approximately 178 m asl (9 m lower than in log 52; Figs. 4, 13;Fitzgerald 599 1982). 600 Radiocarbon dating of wood retrieved from peat and detrital wood overlying the 601 unconformity and paleosol separating SU7 and SU8 at Angus has produced dates of between 602 5,880±25 and 7,510±70 14 C yr BP (6,650 -6,950 and 8,200 -8,400 cal yr BP, respectively) and 603 records the latter part of a lake level rise during the mid-Holocene (Bajc and Rainsford 2010; 604 Table 1; Fig. 14g). No sediments that could be attributed to early-mid-Holocene Mattawa high 605 stands were observed in this study due to separation of the study area and from the rest of the 606 Lake Huron by the Edenvale Moraine (Fitzgerald 1982). SU8 deposits grading to 191 m asl are 607 D r a f t interpreted to record sedimentation in Lake Edenvale during the Nipissing phase high stand in 608 the study area (Figs. 12, 14h, 15;Fitzgerald 1985). 609 The stratigraphic units appear to host an exceptional record of changing depositional 610 environments following deglaciation of the region. The paleogeography of the local area was 611 likely a strong control in determining the style of sedimentation and the deep, elongate valleys 612 excavated during Late Wisconsin ice cover provided the necessary accommodation space for the 613 deposition of sediments associated with each lake stage described herein, despite several tens of 614 meters of water level rises and falls (Fig. 15). 615 616

618
Sedimentological data gathered from exposures along the Nottawasaga River valley 619 provide new insights into the hydrostratigraphic framework and potential for groundwater flow 620 and contaminant transport within the southern Simcoe County region, but the detailed three-621 dimensional (3D) distribution of stratigraphic units cannot be determined from this study alone. 622 Groundwater sapping horizons observed in outcrops along the Nottawasaga River 623 indicate that the majority of water escaping from the outcrop faces is travelling through 16,17). Based on their sedimentological characteristics, stratigraphic position, and 625 spatial distribution, these units are interpreted to be correlative with the 'Lake Algonquin Sand 626 Aquifer' of Sibul and Choo-Ying (1971). Variability in the thickness of the unconfined aquifer 627 (0-14 m) results from the high relief (up to 75 m) on the surface of underlying SU1 deposits that 628 controls the geometry of overlying units (Figs. 4, 17). 629 Past hydrogeological investigations in the study area report groundwater flow in the Lake 630 Algonquin Sand Aquifer from the valley margins toward the Nottawasaga River and downstream 631 D r a f t toward the north (Sibul and Choo-Ying 1971;Hill 1982;Fig. 17). Hill (1982) noted marked 632 decreases in nitrate concentration with increasing depth in the Lake Algonquin sand aquifer 633 (SU4-8), which may reflect widespread distribution of the thicker fine-grained lower portion of 634 SU6 (Figs. 4,5,11,16,17) that impedes downward migration of the surface contaminants. 635 Spoelstra et al. (2017) showed that water extracted from many domestic wells penetrating the 636 Lake Algonquin sand aquifer, and groundwater seeps along the river, contain artificial 637 sweeteners, likely derived from rural septic systems in the sand plain. Further testing of water 638 chemistry in shallow and more deeply-buried aquifers may assist in determining the potential 639 connection between aquifer systems in the region. 640 Although no outcrops were analyzed within the Canadian Forces Base Borden property 641 (Fig. 3a), it is likely that the unconfined aquifer complex represented by SU4-8 extends into the to represent glacial Lake Schomberg and early Lake Algonquin sediments (SU3; Figs. 4, 17). 651

Groundwater flow directions and unconfined aquifer thickness within the Nottawasaga 652
