Chapter 2 The Upper Thames Basin

Introduction

This chapter is devoted to the first of three geographical divisions of the present Thames drainage system, the Upper Thames basin. This comprises the catchment areas of the various streams that combine to form the Thames upstream from the Chiltern escarpment (Figure 2.1). In many ways the Upper Thames is most easily treated as a separate system from the valley downstream of the Goring Gap. Its Quaternary terrace record is extensive, particularly in the valley of the River Evenlode, which appears to have been the main stream until late in the Middle Pleistocene. Unfortunately, there is little preservation of terrace deposits in the area of the Goring Gap and, therefore, no continuity with the succession in the Middle Thames. This has contributed to a long-standing uncertainty over correlation between the Upper Thames terraces and those in the London Basin. The problem of correlation between the Upper and Middle Thames has been addressed in this volume using projections of sediment-body longitudinal profiles between the two areas, reinforced with biostratigraphical evidence for the age of surviving interglacial deposits (see Chapter 1 and (Figure 1.3)).

The Pleistocene sequence in the Upper Thames

Classification of the Pleistocene sequence in .the Upper Thames basin has been somewhat haphazard, with some deposits being named as lithological units and others named after the geomorphological terrace features to which they give rise. Following the trend towards the former method, established in other parts of the Thames catchment (Chapter 1), geological names were applied to parts of the terrace sequence by Briggs et al. (1985) and formal lithostratigraphy was used to classify the high-level Northern Drift deposits by Hey (1986). The nomenclature of these and earlier authors is incorporated in the lithostratigraphical scheme shown in (Table 2.1), the rationale for which is discussed in Chapter 1.

The Pleistocene succession in the Upper Thames basin is readily divisible into two parts:

1. the older, high-level deposits of the Northern Drift Group (Table 2.1), which are primarily preserved parallel to the Evenlode valley and are devoid of calcareous clasts, and
2. more recent, lower-level terrace gravels composed largely of local limestones (see (Figure 2.1), (Figure 2.2) and (Figure 2.3)).

(Table 2.1) Lithostratigraphical classification of Upper Thames deposits.

The Northern Drift remnants contain pebbles of quartz, quartzite, chert and flint derived from the area beyond and to the north of the present Thames catchment, hence the name, which was first applied to these particular deposits by Hull (1855). These high-level deposits are generally unbedded and contain large proportions of silt and clay. For this reason they have, until recently, been regarded as the decalcified remains of till, usually attributed to the Anglian glaciation (Briggs and Gilbertson, 1974). Recent discoveries at Sugworth, near Abingdon, have not only led to the rejection of this interpretation, but have somewhat undermined the distinction between the older and younger Pleistocene deposits. At Sugworth, an interglacial fluviatile channel-fill (see below) was found beneath a spread of high-level unbedded clayey gravel that has been attributed to the Freeland Formation (Hey, 1986). These interglacial sediments yielded biostratigraphical evidence for a Cromerian age and contained pebbles of limestone as well as rock-types characteristic of the Northern Drift (Shotton et al., 1980). This discovery, which was the first occasion on which calcareous gravel had been observed in association with the high-level deposits, demonstrated that Northern Drift material was present in the Upper Thames basin by Cromerian times.

Since the discovery of the Sugworth site, an alternative interpretation of the Northern Drift has gradually gained credence, following the work of Hey (1986). This holds that the high-level material is the degraded and decalcified remains of old terrace gravels of the Upper Thames river system. Some workers continue to believe that glaciation was responsible for the introduction of the Northern Drift pebbles into the Upper Thames basin (Bowen et al., 1986a), but others consider that the Thames catchment was once considerably more extensive, encompassing the source areas of at least some of these rock-types (Bridgland, 1988c). The distribution of the Northern Drift suggests that the Evenlode was the main drainage line at this time.

