STOP 1: OXFORD CUTS

Michael P. Ekberg, Andrine Dell, Barry Miller, Greg Reed, Tom Lowell.

Location, geomorphic setting, prior work

These outcrops lie along Collins Creek, south of Oxford (fig. 1.1). Together we consider them our "reference sections" because they contain the largest number of units in our idealized sequence. The western outcrop has been called "The Bluffs" because of its nearly vertical face. It may be most famous for the numerous logs protruding from it. This area has attracted geologists for quite some time, including Goldthwait (1955), Durrell et al. (1961), Goldstein (1968), Oldfield (1977), and Reddin (1981). Several graduate students from Miami University investigated stratigraphic problems here. Miller (1986) and Dell (1991) have described the mollusk species. Most recently Ekberg et al. (1993) investigated the till genesis and obtained additional radiocarbon ages.

Ekberg (1991) found that the surficial sediment surrounding Oxford (fig. 1.2) consisted primarily of material formed during glacial recession. Thus he simply subdivided the drift in terms of thickness. The thicker deposits are found in two places, within the valleys of Four Mile and Indian Creeks, and in the upland between Indian Creek and Four Mile Creek (fig. 1.2). This thick upland drift has been interpreted as a moraine complex because the number of glacial advances is different on either side of it (Ekberg et al., 1993).

Unit Description

General Stratigraphy

The stop consists of four outcrops along Collins Creek extending upstream from Rt. 27. The close proximity of the outcrops allows correlation and development of a composite stratigraphy. The numbering of the units (fig. 1.3) follows Ekberg et al. (1991). The basal unit, Ordovician shale and

FIGURE 1.1 Location map for the Oxford cuts. Northwestern most portion of Millville, OH 7.5° quadrant.

Units present at the Oxford outcrop shown in shaded

FIGURE 1.2 Surficial deposits in the area surrounding Oxford. All the surface till formed as superglacial debris flows and is simply subdivided on the basis of thickness from interpretation of well logs. Note the curvilinear patterns of the valleys. Since drift older than late Wisconsin is common in them, they were cut prior to the last glacial cycle. From Ekberg, 1991.

.

FIGURE 1.3 Distribution of units at the Oxford outcrops. Twenty-five pebbles used for fabric analysis, S1 and V1 measures according to Lawson (1979). The middle outcrop (B) lies directly upstream from the east outcrop.

limestone (unit 1) crops out on the downstream side of the exposure. Unit 2, a small wedge of compact massive diamicton (till?), is bounded above by unit 3, a colluvial, loose, orange colored, sandy diamicton. The colluvium grades upward to a massive organic-rich loess (unit 4) containing wood and shells of the terrestrial gastropods Vertigo gouldi hannai and Vertigo modesta (Miller 1986; Dell 1991). Two stumps mark the top of the loess. A striated stone lies against the broken top of the first stump. The second stump has been bent toward 230°. These relationships suggest southwest glacial flow during formation of the overlying till (unit 6). The intact blocks of interbedded limestone and shale bedrock, inclusions of the unconsolidated units, wood, and shear planes indicate a deformation genesis. The striated stones argue for some lodgment, so this unit's origin falls between pure lodgment and deformation.

Unit 7 has abundant logs at its base, but only sparse stones thoughout; one log has a sand lens draped over it. Stones have regional sources and a fabric oriented toward the southwest. The massive nature of this unit makes interpretation difficult but we favor a supraglacial meltout till because of

FIGURE 1.4 A small portion of the Oxford west cut briefly exposed in May of 1993. This appears to be a fluvial/collivial unit preserved in slope adjacent to the bedrock. The radiocarbon ages obtained from the organic unit (3) are shown. In this exposure the lateral equivalent of the organic silt is sandy. All ages grouped together represent replicate anlyses on three different in place stumps. For location see Figure 1.3A.

correlation to the sites located downstream. Unit 8 is a compact, clast-rich (20 to 40%) unit. Striated stones show a northeast-southwest orientation. Smudges of shale bedrock and organic silt along with a dominant local lithology, argue for a subglacial lodgment till origin. The uppermost unit, 9, consists of a thick diamicton with numerous interbedded, well-sorted sand lenses. Silt stringers extend through the clast poor (5%) matrix. These properties indicate supraglacial gravity flowtills.

