London - Quaternary
Quaternary deposits, also known as drift or superficial deposits, were laid down in the London district during the last 1.65 million years or so. They provide evidence of an ancient river system, a precursor to the River Thames, a glaciation of Anglian age in the north of the district, and the development of the present River Thames valley. The climatic oscillations that led to the change from glacial to warm temperate conditions are now (tentatively) related to oxygen isotope stages (Table 9).
- 1 PRE-ANGLIAN DEPOSITS
- 2 ANGLIAN DEPOSITS
- 3 POST-ANGLIAN DEPOSITS
- 4 SEQUENCE OF QUATERNARY EVENTS
- 5 Geology of London - contents
The Pre-Anglian deposits were formerly referred to as ‘high level pebble gravels’. They occur as numerous small, and in many cases isolated, patches of sand and gravel at a higher elevation than the main river terrace deposits, and are largely outside the limits of the Anglian glacial deposits. They have not been extensively exploited for aggregate and consequently are known in detail only from small pits and temporary exposures. Nevertheless, many attempts have been made to reconstruct the drainage history of the region based on the evidence of these deposits, most recently by Gibbard (1985, 1995) and Bridgland (1994).
These deposits crop out on the highest ground, mainly in the north of the district (Figure 29a). Most were formerly referred to ‘Pebble Gravel’ or ‘Plateau Gravel’, dependent largely on their altitude; a few outcrops were named ‘Warley Gravel’ (Dines and Edmunds, 1925). Exposures at Stock [685 988] in the extreme north-east of the district were mapped as the Bagshot Pebble Bed of Palaeogene age (Bristow, 1985). They are redefined in this account based on a description of the type area at Harrow Weald Common [147 929] (Bridgland, 1995; Gibbard, 1999).
The deposits almost invariably cap hill tops and give rise to clayey and silty soils containing abundant brown, red and black well-rounded flint pebbles and minor amounts of small, white vein quartz, subangular and nodular flint and rounded Triassic ‘Bunter’ quartzite pebbles. Many of the weathered flint clasts are partially desilicified, forming a white, bleached surface patina that in some cases extends deep into the clast. In the type area most of the flint clasts are between 16 and 150 mm across. Quartz clasts make up 30 per cent of the 4 to 8 mm fraction but only 0.3 per cent of the 16 mm or greater fraction, while Lower Greensand cherts are 10.5 per cent of the 4 to 8 mm fraction but only 0.4 per cent are greater than 16 mm (Moffat, 1986). The matrix consists of orange-brown, pale grey and bright red mottled clay and sandy clay with pockets of coarse sand in places. Cryoturbation structures are almost invariably present in the top 2 m of the deposit in the form of involutions and festoons containing pebbles with long axes orientated vertically and near-vertically. Beds of sand and gravel also occur locally; their thickness is highly variable up to about 5 m. No systematic work has been carried out to compare the pebble content of the widely spread deposits, but there is some regional variation as summarised in the Table 10.
Red staining of clay matrix and flint pebbles is believed to be the result of pedogenic clay enrichment and rubification in a warm humid climate; an event tentatively correlated with development of the pre-Anglian sol lessivé (Moffat and Catt, 1982) that is widespread beneath till in Essex and Suffolk.
The origin of the Stanmore Gravel is uncertain. No absolute or comparative dates have been obtained from the deposits and their correlation remains speculative. An exhaustive review by Bridgland (1994) concluded that they were laid down in south bank tributaries of a precursor of the present Thames, presumed to be earlier courses of the Mole-Wey, Wandle, and Darent. However, the deposits appear to be too widespread to relate to individual river deposits, and a re-evaluation of the altitude of the deposits shows that they lie more or less on an extensive planar surface dipping gently to the north-east (Figure 29a). This surface coincides closely with the base of the Red Crag, constructed from isolated outcrops in the London Basin region combined with data from the more extensive distribution in Essex and Suffolk (Moffat et al., 1986; Mathers and Zalasiewicz, 1988). The outcrop of Well Hill Gravel (see p.57) also falls within the elevation range of this surface. The conclusion drawn from this re-evaluation is that the Stanmore Gravel and Well Hill Gravel may be marine in origin, supporting the view of Wooldridge (1960). It also concurs with the observations of Hey et al. (1971) who claimed that the surface texture of sand grains in the Stanmore Gravel indicate deposition in a fairly low-energy beach environment, although these textures could equally have been formed during deposition of the parent Palaeogene deposit.
Clay-with-flints crops out in the south-east of the district forming a dissected spread largely on the dip slope of the North Downs. The surface of the outcrop is gently inclined, parallel to the regional northerly dip of the chalk. The Clay-with-flints almost invariably overlies chalk, only locally resting on Thanet Sand. A simplified profile through the deposit on the North Downs is shown on Figure 30.
Clay-with-flints is heterogeneous and unbedded. Typically, it has irregular vertical and lateral changes of texture, colour and clast content that are too localised and complex to be shown on the geological maps. It gives rise to a reddish brown silty and sandy clay soil with angular and nodular flints. In some places it is dominantly sandy and in others there are abundant well-rounded flints. Variations between these types may occur within 100 m, but there is no apparent geographical pattern. The lithologies include reddish brown clay with large unworn flint cobbles, yellow fine- to medium-grained sand, reddish brown clayey silt and sandy clay with beds of well-rounded flint pebbles (Catt, 1986). At the base in many places is a dark brown to black, stiff and waxy clay, less than 100 mm thick, and containing relatively fresh nodular flints. These flints may be stained black by manganese precipitated from groundwater and have a green glauconitic cortex, similar to flints at the base of the Thanet Sand. In the main body of the Clay-with-flints the more argillaceous sediments tend to occur at the top (Figure 30). There are three main types of flint:
- very well-rounded, subspherical pebbles derived fromPalaeoene strata
- complete or broken nodules derived directly fromChalk
- frost-shattered angular shards caused by alternatefreeze-thaw of ground ice during periglacial periods
In general, the thickness ranges from 5 to 10 m, but is highly variable over short distances, probably due to dissolution of the chalk which has taken place mostly after the formation of the Clay-with-flints. The base is highly irregular with localised pipes and hollows between 1 and 50 m in diameter.