The highest and oldest of the various limestone-rich terrace gravels, the Hanborough Gravel Formation, differs from the others in that it is, like the Northern Drift, preferentially preserved in the Evenlode valley, suggesting that this river continued as the main drainage line within the Upper Thames system. This view is supported by the reconstruction of the long-profile of this formation, which extends into the present watershed area between the Evenlode and the Stour (a left-bank tributary of the Warwickshire–Worcestershire Avon) and has a gentle gradient, contrasting with those of later formations (Figure 2.3). The Hanborough Gravel has yielded the remains of temperate-climate mammals, on the basis of which it was believed until recently to have formed during an interglacial. Most of the bones have come from the base of the deposits, however. In contrast, molluscan remains from silts interbedded with higher parts of the formation indicate cool conditions (Briggs and Gilbertson, 1973). The aggradation is therefore now attributed to a cold episode, probably within the early part of the Saalian Stage. Arkell (1947a, 1947b) considered the Hanborough Gravel to pass upstream beneath the deposits of the Cotswolds glaciation (Moreton Drift), an interpretation of great importance in correlating the Pleistocene sequences of the Thames basin and the Midlands (see below and Chapter 1).

The next formation in the Upper Thames sequence, the Wolvercote Gravel, is of considerable stratigraphical significance, in that it has long been held to be the first to contain material introduced into the basin by the glaciation of the Cotswolds. This interpretation is generally attributed to Bishop (1958), who recognized the relatively high flint content of the terrace deposits at Wolvercote and traced the formation upstream in the Cherwell valley, linking it with the glaciation of the Fenny Compton area. However, a succession of later authors have contrasted this deposit with the (higher) Hanborough Gravel and concluded that the glaciation of the Cotswolds occurred between the aggradation of the two formations (Briggs and Gilbertson, 1973; Briggs et al., 1985). Moreover, the Wolvercote Gravel has frequently been interpreted as outwash derived directly from this glaciation, an idea that originated with Tomlinson (1929), who noted the importance of the input of fresh flint many years before Bishop. The Wolvercote Formation is best developed in the Cherwell and Evenlode valleys, both of which flow from gaps in the Cotswold escarpment through which outwash is believed to have escaped from the Midlands ice sheet (Bishop, 1958; Briggs, 1973). This glaciation has generally been correlated with the Saalian Stage (Shotton, 1973a, 1973b) and the associated Wolvercote Gravel has been correlated with the Taplow Gravel of the Middle Thames (Gibbard, 1985), which has been firmly dated in the late part of that stage (see Chapter 3). However, recent re-evaluation of the sequence in the Midlands (Sumbler, 1983a, 1983b; Rose, 1987, 1989) has suggested that the ice advance that reached the Cotswolds may have occurred during the earlier Anglian Stage, initiating a major and continuing controversy over the age of this glaciation. The conclusion in this volume (Chapter 1), that the Wolvercote Gravel equates with the Lynch Hill Gravel of the Middle Thames, in no way resolves this problem, as the latter formation is also attributed to the Saalian Stage (Gibbard, 1985; (Table 1.1); see below, Wolvercote). The veracity of the input of flint into the Upper Thames system between the deposition of the Hanborough and Wolvercote Gravels has been called into question in a recent review of the evidence from the Cherwell and Evenlode valleys (Maddy et al., 1991b; see below, Long Hanborough and Wolvercote).

The most important site associated with the Wolvercote Formation, at Wolvercote itself, revealed a large channel cut through the terrace deposits and filled with a sequence of fossiliferous gravels and sands followed by peat and silty clay. The gravels at the base of the channel-fill proved to be the richest source of Palaeolithic artefacts discovered to date in the Upper Thames. Unfortunately, there have been no extensive exposures in these important sediments for many years, so relatively little of their palaeontological potential has been realized. It is known that they represent a cooling sequence following fully temperate conditions, possibly reflecting the end of an interglacial. There has been considerable controversy over the identity of the interglacial represented (see below, Wolvercote), but the correlation of the Thames sequence with the deep-sea record proposed in Chapter 1 suggests that the channel was infilled during Oxygen Isotope Stage 9. This revised correlation raises the possibility that neither of the traditionally recognized post-Anglian interglacials (corresponding with the Hoxnian and Ipswichian Stages) is represented at Wolvercote. Although no conservable remnant of Wolvercote Channel sediments has yet been located, it is hoped that a GCR site in these critically important deposits will be identified in due course.

