Reconstruction of Wisconsinan-age ice dynamics and compositions of southern Ontario glacial diamictons, glaciofluvial/lacustrine, and deltaic sediment

This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier’s archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/authorsrights

Author's personal copy Geomorphology 206 (2014) 421–439

Contents lists available at ScienceDirect

Geomorphology journal homepage: www.elsevier.com/locate/geomorph

Reconstruction of Wisconsinan-age ice dynamics and compositions of southern Ontario glacial diamictons, glaciofluvial/lacustrine, and deltaic sediment W.C. Mahaney a,b,⁎, R.G.V. Hancock c, Alison Milan b, Coren Pulleyblank b, Pedro J.M. Costa d,e, M.W. Milner f a

Quaternary Surveys, 26 Thornhill Ave, Thornhill, Ontario L4J 1J4, Canada Department of Geography, York University, 4700 Keele St., North York, Ontario M3J 1P3, Canada Medical Physics and Applied Radiation Sciences, Department of Anthropology, McMaster University, Hamilton, Ontario L8S 4K1, Canada d Centro de Geologia da Universidade de Lisboa, Faculdade de Ciências da Universidade de Lisboa, Edifício C6, Campo Grande, Lisboa 1749-016, Portugal e Department of Geography, School of the Environment, University of Dundee, Nethergate, Dundee, Scotland DD1 4HN, United Kingdom f MWM Consulting, 182 Gough Ave, Toronto, Ontario M4K3P1, Canada b c

a r t i c l e

i n f o

Article history: Received 1 June 2013 Received in revised form 4 October 2013 Accepted 5 October 2013 Available online 23 October 2013 Keywords: Clast fabric analysis Wisconsinan-age ice dynamics Geochemical sediment sourcing SEM microtextures of glacial grains

a b s t r a c t Macrofabric analysis of till sections in south-central Ontario confirms that clast orientation yields information related to changing ice dynamics during the Wisconsinan glaciation. Test stations in six sections yield unimodal to multimodal macrofabrics that indicate ice flow direction, ranging from SE–NW vectors when ice was thin and flowing radially to variable NE–SW, NNE–SSW, and N–S vectors when ice thickened. Ice loci appear to range from the Lake Ontario basin and southern Quebec (thin ice), Labrador Ungava (thicker ice), and Hudson Bay (thickest ice). The north–south fabric may identify the intergrowth of Keewatin–Labrador ice, presumably the maximum ice thickness of the Last Glacial Maximum (LGM). The preliminary data support the theory that topography directed ice movement during preliminary and closing stages of glaciation in southern Ontario, while thick ice generated flow vectors largely unaffected by underlying topography; hence, leading to clast azimuthal variations reflecting changing ice loci with glacier growth. The fabrics analyzed suggest that inferring difference between ductile and brittle lodgement tills is possible as well as to identifying possible glacial tectonic action/ overburden loading that disturbs the least friction-fit position of clasts in till. The changing dynamics within till sheets are supported, in part, by variations in glacial crushing seen in SEM imagery that depict a range of microtextures from full-scale fractures under brittle conditions to those indicating less viscous transport under ductile regimes. To some degree, changes in flow direction are further supported by geochemical variations that relate to bedrock/regolith up-glacier controlling Ca-dilution and variable concentrations of Rare Earth Elements (REEs). © 2013 Elsevier B.V. All rights reserved.

1. Introduction The glacial geology of the Ontario lobe of the Laurentide ice sheet has been studied by numerous workers (e.g., Karrow, 1967; Prest, 1968; Maclachlan and Eyles, 2013, amongst others) with emphasis on advance and decline over ~90 ka–12 ka. More recently, Occhietti et al. (2011) outlined the growth and decay of the Labrador ice dome and its relation to Keewatin ice. The stratigraphy of two major southern Ontario Wisconsinan lithostratigraphic units–Sunnybrook and Halton– has been the subject of several investigations for over a century, including Hind (1856, 1859), Coleman (1895, 1926), Dreimanis and Terasmae (1958), Terasmae (1960), Karrow (1967), Eyles et al. (1983), Berger (1984), and Sharpe (1987), to mention a few. The question of whether the ‘tills’ were deposited directly by glacial ice or by rainout in a ⁎ Corresponding author. E-mail addresses: [email protected] (W.C. Mahaney), [email protected] (R.G.V. Hancock), [email protected] (P.J.M. Costa), [email protected] (M.W. Milner). 0169-555X/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.geomorph.2013.10.014

glaciolacustrine environment was the center of attention in the 1980s (Sharpe and Barnett, 1985; Eyles and Eyles, 1987) until Hicock and Dreimanis (1989) showed that deposition of till was largely lodgement in character. The addition of macrofabric analysis to the mix of literature on the subject is rather scanty and limited to Ostry (1962), Marsh (1974), Hicock and Dreimanis (1989), and Mahaney (1990a). The further question of whether or not clast fabric analysis could be invoked to shed light on the dynamics of the Sunnybrook and Halton ice has not been asked, possibly because of criticism of fabric analysis in the pursuit of glacial dynamic history reconstruction. In other areas of North America, however, Drake (1974) and Rappol (1985, 1989) demonstrated how till fabrics could be used to reconstruct ice sheet dynamics. In the northern Andes, till fabric has been used to reconstruct changes in flow direction during the early stade of the last glaciation (Early Mérida glaciation stratotype; Mahaney et al., 2001), specifically to establish the growth and demise of an ice cap on the Sierra de Santa Domingo (Mahaney et al., 2010a) and the pre-Mérida glaciation (Mahaney et al., 2010b). In Europe, till fabric analysis appears to have begun with Miller (1884) and has evolved with numerous studies by

Author's personal copy 422

W.C. Mahaney et al. / Geomorphology 206 (2014) 421–439

Author's personal copy W.C. Mahaney et al. / Geomorphology 206 (2014) 421–439

Andrews (1971); Andrews and Shimizu (1966); Andrews and Smith (1970), Andrews and King (1968), Hicock et al. (1996), Bennett et al. (1999), and Larsen and Piotrowski (2003), to name a few. Previous stratigraphic work in several local drainages and at Scarborough Bluffs (Hicock and Dreimanis, 1989; Mahaney, 1990a) indicates that clast azimuths tend to show a pattern, section to section, that reflect thin versus thick ice, with lower beds of each till body containing clasts assuming a SE–NW orientation and with intermediate beds assuming azimuths depicting changing ice loci toward the NE, NNE and N. These studies raise the question of whether or not clast fabrics in two local catchments adjacent to the Scarborough Bluffs might register similar variations reflecting changes in ice loci during stages of the growth and demise of the Halton and Sunnybrook glaciers (Fig. 1A). If the Ontario lobe of the Laurentide ice acted as a separate entity during recession at the end of the last glaciation (Barnett, 1992; Barnett et al., 1998), flow vectors around the western ice margin apparently would have been radial, with a clast fabric signature of NW–SE expected in till sections along the present northwestern shore of Lake Ontario. Presumably, this exit strategy would have been repeated many times in the past; and if the ice exited the basin as a shrinking lobe of the main ice sheet, it might have entered the same way and with the same clast fabric imprint. Fabrics oriented to the NE, NNE, and N associated with intermediate stages of both glacial events likewise bear a similarity to the current understanding of the buildup and demise of Labrador and Keewatin ice (Mahaney, 1990a; Maclachlan and Eyles, 2013). As the ice sheet thickened through accumulation and merging, clast fabric vectors could become less responsive to topographic control and more directly represent the up-glacial region(s) of accumulation. If these clast fabric observations during startup and demise of the LGM indicate radial flow out of the Ontario basin, they are expected to bear a certain similarity to striations interpreted as radial streamlines of flow registered by striae in the marginal areas of glacial lobes as interpreted by Chamberlin (1888). Clast fabric analysis of the Sunnybrook Till (Early Wisconsinan), carried out by Marsh (1974), showed a similar radial pattern during this glacial stade. This observation suggests that clast orientation might shed light on the growth and demise of Wisconsinan ice in south-central Ontario. Hence, the theory that clast fabric might be a tool to reconstruct glacial dynamics is tested here using early and late beds of the Sunnybrook Till and full sections of the Halton Till at several sites along Little Rouge Creek and Rouge River in the newest proposed National Park of Canada. Till fabric represents the least friction fit of inhomogeneous clasts measured at a single station, with mean orientation a response to shear stress. A well-developed till fabric is thought to reflect shearing action at the base of a glacier when till is deposited. Elongate pebbles become aligned parallel with either the strike or dip of inconspicuous shear planes cutting through the till. The fabric thus created is considered penecontemporaneous with deposition of the till, with clast orientation indicating the direction of flow. Subsequent reworking of till following deposition or a shift in ice-movement direction could alter or destroy the original fabric, thus resetting the fabric. In an ice sheet, tension increases proportionally to the sine of the angle of the glacial bed (Flint, 1971). In mountain and in continental ice, density and acceleration from gravity remain the same; only thickness and the sine of the bed angle are different. In continental ice, increasing ice column thickness is offset by a reduced sine of the bed angle, which reduces tension and may explain why continental ice tends to produce more multimodal fabrics compared with mountain ice, a situation that crops up in the data discussed below. While tension is generally reduced, continental ice may also undergo increased localized compression caused by thrusting, stick–slip processes, and differential movement of fast vs. slow ice. Transverse fabrics possibly are generated by these increased forces.

