Quaternary river terraces in England: Forms, sediments and processes

Abstract

Flights of Quaternary river terraces in south and east England have common characteristics involving low-gradient planed or irregular bedrock surfaces and single or multi-storey gravel deposits. Rather than depending on warm–cold or cold–warm transitions, it is suggested that bedrock planation, “working depths” of gravel and later-stage (relatively shallow) aggradations are all dominantly of cold-climate origin. Basal sediments show active incorporation of plucked and periglacially-shattered materials, whilst super-incumbent units incorporating up-catchment and slope-derived materials demonstrate later cold-stage sediment influx and consequent cessation of active bedrock erosion. Channel activity effecting both planation and deposition are reviewed, together with the detailed sedimentology of gravelly sediments which show evidence of both autogenic processes (bar migration, channel switching and infilling, and truncation of upper sedimentation units), cold-climate indicators (turbation, ice-wedge casts, and frozen block transport), and (specifically for the last glacial–interglacial cycle) varying sediment flux as climates changed. Both interglacial and “transitional” activities are believed to be of lesser morphological significance, whilst prior uplift is taken as enabling rather than being a generator of terrace within the timescale of a glacial–interglacial cycle. Variations within cold-stage climates, varying sediment influx and channel-belt bedrock erosion are stressed as dominating mid-catchment and mid-latitude Quaternary terracing at the glacial–interglacial scale.

Full text

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

Author's personal copy
Quaternary river terraces in England: Forms, sediments and processes
J. Lewin
a
, P.L. Gibbard
b,
⁎
a
Institute of Geography and Earth Sciences, University of Aberystwyth, Aberystwyth SY23 3DB, Wales, UK
b
Cambridge Quaternary, Department of Geography, University of Cambridge, Downing Street, Cambridge CB2 3EN, England, UK
abstractarticle info
Article history:
Received 25 August 2009
Received in revised form 29 March 2010
Accepted 5 April 2010
Available online 11 April 2010
Keywords:
River
Terrace
Sedimentation
Quaternary
strath
Flights of Quaternary river terraces in south and east England have common characteristics involving low-
gradient planed or irregular bedrock surfaces and single or multi-storey gravel deposits. Rather than
depending on warm–cold or cold–warm transitions, it is suggested that bedrock planation, “working depths”
of gravel and later-stage (relatively shallow) aggradations are all dominantly of cold-climate origin. Basal
sediments show active incorporation of plucked and periglacially-shattered materials, whilst super-
incumbent units incorporating up-catchment and slope-derived materials demonstrate later cold-stage
sediment influx and consequent cessation of active bedrock erosion. Channel activity effecting both
planation and deposition are reviewed, together with the detailed sedimentology of gravelly sediments
which show evidence of both autogenic processes (bar migration, channel switching and infilling, and
truncation of upper sedimentation units), cold-climate indicators (turbation, ice-wedge casts, and frozen
block transport), and (specifically for the last glacial–interglacial cycle) varying sediment flux as climates
changed. Both interglacial and “transitional”activities are believed to be of lesser morphological significance,