River valley appear to be controlled by the topography of the underlying SU1 and SU3 aquitards. 653 These units are considered as a regional hydraulic barrier through which surface water has 654 D r a f t difficulties penetrating (Fig. 16). However, textural heterogeneities and/or erosion or non-655 deposition of SU1 and SU3 (Figs. 4,6,8,17) led to significant groundwater discharge though 656 discrete coarse-grained interbeds within these aquitard units elsewhere in the region (Fig. 16, 17;657 Gerber and Howard 1996;Gerber et al. 2001;Sharpe et al. 2002). These interbeds form localized 658 conduits for groundwater flow, which may create hydraulic windows connecting surficial and 659 deeply buried aquifer units (Boyce et al. 1995;Gerber and Howard 1996;Desbarats et al. 2001). 660 Assessing the risk of contamination of deep aquifers in the region as a result of surface 661 activities remains problematic due to uncertainties with the distribution and potential 662 connectivity of coarse-grained interbeds within the aquitards (SU1 and SU3;Figs. 16,17). Based 663 on observations of groundwater flow discharging from shallow stratigraphic units within the 664 Nottawasaga River basin, it seems that deep aquifers are reasonably well protected against 665 contamination from surface sources in the former lake plain (Fig. 17). However, the 3D 666 distribution and potential connection of coarse-grained layers within the aquitard units is unclear 667 and should be a focus of future research within the study area. Geochemical analysis of the water 668 in underlying aquifers has revealed a mixing of old (13-25 14 C ka BP) and modern water 669 (Aravena and Waassenaar 1993), although the recharge areas remain uncertain. Previous reports 670 and recent sediment drilling by the OGS have revealed that many aquifers underlying the 671 Newmarket Till (SU1) beneath the paleolake Algonquin plain display upward hydraulic 672 gradients, and artesian conditions are encountered beneath many low-lying areas within the 673 lowlands (Sibul and Choo-Ying 1971; Fig. 17). Future geochemical investigations using tracers 674 (i.e. artificial sweeteners, nitrates and/or tritium) as a proxy for quantification of infiltration of 675 surface water into the deep groundwater system and/or in-situ hydrogeologic testing and 676 monitoring may help to assess potential hydraulic connections between surficial and buried 677 D r a f t aquifer units on a more regional scale. Improving our understanding of the stratigraphic 678 framework and controls on groundwater flow in southern Simcoe County is essential, as this 679 rural area is dominated by heavy agricultural use and is slated for major urban expansion in the 680 near future (Ministry of Public Infrastructure and Renewal 2006). 681

683
Developing a better understanding of the depositional history of formerly glaciated basins 684 is essential for the reconstruction of past environmental changes and the evaluation and 685 preservation of modern groundwater resources. Detailed logging and sedimentological analysis 686 of cut bank exposures along the Nottawasaga River allowed identification of eight stratigraphic 687 units that help establish the major controls on changing sedimentation patterns during the Late 688 Wisconsin. The sediments exposed along the Nottawasaga River provide the local signature of 689 regional-scale events that contributed to environmental changes during the deglacial and 690 postglacial period and give valuable insights into the nature of potentially correlative subsurface 691 successions elsewhere in the Lake Algonquin basin. The data presented here suggest a high level 692 glacial Lake Schomberg developed early during ice retreat (300 m asl). Water levels began to 693 incrementally fall to early Lake Algonquin levels (240-250 m asl), followed by a rapid base level         794-11,214 Lewis andAnderson, 1989 GSC-1983 8,460±180 9,139-9,628 Nipissing phase  outlets shown with black star, red arrows show lake drainage pathways, Fig. 3a shown in yellow 1017 rectangle. Modified from Chapman and Putnam (1984). and the Stratigraphic Units identified in this study. Note that no signature of Mattawa high 1117 stand(s) were observed in this study, but were added to the curve to assist in correlating events 1118 from the study area with regions across the paleolake basins. Refer to Table 1 for radiocarbon 1119 data. 1120 1121 Fig. 16: a)   D r a f t Fig. 4: S-N cross-sectional profile showing distribution of SUs exposed along the Nottawasaga River. Refer to Fig. 3a for log locations. Note the truncation/non-deposition of units where SU1 is exposed at higher elevations and thickening of units when SU1 is at lower elevations. No horizontal scale, but distances between logs shown above. Elevations reported in meters above sea level and have not been corrected for glacial isostatic adjustment.