A loess layer traces between the middle and east sites (Figs. 1.3 B, C); however, the units underlying the loess vary. A compact, sandy silt matrix diamicton, (unit 1) with numerous shear planes dipping north, marks the base of the east section. Blue-green clay stringers and folded inclusions suggest a subglacial deformation till. At the middle site a coarse sand and gravel package with planar cross-beds (unit 2) lies at the base of the exposure; however this alluvium unit pinches out as it rises onto unit 1 at the east outcrop. At the middle outcrop a sandy colluvium (unit 3) overlies the gravel. The lateral geometry of these units indicates that erosion truncates them.

Proglacial lacustrine clays (unit 5) of limited extent rest conformably on the loess and grade upward into unit 6 which contains thin silt and fine sand laminae interbedded with diamictons. Because these sediments grade upward from the clays, we infer the unit to be an ice marginal gravity flowtill deposited in a small proglacial lake. Unit 6 grades upward into a compact unit containing inclusions and streaks of blue clay. One shale bed wraps around a log. The folding and small shear zones suggest a subglacial deformation till (unit 7). The overlying unit, 8, is critical to our interpretation because of a lateral facies change in the Oxford middle outcrop. On the western side unit 8b is massive with low clast concentrations and gradational contacts; the only heterogeneity is one small sand lens. Unit 8b appears similar to other units which occur in many of our outcrops. The key lies on the eastern end of the outcrop where the massive appearance gives way to numerous sand and gravel lenses interbedded within the diamicton matrix (unit 8a). Most of the stratified lenses have lenticular geometry and a few have diamicton stringers within them.

The complex interbedding of the these sediment types argues for a supraglacial gravity flowtill. This lateral geometry implies deposition in a supraglacial position. The small volume and extent of the stratified sediments suggest little running water during deposition of unit 8. Unit 9 is a clast-rich (15%) compact diamicton with well-developed horizontal fissility suggesting lodgment. Unit 10 is a second sequence of supraglacial gravity flowtills. Unit 11 is coarse grained alluvium that cuts into the sequence.

The proximity of the Oxford sections, similar nature and sequence of units, and chronologic data allow direct correlation and suggest three ice advances across the Oxford sites. Unit 2 at the west section, and unit 1 at the east, section provide evidence for the first ice advance. Ordovician bedrock below and an unconformity above bracket its age. The loessial silt is continuous across all three sections and lies directly on either subglacial till, alluvial sand and gravel, or bedrock, indicating a period of considerable subaerial erosion. A second ice advance cut off loess accumulation and caused localized ponding. Following deposition of subaquatic and subaerial debris flows, the ice margin deformed the underlying units. Stagnation of ice and deposition of supraglacial meltout and gravity flowtills followed, but the lack of evidence of running water suggests cold ice-marginal conditions. A readvance of the ice sheet deposited the clast-rich lodgment till during a southwest flow. Final deglaciation ensued during which abundant running water was present, as suggested by the stratified material in unit 9.

Focus on Colluvium/Organic Silt

The Oxford cuts have shown three different expressions of the organic unit, and have provided an opportunity to study it in detail. Based on its variability, we make the provisional interpretation that the package formed from a combination of colluvial, fluvial, and eolian processes.

At Oxford West, stream erosion briefly exposed a layered wedge of sediment (fig. 1.4) For sake of clarity we letter the units temporarily exposed; this entire package constitutes units 2 and 3 described above. Unit A, the basal unit, is a sandy diamicton, massive but with rare lenses. Its matrix coarsens upward. Unit B is a sandy diamicton with common lenses, limestone flag clasts, and rare wood fragments. The lenses pinch out against the bedrock. Unit C is a loose, poorly sorted gravel with a sandy matrix. It grades upward into silty organic rich material. This layer has yielded at least three tree stumps for radiocarbon dating. The unit can be traced several meters downslope, where it overlies colluvium derived from the bedrock. Unit D is a sandy, brown diamicton lens on top of Unit C. It directly underlies unit E, the laminated clay.