whilst prior uplift is taken as enabling rather than being a generator of terrace within the timescale of a
glacial–interglacial cycle. Variations within cold-stage climates, varying sediment influx and channel-belt
bedrock erosion are stressed as dominating mid-catchment and mid-latitude Quaternary terracing at the
glacial–interglacial scale.
© 2010 Elsevier B.V. All rights reserved.
1. Introduction
This paper addresses the problem of extra-glacial Quaternary river
terrace formation in mid-latitudes as a geomorphological process:
how and under what conditions were laterally extensive bedrock
platforms cut, what do the sediments on them indicate, and was there
a prototype sequencing of events within a glacial–interglacial cycle?
We examine specifically middle-course catchment locations in
lowland England where there are elevated flights of valley-side
Pleistocene terrraces, rather than the multi-unit, cut-and-fill aggra-
dation environments of depocentres and coastal prisms downstream
(as in the Thames estuary and East Anglia), or the fill–cut sequences of
smaller valleys and upland environments which are dominated by
post-glacial incision into prior fills, including glacial and solifluction
deposits. We survey river behaviour during a glacial cycle as it affected
bedrock erosion, deposition and terracing in a region of moderate
relief in north-west Europe, using evidence from any particular cycle
that has been well-researched and is relevant, most being available for
the last glacial–interglacial cycle (post-Ipswichian–Eemian–c. MIS
5e). The preserved evidence from cold-stage sediments does not have
the interpretative richness of interglacial biological diversity (Gibbard
and Lewin, 2002), and there are buried erosional features to consider,
so this is not so straight forward.
2. Terraces and their deposits
Flights of terraces can develop in alternative ways as was
appreciated by Bryan and Ray (1940), Leopold and Miller (1954)
and Bull (1991), to mention only a few. Basic concepts are illustrated
in Fig. 1.Leopold et al. (1964) usefully distinguished between strath
(bedrock-cut) and fill terraces (where terrace surface height is
essentially that produced at the end of aggradation episodes). Fig. 1
(a) and (b) shows terrace sequences representing stages of incision
into prior fills (fill–cut), and aggradations within fomerly incised fills
(cut–fill). These are recorded in numerous studies of post-glacial
terraces in upland Britain incised into Pleistocene valley fills (for a
listing of studies, see Macklin et al., 2009,Table 1). Up to six terracing
episodes have been distinguished in the later Holocene alone
(Passmore and Macklin, 2000). The focus in this paper is on older
Pleistocene terraces cut in bedrock and overlain with a thin covering
of sediment. These are essentially of type 1(c). In recent years, they
have been viewed predominantly as recording MIS stages and uplift
(Bridgland and Westaway, 2008), whilst there have been numerous
studies of site stratigraphy (e.g., Collins et al., 1996; Coope et al., 1997,
2002; Maddy et al., 1998; Lewis et al., 2001; Gao et al., 1998, 2007;
Briant et al., 2004, 2005, 2008; Langford et al., 2007; Rose et al., 2000).
Geomorphology 120 (2010) 293–311
⁎Corresponding author. Tel.: + 44 1223 352946.
E-mail address: plg1@cam.ac.uk (P.L. Gibbard).
0169-555X/$ –see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.geomorph.2010.04.002
Contents lists available at ScienceDirect
Geomorphology
journal homepage: www.elsevier.com/locate/geomorph

