Timing and dynamics of Late Wolstonian Substage ‘Moreton Stadial’ (MIS 6) glaciation in the English West Midlands, UK

4.1.1.1. Description

Three sections (A–C) were logged along face 1 in the southeast corner of the Meriden Sand Pit (figure 7a). Face 1 provides a 22 m high exposure (figure 7b) that contains two lithostratigraphic units partially covered by overburden from quarry workings, as shown in figure 7c–e and detailed in table 3. In summary, unit i, a sand that is generally massive (Sm), is exposed centrally in the face, between 0 and 12 m marks of the face, with a maximum thickness of 5 m. Localized cross-bedding in the unit provided a palaeocurrent measurement indicating southward flow. Unit ii, a horizontal sand (Sh) with silty clay bedding (Fl), is consistent across the face, dipping east to west from 0 to 22 m marks of the face, with a maximum thickness of 17.5 m. The upper 2 m of the unit contain quartz and quartzite ‘Bunter pebbles’, as well as involutions.

• Download figure
• Open in new tab
• Download PowerPoint

4.1.1.2. Interpretation

The sedimentary sequence at Meriden is interpreted to record a change in depositional environment from fluvial to glaciolacustrine (table 3). The Meriden Lower Sand (unit i)—mainly coarse, massive sand with occasional thinly laminated fine sands—is attributed to deposition in a moderate-energy fluvial environment dominated by bedload transport (e.g. [103–105]) with a palaeocurrent direction towards the south. It probably represents a moderate-energy fluvial sand, infilling the Blythe Valley in a southwest to northeast direction, though a glacial meltwater source cannot be ruled out. It is consistent with a coarse sand, laid down by meltwater from the same northerly source of ice in the nearby Cornets End Quarry [7,36,73,79]. The overlying Meriden Red Sand probably accumulated within a glacial lake. Prominent horizontal strata within the sand are attributed to cyclic or rhythmic deposition, typical of a lacustrine or seasonally variable environment (e.g. [105]). The matrix-supported massive medium-bedded gravel within the sand suggests a ‘flashy’ regime water source proximal to the sequence, typical of glacial meltwater (e.g. [106]). The Meriden Red Sand is equivalent to a laminated fine sand reported by Shotton et al. [36] and Brown [7] in the nearby Cornets End Quarry, laid down within a proglacial lake. Observed periglacial/cryoturbated structures, within the top 2 m of the sedimentary sequence at Cornets End, post-date their host deposits and have been attributed to an extremely cold, dry periglacial environment during the Devensian Stage [7].

4.1.2.1. Description

The Park Farm Pit is the location of the proposed ‘High Speed Two’ ‘Birmingham International Interchange’ railway station. Two sections (A and B) were logged along Face 1 in the northwest cutting of a sand pit at Park Farm (figure 8a). The face exposes a 13 m deep infilled channel cut into Mercia Mudstone bedrock (figure 8b). The Quaternary succession, which includes the infilled channel in the basal unit (i), can be divided into six lithostratigraphical units (figure 8c,d; table 4). In summary, units i to iv are dominantly sand and gravel, with planar or trough cross-bedding common, imbrication, massive gravel and southward palaeocurrents. Unit v is generally finer grained stratified sand and silty sand, with a prominent horizontal parallel stratification. Capping the sequence, unit vi is a massive fine to medium sand that grades into matrix-supported gravel table 5.