Clay-with-flints is a remanié deposit formed by weathering and solifluction of the original Palaeogene cover and earlier Quaternary deposits and dissolution of the underlying Chalk. Its formation probably commenced during late Pliocene or early Quaternary times during uplift and erosion of the Red Crag and underlying Palaeogene and Chalk. The Clay-with-flints has been subjected to pedogenesis and weathering during periods of warm climate, which has resulted in reddening and clay enrichment, and to several periods of periglacial climate that has produced intensive cryoturbation and solifluction.
The more argillaceous parts of the Clay-with-flints have many small fissures that provide significant permeability (Klinck et al., 1998). This fissuring was probably formed initially in permafrost conditions by small lenses of segregated ice. Some of the larger fissures, particularly those close to the base of the deposit and in karstic cavities, have slickensides that are indicative of slippage, possibly during subsidence which followed dissolution of the chalk.
The Chelsfield Gravel is a newly defined unit. It occurs only in the south-east of the district in small outcrops at the type area at Chelsfield [476 642] and at locations between West Kingsdown [577 612] and Holly Hill [669 630] within the Clay-with-flints outcrop on the chalk dip slope at elevations between 125 and 170 m OD. The deposit consists of well-rounded flint pebbles in a clayey and silty fine-grained sand matrix, which gives rise to a pebbly soil. The gravel in the type area is beyond the main Clay-with-flints outcrop, and is interpreted as a head deposit, partly let down with the Thanet Sand into dissolution hollows in the Chalk. The deposit at Holly Hill, which forms a prominent steep-sided feature, is probably an outlier of Harwich Formation that has been only partly reworked. The other small outcrops were also derived from the Harwich Formation and have been incorporated in the Clay-with-flints.
Pre-diversionary Thames River Terrace Deposits
Deposits of the ancestral Thames river system were laid down in a valley that crossed the north-west of the district, more or less coincident with the present-day Colne. They comprise the Gerrards Cross Gravel and Westmill Gravel, and form gently sloping terrace-like features, degraded by dissection and solifluction. In general, the deposits overlie chalk, with a contact likely to be irregular due to dissolution. These deposits are part of the Kesgrave Sands and Gravels that extend across East Anglia from the Vale of St Albans, and form the largest body of sorted Quaternary coarse-grained sediments on the British land area (Rose et al., 2001). They are characterised by a relatively high proportion of quartz and quartzite pebbles derived from the English Midlands hinterland (Table 11).
The Dollis Hill Gravel and Woodford Gravel (Figure 29b) form hill-top caps that decline in elevation northwards, indicating deposition in south-bank tributaries of the ancestral Thames (see Gibbard, 1985). Information about these gravels and their composition is summarised in Tables 11 and 12.
At Darenth Wood [580 727] and nearby to the south, outcrops of a dissected gravel deposit mapped as River Terrace Deposits undifferentiated occur at an elevation of 65 to 80 m above OD, intermediate between the Stanmore Gravel (see p.52) and the Black Park Gravel. The clasts are mainly rounded black ‘Tertiary’ flints, brown angular flints, and rare quartz and Lower Greensand. The origin of this deposit is uncertain, but it was probably laid down in a river flowing north from the Weald (Figure 29b) and turning east, possibly along the line of the modern Thames (Gibbard, 1994).
Outcrops of high level gravels not assigned to the Stanmore Gravel are the Well Hill Gravel [497 642] and Sand and Gravel of unknown age and origin at Crystal Palace, Norwood and Streatham Common (Figure 29a). Both contain noticeable amounts of Lower Greensand chert, up to 10 per cent in the Well Hill Gravel (Peake, 1982). The Well Hill Gravel lies at an altitude close to the inferred surface on which the Stanmore Gravel was deposited, and therefore the two deposits are interpreted as coeval. The gravels at Crystal Palace are well below the level of the Stanmore Gravel. They are interpreted as a fluvial deposit laid down by a river flowing in an easterly direction, perpendicular to the chalk dip slope. It seems unlikely that such a river would have eroded through the chalk escarpment and then flowed north along a proto-Wandle course as suggested by Macklin (1981) and reviewed by Gibbard (1994).
Details of localities and thickness are given in Table 13.
Outwash from the Anglian ice sheet in the north of the district is mapped as glaciofluvial deposits. These deposits occur in relatively small outliers in the north- west of the district (Figure 31), close to the outcrops of the till and at a similar height. Locally the till overlies the glaciofluvial deposits, for example at Chigwell Row [468 931] and in the River Wid valley [610 975]. The extent beneath the till is not known, but is unlikely to be widespread as there is little evidence for large outwash channels either at outcrop or in boreholes.
The deposits consist of brown and yellow-brown variably clayey sand and gravel with clasts mainly of well-rounded flint (derived from the Stanmore Gravel) and subangular flint, with a few worn nodular flints and a smaller proportion of rounded quartz and quartzite pebbles and cobbles. The deposits at outcrop generally display cryoturbation structures.
These crop out in the east of the district near Chapmans Farm, Upminster [565 890], close to the A127 (Figure 31). They were formerly exposed overlying till in the Upminster brickpits (Dines and Edmunds, 1925). The deposits consist of brown, grey, greenish grey and lilac, thinly interbedded sand, sandy clay, and silty clay. They are thought to have been laid down in a proglacial lake formed at or close to the Anglian ice front at the time of its maximum extent. An interbed of reworked (soliflucted) London Clay (Woodward, 1904) is interpreted as evidence of contemporary periglacial conditions.
Similar deposits of laminated silty clay and sandy silt deposits, up to 1.5 m, thick occur beneath till near Finchley [e.g. 276 901]. They are interpreted as having been laid down in an ice-dammed lake (Gibbard, 1979).
Till, formerly known as ‘chalky boulder clay’, crops out principally in the north-east of the district (Figure 31),with outliers around Finchley [26 91] and in the Colne valley in the north-west of the district. It was laid down at the southern margin of the Anglian ice sheet, and is part of a once-continuous sheet of till that extends from this district northwards across Essex and Suffolk.