Author's personal copy
These have included analyses of lithofacies, palaeontology and
chronology.
In broad terms, these terrace sediments are mobile channel-belt
gravels (Gibling, 2006), with incursions of slope material. Under-
standing is obviously restricted unless the actual nature of erosion and
sedimentation has been established through subsurface investigation.
As Fig. 1 suggests, virtually identical surface features can be produced
by different combinations, and numbers, of aggradation and incision–
planation episodes. The use of surface height data from ground survey
can be adequate for lateral correlations between terrace fragments at
the basin scale, but this reveals only limited aspects of the
geomorphological processes involved. In addition to surface expo-
sures and core information, ground penetrating radar is beginning to
provide much greater understanding of subsurface sediments and
bedrock surface forms (Davis and Annan, 1989; Vandenberghe and
van Overmeeren, 1999; Bridge and Lunt, 2006; Howard et al., 2007;
Gibbard et al., 2008). Sediments may constitute alloformations, or
alluvial architectures, styles or ensembles (Autin, 1992; Miall, 1996)
which are commonly subdivided into facies associations. Together
with biostratigraphy and dating control this allows identification of
formations and members and their interpretation in terms of
environmental change, whilst geomorphological processes may be
deduced from the sedimentology and erosional forms.
Even within the simple model provided by Fig. 1, multi-phase
activity leads to unit insetting, overlapping and intra-terrace sediment
truncations (that is, within the sediment packages beneath the levels
shown in Fig. 1)—both autogenically as channels relocate laterally or
vertically, and allogenically as external factors (such as changes in
river discharge or sediment supply) remove at least some of the
primary-phase materials related to terrace-cutting, and then deposit
others. The point to emphasise is that a sedimentological record of
much geomorphological activity is likely to be missing. Fig. 2
represents three possible situations affecting preservation. The
process of vertical stacking (Fig. 2a) quite commonly also eliminates
upper facies in older alluvial materials. This is the autogenic norm for
braided rivers where migrating channels operate at different depths
and energies within an active alluvial system even without any
necessary vertical system tendency, and there is continous over-
printing of the total sediment package as channels shift. Stepped
incision (Fig. 2b) may preferentially preserve (and in a sense over-
represent in the lithostratigraphic record) valley-margin colluvial
materials and deeper, confined scour-pool fills in the stratigraphic
sequence (Lewin and Macklin, 2003). Preserved terrace remnants
can be broad or narrow, with alluvial units overlapping bedrock
benches where incision amounts are small (e.g., Antoine et al., 2000).
Unidirectional shifting without pronounced incision (Fig. 2c) may
Fig. 1. Terrace formation sequences producing the same surface appearance: (a) by a set of incisions into prior valley fill, (b) by coupled cut-and-fill episodes, and (c) by progressive
incision into bedrock to produce strath terraces.
Table 1
Alluvial plains, channels and channel patterns for contemporary Arctic and British gravel-bed rivers.
River Plain width
(km)
Channel width
(m)
Slope
(m/km)
Pattern style and sequence
Cold-climate
Colville 69°18′N 152°24′W 1.48 250 0.88 mWbmd
Toolik 69°47′N 149°36′W 3.19 130 0.74 Ma
Sagavanirkok (Lunt and Bridge 2004) 69°50′N 148°43′W 2.4 260 2.12 m/bB
Babbage (Forbes 1983) 69°09′N 138°20′W 2 110 1.25 mM
Albany (Martini et al., 1993) 52°11′N 81°53′W 3.91 1160 0.55 aA
Tana 69°26′N 25°41′E 0.79 90 1.12 M
Mongocheyakha 72°06′N 79°01′E 4.34 110 0.05 mM
Lenivaya 75°05′N 89°38′E 1.3 160 0.35 mW
Yana 70°21′N 133°00′E 3.94 230 0.98 mAm
Khroma 70°27′W 147°34′E 11.46 750 0.03 mAm
Regtymel' 69°25′N 174°48′E 1.48 290 1.2 bBb
United Kingdom (historical change sites)
Spey (Lewin and Weir 1977) 57°39′N 3°05′W 1.4 50 4 B/W
Feshie (Werritty and Ferguson 1980) 57°39′N 3°54′W 0.2 40 11.56 W/B
South Tyne (Macklin and Lewin 1989) 54°56′N 2°31′W 0.18 30 3.84 W/B
Severn (Brewer and Lewin 1998) 52°28′N 3°26′W 0.14 30 2.5 M/B
Ystwyth (Lewin et al. 1977) 52°20′N 3°54′W 0.18 20 8.47 B/W
Pattern styles: m, meandering; b, braiding; w, wandering; a, anastomosing; and d, deltaic. Pattern at data site in capitals.
294 J. Lewin, P.L. Gibbard / Geomorphology 120 (2010) 293–311