• Download figure
• Open in new tab
• Download PowerPoint

4.1.2.2. Interpretation

Overall, the sedimentary sequence at Park Farm is interpreted to record a change in depositional environment from a proglacial braided river to a glacial lake (table 4). The base of unit i has a scoured contact with the Mercia Mudstone Formation, which is typical of a fast-flowing river scouring bedrock [108,109]. The coarse, sandy to gravelly texture of units i to iv suggests relatively high-energy depositional conditions, and the trough cross-bedding, planar cross-bedding and imbrication are common in braided rivers [110]. The material infills the Blythe Valley in a southward direction and is equivalent to the glaciofluvial gravel reported by Cannell [14], which contains quartzite-rich, quartz and sandstone, mudstone, granite, rhyolite, andesite and coal of Welsh or Lake District provenance and infilling the Blythe Valley to the north [7,14]. Collectively, these three lower units are equivalent to the Wolston Sand and Gravel [5–7,26,36,60,61,73], which has been interpreted as outwash sand from the Thrussington glacier to the north of the region [6,61]. The massive gravel in unit i is suggestive of deposition as a longitudinal bar or as a gravel sheet [108]. The sandy and silty texture together with the horizontal planar parallel stratification suggest that the Park Farm Upper Sand (unit v) accumulated by rhythmic deposition as a proglacial lake deposit (cf. [106,111,112]), subject to episodic influxes of coarse sand and gravel from a nearby meltwater source (e.g. [113–115]). These sediments are equivalent to the Upper Wolston Clay of Shotton [1] and attributed to deposition within a glacial lake impounded by ice in the Severn catchment to the south and Oadby glacier to the north [6,61]. The Park Farm Disturbed Sand (unit vi) is interpreted as a glacial lake deposit that has been periglacially disturbed and soliflucted downslope. Periglacial disturbance is evidenced by involutions within the matrix-supported gravel. Similar observations were made in the nearby Cornets End Quarry by Brown [7].

4.1.3.1. Description

Gravel pit extraction has occurred around the Waverley Wood sites for several decades. The present excavation is that at Glebe Farm Pit (figure 9a). The sediments exposed here all form part of the Wolston Glacigenic Formation [1,10,11,20,26,36,49]. Six sections (A–F) were logged at Glebe Farm. Sections A–E were along face 1, and section F along face 2 of the working pit (figure 9a,b; table 5 electronic supplementary material, S2). Above Mercia Mudstone bedrock, the Quaternary sequence comprises three lithostratigraphic units. Unit i is a matrix-supported gravel with a coarse sand matrix. Clasts are predominantly quartz and quartzite ‘Bunter pebbles’. The sand matrix fines toward the top of the unit into fine to medium sand. The unit thins towards the southwest, with an average dip between 8° and 12°. Unit ii features a bedded medium to fine sand, within which the lowest 60 cm appear to be massive. The sand dips along the surface of unit i. After 60 cm the sand shows cross-bedding, though the uppermost part of the sand is horizontally bedded. Unit iii is a massive red–brown sandy diamicton, which is discontinuous across the face and has a sharp undulating contact with the underlying sand.

• Download figure
• Open in new tab
• Download PowerPoint

4.1.3.2. Interpretation

The sedimentary sequence at Glebe Farm is attributed to deposition in a braided river system, shifting from higher to lower energy, before being overridden by glacial ice. Unit i (Baginton–Lillington Gravel) with its matrix-supported gravel and mainly sub-rounded to rounded clasts is typical of a bedload-dominated, high-energy fluvial system (e.g. [108,115–117]). It is equivalent to the Baginton–Lillington Gravel described by Shotton [1,3,6], Crofts [118] and Old et al. [11] at the Waverley Wood site. Unit ii (Baginton Sand), characterized by trough cross-bedded medium to fine sand, with occasional massive sand in pockets (sections B and C; figure 9c, electronic supplementary material, figure S2), is thought to have been deposited by water in a lower energy environment, within a series of channels, forming trough cross-bedded sand as transverse bars and dunes typical of fluvial deposition. Palaeocurrent directions (northeasterly, north and east) indicate flow was predominately to the northeast. It is equivalent to the Baginton Sand described by Shotton [1,3,6], Crofts [118] and Old et al. [11]. Unit iii (Thrussington [Till] Diamicton), which contains striated clasts, is interpreted as a subglacial traction till. The diamicton is equivalent to the Thrussington Till, discussed by Rice [60,61], Shotton [6], Crofts [118] and Old et al. [11]. Within it, erratics of Leicestershire diorites, Coal Measures ironstones and Lower Carboniferous limestone indicate that the glacier moved across the region from the northwest.