The till is a hetereogeneous deposit consisting mainly of firm to stiff, pebbly, variably silty and sandy clay. At the surface it weathers to yellow-brown and ochreous colours; below about 2 m, where less oxidised, the till is pale grey becoming darker grey with depth. Clasts in the till are typically of chalk and flint with subordinate rounded Triassic ‘Bunter’ vein quartz and quartzite, minor Jurassic and Carboniferous limestones and sandstones, and rare granite and basalt (Moorlock and Smith, 1991). The chalk clasts are leached out of the weathered near-surface deposits in the top 2 m or so, but locally, for example around Battles Hall [496 959] near Stapleford Abbotts, the leached zone is less than 0.5 m. A relatively high proportion of well-rounded flint pebbles, derived from the Stanmore Gravel, are present in till outliers particularly south and east of Brentwood. The clay matrix is almost certainly derived from Jurassic as well as Palaeogene clays and, where not decalcified, typically contains a significant proportion of chalk flour.
The distribution of till and the altitude of the base of the deposit indicates the form of the ice sheet at the maximum extent of the Anglian glaciation (Figure 31). It is possible that two ice advances are represented in the district: an earlier one in the Finchley area, and a later one with ice tongues moving down what are now the Lea, and the Roding valleys and a narrow tongue extending to Hornchurch [550 879] (Baker and Jones, 1980; Bridgland, 1994).
Post-diversionary River Terrace deposits
Evidence suggests that the River Thames was diverted into its present valley late in Anglian times about 500 000 years ago (for overviews see Bridgland, 1994; Gibbard, 1985), thus marking the start of the deposition of extensive river terrace gravels in the district.
The deposits occur in a sequence of river terraces that are differentiated on the basis of altitude. The history and extent of research on the sequence, including the development of the nomenclature, is summarised in Bridgland (1994).
The terrace gravels occur at progressively lower elevations above the modern floodplain, a situation now considered to be a response to continuing neotectonic uplift (Maddy, 1997). The bulk of the gravels were deposited on a broad river braid plain during colder periods when periglacial activity made available the greatest volume of sediment. In the lower reaches of the Thames, aggradation of some terrace deposits was complicated by rising sea level, which increased the space available to accommodate sediment, and led to deposition of mainly fine-grained inner-estuary sediments.
Deposits from more than one cold climate episode may be represented in a single river terrace deposit. A basal cold climate gravel may be overlain by interglacial deposits that in turn are overlain by more cold climate gravels (Figure 32).
Sites with fossils and Palaeolithic artefacts are known from numerous localities in the river terrace deposits. These contain evidence for both cold and interglacial intervals, and allow the terrace deposits of the district to be correlated with oxygen isotope stages (Bowen, 1999; Table 9).The relationship of these sites to the river terrace deposits is shown in Figure 34.
The earliest systematic subdivision of the river gravels, into Boyn Hill Gravel, Taplow Gravel and Floodplain Gravel, was made on the Geological Survey maps of the district published around the beginning of the 20th century. Hare (1947) made the next significant advances. He identified the (highest) Black Park terrace, formerly mapped by the Geological Survey as Glacial gravel, and the Lynch Hill terrace at an intermediate height between the Boyn Hill and Taplow Terraces. The gravels that underlie all these terrace features have subsequently been mapped as River Terrace Deposits.
Alterations to the terrace nomenclature, including the introduction of a numbering system (Table 14), were made as a consequence of a revision survey of the south London district, the results of which were incorporated on the 1:50 000 scale map published in 1981. An additional terrace deposit, numbered 3a, mapped between the Taplow and Lynch Hill Gravel, has been named Hackney Gravel on the most recent maps.
In a review of the Middle Thames terrace sequence, Gibbard (1985) correlated the terrace gravel deposits between Reading and central London. He extended the scheme erected by Hare, introducing the terms Kempton Park Gravel to replace the Floodplain Gravel, the Shepperton Gravel for even younger gravels that underlie the alluvium in the London district, and Staines Alluvial deposits for the fine-grained river floodplain sediments. His terminology, with the exception of the Staines Alluvial deposits, has been adopted in this account (Table 14) and on the geological maps of the district. The six principal river terrace deposits that have been mapped correspond in general to those of earlier workers, with minor additions and renaming as a consequence of analysis of new borehole data in combination with topographical evidence. In addition, one small outcrop of an additional terrace deposit, the Finsbury Gravel, at an altitude intermediate between the Boyn Hill and Lynch Hill Gravels, has been mapped.
Gibbard (1994) reviewed the sequence of terrace deposits in the Lower Thames valley, downstream from central London. In this area he introduced a series of local names whose equivalence with the Middle Thames sequence is shown on Table 14.
Correlation of the terrace deposits in this district is illustrated on Figure 34, which shows the base and surface of each deposit projected onto a notional river-centre line.
In many places the River Terrace deposits form a bench or terrace feature that is bounded by a concave break of slope on the margin farthest from the contemporary river channel, and a convex slope on the margin nearest the river. These terrace features are particularly well developed on the north side of the Thames, for example between Southall and Teddington and Stoke Newington and Bow. The terraces are also apparent in the centre of London, for example between Soho [296 811] and the River Thames at Whitehall [304 800]. In many places, however, where different terraces are adjacent they merge into a single planar or concavo-convex slope. The outcrop of the base of the deposits is defined clearly where it rests on London Clay. In some places head covers this contact (see for example Figure 33), and the boundary is determined on borehole data. The basal surface of the deposits may be irregular, either due to channelling or to dissolution where Chalk forms the bedrock.
Many of the older terrace outcrops are on hilltops, for example Wimbledon Common [235 735], Islington [314 838] and Dartford Heath [52 73], and in general have been dissected by erosion to a greater extent. The outcrop of the base of these deposits is commonly masked by downwash, and is more accurately determined on borehole evidence.
River Terrace Deposits consist of variable proportions of sand and gravel. They were deposited in a braided river system, an estimated 5 km wide. Gibbard (1995) in a review of the sedimentary features of the deposits identified gravel-dominated beds, generally less than 2 m thick, characterised by horizontal bedding with little internal structure and rare imbrication. These are cut by broad, shallow channels, which are infilled with tabular cross-bedded gravelly sand in fining-upwards units. Trough cross-bedded, sand-dominated beds, up to 3 m thick, are cut through the other units. Fine-grained sediment occurs locally as impersistent beds, less than 1 m thick; it consists of clayey and silty sand. In the east of the district, thicker sequences of fine- to coarse-grained sands occur and are interpreted as deposits laid down in estuarine conditions. The river terrace deposits have yielded vertebrate remains and Palaeolithic flint artefacts at numerous localities (see Gibbard, 1985,1995; Wymer, 1999).