423

The most common till fabric is parallel to ice movement, with a majority of pebbles plunging at a slight angle in the up-ice direction. Fewer pebbles plunge slightly down-ice. Still fewer pebbles may lie approximately transverse to ice flow which in water-saturated conditions, is thought to produce Jeffrey-type clast rotations that appear nearly girdle-like on stereo plots (Lian et al., 2003). The transverse till fabric is rarer but still appears often enough to warrant discussion. Of the transverse type, occasional pebble axes lie perpendicular to presumed flow within a vertical string of stations and are divided about equally between near-horizontal and very slight declivities. Mixed or ambiguous till fabrics also occur but are of limited use for establishing ice movement direction. Prior research suggests that clast orientation may prove useful in reconstructing the growth and demise of Wisconsinan ice in southcentral Ontario and, more particularly, may enhance current understanding of the separating/thinning of Ontario basin ice and the thickening of Labrador and Keewatin ice. If consistent with other analyses such as geochemical and SEM microtextural methods, it is possible that the radial and variable NE/N patterns could indicate changing thickness and provenance of regional ice and can be used as a tool for further investigations of regional ice dynamics. Therefore, the question of whether the observed SE–NW fabrics correspond to thin ice flowing radially from the Ontario basin during early and late stades of each glaciation while variable NE-N vectors represent the thickening of ice in Labrador and the eventual merging of Labrador and Keewatin ice is tested here using a combination of fabric-geochemical controls on source areas (Fig. 1B) and SEM microtextures of the sand fraction. Six sites along Little Rouge Creek and Rouge River containing early and late beds of the Sunnybrook Till and full sections of the Halton Till were examined, making it possible to compare fabrics associated with the two advances.

2. Regional geology In most sections (R45, R48, R49, R61, R15B, Fig. 2) weathering, either of till or lag gravel (i.e., R61; Mahaney and Sanmugadas, 1986) and/or fluviatile deposits (Mahaney and Hancock, 1993) in the upper beds, is considered to have started at the end of the Halton Glaciation when the climate was briefly cold/dry, later reverting to warmer temperatures during the Holocene. Thus, these pedons are relict paleosols, the product mostly of cold and warm climates. The soil/buried soil profile in R47 is younger than Pleistocene and hence is a ground soil complex (Mahaney and Hancock, 1993). Within the Late Pleistocene stratigraphy, deposits are ordered from oldest to youngest: Scarborough Formation, Sunnybrook Till, Thorncliffe Formation, glaciofluvial sand/breccia interbedded within Halton Till and Halton Till. The Scarborough Formation of deltaic sediment is considered along with Sunnybrook Till to represent the Early Wisconsinan glaciation (~90 ka–~60 ka) equivalent to marine isotope stage 4. The onset of glaciolacustrine events represented by the Thorncliffe Formation of Middle Wisconsinan age (marine isotope stage 3), while poorly dated (Barnett, 1992), is known to have ended ca. 25 ka with the onset of colder climate and ingress of Halton ice representing marine isotope stage 2. Quaternary stratigraphic investigations in southern Ontario traditionally have used the mineralogy and physical/chemical characteristics of sediment (Karrow, 1967; Mahaney and Hancock, 1993) and fabrics (Hicock and Dreimanis, 1989; Mahaney, 1990a); to determine lithostratigraphic differences and provenance of different diamictons in the glacial succession. Although a legion of researchers have studied the succession of diamictons at Scarborough Bluffs (Coleman, 1895; Karrow, 1967; Eyles et al., 1983; Mahaney, 1990a) and others have concentrated on areas around Oshawa at Port Hope (Martini et al., 1984) and in the Rouge River basin (Westgate et al.,

Fig. 1. (A) Location of the Lake Ontario Basin and adjoining ice source areas with generalized azimuthal directions of ice streams from major source areas. Map constructed by GIS; (B) Geologic map showing relative distributions of carbonate (Ca track) and PC (PC track) rock in source areas and idealized glacial flow patterns (base map after Douglas, 1970).

Author's personal copy 424

W.C. Mahaney et al. / Geomorphology 206 (2014) 421–439

Fig. 2. GIS topographic map of the study area along the northwestern shore of Lake Ontario.

1987; Mahaney, 1990a), none have concentrated on combinations of fabric-particle size-geochemical-SEM image analysis of the diamictons.

3. Materials and methods Clast fabrics were taken per station at each site as shown on the following figures, with each station comprising a vertical slice of till some 15–20 cm thick. Some sections required stepwise excavation through fan debris from crest to stream level to open up a complete vertical exposure. Sections were cleaned to variable depths and cut back ~0.5 m to obtain undisturbed sediment samples. Fabrics were collected immediately to insure that clast ‘artifacts’ did not develop from dewatering on drying out after exposure to the atmosphere (Dowdeswell and Sharp, 1986). Deposits and overlying paleosol profiles were checked for color variations using the soil color charts of Oyama and Takehara (1970). Soils (paleosols) in the surface of sections were described according to the NRCS (2004) but are not fully analyzed here. Clast weathered/ fresh states were described as indicated at each till sampling station. A genetic code for diamicts was assigned to deposits in some sections following the system of Krüger and Kjær (1999). Approximately 250 g samples were collected to obtain particle size and chemical and mineral analyses. For the most part these beds, upon observation, showed little sign of deformation (van der Meer et al., 2003; Menzies et al., 2006), but micromorphological samples were not collected. The long axis orientation of elongated pebbles embedded within till were measured (typically 50 clasts per station; Benn, 1995), although other workers use 25 clasts (Lawson, 1979) or 30 (Hart and Smith, 1997). The azimuths and inclinations for each collected pebble were taken with a Brunton compass. Only clasts with the a-axis 2.5 × the baxis were used in the analysis. Limestone was encountered approximately 80% of the time (with ±10% variations in some sections), silicates the remainder and the carbonate fraction increasing slightly at the beginning and end of each glaciation.

Once in the laboratory, samples were separated into pebbles for lithic analysis and b 2 mm fraction for particle size, mineral, and chemical analyses. Samples were oven treated to obtain moisture content followed by determination of the air dry weight of the oven dry sample required for particle size analysis (Day, 1965; Mahaney, 1990b). Particle size analysis was completed by hydrometer for the b 63 μm fraction; the 63–2000 μm fractions were determined by dry sieving. The collected fabric data was processed using the Stereo 32 orientation program developed by K. Roller at Institut für Geologie, Mineralogie und Geophysik, Ruhr-Universitat Bochum, Germany. Stereo 32 offers different representations of orientation data (data table, equalangle and -area projections, density plots, rose diagrams) and options to sort and manipulate (rotate) orientation data. The unregistered version may be used freely by students or members of educational institutions. For each fabric, clast axis measurements were plotted as points on a stereonet. The data points were then contoured to show the pattern of data density. Density plots on polar stereo plots map fabric measurements as primary and secondary vectors, with contour spread indicating degrees of tightly constrained or unconstrained orientations. The closeness of contours to the pole on the stereo net indicates the degree of clast inclination. The data were also plotted on traditional rose diagrams to depict primary and secondary orientations. Microscope analysis was carried out using the binocular microscope and the scanning electron microscope (SEM) following procedures outlined by Vortisch et al. (1987) and Mahaney (2002). Across the six sections, subsamples from each particle size station were selected for image analysis of the fine (63–250 μm), medium (250–500 μm), and coarse (500–2000 μm) sand fractions. Geochemical analysis of the samples was performed by instrumental neutron activation analysis (INAA) carried out in the SLOWPOKE Reactor at the University of Toronto (Hancock, 1984). Appropriate standards were employed to calibrate the equipment (Harrison and Hancock, 2005).

Author's personal copy W.C. Mahaney et al. / Geomorphology 206 (2014) 421–439

4. Results 4.1. R45 and R61, Little Rouge Creek Collected at separate locations along Little Rouge Creek, the R45 and R61 sections represent the upper (R45) and lower (R61) beds in the Halton Till. The R61 beds collected lie immediately above the Thorncliffe Formation and, hence, represent the onset of the Halton Glaciation. The R45 beds lie just below the tableland surface of the Halton and, given

425

the thickness of 15 m (usual Halton Till thickness is ~25 m), probably represent part of the middle and upper members of the Halton. Thus, in these two sections and in the other four, the Halton is subdivided into lower, middle, and upper members of the Halton Till. In this instance clast fabric collected from select beds in the lower Halton Till (R61; Fig. 3) can be considered alongside fabrics from the middle to later member of the Halton (R45; Fig. 4) to test the theory that clast fabric can be used to interpret changes in ice sheet thickness in southern Ontario prior to and during the LGM.

Fig. 3. The R61 section with various members and fabric from lower beds of the Halton Till in the upper 4 m.

Author's personal copy 426

W.C. Mahaney et al. / Geomorphology 206 (2014) 421–439

Fig. 4. R45 section showing fabric stations and fabric density and rose diagrams below intertill subglacial beds of sand and gravel. Above the stratified beds, till extends to within 2 m of the land surface.

The base of the Halton drift in R61 is a boulder/cobble bed comprised of 75% limestone and 25% mixed granite and gneiss in a clast-supported sediment ±20 cm thick (4.5 m depth). The boulder bed fabric (Fig. 3), based on a population of 35 clasts, depicts primary azimuths of WSW, NW, and a smaller secondary orientation NNW and suggests possible dewatering effects, and/or shifting dynamics (Jeffrey-type rotation) at the ingress of Halton ice. Immediately above the boulder/cobble bed at 4.1 m depth, clast fabric becomes tightly constrained with a unimodal azimuth of 035°, confirmed by closely packed contours and little azimuthal deviation among the 50 clasts measured. Assuming no erosion or loss of till, the clast fabric swings from NW–SE to NE–SW indicates a rapid thickening of ice emanating from southern Quebec. Twentyfive centimeters above, at 3.85 m depth, fabric becomes quadrimodal but with one very strong heading of 045°, suggesting a slight southward swing of primary ice activity. The near 360o contour spread, a full girdle, indicates a secondary clast orientation is approaching randomness. This distribution may be attributable to Jeffrey-type rotations or possibly to glacial tectonics/dewatering effects. Farther above in R61, at ~2m depth, clast fabric is again quadrimodal but with one strong azimuth of near north, a considerable departure from the mean fabric at 3.85 m depth (Fig. 3). The density diagrams

again show widely spaced contours but with ~70% of the population of clasts measured showing a north orientation. The minor transverse fabrics with headings of W–E and NW–SE, although of little use in determining ice flow direction, may result from compression in selected beds within the ~25 cm slice of till from which measurements were taken. Immediately below the lag gravel (~1.4 m depth), a pentagonal clast fabric dominates but with one preferred azimuthal orientation of NNE, implying a shift to the south of the locus of thickest ice. The minor azimuths recorded here may be related to glacial tectonics, dewatering, or possibly to overburden pressure during emplacement of the lag gravel during the early stage of valley construction at the end of the Pleistocene/beginning of the Holocene, the event following the slow recession of Glacial Lake Iroquois. Two additional sampling stations in R61 (Fig. 3) were used to collect clast fabric data, from the Sunnybrook Till (6.1 m depth) and from a dropstone bed in a glaciolacustrine deposit sandwiched between members of the Sunnybrook Till (8 m depth). Despite the multimodal fabric and two preferred orientations in the glaciolacustrine bed, clast fabric extends 360o producing a girdle, with contours on the density diagram indicating randomness and near horizontal clast inclinations. In the upper bed of the Sunnybrook Till, the rose diagram depicts a unimodal