4.1.4.1. Description

The Wolston site forms the stratotype of the Wolstonian Stage in the British Isles [1,38]. Within the Wolston SSSI site boundary, one borehole (Wolston SSSI borehole 1; SP41046 74620) and one section (Wolston SSSI section A; SP41064 74625) were logged along a slumped face identified in Old et al. [11] and Shotton [119] (figure 10a). Three lithostratigraphical units are identified using terms previously set out by Shotton [1,36,119], Sumbler [10], Old et al. [11] and Lewis [20] (table 6). Unit i (the Lower Wolston Clay) consists of laminated to massive clay with sandy to gravelly interbeds. Unit ii (the Wolston Sand and Gravel) is massive sand and matrix-supported gravel, grading upwards into unit iii (Dunsmore Gravel), a largely matrix-supported gravel.

• Download figure
• Open in new tab
• Download PowerPoint

4.1.4.2. Interpretation

The sedimentary sequence at Wolston is attributed to deposition within a proglacial lake that was subsequently buried by glaciofluvial outwash [1]. The laminated clay of unit i (Lower Wolston Clay) is typical of a lake deposit, and the interbeds of silt, sand or gravel suggest sediment sources of varying texture and/or fluctuations in energy conditions, all features common to glacial lakes [106]. The gravel bed at 3.35 m depth (figure 10b) may indicate a glacial meltwater source into the lake, and the overall coarsening-upward texture of unit i may record a meltwater source that became increasingly proximal through time (e.g. [20,107]). The likely source of meltwater was the glacier that deposited the Thrussington [Till] Diamicton in the local area (table 6). Similar observations were made by Douglas [120] around Market Bosworth. Furthermore, Lewis [20] reported small clasts in the Lower Wolston Clay resembling those from the Thrussington [Till] Diamicton (Bromsgrove (Keuper) sandstone, coal and erratics from Leicestershire and Nuneaton igneous rocks). The sand and gravel of units ii (Wolston Sand and Gravel) and iii (Dunsmore Gravel), which include sub-angular to angular clasts, are typical of glaciofluvial deposits [110], possibly subject to some periglacial reworking (e.g. by solifluction; [119]).

4.1.5.1. Description

At Seisdon, two sections (A and B) and one face [1] in the current working pit were logged (figure 11a). Two lithostratigraphical units are identified in the present study, though due to complexity of the deposits at Seisdon, Morgan [17] found it impossible to correlate individual units in the pit. Unit i (Seisdon Sand) is a discontinuous, horizontal, thinly bedded, fine to medium sand containing some silty clay and coarse sand. The sediments dip between 22 and 25°. Unit ii (Seisdon Gravel) is a horizontally bedded, matrix-supported gravel with horizontal, thinly bedded fine to medium sand.

• Download figure
• Open in new tab
• Download PowerPoint

• Download figure
• Open in new tab
• Download PowerPoint

4.1.5.2. Interpretation

Overall, both lithostratigraphical units at Seisdon are interpreted as fluvial or glaciofluvial in origin. The generally coarse-grained nature and stratification of the deposits, and the rounded to well-rounded clasts, are consistent with deposition in a relatively high-energy braided or anastomosing river system (e.g. [108]). The sandstone clasts of local Keuper Sandstone and predominantly ‘Bunter pebbles’ lithologies in the Seisdon Gravel imply a local river system in the valley, as the material is locally sourced, although it is not possible to exclude a glaciofluvial source. Deformation of the Seisdon Sand has resulted in some tilting of some strata, though the cause is not known.