The thickness of the deposits varies considerably as indicated by the range of maximum values given in Table 15. Individual River Terrace Deposits are lithologically indistinguishable although there are minor local variations in clast lithologies and proportions (Gibbard, 1985, pp.96, 146; Gibbard, 1994, pp.119, 216; Bridgland, 1994, p.181). For example, the Black Park Gravel contains most erratics and shows the greatest variability of clasts, presumably because they are derived from the glacial deposits. A higher proportion of Lower Greensand clasts are found in terrace deposits of the rivers Mole and Darent, and there is a consequent increase in these clasts in the Thames gravels close to the confluences. There is also evidence (Gibbard, 1985, 1994) of a minor decrease in the proportion of quartz and quartzite clasts and an increase in angular flints in the more recent gravels.
The Black Park Gravel (Figure 29c) is generally recognised as the oldest deposit laid down by the Thames in its current post-diversion valley. Gibbard (1985) traced the gravel from its type area west of this district to Wimbledon Common [235 735], and tentatively correlated it with the deposit capping Hangar Hill [182 819] (Figure 34).
The Boyn Hill Gravel (Figure 29d) was first defined near Maidenhead (Bromehead, 1912). It occurs in the main Thames valley and in the tributary valleys of the rivers Lea, Roding and Wey.
The Finsbury Gravel, first identified by this resurvey, crops out only at Finsbury [315 829]. It may be related to a phase of deposition of the Lynch Hill Gravel close to the confluence of the rivers Lea and Thames.
Much of the area now mapped as Lynch Hill Gravel (Figure 29e) was formerly shown on maps as Taplow Gravel. Hare (1947) first recognised it as intermediate between the Boyn Hill and Taplow gravels, and it was identified on the 1981 edition of the geological maps as Terrace 3b.
The Hackney Gravel (Figure 29f) formerly known as the Taplow or ‘Middle’ terrace (Bromehead, 1925), was identified as a separate deposit (Terrace 3a) on the 1981 edition of the geological maps. The altitude range of the Hackney Gravel is similar to that of the Lynch Hill Gravel, but as the base is generally lower it may represent the lower part of a single deposit encompassing both gravels, although this is not proved.
The Taplow Gravel (Figure 29g) is correlated from the type area [SU 916 816] near Maidenhead eastwards to London. Extensive deposits of Taplow Gravel occur in the Thames valley and in the lower parts of the Brent, Wandle, Lea, Cray and Darent valleys. The outcrops are less extensive than formerly shown on geological maps because parts of them have been renamed.
The type section of the Kempton Park Gravel (Figure 29h) is at Kempton Park [118 703] (Gibbard and Hall, 1982). The deposits equate with the ‘upper floodplain gravel’ of Dewey and Bromehead (1921) and much of the Floodplain gravel of Bromehead (1925). They constitute the lowest terrace deposit above the floodplain of the Thames, but east of Woolwich [405 784] and Poplar [38 81] the deposits are concealed beneath alluvium.
Some anomalies of correlation of the River Terrace Deposits are illustrated in Figure 34. Black Park and Boyn Hill gravels, in particular, have been the subject of a long- standing, and as yet unresolved debate (summarised by Bridgland, 1994, and Wymer, 1999), concerning the relationship between the height and age of the deposits. The surface of the Boyn Hill gravel deposits at Dartford Heath [52 73] is at about 40 to 42 m above OD and the base is channelled down to about 25 to 28 m above OD. At Swanscombe, the surface level of these terrace deposits is about 8 m lower than at Dartford Heath, but the base is at a similar level to the floor of channels at Dartford Heath. The elevation of other outcrops of the Boyn Hill Gravel, at South Ockendon [595 834] and Orsett Heath [64 80], corresponds closely to that at Swanscombe. Bridgland (1994) put forward evidence that all the deposits should be regarded as Boyn Hill Gravel, the gravels of highest elevation being a ‘feather edge’ remnant of the deposit. However Figure 34 indicates this is not the case, and that the height range of the Black Park and Boyn Hill gravels overlap. This may be because the deposits at Dartford Heath include gravels equivalent to both the Boyn Hill and Black Park gravels (Gibbard, 1995 p.18; Wymer, 1999, p.72), with the younger Boyn Hill Gravel at the surface, as shown on the geological maps.
The relationship of terrace gravels to till in the Thames valley is problematic for terrace correlation. The Black Park Gravel has been recognised as the first postdiversionary river gravel in the Thames valley, and was therefore deposited after the till. The base level of the river in which the Black Park Gravel was deposited would have cut down lower than the base of the till, which is at 25 m above OD in the Hornchurch railway cutting [547 874] (Figure 34). However, at Dartford Heath and Swanscombe, Hoxnian interglacial deposits that postdate the Anglian Till and Black Park Gravel are within the altitude range of the Black Park Gravel rather than below it as might be expected (Bridgland, 1994). This brings into doubt the relative ages of the till, Hoxnian deposits and the Black Park Gravel.
Deposits in deep depressions
Irregularities in the rockhead surface beneath Quaternary deposits are generally less than 5 m in amplitude, but locally beneath the Kempton Park Gravel, for example in central London between Battersea and Greenwich, there are enclosed depressions in the rockhead surface. These are known as ‘scour hollows’, which were documented by Berry (1979) and whose locations are shown on Figure 35. Nearly all are eroded into London Clay, but some cut through into the underlying Lambeth Group and, exceptionally, the Chalk. In some cases the bedrock beneath the depression appears to have been uplifted and the underlying strata reduced in thickness. This is particularly apparent where the underlying strata is the London Clay Formation. The depth of ‘scour hollows’ is usually 5 to 15 m but exceptionally more than 33 m are recorded at Battersea and 60 m at Blackwall (where the base is not proved). In plan, the depressions may be irregular, roughly circular or boat-shaped, and vary from about 90 to 475 m wide. The sides are steep (locally with ‘cliff-like walls’; Berry, 1979), many with slopes of 20° or greater. Infill deposits consist mainly sand and gravel with some clayey beds. The clasts are predominantly flint, but where chalk forms the rockhead it also occurs within the fill, ranging from silt to large boulder grade. The deposits are generally stratified, but may be disturbed by soft sediment deformation as illustrated by core drilled through the deposits at Blackwall (Figure 36). Upward injections of London Clay that penetrate the deposits from the base have also been recorded. The deposits accumulated during the Devensian transition from cold to warm climate, and some contain palaeontological and palynological evidence of interglacial conditions (Berry, 1979).