The truncation of these deposits against the bedrock slope, and the inclusion of bedrock flags and lenses, argue that a combination of weak fluvial or colluvial processes produced this unit. These processes were stripping the landscape and moving sediment around before wind draped the organic silt across the adjacent bedrock.

A second expression of the organic bed occurs at the small Oxford lowbank exposure (fig. 1.5). At the base several limestone flags suggest close proximity to the bedrock (which is exposed just 30 m upstream). The bulk of the section is a faintly layered silt with weak organic layers. The magnetic susceptibility (fig. 1.5) shows several peaks and troughs reflecting these layers. The upper part of the organic unit is a clay-silt including well developed organic layers. The magnetic susceptibility of this unit is low, reflecting its smaller grain size. A thin, fractured, highly oxidized diamicton caps the outcrop. The heavy alteration occurs because the top of the unit is a stream terrace surface.

The slope of the silts and sediment argue for deposition by an overbank/colluvial system. The finer grained sediments at the top of the sequence appear to be from a shallow lake environment. Radiocarbon dating of this sequence shows that the unit is older than other organic material at Oxford.

The third investigation, at the east outcrop, involves detailed magnetic susceptibility measurements (fig. 1.6). As Figure 1.3C shows, the organic layers slope down and disappear under slump.

FIGURE 1.5 The Oxford lowbank exposure, Fall, 1993. Results of radiocarbon analysis shown. Note magnetic susceptibility profile plotted directly on units measured.

Depending on outcrop conditons, the silt may or may not be exposed at the middle outcrop. We have picked a point on the slope for detailed magnetic susceptibility measurements. The lower till has a relatively low magnetic susceptibility, but the poorly sorted material on the contact has values nearly 4 times higher. The overlying unit is a non-descript silt that grades upward into an organic rich silt. The magnetic susceptibility of the organic material is high. The source of the magnetic signal is not clear, but unless some concentration occurred, it cannot have come directly from the lower till. It may be that erosion of a well drained soil would release authigenic magnetic minerals, and allow them to be deposited. If so, we would argue that the upper silt and the upper till also carry some of that signal .

At all three sites we see that higher positions on the landscape coincide with thin units, whereas

lower positions have thicker sequences. These reflect the role of colluvial processes, and strongly support a partially colluvial origin for some of the lower sediments. The gradational upper contact indicates that colluviation shut down and gave way to silt deposition. Since the silt does not retain bedding structure, is organic rich, and most important, drapes over topography, we feel that the bulk of it is wind blown.

Additional insight into the environment of deposition comes from the study of snail assemblages. Dell (1991) reported species typically found in wet leaf litter, floodplain, and grassy meadow habitats (Table 1.1). Quantitative analysis, including gastropod counts from other sites in southern Ohio, showed that the assemblages from the Oxford sites formed a distinct cluster. The reported species occur today where the average summer temperature is 15°C.

Chronology

At Oxford, organic materials are only present in three successive units; the organic rich silt, the deformation till, and the supraglacial sediments. These three all relate to the second advance. Within the organic-rich silt are two ages of 20,430±160 yr B.P. (Beta-12580) and 19,880±130 yr B.P. (TO-2069), from wood fragments. At the top of the silt, two different in situ stumps provided four ages: 20,030±140 yr B.P. (PITT-0625); 20,620±180 yr B.P. (PITT-0624) and 21,240±150 yr B.P. (PITT-0765); 21,390±200 yr B.P. (PITT-0764). More recent dating has yielded ages centered near 20,800 yr. B.P. (see ISGS-2757, ISGS-2758, ISGS-2760, ISGS-2761, ISGS-2763, ISGS-2762, Table 1). Even though the stumps repre

sent the same period of tree growth, they have 1,500 radiocarbon years variation between them. Taken together, however, these ten ages point to an average age of 20,800 yr B.P.