Author's personal copy W.C. Mahaney et al. / Geomorphology 206 (2014) 421–439

fabric which is shown to include a wide spread of clast orientations with slight inclination in the density diagram. The lower portion of the R45 section (Fig. 4) continues the upsection progression seen in R61. Three fabrics were taken in the lodgement till beneath a 2-m thick bed of subglacial gravel and sand seen in the section. The bottommost fabric at 10.5 m depth shows a dominant azimuth of 035° with a tight range of lesser orientations, ranging from 355° to 065°, while the fabric above (9 m depth) is mainly bimodal with several deviant minor orientations. These minor departures of anomalous clast orientations widen the spread of contours. Below the subglacial gravel beds, fabric remains to the north and east, with a multimodal orientation showing dominant azimuths every 05°, ranging from 05° to 045°. Some of these minor clast fabric orientations may possibly relate to perturbations resulting from loading of overlying breccia beds. Prior to deposition of the subglacial gravel/sand beds above, fabrics in the 3-m-thick bed below show preferred orientations of a majority of the clasts measured to oscillate between 30o and 40o between near north to NE, with multimodal fabrics dominating closer to the subglacial gravel/sand beds. Because clasts within the subglacial gravel beds were slightly imbricated and rounded, no fabric was measured, as the degree of rounding fell short of the 2.5:1 ratio of the a:b axis requirement (Lawson, 1979) that is unsuitable for reliable fabric measurement. Sands show a degree of rounding similar to the gravel beds. Above the subglacial gravel and sand beds, three additional fabrics were spaced at intervals of ~2m. The first (4.5m depth) shows azimuthal headings ranging from N-NNE with tight contour control. Above, at 3 m depth, the fabric becomes multimodal with the density diagram showing widely spaced contours that indicate trimodal conditions from NW to near N. The fabric in station six above at 1.5-m depth depicts a clast fabric shift to a unimodal pattern with very steep inclinations indicated by contours reaching the pole in the density diagram. The fabric progression upward from the subglacial beds depicts biand trimodal fabrics becoming unimodal in the upper station of the section, suggesting a rapid shift of ice direction near the end of the LGM. Tills overlying the glaciofluvial beds of R45 carry fabrics indicating NNE–SSW and N–S orientations indicating the shifting of clast fabric equated to shifts in glacial loci, schematically drawn by Flint (1971, p. 92, diagram ‘b’). The unimodal NW–SE fabric at 1.5-m depth depicted in Fig. 4 is typical of late stage glacial fabrics observed in several sections. 4.2. R47 section, Little Rouge Creek The R47 section (Fig. 5) depicts clear variations in the orientation and episodic variations of ice flow, within the Halton and the Sunnybrook glacial diamictons. The section is principally silty with pronounced coarse particle size spikes at the very top and bottom of the section. At base (33m depth), the Sunnybrook Till displays a clast-supported matrix of pebbles and cobbles. This is overlain by deformed beds of partly stratified gravelly sand making a sharp contact with lodgement till, the latter yielding well-formed columnar and prismatic features on the surface typical of the Sunnybrook (Karrow, 1967). These features disappear as the dried surface is removed to reveal a darker and moist till matrix with a lower frequency of prismatic/columnar forms. The bottom bed of the Sunnybrook depicts fabric showing a unimodal NNE orientation at 015°, with some contours indicating clasts with high inclinations. The deformed beds of sand and gravel above this station are similar in kind to the subglacial gravel and sand seen in R45 (Fig. 4), although with less stratification. Presumably this 22-m-thick section of Sunnybrook Till does not reach to the startup of glaciation. A clast fabric sample in R47 taken at station 9 at a 26m depth depicts a tightly constrained bimodal fabric oriented NW–SE, and the density showing a wide dispersal of clast orientations with less inclination than in the underlying sample (Fig. 5). The upper portion of the Sunnybrook (station 9i at 17 m depth) displays widely dispersed

427

azimuths of NE to NNE exhibiting a weakly constrained bimodal fabric with a dominant azimuth at 020° and increased clast inclination that convincingly indicates a change in glacial dynamics. The NW–SE fabric at station 9 (26 m depth) may relate to a thinning of the Sunnybrook ice as indicated in the R61 section where Sunnybrook Till is separated into lower and upper members. If so, then the NW–SE clast orientation may relate to the upper or lower beds of the Sunnybrook (cf., depths 9.5 vs. 7.5 m in R61; Fig. 3). Above the unconformity separating the two tills (Fig. 5; 13.5 m depth), two fabrics were taken from the Halton Till at stations 12 and 19, both with orientations to the NW. The lower station, at ~13 m depth, shows a multimodal spread in clast direction with prominent azimuths ranging from 305° to 335°. The roughly trimodal fabric at the top of the section indicates similar ice flow, with significant azimuths at 335°, 315°, and 345°, respectively. As shown in the density diagrams, both fabrics yield clasts with low inclinations. From the bottom of the Halton, particle size distributions become somewhat coarser upward in the section. 4.3. R48 Section, Little Rouge Creek Collected along Little Rouge Creek, the R48 section (Fig. 6) was sampled to analyze glacial dynamics in the Rouge basin during the terminal stage of the Early Wisconsinan Sunnybrook glaciation and the LGM. The one fabric from the upper Sunnybrook bed at station 31 (29.2 m depth) yields a NW fabric oriented at 325o with considerable clast orientation spread indicated by the density diagram. Particle size ranges from fine to considerably coarse within both the Sunnybrook and the Thorncliffe; whereas sand, silt, and clay fractions remain relatively constant in the Halton above. Above the Thorncliffe Formation up-section, stations in the Halton Till yield more dominant NW fabrics with greater proportions of steeply inclined clasts compared with the Sunnybrook below. Station 24 (24.1 m depth) displays a WNW unimodal fabric with wide dispersal of clast orientations from near W to NW, some with steep inclinations. At station 23 (23 m depth) the fabric is multimodal, the majority of clasts oriented NW and W, with a slight subordinate orientation of NNE, the latter fabric almost certainly a typical Jeffrey-type rotation. Station 14 (12.5 m depth) yields a unimodal fabric centered at about 320o, again with secondary headings near NW and a minor orientation NE. At station 07 (6.2 m depth) a strong multimodal fabric oriented at 310– 330o also exhibits two minor departures oriented near W and NE. All fabrics yield some clasts with steep inclinations but none meet the inclination record at station 02 near the top of the section and close to the end of the LGM. The orientation here, however, is unimodal with clasts oriented NNE, with a subordinate departure oriented ENE. The fractures (shown in Fig. 6) proved ephemeral as a year after the section was described they disappeared. They were most likely dewatering features. 4.4. R49 section, Rouge River Comprised of a thick segment of the Halton Till (17 m), the R49 section (Fig. 7) from the Rouge River is similar to R47 in Little Rouge Creek, its clast fabric following a similar pattern possibly attributed to similar glacial dynamics. The fabric taken at the bottom of the section (17 m depth) is unimodal with clast orientation tightly constrained around a dominant 330° vector. Five samples taken just above this 17 m station show dissimilar patterns, but generally trending toward NE. The first fabric, taken at 16 m depth (station 14), is tightly constrained and centered around 040°. The density diagram indicates some steeply inclined clasts. Clast fabric becomes multimodal at station 10 (10 m depth) shifting farther E. Comparatively, the fabric at station 7 (7.5 m depth) is more tightly constrained but still oriented NE, and the fabric centers around 055° with a range of strong azimuths from 045° to 085°. Above 6 m depth (station 6), the multimodal fabric taken shows a

Author's personal copy 428

W.C. Mahaney et al. / Geomorphology 206 (2014) 421–439

Fig. 5. The R47 section along Little Rouge Creek, a near intact log of Halton Till deposition, makes a sharp contact with Sunnybrook Till below and alluvial gravels above, the upper zone of which houses a relict Holocene soil.

relatively unique orientation, with the bulk of clasts oriented nearly due E and with significant azimuths at 095°, 085°, 015°, and 060°, respectively, as well as two clasts with seemingly anomalous 305° azimuths. The succeeding fabric (station 5) at roughly 4.5 m depth is taken at

the bottom of a silty bed and contains a fabric orientation trending toward NE, with a strong dominant azimuth at 045°. The topmost fabric (1.5 m depth) returns to a NW orientation, with its primary azimuth at 325° and minor azimuths ranging from 295° to 05°.

Author's personal copy W.C. Mahaney et al. / Geomorphology 206 (2014) 421–439

429

Fig. 6. The R48 section located between R47 and R61 in Little Rouge Creek maintains what appears to be a complete Halton Till log but with a variable NW–SE fabric suggestive of the late stage of the Halton, presumably with earlier beds eroded.

4.5. R15B Section, Rouge River Section R15B (Fig. 8) includes the bottom of the Halton, comprised here of a boulder bed lodged into a deformed sand member of the

Thorncliffe Formation, a stratigraphic situation similar to what is described at the base of the Halton Till at R61 (Fig. 3). The fabric retrieved from the boulder bed was recovered from a matrix material that showed minor folds, one of the few fully deformed beds of all examined.