Formerly, these depressions were thought to have been formed by the scouring action of seasonal, glacial meltwater in periglacial conditions (Berry, 1979). However, this mechanism does not entirely explain the formation of the deeper hollows, and some of the associated features such as the apparent bulging of underlying strata. Thus, it is now thought that the depressions were originally at the sites of open-system pingos, where massive bodies of ground ice grew by drawing water from unfrozen groundwater beneath the permafrost (Hutchinson, 1980); this could account for the bulging of underlying strata. During interglacial or interstadial times, the pingos melted, resulting in collapse of the overlying deposits into the hollow made by the melting ice. At the same time a sudden release of high hydrostatic pressure in the confined Chalk and lower ‘Tertiary’ sand aquifers would have caused a rapid ejection of water carrying with it Quaternary sediments and bedrock debris, and further deepening the existing hollow. The presence of large chalk blocks in the deposits at Blackwall, at least 15 m above the level of the Chalk bedrock (Figure 36) provides evidence of the powerful upward forces involved.
Deposits formerly mapped as ‘brickearth’ are divided into six units (Table 16) whose outcrops are shown on Figure 37. They consist predominantly of silt and occur mainly on gentle slopes overlying river terrace gravels. In some places (for example at Hounslow West [11 77]) there is a veneer of this type of deposit extending across more than one terrace. In others, particularly at the back of terraces farthest from the present-day river, the deposits overlap onto bedrock. Where there is a relatively steep bedrock slope the deposits may be locally banked up against it, for example at Southall [12 81], Stoke Newington [33 86] and Crayford [514 767]. ‘Brickearth’ is generally less than 3 m thick (see Table 17). Formerly it may have been more extensive than shown on the geological maps, but much has been removed for brickmaking since Roman times, particularly in areas now built up. There are no significant lithological differences between the six units, and they are separated simply on the basis of their distribution rather than age.
The deposits consist of very fine-grained sand, silt, and clayey silt, which is brown to orange-brown in colour. They are unstratified with characteristic vertical columnar desiccation cracks. The top metre or so is decalcified but below this there are scattered irregular-shaped nodules of reprecipitated calcium carbonate (race), sometimes mistaken as chalk. Scattered angular flint fragments are common and there are sporadic pebbly seams. Gibbard (1985, 1994) concluded that the silt-grade grains are loessic (windblown) in origin, but gravel in the basal parts of the deposit was probably transported by solifluction; lamination, also present in the basal beds, indicates fluvial deposition.
Locally, laminated and cross-bedded sand, usually in association with interglacial sediments, beneath the typical ‘brickearth’ deposits is included in the thickness of the deposits classified as ‘brickearth’ (Table 17).
Interglacial lacustrine deposits are mapped only around Peckham in south London [340 767]. Deposits of a similar origin are recorded from at least 25 sites in the London region (Figure 38; Table 18), in association with River Terrace Deposits and the ‘brickearths’. They are fossiliferous, providing evidence of deposition in both cold and warm (interglacial) climates, and a key for correlation of Quaternary sequences in Britain.
The mode of formation and preservation of interglacial sediments and their relationship to river terrace deposits is not clear in all cases. The sediments occur mainly at or close to the back of River Terrace Deposits (see Figure 34) where they are overlain by head deposits derived from the contemporary river bluff. At several localities they form a channel fill in the river terrace surface (for example, at Aveley, Figure 39); in others they fill a channel cut into bedrock, beneath the river terrace gravels. Particularly in the east of the district, the relationship of interglacial deposits and associated terrace deposits is complicated because of estuarine influence on deposition.
Differentiation between interglacial deposits of different ages is based on several methods including absolute and relative dating. C14 absolute dating is the only one that has proved successful so far, but is of value only for sediments of less than 40 000 years old. Work on the amino-acid geochronology of shells from Quaternary deposits at interglacial sites in this district has yielded mainly reliable results (Bowen et al., 1989; Bowen et al., 1995). Other relative dating is based on the fauna and flora of the Quaternary sediments (for a review see Wymer, 1999). Pollen is used to a limited extent to separate interglacials. Mammal faunas differ from glacial to interglacial stages and some species occur in deposits of a particular age. For example the giant beaver became extinct after the Hoxnian (OIS 11), hippopotamus and horse bones have been found in the Ipswichian (OIS 5e) but not the Hoxnian (OIS 11) whereas the extinct water vole Arvicola cantiana is found in deposits interpreted as OIS 11, but not in earlier interglacial deposits. Numerous Palaeolithic flint implements have been collected in the district. These are also valuable for the determination of the relative dates of deposits (Wymer, 1999).
Alluvium forms a nearly flat surface in valley floors. It occurs principally in the River Thames valley, its main tributaries and also in minor valleys where there is a distinctive floodplain developed (Figure 40). In the Mar Dyke area, north of Thurrock [61 82] there are some surface irregularities in alluvium, perhaps related to drainage and consequent desiccation and shrinkage of organic-rich layers.
The base of the alluvium in the Thames valley falls regionally from about sea level in the west of the district to lower than 10 m below OD in the east. The deposits rest unconformably on river gravel with a gently undulating boundary, particularly in the east, probably caused by channelling similar to that identified at Westminster (Wilkinson et al., 2000) and Plumstead Marshes (Devoy, 1979).
The thickness of alluvium in the Thames valley varies from less than 1 m in places in the west to around 15 m at Tilbury (Table 19). It is less than 3 m thick, and absent locally, below the present-day river channel west of Docklands. In tributary valleys, it is generally 2 to 5 m thick.