4.2.1.1. Meriden

At the Meriden site three samples were collected for luminescence dating: two from the Meriden Lower Sand (Shfd15022 and Shfd16017) and one from the Meriden Red Sand (Shfd15023). Quartz OSL D_e replicate distributions for all three were multimodal, which was taken to indicate only partial bleaching had occurred prior to burial. For partially bleached sediment, ages would normally be based on lowest D_e values (assumed to contain the best bleached grains and therefore closest to true burial dose). Such an approach when applied here using the lowest FMM D_e component produced ages which fell within the Ipswichian Interglacial Stage (ca MIS 5e; figure 12). Given the glacial depositional context of the sediments (described above) quartz age underestimation is suspected as reported elsewhere (e.g. [121]). This may be due to the dim nature of the OSL signal but also may be attributable to a hard-to-bleach feldspar contribution to the OSL signal not removed after the 40 s IR wash at 50°C within the SAR measurement. Accepting the dominant FMM D_e component yielded quartz OSL ages of 192 ± 17 ka (Shfd16017), 175 ± 17 ka (Shfd15022) and 150 ± 12 ka (Shfd15023; table 2). These ages increase in antiquity with depth and produce ages for Shfd16017 and Shfd15022, which were taken from the same unit, within errors of each other. Sample Shfd15022 also produced a feldspar IRSL₅₀ age of 211 ± 10 ka and an IRSL₂₂₅ age of 276 ± 20 ka (table 2). The much higher age for the IRSL₂₂₅ measurement compared with both the IRSL₅₀ and OSL ages suggests this sample had some sunlight exposure but insufficient to reset the harder to bleach IRSL₂₂₅ signal, leading the IRSL₂₂₅ age to be an overestimate. In summary, the sediments from the Meriden site were probably deposited between 150 ± 12 ka and 211 ± 10 ka, with ice-proximal deposition (Oadby Till) after 175 ± 17 ka.

4.2.1.2. Park Farm

Two luminescence samples were collected from the Park Farm. Sample Shfd15024 from the outwash or a subglacial meltwater channel deposit of the Park Farm Sand and Gravel and Shfd15025 from the glaciolacustrine Park Farm Upper Sand. For Shfd15024 the OSL D_e replicates were unimodally distributed (figure 12), suggesting this sample was fully bleached at deposition. It returned an age of 180 ± 12 ka (table 2). Sample Shfd15025 had a multimodal D_e replicate distribution when measured by OSL (figure 12). FMM isolated two components. Just as with the Meriden samples, the lowest D_e component returned an implausibly young age within the Ipswichian Interglacial Stage (ca MIS 5e) and are assumed to have suffered from a dim contaminated OSL signal. As a result the age of 181 ± 12 ka from the larger FMM component was preferred. The overlapping of both ages is concordant with the sedimentological evidence for near-continuous deposition at this site. In summary, the age of the Park Farm Sand and Gravel and Park Farm Upper Sand appears to be 180 ± 12 ka.

4.2.1.3. Glebe Farm

Sampling for luminescence targeted the cross-bedded and massive medium glacial sand of the Baginton Sand Member of the Wolston Glacigenic Formation. Two samples (Shfd17172 and Shfd17173) were collected and both had a multimodal D_e replicate distribution when measured by OSL (figure 12; table 2). FMM isolated two and three components in the dim OSL data respectively but all returned implausibly young ages of Ipswichian Stage Interglacial or younger so have been disregarded. The IRSL₅₀ ages produced a young age for Shfd17173 which is seen as an underestimate and an age of 142 ± 5 ka for Shfd17172 (table 2). The IRSL₂₂₅ measurements yielded ages of 207 ± 11 ka (Shfd17172) and 143 ± 7 ka (Shfd17173; figure 12; table 2) but the errors on these do not overlap and they are stratigraphically reversed. Given the coincidence of the IRSL₅₀ for Shfd17172 and IRSL₂₂₅ for Shfd17173, these ages have been tentatively accepted (assuming the older IRSL₂₂₅ age for Shfd17172 is due to less bleaching). This gives a combined age (combined using OxCal v. 4.4.3; [122]) of 143 ± 4 ka (Late Wolstonian Substage, ca MIS 6a). However, given the uncertainty within the data it is suggested that this is used only as a minimum age for the deposition of the Baginton Sand Member and the Thrussington Till.