Author's personal copy 430

W.C. Mahaney et al. / Geomorphology 206 (2014) 421–439

Fig. 7. The R49 section along the Rouge River is an unconformable sedimentary log housing Scarborough Formation overlain with the Halton Till yielding a near complete log with a typical Brunisol (Inceptisol) weathered into the upper beds. The Thorncliffe Formation and Sunnybrook Till are is missing in this sequence.

Author's personal copy W.C. Mahaney et al. / Geomorphology 206 (2014) 421–439

431

Fig. 8. The R15B section lies within the ~15 m terrace (Rouge formation; Mahaney and Sanmugadas, 1986), the R15 soil weathered into lag gravel and sand deposited during the initial fluvial equilibrium stage of stream erosion following glacial retreat. Overlying the Scarborough Formation, as in R49, glaciolacustrine beds merge with a boulder/cobble bed at the base of the Sunnybrook Till, a sedimentary complex similar to what is recorded in R61 for the onset of the Halton advance.

The fabric is remarkably unimodal, oriented to about 030º, with some clasts showing very high inclinations. Above this, station 3 (3.5 m depth) shows a quadrimodal fabric ranging from 350o to 040o with minor departures oriented NW and E, the latter presumably transverse to the ice stream direction. As indicated by the density diagram, the main fabric is oriented near N. At 3.3 m depth, the till makes a sharp transition with noticeable changes in grain size and color, the upper beds exhibiting the usual light 10 YR colors in contrast with the normal 5Y hues lower down, representing a transition between beds with different redox histories in the Halton. Above, in (station 2, 2.2 m depth), the primary fabric swings to 325o with subordinate orientations of 355o, 015o, and 030o. The prime orientation of 325o is taken to indicate ice flow to the NW. At 1 m depth (station 4), the fabric is similar (multimodal) with the density distribution a near carbon copy to that of the underlying station 2 and indicating very high clast inclinations.

The bottommost fabric, in the boulder pavement, shows ice flow directed from NNE. The sample above shows a change in glacial dynamics where the fabric becomes multimodal, with azimuths ranging throughout the northern hemisphere, predominantly directed N/NNW, signifying the beginning of the Halton event. The two fabrics taken in the Halton at the top of the section have very similar orientations, most strongly to the NNW. This range of secondary azimuths, from N to NE, coupled with the primary azimuth in the NNW suggests Jeffrey-type rotations, compression effects, or glacial tectonics. 4.6. Particle size Particle size trends follow Link (1966) and are shown in Figs. 3–8. As reported previously by Karrow (1967), the Halton Till contains higher sand content (~30%) compared with the Sunnybrook Till (~20%). Clay

Author's personal copy 432

W.C. Mahaney et al. / Geomorphology 206 (2014) 421–439

content differs between the two tills, offset by ~10% on average, although both units show anomalies at various depths. Thin sand beds in R47 and R48 in the Sunnybrook and Halton deposits are offset somewhat by the thicker and variegated upper to lower flow regime breccia and sand subglacial beds in R45 (Fig. 9). Comparatively, these sandy/breccia beds are joined by occasional silty anomalies suggestive of possible meltout facies or local variations of source materials. Variations in particle size distributions (cf. R49 with anomalously low clay vs. R47 with higher percentages) are indicative of source area sediment/bedrock differences affecting glacial loads or different porosities/surface to groundwater transfers affecting rates of diagenesis. Variations in the clay mineral composition of Sunnybrook and Halton tills (Mahaney and Hancock, 1993) indicate slightly elevated kaolinite, metahalloysite, illite, illite–smectite, vermiculite, and chlorite in the older Sunnybrook Till compared with the younger Halton. This distribution, although tentative, is suggestive of the redistribution of bedrock/weathered regolith source materials from previously weathered Sangamonian-age surfaces and/or diagenesis. Higher amounts of illite–smectite and illite in the Sunnybrook strongly indicate contact with underlying shales (see Table 6 in Mahaney and Hancock (1993) for chemical composition), just as increased metahalloysite and kaolinite in the oldest deposit may relate to Sangamonian paleosol sediment reworked by Early Wisconsinan ice. 4.7. SEM imagery The SEM imagery was inventoried through all six sections with a view to testing the degree of glacial crushing correlated with fabrics ranging from NW–SE (basin ice) through NE–SW thicker ice to N–S (thickest ice). In addition, glacial and glaciolacustrine/deltaic microtextures were compared to determine if fractures/abrasion microfeatures could be used to distinguish the different sedimentary environments.

Because sands, like pebbles, are inhomogeneous inclusions in till subjected to shear stresses and stick–slip processes occurring within the ice sheet as well as contact with other mineral grains with various moduli of elasticity properties, there is the factor of random chance involved in how much damage a grain may receive (Mahaney, 2002). As indicated in this analysis, some grains appear to sojourn in the ice without receiving much or any damage aside from distinct fracture faces that probably result from bedrock release at some point between the locus of ice accumulation and position in a deposit. Such occurrences are well documented elsewhere (Mahaney, 1995, 2002). Perhaps 10% of all grains analyzed fall in this category. Another 15–25% of all grains receive less damage to perhaps 30–50% of grain surfaces. This leaves ~60% of all grains in the till sheets displaying severe crushing to more than ~50% of grain surfaces. Deltaic sediments, in three out of six sections (R15B (Fig. 10), R49 and R61), show the highest degree of weathered or preweathered grains, a high ratio of subround/subangular plagioclase, and occasional mica, the latter not generally seen in the population of glaciolacustrine, glaciofluvial and glacial grains. Approximately 30–40% of grains analyzed exhibit v-shaped percussion cracks, a characteristic microtexture indicating high velocity aqueous flow. Adhering particles within the deltaic grain population are noticeably reduced compared to grains from other proglacial and glacial environments, most probably due to minimal glacial crushing (Mahaney, 2002). Younger glacial grains from the Sunnybrook Till show a microtexture assemblage similar to that of the still younger Halton Till, with the exception that the limited population of samples available in R61, R47, and R48 display minor weathering effects (principally etching and dissolution effects presumably from diagenesis). While fabric changes through these sections show similar variations of NW–SE ice movement in the beginning and waning stages of the Sunnybrook, there is little change in glacial crushing microfeatures that could be used to support the fabric analysis.

Fig. 9. Particle size distributions of till and subglacial gravel/sand beds in the R45 section. The lower three till stations (8–11.5 m) are clay rich; upper till stations (1.5–4.5 m) contain less clay (~10%) and more silt. The till distributions are nearly linear and similar to particle size distributions of tills in all other sections. The upper subglacial gravel bed has a bulk b2 mm particle size similar to the tills. The lower subglacial gravel and sand beds are typical fluviatile parabolic grain size curves. Slight variations in the tails of the curves represent clay fluctuations that are either related to source material variations or to post-depositional weathering.

Author's personal copy W.C. Mahaney et al. / Geomorphology 206 (2014) 421–439

433

Fig. 10. Overview of quartz clasts in the R15B Section represent the degree of damage inflicted from deltaic, through glaciolacustrine environments, to the end (Halton) stage of Wisconsinan glaciation. (A) Grains from the Scarborough Formation sand member, including samples of well-preserved mica (e.g. center top). The majority of grains are subangular and fairly fresh. Highly weathered samples are calcite. (B) Extensively abraded subangular quartz grain (medium grade, center left) from lower Thorncliffe Formation sand member. Edge rounding and v-shaped percussion cracks prominent on upper fracture face indicate water processing. Two smaller, very fine quartz grains show variable dissolution stages (advanced center, fresher right). Grain to right contains deep troughs that could be glacially crushed. (C) Glacial grain retrieved from the base (3.7 m depth) of the Halton Till. Curved grooves, abraded faces, and high frequency fractures that grade into linear steps suggest significant levels of grinding during the early stage of deposition. Preweathered edges and dissolution cavities (upper left) indicate a complex history. (D) Quartz grain retrieved from the 2.2 m bed in Halton Till, completely surfaced with v-shaped percussion cracks, suggests either a previous fluvial history followed by a sojourn in the ice with little glacial crushing or a moulin history.

Middle Wisconsinan glaciolacustrine grains on the other hand show a marked difference sandwiched between Early and Late Wisconsinan tills. As distilled from the analysis of sands in R48, R61, and R15B, it is possible to use a heightened frequency of occurrence of edge roundness, sphericity, v-shaped percussion cracks, and reduction in occurrence of adhering particles to distinguish these clasts from their glacial neighbors. These glaciolacustrine grains bear a similarity with glaciofluvial grains recovered within tills at R47 and R45 (Fig. 11) and with a sand lens in R48 (station 12). Noticeably different is the character of the upper and lower breccias in R45 when compared with the sand beds there and in R48 and R47 (Figs. 12 and 13), with the breccias carrying higher angularity and fewer v-shaped percussion cracks. A summary of SEM microtextures on grains analyzed within the Halton Till sections reveals a similar pattern to that observed in the Sunnybrook population, the only difference being a somewhat fresher collection of sand in the younger deposit, one subjected to less diagenesis. Boulder/cobble bed pavements at the base of the Halton Till in R15B and R61 reveal 10–15% greater abrasion when compared with grains retrieved from higher in the sections, but otherwise grain damage from glacial crushing is similar in these contact sediments. Rising upward in sections R61, R47, R49, and R15B-above respective contacts with the underlying deltaic, glaciolacustrine, or till beds- glacial quartz grains analyzed carry the usual assemblage of conchoidal and subparallel fractures, thin striae, deep grooves, sharp edges, variable abrasion and increased adhering particles, all hallmarks of glacial crushing (Mahaney, 2002). The majority of grains cluster within the subangular class. As fabrics in these beds most often show a NW orientation, suggesting radial flow and the presence of thin ice, it is plausible that grains deposited in these beds were subject to lower shear stress and reduced

tendency for stick–slip processes to severely damage grains. An attempt was therefore made to determine if such changes in damage might be visible on the populations studied. Outside of slightly lower abrasion, fewer fractures, absence of upturned plates, general lack of striae and grooves, and lower percentages of adhering particles in the lower beds of the Halton, no other major differences were recorded. Moving upward into these till bodies, the normal trend involved increased frequency of crushing microfeatures, especially notable increases of adhering particles, increased abrasion and fracture counts, along with occasional upturned plates, all suggesting adjustments to increased shear stress and possibly stick–slip processes at the ice/bed contact. 4.8. Geochemistry Samples from the six sections were pooled under the labels Halton, glaciofluvial, Thorncliffe, Sunnybrook, and deltaic sediments; and the elemental concentration means and standard deviations were calculated to identify chemical element differences between groups that might be used to determine source areas, degrees of mixing, and dilution between the tills and associated glaciofluvial and deltaic deposits. The summary data (group means and standard deviations) are presented in Table 1. Major and minor elements are listed first, followed by trace elements (all in alphabetical order) and then by the rare-earth elements (ordered by atomic number). Clearly, the primary source of differences in the numbers arises from CaCO3 dilution effects: as Ca contents increase, the concentrations of most other elements decrease. The only other element that behaves somewhat akin to the Ca is Sr, implying that a significant amount of Sr is associated with additional CaCO3 in the sediments.