Alluvium consists largely of silty clay and clayey silt with locally developed beds of fine- to coarse-grained sand mainly less than 1 m thick but locally up to 4 m, for example in the vicinity of the Blackwall tunnel (Bates and Barham, 1995). There are also sporadic beds with scattered pebbles and granules. In the Thames valley from Erith eastwards, and to a lesser extent in the Cray and Darent valleys, chalk clasts of silt to granule grade derived from the underlying bedrock may occur in the basal part of the alluvium. East of Tilbury the succession contains a relatively high proportion of silt to fine sand.
Interbedded peat occurs in the east (Figure 40), with four main horizons recorded between Swanscombe and Tilbury (Devoy, 1979); the most widespread of these extends west to the Rotherhithe tunnel [355 805]. The total thickness of peat beds exceeds 2 m in large areas between the confluence of the rivers Thames and Lea and Tilbury (Bingley et al., 1999).
Tidal River or Creek Deposits
These are mapped only in areas between man-made sea defences and the mean low-water mark shown on the Ordnance Survey maps. They occur downstream of a point about 1.5 km west of the Thames Barrier [415 795] where there is a high suspended load and much sedimentation related to the position of the salt-water and freshwater mixing zone (Prentice, 1972). The preservation of these deposits is, in many areas, only temporary. The veneer of mud deposited during one tidal cycle is typically scoured off during another cycle, but in some places there may be a net accretion between the high and low water marks. However, this sediment eventually becomes unstable and slides towards the river channel where it is taken up in suspension.
Peat crops out only locally, as it is mostly interbedded with alluvium (see above). It occurs, for example, in a small valley fed by springs [685 695] emanating from the Harwich Formation, associated with abandoned river courses in the Kempton Park Gravel near Camberwell [323 792; 362 774] and on the spring line at the base of the Claygate Member of the London Clay near Herongate [6408 9022]. It is forming at present only in areas of permanent wet land, for example in ponds fed by springs. Thin beds of peat have been identified as Devensian interstadial deposits of some rivers, for example interbedded in the Lea valley alluvium (see p.76 — Lea Valley Arctic Bed). More recent peat deposits occur in the fen areas of the Mar Dyke valley [e.g. 62 85], which were poorly drained in former times.
Head is defined as material that has moved downslope by solifluction (mass movement under periglacial conditions). It formed principally beyond the ice limit during the glacial stages of the Pleistocene, when mass wasting was accelerated because of the arctic climate and lack of vegetation. When snow melted in the spring, debris of frost-weathered material formed a slurry, which gradually flowed downhill to form a poorly bedded deposit of variable character.
Periglacial solifluction has been the most potent agent of erosion in the district. As a consequence most of the hillslopes are essentially relict periglacial landforms that were largely formed during successive Pleistocene cold stages. The slopes have been modified only to a limited extent under temperate interglacial conditions (Ballantyne and Harris, 1994). The district has been beyond the glacial limit and thus subjected to longer periods of periglacial conditions than farther north. Consequently head deposits are widespread and almost certainly occur more extensively than are shown on the geological maps. However, as it has proved impractical to map solifluction deposits separately, the head that has been mapped includes these and surface wash and creep material.
In general, head occurs beneath concave slopes on the flanks and floors of valleys. Because of its local derivation, it is extremely variable in lithology, but composition closely reflects the source. For example, head derived from London Clay is clayey, and that derived from River Terrace Deposits is gravelly and sandy. Head derived from Chalk is generally calcareous, but may be locally decalcified depending on the acidity of ground and surface water.
The majority of the deposits are clay-dominated, derived from London Clay. Generally less than 2 m thick, they probably accumulated in shallow mudslides of softened and brecciated bedrock in the active layer. They consist of soft ochreous brown silty clay with blue-grey mottling in places and angular, frost-shattered fragments of flint occur sporadically throughout. At the base of these deposits and interbedded in places, there is a bed of pebbly clay, generally less than 0.2 m thick, with well-rounded flint pebbles derived from nearby outcrops of ‘high level’ gravel such as Stanmore Gravel. Beneath the head, there may be a low-angle shear surface or series of shears in the top part of the London Clay. Head derived dominantly from London Clay is mapped extensively south of Brentwood [60 90], around Wimbledon Common [23 72] and Richmond Park [20 73]. In these three areas there was considerable dissection in periglacial conditions. Head is likely to be widespread also in the areas of dissected topography in the north-west of the district, although it has not been mapped. Table 20 gives the location of the thicker Head deposits known in the district.
Head deposits are extensive on the almost featureless and gentle concavo-convex slopes in the Mar Dyke valley between Upminster [56 87] and Bulphan [64 86]. These too consist of remobilised London Clay with a basal gravelly clay layer. Solifluction sheets collected in this area from the surrounding high ground to the north and east. The broad Mar Dyke depression, formed after deposition of the majority of the River Terrace Deposits, may have been a large nivation hollow in Devensian time.
Head in the dry valleys on Chalk bedrock consists mainly of sandy silt and angular to subangular flint derived from dissolution of the chalk and from the Clay-with-flints. It is likely to have been reworked periodically by ephemeral streams that flow after high rainfall and when the water table rises above the valley floor. Locally the head on Chalk bedrock consists almost entirely of chalk clasts that may be cemented in places (formerly known as Coombe Rock). It occurs around Greenhithe and Swanscombe where deposits up to 3 m thick appear to be in the form of a fan overlying a steep Chalk bedrock surface, and between Purfleet [56 78] and Grays [62 78] (Plate 4). The outcrops at the latter localities are small, mainly because much of the deposit has been removed by quarrying for chalk. Similar deposits of soliflucted chalk debris may also occur beneath the larger dry valleys in the south of the district.
Small patches of head occur on interfluves, for example near Longfield [59 68] and Meopham [655 665]. They are derived from Thanet Sand and Clay-with-flints and consist mainly of fine-grained sand with variable amounts of clay and a mixture of angular, subangular and well-rounded flint pebbles.
These deposits are formed by solifluction and downwash of gravel-dominated deposits. In the area around Brentwood they occur generally on interfluves downslope from outcrops of Stanmore Gravel. Outcrops on the chalk in the south-east of the district are a mixture of fine-grained sand derived from the Thanet Sand, flinty clay derived from Clay-with-flints and well-rounded black flint pebbles derived from the Harwich Formation; many of the outcrops are in shallow dissolution depressions in the chalk.