The deformation till (unit 6, fig. 1.3), directly above the silt, has six age estimates: 19,535±655 yr B.P. (OWU-490), this sample was later analyzed as DAL-5 and gave 25,100±1600 yr B.P.; 19,770±110 yr B.P. (PITT-0623); 19,800±175 yr B.P. (PITT-0627); 19,980±500 yr B.P. (W-92); and 19,970±140 yr B.P. (PITT-0626). Disregarding the DAL-5 age, because of its large error and large difference compared to the other samples, we obtain an average age of 19,836±76 yr B.P.. Although transported, they are within a deformation till which includes many blocks of local materials, suggesting a short (< 0.5 km) transport distance. The difference in age between the in-place stumps and the transported wood is problematic. Although the samples within the silt could be older than the age of the advance, the broken stumps at the top of the silt must be the same as the age of the advance, yet they date 1000 yr older. To assign an age to this advance we have three options: 1) adopt the ages from only the in place stumps and put aside the younger ages from the overlying till, 2) combine the two data sets and compute an average age on

all samples. 3) adopt the ages from the deformation till, but this would ignore the older ages of the in situ stumps that, in theory, provide the tightest stratigraphic control. We favor option 1 and suggest an age of 20,800 yr B.P. for the advance.

The third unit providing ages at Oxford is the supraglacial sediment directly above the deformation till which bears ages of 20,840±110 yr B.P. (PITT-0769); 21,550±150 yr B.P. (PITT-0768); 21,500±170 yr B.P. (PITT-0767); and 21,680±170 yr B.P. (PITT-0766). Averaged, this data set gives 21,274±70 yr B.P. Considering the ages alone, one can see agreement with Gooding's (1975) suggestion of an advance at about 21,000 yr B.P. However, recalling that these ages are in retreat, not advance sediments, and that they are in a stratigraphically younger position than the advance sediments, we conclude that the transported clasts cannot reflect the time of deposition. Because the logs are in good condition, with branches and bark, and occur in meltout till, we suggest transport took place in an englacial position.

The additional data from the Oxford sites puts us in a quandary. We have recently argued (Ekberg et al. 1993) that this site was covered about 20,200 yr B.P. If the 20,200 yr BP age is correct, it is close to the age of 19,670 yr BP for an advance at the Sharonville site (Stop 5). But with the new data, we find a 1000 yr difference between the two sites. Is a time difference of 1000 yr enough to invoke two different ice-margin advances or should these two sites be lumped into one advance?

TABLE 1.1 Oxford gastropods

FIGURE 1.6 Magnetic susceptibility properties vs. the lithology. Measurements taken on 2 cm spacing. Mean (X) and standard deviation (S.D.) for each diamicton shown. Both interpreted as basal tills. For location of study see Figure 1.3 C. Contour interval is 20 units.

Interpretive issues

To promote discussion and have some fun, we will propose that the units at this sequence record five major environmental changes in the late Wisconsin. First came the primarily fluvial erosion that stripped the pre-late Wisconsin units. Second was the transition from colluvial deposition to eolian silt deposition. It does not appear that much water ran across this later landscape, since the silt lies on fluvial units. A colder, dryer climate could have caused this. Such conditions probably incited the glacier to advance. Thus the cooling started sometime before 20,800 yr BP. Third was the cessation of ice movement and partial retreat; it requires a drop in accumulation or an increase in ablation to stop a glacier. The scarcity of fluvial channels implies cold conditons when the glacier stopped. A return to colder conditions, the fourth change, reactivated the glacial margin. Fifth, massive warming destroyed the glacier. The large number of channels indicate abundant meltwater. This notion of a cold vs. warm retreat for the ice margin might be worth some debate.

As stated in the opening, Oxford is our best reference section where the sediments and their stratigraphic associations can be viewed. Two differences complicate the ideal sequence. First, the Oxford cuts record two advances within the late Wisconsin portion of sequence. Thus, the advance and retreat units are repeated. Second, the topographic setting here precluded the formation of lakes.