Author's personal copy 434

W.C. Mahaney et al. / Geomorphology 206 (2014) 421–439

Fig. 11. Grain microtexture differences from till to glaciofluvial beds are represented by clasts in section R45 (Fig. 4), upper 11 m of Halton Till with glaciofluvial member. (A) modified angular grain with sharp edges but otherwise little glacial crushing damage retrieved from the 3.0 m bed. Precipitate and adhering particles are concentrated on upper half of the grain suggesting this clast traveled in ice with only minimal contact with neighboring grains. (B) Triangular faceted grain with adhering particles, partial uplifted plate and abraded conchoidal fractures recovered from the upper breccia bed (5.1 m depth). (C) Glaciofluvial grain retrieved from middle sand bed (6.0 m depth). Slight sphericity, edge rounding, and v-shaped percussion cracks are indicative of glaciofluvial transport. Etching is reduced on central top portion of grain, suggesting differential exposure to moisture post-deposition. (D) Triangular faceted grain from the 10.5 m bed with curved groove (top right), and wide troughs (lower left and right) and uplifted plate remnants (below center) are comparatively fresh with respect to preweathered facies (minor dissolution, lower facet).

The next point of interest in Table 1 is the spread of Ca concentrations within each group, most being relatively low spreading. The deltaic material is the lowest in Ca with a range of 1.0–3.9%. This is followed by Thorncliffe at 7.6–12.3% Ca and Sunnybrook at 9.2–14.5% Ca. The glaciofluvial sands have a high Ca spread at 14.8–26.4%. In this sequence, it is the Halton samples that exhibit anomalous behavior, with Ca varying from 0.9% up to 20.9%. Table 2 shows that most of the very low Ca contents (b 5%) occur at sites R49 and R61. Medium Ca concentration range samples (8–14%) occur at R45, R47, R49, and R61. Highest Ca (N14%) occurs at R15B, R45, R47, and R48. The deltaic deposits represented by samples (n= 10) from the R15B, R49, and R61 sections show the lowest Ca of all groups studied. This supports the contention that these materials originated from the Canadian Shield to the north (Fig. 1A), which would have provided little access to Ordovician limestone source rock. In support of this interpretation are the low mean concentrations of Sr (an important accessory element for aragonitic limestone) and somewhat elevated concentrations of Hf (an accessory element for felsic igneous sources). Within the REEs the light elements (La–Eu) are within the bracketed concentrations of the younger deposits and somewhat elevated within the heavy elements (Eu–Lu), probably a shield signature. The Sunnybrook Till crosses the Ca threshold with a population of 26 samples showing elevated Ca concentrations accompanied by an increase of the mean concentration by half an order of magnitude to 10.4% with a narrow standard deviation of 1.2%. With the increase in percentage of Ca, the mean concentration of Sr, compared with the Scarborough Formation, doubles to 310 ppm and Hf drops slightly, presumably in response to Ca dilution. With very slight perturbations in concentrations of light and heavy REEs, and despite significant

variations of particle size (deltaic sand high; till lower by 20%) and shifts in orientation of fabric within the Sunnybrook, both Early Wisconsinan deposits are in synchronization with one another. The Thorncliffe Formation as a stand-alone entity within the Middle Wisconsinan interstade yields concentrations of Ca and Sr similar to the Sunnybrook composition, only Hf showing a slight increase in mean concentration. This is possibly a reflection of a minor Ca dilution and hydrolysis of silicates leading to loss of K–Mg metals as hydroxides with a slight increase in silica. Compared with the older deposits, the reduction in concentration of the light and heavy REEs is probably related more to lower clay in the glaciolacustrine beds of R15B, R48, and R61. The Halton Till data set consisting of 69 samples from all six sections represents the widest sampling net of all deposits in the succession. The highest mean concentration of Ca recorded when compared to the other groups is accompanied by the highest standard deviation, the latter reflecting a variety of different source materials as well as a greater proportion of samples collected from ice streams emanating out of the St. Lawrence and Ontario basins floored with limestone. Of the 69 samples analyzed ~40% are from beds with a NW–SE flow vector, presumably ice spreading radially from the Ontario basin. The REEs are within the range of concentrations in the older deposit groups excepting the deltaic samples of the Scarborough Formation which are slightly elevated. Glaciofluvial beds (breccia and sand) interbedded within the Halton Till of the R45 section (Fig. 4) are a case in point and bear some relation to thin sand beds in R47 (Fig. 5). Both contain the highest mean concentration of Ca but with an equally high standard deviation and expected inverse relationship between Sr and Hf, which reaches 410 ppm and 3.5 ppm, respectively. Within these glaciofluvial beds the light and heavy REEs are similar or depressed principally following variations in

Author's personal copy W.C. Mahaney et al. / Geomorphology 206 (2014) 421–439

435

Fig. 12. Range of damaged clasts within the till member of Section R47. (A) Well-weathered grain with glaciofluvial characteristics and minor glacial crushing (upper left) retrieved from the base (34 m depth) of the R47 Section. Enlargement (5×) shows comparatively fresh fracture faces (center right) with minor glacial crushing (lower right). Some fracture remnants contain v-shaped percussion cracks. Grains at top right and lower left background show fresh fractures, abrasion, glacial grooves, and sharp edges. (B) Partially abraded angular grain with multidirectional striations, indicating orientation of motive force changes within the till body when crushing was produced. Change in particle size with high sand percentage suggests upper flow regime glaciofluvial transport. Enlargement (10x) shows highlights including the abundance of adhering particles characteristic of this and other sand members within the recovered till. (C) a mix of quartz (Q) and potassium feldspar (KF) grains with representative glacial crushing/abrasion from top (site 9i, 16.5 m depth) in Sunnybrook Till. (D) Angular grain retrieved from top of the Sunnybrook Till. Multiple upturned plates are indicative of substantial forces possibly generated by stick–slip processes. Enlargement (10×) shows detail of fracture pattern, abundant adhering particles and preweathered surface (enhanced enlargement) that probably weakened the grain fabric.

Fig. 13. Quartz record in section R48. (A) Subangular grain retrieved from just below lag gravel at 1.0 m depth. Abundant adhering particles, crescentic gouges, and partial abrasion are characteristic of glacial grinding. Crescent-shaped gouges may be archives of previous mechanical weathering. Minor dissolution features top center are suggestive of diagenesis in upper beds. Enlargement is 10×. (B) Complex subangular grain characteristic of well abraded clasts in various sand lens. Bulbous edges suggest a previous aeolian history, while v-shaped percussion scars at lower left and few adhering particles indicate subsequent glaciofluvial processing. (C) Well-fractured grain from till immediately below sand lens (site 13, 11.4 m depth). Radial and conchoidal fractures grade to crescentic steps. Edges show minor weathering, possibly caused by water percolating through an overlying bed of a subglacial meltwater channel. (No v-shaped percussion scars) Enlargement is 10×. (D) Probable glaciolacustrine grain overwritten with minor crushing retrieved from the Halton just above contact with Thorncliffe Formation (24.5 m depth). Striations and linear steps (top center) indicate glacial grinding, while bulbous edges, sphericity, and v-shaped percussion cracks suggest water transport antedated by aeolian processing. Microfeatures in ‘D’ are similar to those in the R48 sand lens (B above). The aqueous history suggested from grain microtextures may explain the multimodal fabric described.

Author's personal copy 436

W.C. Mahaney et al. / Geomorphology 206 (2014) 421–439

particle size (till with elevated clay at ~20%; glaciofluvial beds with clay at ~5%).

Table 2 Distribution of Ca concentrations in samples from different sites. Ca Percentage

5. Discussion As discussed by Flint (1971), quoting Chamberlin (1888), individual striations, like individual clast fabrics, should ideally be radial near the margin of ice, especially thin ice housed in a structural basin (see Fig. 5.3 (c) in Flint, 1971). A perfect radial pattern is not to be expected, as meltout, postulated to be more frequent in the ablation zone or glacial thrusting over dead ice, may interfere with clast orientation. These, allied with variable geothermal release and dewatering effects in postglacial time, may also affect glacial fabric, producing multimodal orientations, sometimes to near random clast distributions. Because flow can be altered by minor topographic irregularities (Demorest, 1938), in this case with topographic roughness across the Ontario basin, it is understandable that some fabric might depart from the main ice flow vectors changing over time, even varying within an ~25-cm thick slice of till. Like striae, clast fabric has been used for over a century and half to register the least friction fit for clasts carried as load in glaciers. Clast fragments within the ice are relatively well preserved and are likely to suffer fracture or abrasion because of mutual contact or contact with bedrock. As foreign bodies within the ice, clasts form misfitting inhomogeneous inclusions that respond to stresses and strains from far upglacier and become elastic inhomogeneities in an applied stress field (Mahaney, 1995). Depending on the fit of clasts within the ice, the data reported here show they may contact one another and generate high stress leading to fracture, and may assume a least friction position approximating direction of ice flow. Fabric analysis as used here identifies and interprets clast orientation from tightly constrained unimodal to widely oscillating multimodal groupings, both depicted by rose and density diagrams. Reducing a three-dimensional orientation net to two dimensions, produces a

Table 1 Summary data for Rouge River deposits. Calcium concentrations are presented in italics to provide relative diluting effects unit to unit.