SEQUENCE OF QUATERNARY EVENTS
The broad outline of Quaternary events in the district has been understood since the early work of the Geological Survey (see for example Whitaker, 1884; Prestwich, 1891). A modern account of the regional setting for these events is given by Sumbler (1996). Because of the fragmentary distribution of the Quaternary deposits, the correlation between many of them has been the subject of a long-running and continuing debate. There are also conflicting views about the age of some of the deposits in relation to the oxygen isotope stages (Table 9).
The oldest Quaternary deposits, the Stanmore Gravel, are interpreted as inshore or beach deposits, equivalent in age to the marine sands of the Red Crag that occur to the north-west of the district in Essex and Suffolk. These deposits were subjected to neotectonic uplift in early Quaternary times, which led to the removal by erosion of much of the Stanmore Gravel, and, in the south of the district, the exhumation of the Palaeogene deposits on the dip slope of the North Downs. Subsequently, Clay-with-flints accumulated on this exhumed surface.
The earliest major river system in the district drained northwards towards an ancestral Thames that flowed across East Anglia; this has been dated as Oxygen Isotope Stages (IOS) 16 to 14. One tributary flowed along the present-day Colne valley and another across the western part of the district. Deposits of these rivers were the last to be laid down before the drainage system was radically modified as a result of ice advance during the Anglian glaciation, about 500 000 years ago. The Anglian ice blocked the rivers, which were diverted into what is now the Thames valley. The mechanism of this diversion is complex with four ice advances being recognised (Cheshire, 1981) in the Vale of St Albans to the north of this district. One of the earlier ice advances may have diverted the Thames into a spillway along the line of the present-day Lea valley, causing major erosion. This mechanism provides an explanation of the abnormal size of the Lea valley in relation to its current small catchment area. Later ice advances blocked the Lea spillway and diverted the Thames into its present valley. The last pre-diversionary Thames river deposit was the Westmill Gravel, preserved in the Colne valley in the north-west of the district. Small tongues of ice extended from the main ice sheet, to the north of the district, southwards along pre-existing valleys including the Colne in the north-west, the Roding at Hornchurch, and the ‘Finchley depression’ [26 91] (Figure 31). It is unlikely that these valleys were formed by ice scour because they were occupied by relatively small, marginal ice lobes with only minor erosive potential.
The relationship between the Anglian till and underlying gravels is uncertain. Two lines of evidence suggest that there was a considerable period of time between them. First, a palaeosol developed locally in the top part of the Gerrards Cross Gravel (Bridgland, 1994, p.121) may correspond to the widespread interglacial palaeosol developed on the correlative Kesgrave Sands and Gravels farther north and east. Second, the topography across which the valley ice tongues advanced was in several places significantly lower than the base level of the gravels. The latter evidence indicates a major period of erosion in the Thames valley before the ice advanced. It indicates also that there was already a depression in the present Thames valley prior to its occupation by the diverted River Thames.
It is generally agreed (for overviews see Gibbard, 1985; Bridgland, 1994) that the River Thames was diverted into its present valley late in Anglian times. This event led to the deposition of extensive river gravels that now form the River Terrace Deposits. Post-Anglian times were characterised by climatic oscillations, which were associated with marked changes of sea level that significantly influenced the depositional environment.
During cold periods rapid erosion and the downslope movement of vast quantities of material occurred in periglacial conditions. Much of the debris produced was transported in the contemporary river systems and laid down as river terrace deposits. Remnants of the periglacial deposits are mapped as Head and Head Gravel. In intervening warm periods, interglacial sediments accumulated in abandoned meanders that were cut into the River Terrace Deposits or in channels cut into bedrock, but because of later erosion these are preserved only locally. Interpretation of the relationship between River Terrace Deposits and Interglacial Deposits in the Thames valley suggests that there are cold fluvial deposits and warm interglacial deposits representative of all the oxygen isotope stages from IOS 11 and IOS 3 (Figure 41).
The oldest post-diversionary terrace deposit, the Black Park Gravel, is generally accepted as late Anglian (IOS 12) in age. The next oldest terrace deposits, the Boyn Hill Gravel, contains interglacial deposits at Swanscombe which are internationally important because of the discovery of the ‘Swanscombe Skull’ and associated flint artefacts. Although there are some uncertainties because of the lack of pollen data, the Swanscombe sediments are attributed to the first inter- glacial after the Anglian, now generally regarded as OIS 11 (Bowen, 1999). The subsequent Lynch Hill Gravel includes interglacial sediments at a number of sites, principally at Purfleet and Grays. These sediments include laminated beds of estuarine origin, recording an episode of high sea level. The fossil plants and animal remains are not diagnostic of any interglacial period but their position in the terrace sequence suggests deposition in the second of the four post-Anglian interglacials, generally ascribed to IOS 9 by Bridgland (1994), but regarded as earlier by Gibbard (1999).
Interglacial deposits have been found associated with the Taplow Gravel in several places, notably at Uphall Pit Iford, Aveley and Ockendon, again with evidence for estuarine conditions in part. These sites have been attributed to the last (Ipswichian) interglacial on the basis of pollen analysis, a view challenged by palaeontologists working on mammals and molluscs, who refer the deposits to the third of the post-Anglian interglacials, denoted as OIS 7, and termed the Ilfordian by some workers.
The last interglacial (Ipswichian sensu stricto) is now generally correlated with IOS 5. Sediments of this age have not been found downstream from London as they are now buried beneath the modern floodplain. Deposits referred to the IOS 5e occur at Trafalgar Square and Kew, within the Kempton Park Gravel.
The Devensian glacial stage, from about 122 000 to 10 000 years before present (BP), incorporates at least three short episodes of milder climate (interstadials). The first two of these, known as Chelford and Upton Warren interstadials, occurred before the main Devensian glaciation; the latter is represented by deposits in the top part of the Kempton Park Gravel at Isleworth, Twickenham and Kempton Park.