The next terrace in the Upper Thames sequence, the Summertown-Radley Terrace, is the most extensively preserved. The terrace surface is underlain by a complex aggradational sequence (here classed as a formation) apparently spanning two temperate episodes and one full cold interval, with parts of two others. A complex stratigraphy was recognized in early research on the Summertown-Radley deposits, a cold-climate mammal fauna being identified in the lower part of the sequence, with richer fossiliferous sediments in the upper part containing mammals and molluscs indicative of temperate conditions (Sandford, 1924, 1926). These upper (temperate-climate) sediments, recently given the name Eynsham Gravel (Briggs et al., 1985), have generally been attributed to the last (Ipswichian) interglacial (see below).

The cold-climate Stanton Harcourt Gravel, underlying the Eynsham Gravel, probably constitutes the largest part of the Summertown-Radley Formation. In a large working quarry at Stanton Harcourt, a channel cut in Oxford Clay and filled with richly fossiliferous sediments was found beneath typical cold-climate gravels of the Stanton Harcourt Member. The fauna and flora from this channel is of fully interglacial character, but differs significantly from the assemblages from the Eynsham Gravel. The Stanton Harcourt Channel Deposits have thus been interpreted as the product of an earlier temperate event, possibly one that is as yet undefined, chronostratigraphically, in Britain. This event has been correlated with Oxygen Isotope Stage 7 of the deep-sea record (Shotton, 1983; Bowen et al., 1989; Chapter 1). It is suggested below that many of the historical records of interglacial sediments within the Summertown-Radley Formation that have previously been ascribed to the Eynsham Gravel (Briggs et al., 1985) should be equated instead with the Stanton Harcourt Channel Deposits (see below, Stanton Harcourt and Magdalen Grove). Included amongst these reinterpreted sites is a pit in the grounds of Magdalen College, Oxford, which was re-excavated in 1984 as part of the GCR programme (Briggs et al., 1985).

The Pleistocene sequence in the Upper Thames is completed by the 'Floodplain Terrace' (Sandford, 1924, 1926), otherwise known as the Northmoor Terrace (Arkell, 1947a). This aggradation, here termed the Northmoor Gravel Formation, reaches only c. 2–3 m above the alluvium of the modern valley floor and lies 6–10 m below the level of the Summertown-Radley Terrace (Figure 2.2). Recent mapping has suggested that the floodplain surface is a composite morphological feature (Robson, 1976; Harries, 1977). The aggradational sequence forming the Northmoor Terrace comprises well-bedded limestone gravels, with abundant evidence of a contemporaneous periglacial climate, and is considered to have been laid down during the Devensian (last glacial) Stage (for example, Briggs et al., 1985). Beds and lenses of organic material, widely preserved within the formation, have been found by radiocarbon dating to represent three distinct periods. Plant macrofossils, pollen and the remains of large mammals, molluscs and insects have been found in these various organic sediments. The oldest and lowest occurrences are attributed to cold-climate phases within the latter part of the mid-Devensian, with no evidence of deposition much earlier than 40,000 years BP (these sediments appear to post-date the mid-Devensian (Upton Warren) interstadial of Coope et al. (1961)). A further group, higher within the sequence, relates to the Late Devensian late-glacial interstadial; the remainder, from channels cut into the terrace surface, are of early Holocene age (Briggs et al., 1985). The last mentioned coincides with a sedimentary change from (predominantly planar-bedded) gravels to fine-grained silts and peat, thought to reflect an adjustment from a braided gravelly floodplain to a single channel river. A site has come to light recently at Cassington, in the valley of the main Thames, where earlier Devensian sediments may be represented. Early indications, principally from molluscan remains, suggest that channel-fills of early or early mid-Devensian age are represented here within a sequence that is mapped as Northmoor Gravel (D. Maddy, pers. comm.; see below, Stanton Harcourt and Magdalen Grove).

Numerous commercial workings have operated in and continue to exploit the deposits of the Northmoor Formation, but the sections are consistently below the water table, so that abandoned workings rapidly become flooded. This makes the conservation of sections a difficult proposition and no sites in the Northmoor Formation have been selected in the GCR. Devensian and Holocene fluvial deposits are invariably situated at low levels in valleys, so that this problem applies quite generally. New exposures are, however, frequently opened, allowing temporary access to these widely distributed Upper Pleistocene sediments. Devensian fossiliferous deposits are represented within the GCR Pleistocene coverage of the Thames basin at Great Totham, in Essex (see Chapter 5).

References