Al Ca Fe Mg K Na Ti Ba Co Cr Cs Hf Mn Rb Sc Sr Ta Th U V La Ce Nd Sm Eu Tb Yb Lu

% % % % % % % ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm

Halton

Glaciofluvial

Thorncliffe

Sunnybrook

Till

sand

Formation

Till

Deltaic

69 samples

4 samples

9 samples

26 samples

10 samples

4.6 ± 1.6 13.1 ± 6.6 2.4 ± 1.2 1.8 ± 0.6 1.5 ± 0.3 1.2 ± 0.3 0.26 ± 0.11 450 ± 100 7.9 ± 6.1 50 ± 64 1.4 ± 0.9 5.8 ± 2.3 540 ± 190 53 ± 19 8.5 ± 5.9 320 ± 130 0.57 ± 0.36 4.5 ± 2.2 1.3 ± 0.8 56 ± 39 21 ± 7 42 ± 13 16 ± 5 3.7 ± 1.1 1.1 ± 0.4 0.58 ± 0.24 2.3 ± 1.0 0.35 ± 0.14

2.8 ± 0.6 21.0 ± 5.4 1.7 ± 0.3 – 0.9 ± 0.2 0.8 ± 0.3 0.19 ± 0.05 310 ± 80 5.4 ± 0.7 31 ± 20 0.5 ± 0.3 3.5 ± 2.2 510 ± 60 32 ± 6 5.2 ± 0.9 410 ± 20 0.31 ± 0.14 2.8 ± 1.1 0.9 ± 0.3 34 ± 6 15 ± 4 32 ± 7 18 ± 3 2.8 ± 0.3 0.7 ± 0.1 0.55 ± 0.05 1.7 ± 0.7 0.24 ± 0.08

4.6 ± 0.5 9.5 ± 1.4 2.1 ± 0.4 1.9 ± 0.3 1.7 ± 0.1 1.3 ± 0.2 0.24 ± 0.04 470 ± 40 6.6 ± 1.4 35 ± 8 1.1 ± 0.4 6.8 ± 1.3 510 ± 70 51 ± 6 6.7 ± 1.1 300 ± 20 0.52 ± 0.08 3.9 ± 1.2 1.1 ± 0.8 43 ± 6 20 ± 4 38 ± 7 16 ± 5 3.8 ± 0.6 1.0 ± 0.2 0.57 ± 0.13 2.1 ± 0.2 0.31 ± 0.03

5.2 ± 0.4 10.4 ± 1.2 2.6 ± 0.4 2.3 ± 0.4 2.0 ± 0.2 1.2 ± 0.1 0.28 ± 0.04 540 ± 50 9.1 ± 1.5 48 ± 8 2.0 ± 0.5 5.1 ± 0.8 570 ± 60 73 ± 12 8.5 ± 1.1 310 ± 40 0.59 ± 0.07 6.1 ± 1.4 1.5 ± 0.3 58 ± 10 27 ± 4 52 ± 8 26 ± 6 4.8 ± 0.7 1.1 ± 0.1 0.73 ± 0.12 2.3 ± 0.2 0.34 ± 0.05

6.8 ± 1.4 1.9 ± 1.0 4.1 ± 1.3 2.2 ± 0.5 1.8 ± 0.5 1.6 ± 0.3 0.41 ± 0.15 470 ± 150 15.4 ± 6.4 93 ± 35 2.3 ± 0.9 6.2 ± 1.4 770 ± 250 84 ± 22 15.7 ± 5.2 140 ± 120 1.1 ± 0.4 7.9 ± 2.6 2.3 ± 0.7 94 ± 30 28 ± 9 55 ± 14 18 ± 6 4.7 ± 1.3 1.8 ± 0.4 0.84 ± 0.28 3.6 ± 1.0 0.46 ± 0.12

Site R15B R45 R47 R48 R49 R61 Total

0–5 – – – – 14 2 16

5–10 – – – – 1 1 2

10–15 2 6 2 – 2 – 12

15–21 2 4 6 24 – 3 39

softened database but one capable of reconstructing mean ice flow vectors within a glaciation nonetheless. Sampling within a mean vertical distance of b 20–25 cm (per station) as reported here is subject to about the tightest time control possible to obtain information on what direction ice flowed during the time the bed in question was deposited. Because it is not always possible to determine unconformities within glacial sections, it is possible that the time frame between sampling stations is anything but conformable, as outlined in this database. Till fabric observations carried out in conjunction with other stratigraphic analyses in the Rouge Basin indicate that while clast fabric in the Sunnybrook and Halton tills is preferentially ordered N–S, NNE– SSW, and NE–SW, the beginning and end of both glaciations are different, with clasts carrying a NW–SE orientation, one that might reflect thin ice flowing radially out of the Lake Ontario basin. Previous work carried out by Ostry (1962) and Marsh (1974) also suggested similar patterns, but the study of sections carrying both Sunnybrook and Halton deposits awaited future investigation. Other later investigations showed either that fabric alone with SEM microtexture analysis (Mahaney, 1990a) or fabric compared with clast striae (Hicock and Dreimanis, 1989) could yield important information about glacial vectors in the local area. As a limited test of clast fabric to respond to a low viscous fluid medium with hydrostatic pressure, fabric was taken from a matrixsupported glaciolacustrine dropstone deposit (R61; Fig. 3). Despite a number of clasts accounting for a preferred orientation NW–SE, the remainder present a random fabric with widely spaced contours across 360o of the stereo net and with little inclination, as might be expected of clasts settling out of a fluid medium. The Sunnybrook Till, where present (R61, R47, R48), exhibits only near full sections at R61 and R47 and upper beds at R48. Where sampled, the base or top of the Sunnybrook revealed NW–SE fabrics (R61, R48), consistent with radial flow out of the Ontario basin as predicted by Marsh (1974). The R47 section presents some interesting fabric variations from base to contact with the Halton Till. The NNE–SSW fabric at the base of R47 is consistent with flow from the main Labrador ice and probably does not represent clast fabric orientation at the base of the ice, such beds being unrecoverable. The NW–SE multimodal fabric above at 26 m depth might relate to thinning of the ice with underlying topography directing flow vectors or to meltout, contours in the density diagram depicting nearly half a hemisphere of variation. The fabric at 17 m depth is bimodal, NNE–NE-depicting flow from the full Labrador ice. Fabric in the upper Sunnybrook of R48, juxtaposed with the lower Thorncliffe beds, is trimodal with a strong preferential orientation NNW–SSE, which fits a radial flow model for the Late Sunnybrook in R47 and R48 that is close to the findings of Marsh (1974). Analysis of the Halton succession shows a reasonable fit with a NW– SE flow during early and late stages of deposition, although complicated with multimodal fabrics that could be related to possible meltout processes producing clast rotation in a flowing till matrix, thrusting, dewatering effects, and/or compression effects. However, taking the data at face value, R61 (Fig. 3) depicts an overall NW–SE fabric in the boulder bed at the deposit base, followed by fabric showing clast

Author's personal copy W.C. Mahaney et al. / Geomorphology 206 (2014) 421–439

orientations to the NE, eventually residing N–S, which is similar to fabrics at Scarborough Bluffs (Mahaney, 1990a). At R45 (Fig. 4), this N-NE fabric orientation continues in the lower part of the section, sometimes with tightly constrained unimodal fabric and other times with multimodal orientations or transverse fabrics, which might be mixed beds not useful for direction analysis. This trend continues up-section to a 2-m-thick bed of subglacial gravel and sand that could be moulin channel deposits or subglacial stream channel sedimentation of unknown length but with an E–W orientation. Above the subglacial beds, till resumes with clast fabrics oriented approximately north, varying to NW–SE at the very top of the section. The R47 section contains NW–SE fabrics near the bottom and top of the section, totally in agreement with radial flow out of the Ontario basin. The R48 section, on the other hand, is a peculiar case in point where clast fabric orientations from the base of the Halton at 24 m to near the top at 6 m depth show a multimodal print of approximately NW–SE fabric orientations that are difficult to explain. They may be transverse fabrics, or they may represent only the last phase of Halton ice movement. The R49 section is the perfect model for fabric depicting a radial flow at inception and end of the Halton event, both nearly exact replicas of one another oriented NW–SE with tightly constrained clast azimuths. Fabrics in mid-section are either tightly constrained or multimodal with densities suggesting either transverse fabrics or possible divergent ice stream fluctuations across ±25–30-cm-thick slices of section. The switch back to a NW signature at the top of the section coupled with a sharp increase in silt and decrease in sand just below is indicative of a change in glacial dynamics that may have incorporated more liquid water resulting in intense weathering. The single anomalous station fabric trending E could represent an unexplained ice buildup or compression at the till bed from ice moving out of the north. Somewhat differently, the boulder bed in R15B yields a NNE–SSW fabric, which may indicate that incoming Halton ice thickened rapidly out of the Ontario basin or that part of the record was truncated prior to the ingress of ice from the main Labrador ice sheet. The R15B fabric succession above the Halton pavement is multimodal with the strongest orientation registering NW–SE and probably represents a transverse signature. The above analysis assumes that till is deposited incrementally from bottom to top but cannot account for deposition followed by erosion events producing unconformities that would be difficult to observe. It is only possible to assume that some sediments may be lost due to ice/ meltwater activity during the actual glacial event deposition process, leaving only a net thickness of till following deglaciation. A case in point among the six sections analyzed is R48 where fabric from above the contact with the Thorncliffe Formation suggests a NW–SE orientation with ice moving out of the Ontario basin through the entire Late Wisconsinan stade. Because we have no direct means of dating the base and top of the Halton Till deposit, it is possible (assuming that the fabric analyses are correct) that the entire section or most of it was deposited at or close to the end of the Halton event. Looking to the SEM analysis for support of this assertion, it is only possible to invoke reduced crushing effects, particularly lower fracture counts, reduced striae, and lower frequencies of adhering particles to support a massive deposition of till during the later stages of the Halton Glaciation at R48. To some degree, ice flow direction inferred by clast fabric orientation is also supported by drumlin orientation in southern Ontario and western NY State (Maclachlan and Eyles, 2013). End moraine shapes reported by Muller and Calkin (1993), especially in the Vinemount and Fort Erie regions, reveal that the likely ice vector in the terminal areas was to the SW. Detailed clast fabric is not available; however, these regions might likely yield a clast fabric record similar to what is reported here. Recent bedform analysis (Hess and Briner, 2009) west of the Finger Lakes in NY State reveals a southwesterly orientation that might correlate with clast fabric when it becomes available.