The main part of the Devensian Stage (IOS 4 to 2) corresponds to the period between about 70 000 and 13 000 years BP. Cold climate gravel deposits that form the later part of the Kempton Park Gravel (Bridgland, 1994), and the early part of the Shepperton Gravel (Gibbard, 1985) were probably laid down during this stage. The latter occurs beneath the present Thames river course and is not exposed in the London district (see Figure 34). In the latest Devensian, during the Dimlington Stadial, 25 000 to 13 000 years BP, sea level was as much as 120 m lower than at present. The Thames river channel cut at that time now lies 5 m below OD in central London falling to 20 m below OD in the east of the district. As sea level rose again towards the end of the Devensian, this channel was infilled, initially, by sand and gravel deposited in a braided river. Lenses of organic silt and clay giving radiocarbon dates ranging from 28 000 to 21 000 years BP, are interbedded with and overlie these gravels beneath the floodplain of the River Lea at Ponders End [360 950] (Wymer, 1985; Gibbard, 1994). They are known as the ‘arctic bed’ and contain remains of mosses and dwarf birch and willow that indicate a severely cold climate. Deposition of the gravels appears to have ceased after about 15 500 years BP, probably as a result of declining river flow. The main river changed character and flowed in more restricted channels, with sand being deposited in point bars. Some of these sands crop out as islands or eyots within the alluvium, for example at Westminster [302 795] and Bermondsey [335 799]. The river maintained this low flow regime probably until about 7000 years BP at Tilbury and 3500 years BP in central London.
The final part of the Devensian stage began with the Windermere Interstadial, dated between about 13 000 and 11 000 years BP. Sediments of this age are recorded from the district only at Bramcote Green, South Bermondsey [350 785]. Here, deposits at the base of the alluvium have yielded birch pollen (Thomas and Rackham, 1996). Cold conditions returned briefly during the Loch Lomond Stadial from about 11 000 to 10 000 years BP. Organic deposits of this age occur in a shallow channel at the base of alluvium at Silvertown [401 805] (Wilkinson et al., 2000).
The climate finally ameliorated at the start of the Holocene about 10 000 years BP; this marks the start of the Flandrian stage of the British Quaternary. Sea level continued to rise through much of the Holocene resulting in the deposition of alluvial mud. Lenses and beds of peat within the sequence, mainly to the east of Docklands [39 81], are thought to indicate temporary minor falls in sea level (Devoy, 1979) or a slowing in the rate of sea level rise. Distribution of the peat was governed partly by a supply of fresh water from springs along the valley sides to emergent mudflats. A maximum of five peat beds have been recognised in the vicinity of Tilbury (Devoy, 1979) where the alluvial sequence is thickest (Figure 40).
The basal peat at Tibury is dated about 8300 years BP. Subsequently, as sea level rose rapidly (about 13 mm per year) the tidal head of the estuary probably reached Tilbury by around 7700 years BP. A regression at 7000 years BP is marked by a second bed of peat, which became submerged at 6600 years BP. Commencing at about 6200 years BP and lasting until about 2500 BP, the area of estuarine intertidal mudflats was reduced in size and there was a seaward expansion of substantial tracts marsh deposits along the river margins and in the tributary valleys. This coincided with the development of the thickest peat in the alluvial succession (see Long et al., 2000). Following this, sea level continued to rise slowly, and estuarine conditions once again migrated upstream so that in central London there was a gradual change from a freshwater system to an estuarine one. In the east of the district there was a return to brackish marine conditions in the estuary, and the channels cut into the peat were infilled with intertidal sediments. The uppermost peat in the alluvial succession marks a regression during the third and fourth centuries AD. This was followed by further sea level rise that is continuing at present.
Present-day sea level change and ground movement
Archaeological evidence indicates that the River Thames was not tidal in Roman times and occupation levels in London were at least 2 m below current high tide level. Today the River Thames is tidal as far as Teddington, and over the last two centuries there has been an increased tidal range caused by a decrease in tidal friction in the Thames estuary. This has been brought about by removal in 1830 of the Old London Bridge, which had always acted as a partial barrage. Also, the River Thames was extensively dredged in conjunction with the expansion of the London docks in the late 19th century. These two factors led to an increase in the tidal range from 4.6 to 6.3 m by 1877.
In more recent times, high quality tide gauge records have been used to produce mean sea level trends at coastal tide gauges, which form part of the National Tide Gauge Network (Woodworth et al., 1999). However, the only high quality tide gauge records available for the Thames estuary are those at Tilbury, east of London, and the coastal sites of Southend and Sheerness. Published estimates for the mean sea level trends showed a rise in relative mean sea level at all three sites. The trends and their standard errors were:
- +1.58 ± 0.91 mm/yr for Tilbury for the period from 1961 to 1983
- +1.22 ± 0.24 mm/yr for Southend for the period 1933 to 1983
- +2.14 ± 0.15 mm/yr for Sheerness for the period 1901 to 1996
Superimposed on these trends is a rise in global sea level. In 1990 and 1995, the Intergovernmental Panel on Climate Change (IPCC) reviewed the published evidence on the influence of global warming on sea levels (IPCC, 1995). They found that global sea level had risen by 100 to 200 mm over the past century. This is equivalent to a linear rise in the order of 1.0 to 2.0 mm/yr. In contrast, a rise in sea level of 8 mm per year was indicated by the height of storm tide surges and records from a Greater London tide gauge. An estimated 40 to 75 per cent of the difference (3 to 6 mm per year) could be due to the increase in tidal range, but the balance, up to 4 mm per year rise in water level, could be accounted for by changes in ground level in the Thames estuary and Greater London (Muir Wood, 1990).
To add to the uncertainty, there are few historical geodetic observations of changes in ground level in Greater London. In the Second National Geodetic Levelling of Great Britain carried out by the Ordnance Survey (OS) Greater London was surveyed between 1946 and 1951. The Third National Geodetic Levelling of Great Britain was carried out between 1952 and 1959. Not surprisingly, considering the precision of geodetic levelling at the time, Kelsey (1972) reported no significant change in the relative heights of OS benchmarks in Greater London between them. Later, the results from a north–south levelling traverse across London, carried out by the OS in the 1960s, indicated a consistent sinking of Central London of approximately 2 mm/year compared to ‘stable’ points to the north and south.