4.2.1.4. Wolston

Luminescence sampling targeted the Wolston Sand and Gravel Member of the Wolston Glacigenic Formation with two samples collected. OSL measurement indicated the samples were dim and poorly bleached based on the scatter of D_e replicates. FMM isolated two components for both samples but just as with the Meriden samples the lowest component returned ages within the Ipswichian Stage Interglacial as did the older FMM D_e component for Shfd17174. Sedimentologically, deposition of the sand and gravel member during an interglacial was unlikely (e.g. [123]). The older FMM D_e components for Shfd17175 returned an age of 137 ± 22 ka but given the problematic nature of the OSL results could not be accepted without corroboration. IRSL₅₀ ages for these samples produced ages of 148 ± 6 and 148 ± 6 ka. pIRSL₂₂₅ measurements yielded ages of 154 ± 7 ka (Shfd17174; figure 12) and 151 ± 14 ka (Shfd17175; figure 12; table 2). The consistency of the ages from samples taken in close proximity from the same sedimentary Member and the concordance with a sedimentological non-interglacial time interval is taken to indicate these are true burial ages. The quartz OSL age for Shfd17175 also is within errors of these. Based on this the Wolston Sand and Gravel Member has a combined age (combined using OxCal v. 4.4.3; [122]) of 150 ± 3.3 ka placing it and the ice associated with these sediments to within the Late Wolstonian Substage.

4.2.1.5. Seisdon

Two samples for luminescence dating were collected from the Seisdon Sand (Shfd15020) and Seisdon Gravel (Shfd15021). The glacial depositional context of the Seisdon Sand leads to a high probability of partial re-setting of the luminescence signal during deposition [124,125]. The site also underwent post-depositional deformation and tilting during subsequent late Devensian glaciation (ca MIS 2) [126]. FMM analysis of the dim quartz OSL results showed multiple D_e components for both samples, suggesting incomplete bleaching or disturbance. The lowest FMM D_e components for both samples yielded Ipswichian Stage Interglacial (ca MIS 5e) ages which are implausible based on the site sedimentology and its glacial nature (table 7; figure 12). Use of the larger FMM D_e component found in each sample produced ages of 209 ± 15 and 236 ± 22 ka. These agree well with the IRSL₂₂₅ age for sample Shfd15020 of 202 ± 11 ka. Combining these ages (using OxCal v. 4.4.3; [122]), the best estimate of the age of the Seisdon sediments and associated proximal ice is 211 ± 7 ka (ca MIS 7b of the Late Wolstonian Substage).

4.2.1.6. Luminescence dating summary

Luminescence dating at Seisdon shows the glacier was proximal to the site at 211 ± 7 ka (ca MIS 7b) and proximal at the Park Farm site around 180 ± 12 ka (ca MIS 6e), within the Late Wolstonian Substage. The Meriden site indicates the deposition of the Oadby Till after 175 ± 17 ka (ca MIS 6e) and deposition of the Baginton Sand Member and the Thrussington Till at Glebe Farm before 143 ± 4 ka (ca MIS 6a). At Wolston the ice associated with the Wolston Sand and Gravel Member has an age of 150 ± 3.3 ka (ca MIS 6b). Analysis of the resultant age dataset in OxCal showed this dataset could not be combined into a single phase. FMM analysis of the ages with a sigma-b set at 0.07 to reflect the average uncertainty of individual ages identified two phases. OxCal defined these as an early phase at 199 ± 5 ka (ca MIS 7a) and a later one at 147 ± 2.5 ka during the Late Wolstonian Substage (ca MIS 6b; figure 13).