437

A familiar criticism of till fabric analysis is that databases are often not consistent (Bennett et al., 1999), sample populations are sometimes fewer than 50 samples, and investigators might wish for more fabrics per section to fully understand flow vector variance bed by bed. Sample size and sampling frequency within sections are two criticisms that make it difficult to correlate between sections in different areas. However, if unencumbered with high clast frequency at the glacial bed, the stresses and strains inflicted on inhomogeneous inclusions (rock fragments) should theoretically produce a least friction fit (azimuth) parallel with the ice flow, one that is transmitted to the depositing sediment. With the normal till sampling station of~10–30cm vertical extent, a collector cannot be certain of uniform flow vectors or basal temperature fluctuations that could produce meltout or flowtill variants. However, while rose diagrams have been used for the last several decades to depict variations in glacial flow, the new orientation software (Stereo32) demonstrated here provides sample distribution plots, rose diagrams, and density plots with contours, the latter marking a discrete orientation (unimodal fabric) or orientations (bi- or multimodal). While there are still problems with the interpretation of clast fabric and the use of such to discriminate glacigenic facies, some might say impossible (Bennett et al., 1999), data presented here show that tightly constrained uni- to trimodal fabrics with moderate inclinations probably reference consistent flow direction with little azimuthal variation for the time required to deposit the beds analyzed. A multimodal till fabric is a conundrum in that it may well represent changing ice vectors over short time periods, meltout facies or Jeffrey-type rotations; or it could relate to multiple thrusting events over the vertical extent of the till body analyzed. It may also represent dewatering effects, which are dependent upon water content at time of deposition and clay content. Lastly, if compression dominates over tension, clasts may be left nearly normal to flow. Transverse fabrics reported here, some with very wide girdles, bear certain similarities to Jeffrey-type rotations (Davies, 2009), which may suggest clast rotation in a ductile water-saturated till matrix (Lian et al., 2003; Neudorf et al., 2013). Perhaps deforming beds (Menzies et al., 2006) are more common in Laurentide sections than Piotrowski et al. (2001) believed. The Ca-dilution, which affects the chemical element distributions in the till, glaciolacustrine, and deltaic facies analyzed here, also provides the location of targeted source areas for various ice streams entering and leaving the eastern end of the Lake Ontario basin. Initial and closing stages of the Sunnybrook and Halton glaciations were accompanied by deposition of till with high Ca concentrations that could only be derived from movement over a long stretch of limestone. This supports an approximate NE–SW ice movement from the St. Lawrence into the Ontario basin culminating in a radial flow of ice at the western end of the trough. The high Ca concentration in the glaciofluvial beds of R45 suggests that the sediment load came largely from bedrock sources off to the NE during the waning stage of the Halton Glaciation. Similar Ca-concentration spikes in sand beds in R47 and R48 carry like signatures. The inverse relation between REEs plus Hf against Ca/Sr to distinguish shield vs. limestone sources needs additional work for verification. In all, the fabrics correlate well with what is known of ice streams—fast ice reconstruction of the Laurentide Ice Sheet (Winsborrow et al., 2004).

6. Conclusions The fabric analysis discussed here supports the existing interpretation of the recession that occurred at the end of the last glaciation, whereupon the Wisconsinan ice sheet separated into separate lobes, the Ontario lobe receding from the Niagara Escarpment with the Interlobate Moraine extending northeastward toward the St. Lawrence. Clast fabric orientations are in complete agreement with the recession phase reconstructed by other researchers. The path of the ingress of ice during the Last Glacial Maximum (Halton Glaciation) is unknown, but clast fabric analysis from beds at the base of the Halton Till suggests

Author's personal copy 438

W.C. Mahaney et al. / Geomorphology 206 (2014) 421–439

that the Ontario lobe preceded the growth of the entire ice sheet as it enveloped southern Ontario. The growth and demise of the Sunnybrook ice are less certain, but the end phase of glaciation appears to have followed a similar pattern of radial deposition (R48 and R61). The number of multimodal fabrics, if real models of compression effects and/or stone rotation in a ductile till matrix, open up questions related to stress–strain processes at the base of both ice sheets as well as geothermal heat fluctuations that might have accelerated meltout and stick–slip events. The presence or absence of multimodal fabric appears unrelated to particle size as shown in the figures. The presence of subglacial gravel and sand beds in R45 (Halton) and thinner sand beds in R47 (Sunnybrook) and R48 (Halton) register deformation in some cases and the presence of variable upper and lower aqueous flow regimes in others, likely corresponding to the beginning (R47) and end (R45) when ice was presumably thin. At the very least, these multimodal fabrics suggest that tightening of the vertical slice of till examined in the field to b10 cm might tend to produce more tightly constrained fabrics. Image analysis by SEM of grains recovered from six sections shows diagenesis/weathering and fracture/abrasion trends through major Wisconsinan-age deposits in south-central Ontario—Scarborough Formation, Sunnybrook Till, Thorncliffe Formation, and Halton Till. The major microtexture characteristics of sand grade size sediment within these units include a tendency for angularity, abrasion, and frequency of fracture microfeatures to increase with glacial crushing and significant attenuation of these features in the glaciolacustrine and deltaic facies. The geochemistry shows not only distinct variations between different deposits in the stratigraphic succession but clearly indicates that additional carbonates affect the concentrations of other elements in the chemical matrix. The Ca dilution unit to unit is strongest within the Halton, illustrating the range of source materials with higher Ca in the opening and closing stages of ice ingress into the Ontario basin. The lower mean concentration of Ca in the Sunnybrook vs. the Halton probably reflects diagenesis and greater dilution over time. The REE distributions between the tills show overall higher mean concentrations in the Sunnybrook compared with the Halton, presumably reflecting the higher clay concentration in the former. Acknowledgments This research was funded by Quaternary Surveys, Toronto and by minor research grants to WCM from York University. The GIS map was constructed by Andrew Stewart (Strata Consulting, Toronto). We thank Jaap van der Meer (University of London), two anonymous reviewers and Richard Marston (Kansas State U.) for their critical comments that greatly improved the manuscript. We gratefully acknowledge the assistance of Larry Gowland during field work and of Klaus Fecher (deceased, Geology Inst., Marburg University, Germany) for producing the imagery in Figs. 12 and 13. References Andrews, J.T., 1971. Techniques of till fabric analysis. British geomorphological. Res. Group Tech. Bull. 6 (43 pp.). Andrews, J.T., King, C.A.M., 1968. Comparative till fabrics and till fabric variability in a till sheet and drumlin: a small scale study. Proc. Yorks. Geol. Soc. 59 (2), 435–461. Andrews, J.T., Shimizu, K., 1966. Three-dimensional vector technique for analyzing till fabrics: discussion and FORTRAN program. Geogr. Bull. 8 (2), 151–165. Andrews, J.T., Smith, D.I., 1970. Statistical analysis of till fabric: methodology local and regional variability (with particular reference to the north Yorkshire till cliffs). Q. J. Geol. Soc. Lond. 125, 503–542. Barnett, P.J., 1992. Quaternary geology of Ontario. In: Thurston, P.C., Williams, H.R., Sutcliffe, R.H., Stott, G.M. (Eds.), Geology of Ontario, Special Volume 4, Part 2, Ontario Geological Survey. Ontario Ministry of Northern Development and Mines, pp. 1011–1088 (Chapter 21). Barnett, P.J., Sharpe, D.R., Russell, H.A.J., Brennand, T.A., Gorrell, G., Kenny, F., Pugin, A., 1998. On the origin of the Oak Ridges moraine. Can. J. Earth Sci. 35, 1152–1167.