237x157mm (300 x 300 DPI) D r a f t D r a f t Fig. 8: SU3: a) scattered clasts within SU3; beds dipping gently to the right in response to underlying topography of SU1; selected clasts circled in black, knife is 30 cm long; b) silt rhythmites and sand interbeds within SU3. Note continuity of sand bed (light coloured bed at knife tip) along entire outcrop face; knife is 30 cm long; c) dipping beds and a clast-rich horizon (pebble band) within silt and clay rhythmites (knife is 25 cm long); d) rounded limestone clast deforming underlying beds, while overlying beds drape clast surface, and are truncated by planar sandy lamine approximately 2 cm above the clast. Knife blade is 3 cm wide.

573x542mm (96 x 96 DPI)
D r a f t Fig. 9: SU4: a) large-scale soft sediment deformation structures involving thick silt beds that both underlie and drape the sand body. Compass is 12 cm long; b) Type-A climbing ripples passing upward into Type-B ripples with increasing climb angle, fingertip for scale; c) gradational transition from planar and ripple-scale cross-laminated land (SU4) into highly fossiliferous trough cross-bedded medium to coarse-grained sand (SU5). Shovel handle is 50 cm long.
269x735mm (96 x 96 DPI) D r a f t D r a f t Fig. 14: Late Wisconsin -Holocene paleogeographic reconstructions of central Ontario overlain onto regional DEM viewed obliquely from the south (study area outlined in black): a) separation of ice lobes during deglaciation and formation of Oak Ridges Moraine (ORM) and kame terrace deposits along Niagara Escarpment to the west; b) extent of glacial Lake Schomberg following ice retreat; c-f) extent of glacial Lake Algonquin during several distinct phases; g) Stanley low phase; h) Nipissing phase high stand in the Lake Huron basin. Data for water plane elevations and dates from multiple sources ( Fig. 15; Table 1); shorelines were mapped by the author, compared to elevations reported by previous workers, then lake extents were manually digitized onto the hillshaded relief map; see references within text.
187x234mm (300 x 300 DPI) D r a f t Fig. 15: Lake level curve for the study area. Local water plane elevations were calculated based on regional shoreline mapping, timing of lake level fluctuations based on correlation of dated organic material within the Huron basin. Elevation data have not been corrected for isostatic effects. The curve is tied to ice margin positions during deglaciation, changing drainage outlets, and the Stratigraphic Units identified in this study. Note that no signature of Mattawa high stand(s) were observed in this study, but were added to the curve to assist in correlating events from the study area with regions across the paleolake basins. Refer to Table 1 for radiocarbon data.
234x86mm (300 x 300 DPI) D r a f t Fig. 16: a) surface seepage zones identify preferential flow paths of groundwater through coarser-grained interbeds within SU1 (dark tones shown with red arrows); b) small piping feature within silt rhythmites of SU3; c) photo of northern portion of section 37, showing planar tabular geometry of SU3-6 and continuity of seepage horizons (dark coloured areas) along entire outcrop face; section is 14 m high, location of d shown with black rectangle; d) groundwater seepage horizons (arrowed) within SU4 sediments, section 37 (see Fig.  4a for location); e) active erosion of outcrop from groundwater sapping within SU6. Note large blocks of sediment that have fallen from undercutting due to seepage of water along the upper contact on finegrained sediments beneath; f) large amphitheatre-shaped piping scar within deltaic sediments of SU7 southeast of Angus.
565x723mm (96 x 96 DPI) D r a f t Fig. 17: Conceptual model of groundwater flow through shallow subsurface sediments identified in this study. High flow rates are interpreted for SU4-8 (large blue arrows). SU1 and SU3 are interpreted to form hydraulic barriers (thick black arrows) that impede downward flow of groundwater. Note that coarse-grained interbeds within SU1 and SU3 may create thin, highly conductive flow paths (blue arrows) that may provide hydraulic connection between surface and buried aquifers. SU2 not shown. Depending on elevation within the lowlands and proximity to elevated recharge areas, shallow groundwater may infiltrate downward into deeper confined aquifers. Elsewhere, strong upward gradients exist, which create artesian conditions in many areas.