• Download figure
• Open in new tab
• Download PowerPoint

5.2.2. Neighbouring regions

In the English East Midlands, glacial ice advancing from the northeast deposited the Wragby Till Member, equivalent to the Oadby Till Member in the English West Midlands [64]. The associated glacially derived fluvial River Trent terrace deposits of the Eagle Moor Sand and Gravel Member near Lincoln [141] are contemporaneous with the Wragby Till. White et al. [64] suggested that the Wragby ice in the East Midlands provides evidence for Middle Wolstonian Substage (ca MIS 8) glaciation in lowland Britain, based on the relationship of the glaciogenic sequence to the overlying temperate deposits. However, the correlation is questioned by Gibbard and West [130] who assign these deposits to the Ipswichian (=Eemian) Stage Interglacial (ca MIS 5e). Evidence for Middle Wolstonian Substage glaciation has been reported across continental Europe [142–146]. However, glaciation during this stage was significantly less extensive than that during the Late Wolstonian Substage (ca MIS 6) [147]. Schwenninger et al. [148] reported luminescence ages from East Leake, Loughborough (figures 1 and 14). This unit was correlated with an outwash gravel deposit derived from the Oadby Till (Wragby Till) ice by Bridgland et al. [149]. It formed during the deposition of the downstream Eagle Moor Terrace of the River Trent. This critical sequence is highly relevant because the luminescence ages from the site studied by Schwenninger et al. [148] place deposition to a minimum age of ca 178 ± 19 ka (mean of 212 ± 22, 174 ± 17 and 147 ± 17 ka). This corresponds closely to the findings in this study. It provides a clear constraint on the post-glacial deposition of the Eagle Moor deposits within the Trent catchment, and correlates to equivalent glacial advances in the English West Midlands during the Moreton Stadial (figure 14b).

Evidence of glaciation to the southeast in the Fenland Basin during the Late Wolstonian Substage is indicated by ice-contact, glacio-marginal delta fan deposits in proglacial lakes impounded by ice [150]. The deposits form part of a series of local glaciofluvial delta fans first identified at Tottenhill, Norfolk, that extended into proglacial lakes [54,55,149]. Ice advance is also demonstrated at Warren Hill (Three Hills), High Lodge, Lakenheath, Feltwell and Shouldham Thorpe [40,132,150,151–153] (figure 14). Here Feltwell Formation meltwater deposits, comparable to those at Tottenhill, accumulated [54]. The Fenland Basin was entirely filled by a glacier during the Late Wolstonian Substage. This ice lobe formed the eastern sector of the northeast Oadby ice in the English West Midlands and advanced down the east coast of England, entering the Fenland Basin and impounded itself against the higher ground of the Fenland margins [40,133,154]. It re-organized the rivers Nene, Ouse and Cam, and dammed several proglacial lakes within the Fenland Basin, and the Nene and Welland catchments [54,132,133,155,156]. Six luminescence ages from Shouldham Thorpe and Watlington in Norfolk and Maidscross Hill in Suffolk between 212 ± 9 and 169 ± 6 ka have been reported by Gibbard et al. [156] as representing ice-marginal delta and alluvial fan deposits within the Fenland Basin as ice advance during the Late Wolstonian Substage. A further seven luminescence ages from the Tottenhill sand and gravel have ages between 200 ± 33 and 125 ± 14 ka (mean of ca 174 ± 38 ka), constraining the maximum advance of the Fenland Basin ice to the Moreton Stadial of the Late Wolstonian Substage [157,158]. Finally, two luminescence ages from Lynford in the Wissey Valley on glacio-fluvial sands have yielded ages of 175 and 169 ka [132]. In addition, two luminescence ages within an ice-proximal depositional environment formed of outwash sands and part of the Britons Lane Sand and Gravel Member at Stiffkey in North Norfolk, have returned ages of 164 ± 11 and 141 ± 9 ka, establishing glacial margins of the Moreton Stadial ice along the North Norfolk Coast and the Fenland Basin region [154]. Reported luminescence ages at Tottenhill, Shouldham Thorpe, Watlington, Maidscross Hill and Lynford in the Fenland Basin [156–158] and Stiffkey on the North Norfolk Coast [154] support the maximum extent of the glacier during the Moreton Stadial in the region and correspond well with dating of the English West Midlands. Overall, the increasing lithostratigraphical and chronostratigraphical evidence across the English West Midlands and East Anglia strongly supports the occurrence of glaciation during the Late Wolstonian Substage (i.e during MIS 6) in the British Isles.