Benn, D.I., 1995. Fabric signature of till deformation. Breidamerkurjökull, Iceland. Sedimentology 42, 735–747. Bennett, M.R., Waller, R.I., Glasser, N.F., Hambrey, M.J., Huddart, D., 1999. Glacigenic clast fabrics: genetic fingerprint or wishful thinking? J. Quat. Sci. 14, 125–135. Berger, G.W., 1984. Thermoluminescence dating studies of glacial silts from Ontario. Can. J. Earth Sci. 21, 1393–1399. Chamberlin, T.C., 1888. The rock-scorings of the great ice invasions. United States Geological Survey, 7th Annual Report, 1885–86 147–248. Coleman, A.P., 1895. Glacial and interglacial deposits near Toronto (Ontario). J. Geol. 3, 622–645. Coleman, A.P., 1926. Ice Ages Recent and Ancient. , 2. The Macmillan Company, New York. Davies, B.J., 2009. British and Fennoscandian Ice-Sheet Interactions during the Quaternary. Department of Geography, Durham U, Durham, UK (Ph.D. Thesis, 502 pp.). Day, P.E., 1965. Particle fractionation and particle size analysis. In: Black, C.A. (Ed.), Methods of Soil Analysis. American Society of Agronomy, Madison, Wi, pp. 545–567. Demorest, M., 1938. Ice flowage. J. Geol. 46, 700–725. Douglas, R.J.W., 1970. Geology and Economic Minerals of Canada. Geological Survey of Canada, Ottawa, Ontario 838 pp. Dowdeswell, J.A., Sharp, M.J., 1986. Characterization of pebble fabrics in modern terrestrial glacigenic sediments. Sedimentology 33, 699–710. Drake, L.D., 1974. Till fabric control by clast shape. Bull. Geol. Soc. Am. 85 (2), 247–250. Dreimanis, A., Terasmae, J., 1958. Stratigraphy of Wisconsin glacial deposits of Toronto area, Ontario. Geol. Assoc. Can. Proc. 10, 119–135. Eyles, C.H., Eyles, N., 1987. Sedimentation in a large lake: a reinterpretation of the late Pleistocene stratigraphy at Scarborough Bluffs, Ontario, Canada. Geology 11, 146–152. Eyles, N., Eyles, C.H., Miall, A.D., 1983. Lithofacies types and vertical profile models: an alternative approach to the description and environmental interpretation of glacial diamicton and diamictite sequences. Sedimentology 30, 393–410. Flint, R.F., 1971. Glacial and Quaternary Geology. Wiley, NY 892 pp. Hancock, R.G.V., 1984. On the source of clay used for Cologne Roman pottery. Archaeometry 26, 210–217. Harrison, T.P., Hancock, R.G.V., 2005. Geochemical analysis and sociocultural complexity: a case study from Early Iron Age Megiddo, (Israel). Archaeometry 47 (4), 705–722. Hart, J., Smith, B., 1997. Subglacial deformation associated with fast ice flow, from the Columbia Glacier, Alaska. Sediment. Geol. 111, 177–197. Hess, D.P., Briner, J.P., 2009. Geospatial analysis of controls on subglacial bedform morphometry in the NY drumlin field-implications for Laurentide ice sheet dynamics. Earth Surf. Process. Landf. 34, 1126–1135. Hicock, S.R., Dreimanis, A., 1989. Sunnybrook drift indicates a grounded early Wisconsin glacier in the Lake Ontario basin. Geology 17, 169–172. Hicock, S.R., Goff, J.R., Lian, O.B., Little, E.C., 1996. On the interpretation of subglacial till fabric. J. Sediment. Res. 66, 928–934. Hind, H.Y., 1856. On the blue clay of Toronto. Canadian Journal, reported without volume and page numbers in Elson, J.A., Early discoverers XXIII, 1966. Till-stone orientation, Henry Youle Hind (1823–1908). J. Glaciol. 6 (44), 303–306. Hind, H.Y., 1859. A Preliminary and General Report on the Assiniboine and Saskatchewan Exploring Expedition. , 19. Legislative Assembly Journal, Canada (Appendix 36). Karrow, P., 1967. Pleistocene geology of the Scarborough area. Geological Rept, 46. Ontario Dept. Mines, Toronto, On (108 pp.). Krüger, J., Kjær, K.H., 1999. A data chart for field description and genetic interpretation of glacial diamicts and associated sediments—with examples from Greenland, Iceland, and Denmark. Boreas 28, 386–402. Larsen, N.K., Piotrowski, J.A., 2003. Fabric pattern in a basal till succession and its significance for reconstructing subglacial processes. J. Sediment. Res. 73 (5), 725–734. Lawson, D.E., 1979. A comparison of the pebble orientations in and deposits of the Matanuska Glacier, Alaska. J. Geol. 87, 629–645. Lian, O.B., Hicock, S., Dreimanis, A., 2003. Laurentide and Cordilleran fast ice flow: some sedimentological evidence from Wisconsinan subglacial till and its substrate. Boreas 32, 102–113. Link, A.G., 1966. The textural classification of sediments. Sedimentology 7 (3), 249–254. Maclachlan, J.C., Eyles, C.H., 2013. Quantitative geomorphological analysis of drumlins in the Peterborough drumlin field, Ontario, Canada. Geogr. Ann. Ser. A 95, 125–144. Mahaney, W.C., 1990a. Macrofabrics and quartz microstructures confirm glacial origin of Sunnybrook drift in the Lake Ontario basin. Geology 18, 145–148. Mahaney, W.C., 1990b. Ice on the Equator. Wm Caxton Ltd, Ellison Bay, WI (386 pp.). Mahaney, W.C., 1995. Pleistocene and Holocene glacier thickness and/or transport histories inferred from microtextures on quartz particles. Boreas 24, 293–304. Mahaney, W.C., 2002. Atlas of Sand Grain Surface Textures and Applications. Oxford University Press, Oxford, UK (237 pp.). Mahaney, W.C., Hancock, R.G.V., 1993. Late Quaternary stratigraphy and geochemistry of the R47 section, Rouge River Basin, south-central Ontario, Canada: correlation with Scarborough Bluffs. J. Quat. Sci. 8, 167–178. Mahaney, W.C., Sanmugadas, K., 1986. Soil development as a function of time in the Rouge River basin, south-central Ontario. Géog. Phys. Quatern. 40, 207–216. Mahaney, W.C., Kalm, V., Bezada, M., Hütt, G., Milner, M.W., 2001. Stratotype for the Mérida glaciation at Pueblo Llano in the northern Venezuelan Andes. J. S. Am. Earth Sci. 13, 761–774. Mahaney, W.C., Kalm, V., Menzies, J., Hancock, R.G.V., Milner, M.W., 2010a. Reconstruction of the pre-Mérida glaciation, northwestern Venezuelan Andes. Sediment. Geol. 230, 10–20.

Author's personal copy W.C. Mahaney et al. / Geomorphology 206 (2014) 421–439 Mahaney, W.C., Kalm, V., Menzies, J., Milner, M.W., 2010b. Reconstruction of the Early Mérida, pre-LGM glaciation, northwestern Venezuelan Andes. Sediment. Geol. 226 (1–4), 29–41. Marsh, P., 1974. Fabric Analysis of the Sunnybrook Till at Scarborough Bluffs, Ontario. York University, Toronto (MA Thesis). Martini, I.P., Brookfield, M.E., Gwyn, Q.H.J., 1984. Quaternary stratigraphy of the coastal bluffs of Lake Ontario east of Oshawa. In: Mahaney, W.C. (Ed.), Quaternary Dating Methods. Elsevier, Amsterdam (427 pp.). Menzies, J., van der Meer, J.J.M., Rose, J., 2006. Till — as a glacial ‘tectonic’, its internal architecture, and the development of a ‘typing’ method for till differentiation. Geomorphology 75, 172–200. Miller, H., 1884. On boulder-glaciation. Proc. R. Phys. Soc.Edinb. 8 (16), 156–189. Muller, E.H., Calkin, P.E., 1993. Timing of Pleistocene glacial events in New York State. Can. J. Earth Sci. 30, 1829–1845. Neudorf, C.M., Brennand, T.A., Lian, O.B., 2013. Till-forming processes beneath parts of the Cordilleran ice sheet, British Columbia, Canada: macroscale and microscale evidence and a new statistical technique for analysing microstructure data. Boreas. http:// dx.doi.org/10.1111/bor.12009. NRCS, 2004. Soil survey laboratory methods manual. In: Burt, R. (Ed.), Soil Survey Investigations Report No. 42, Version 4.0, U.S.D.A. (700 pp.). Occhietti, S., Parent, M., Lajeunesse, P., Robert, F., Govare, É., 2011. Late Pleistocene–Early Holocene decay of the Laurentide ice sheet in Québec–Labrador. In: Ehlers, J., Gibbard, P.L., Hughes, P.D. (Eds.), Developments in Quaternary Science. , Vol. 15. Elsevier, Amsterdam, The Netherlands, pp. 601–630. Ostry, R.C., 1962. An Analysis of Some Tills in Scarborough Township and Vicinity. Dept of Geology, University of Toronto (M.A. Thesis, 65 pp.).

439

Oyama, M., Takehara, H., 1970. Standard Soil Color Charts. Japan Research Council for Agriculture, Forestry and Fisheries. Piotrowski, J.A., Mickelson, D.M., Tulaczyk, S., Krzyszkowski, D., Junge, F.W., 2001. Were deforming subglacial beds beneath past ice sheets really widespread? Quat. Int. 86, 139–150. Prest, V.K., Grant, D.R., Rampton, V., 1968. Glacial Map of Canada. Map 1253A, scale 1:5 000 000, Geological Survey of Canada, Ottawa, On. Rappol, M., 1985. Clast-fabric strength in tills and debris flows compared for different environments. Geol. Mijnb. 64, 327–332. Rappol, M., 1989. Glacial history and stratigraphy of northwestern New Brunswick. Géog. Phys. Quatern. 43, 191–206. Sharpe, D.R., 1987. Quaternary Geology of Toronto Area, Ontario. Geological Society of America Centennial Field Guide, Northeastern Section 339–344. Sharpe, D.R., Barnett, P.J., 1985. Significance of sedimentological studies on the Wisconsinan stratigraphy of southern Ontario. Géog. Phys. Quatern. 39, 255–273. Terasmae, J., 1960. A palynological study of Pleistocene interglacial beds at Toronto, Ontario. Geol. Surv. Can. Bull. 56, 24–40. van der Meer, J.J.M., Menzies, J., Rose, J., 2003. Subglacial till: the deforming glacier bed. Quat. Sci. Rev. 22, 1659–1685. Vortisch, W., Mahaney, W.C., Fecher, K., 1987. Lithology and weathering in a paleosol sequence on Mt. Kenya, East Africa. Geol. Paleontol. 21, 245–255. Westgate, J.A., Chen, F.-J., Delorme, L.D., 1987. Lacustrine ostracodes in the late Pleistocene Sunnybrook diamicton of southern Ontario, Canada. Can. J. Earth Sci. 24 (22), 2330–2335. Winsborrow, M.C.M., Clark, C.D., Stokes, C.R., 2004. Ice streams of the Laurentide ice sheet. Géog. Phys. Quatern. 58, 269–280.