Author's personal copy
create so-called “row terraces”at similar elevation but with a lateral
ageing sequence. This may occur at a bedrock base, either forming or
sliding across a surface which was previously created, or above an
older basement-sediment “socle”as in the Holocene sequences of the
Main, Donau and Rhine (Schirmer, 1995; Erkins et al., 2009). This
progressive lateral shift on larger rivers proceeded on a millenial
timescale and it has probably been under-reported in Britain where
the Holocene mobility of major rivers may also have been less. This is
in contrast to British upland channels which have laterally reworked
valley floors on a scale of centuries.
On a smaller scale, and within the active alluvial assemblages of
Fig. 2, there is also preferential preservation, specifically of deep
scour-pool infills (Huggenberger and Regli, 2006). Pleistocene gravel
exposures normally exhibit multi-storey characteristics which may be
created at three levels: active channel and bedform migration at the
reach scale (Bridge and Lunt, 2006), migration of the channel-belt as a
whole across a braid or meander plain, and lastly allogenic
transformations of channel style involving also an aggradation–
degradation balance. It is important not to confuse these. For example,
scour-pools, deepened especially at confined valley margins and
channel junctions, may be partly filled at flood-waning stage with
poorly sorted bimodal gravels, in contrast to bar surface openwork
and well-sorted gravel sheets which remain intact. Both can be parts
of the same autogenic system and do not indicate an allogenic
environmental change, and both may be truncated by further lateral
channel migration —again, such truncations may be autogenic rather
than allogenic in origin.
In all these circumstances, neither terrace surface elevation nor
aggradation materials may be closely related to terrace base or strath.
It can be no surprise that the examination of limited exposures or
surface forms yields mixed messages. Also in published discussions,
all terrace sediments may be termed “aggradations”, with the possible
implication that the planed-off bedrock features beneath have been
buried by, rather than formed in conjunction with, the sediments that
lie on them. As these sediments are commonly cold-climate in origin,
it has sometimes been presumed that the incision seen as the
antithesis of aggradation must be interglacial or “transitional”(e.g.,
Hancock and Anderson, 2002; Bull, 2007; Bridgland and Westaway,
2008; Vandenberghe, 2008). Some stacked terrace sequences can
indeed be aggradations (including channel fills, and surface depres-
sion “pockets”of later interglacial fluvial, wetland and colluvial fills),
but other terrace sediments may more properly be regarded as
remnants of a co-formational “working depth”of materials in transit
as broad valley floors were being simultaneously widened and
deepened as bedrock straths or basal boundaries were being
created and lowered (Gibbard and Lewin, 2009). The thickness of
such migration plain sediments is approximately that of deeper
channels at the time, though channel forms themselves are often
poorly preserved within channel-belt sediments (Gibling, 2006). The
concept of incision episodes involving broad valley-floor bedrock
planation surfaces being “sediment free”is unreal. Furthermore, even
where there are thick aggradations these may be underlain by
extensive bedrock surfaces, and there is a very wide lateral erosional
surface beneath sedimentation units to explain.
Idealized terrace components are illustrated in Fig. 3.Itis
suggested that four basic geomorphological elements are involved:
the terrace base, its sediments, its surface form and its margins. In the
simplest case, the total sediment package or unit produced without a
system change is bounded by what Miall (1996), as a sedimentologist,
called “sixth-order bounding surfaces”(the first five orders defining
unit sets of increasing size and combination, from ripples to macro-
forms like point bars and channel deposits) or what Autin (1992),asa
stratigrapher, called “alloformations”. For terraces, the base and
valley-side margin are co-formational (being created together with
and alongside the transiting presence of temporarily stored erosion
products), while the down-slope lateral margin is a terrace-
consuming one developed under later conditions. Sediments may be
divided into those with genetic linkage to underlying strath formation
(Fig. 3, I), and those of subsequent deposition, multi-storey aggrada-
tion or incision, some episodes of which may considerably post-date
the primary phase within glacial–interglacial cycles. Two are
illustrated in the model (II–III) but there may be more. The terrace
margins reflect onset–termination episodes: termination times and
thereafter (IV) can be accompanied by late phases of surface
accumulation (an upper bounding surface may, for example, be
modified by loess or colluvial sedimentation and depression-filling
over an extended period), as well as marginal incision and unit
destruction. The latter literally circumscribes preserved terraces,
whilst the remnants that are left are considerably affected by later
degradation or burial under potentially quite different environmental
conditions. It should be noted that whilst (I) is integral to strath
formation, (II–III) may prevent it, although it is they that may largely
determine terrace surface elevation, as in Figs. 1b and 2a.
Geomorphologically, therefore, both surface, internal sediments
and bedrock forms are of interest; this includes microfeatures as well
as steps and gradients. For bedrock surfaces, as with most erosional
forms, there may be only residual and accidental evidence for how
material has been eroded. Fig. 3 illustrates a two-stage process of
bedrock planation as well as sedimentation, the rock surface showing
evidence of direct river processes like abrasion and quarrying, but also
the weathering state of the bedrock itself. Also included are the lateral
limits to planation in the form of upper terrace margins. Confined
channels may leave characteristic traces at valley-floor margins
(Lewin & Brindle, 1977), with arcuate terrace fronts defining
outward-growing meander loops, or straight ones created by down-
valley shift of translating loops or braid-channel migration. Late-
glacial large-scale meander scars are well known (e.g. Panin et al.,
1999; Sidorchuk et al., 2003; Leigh, 2006), whilst linear terrace fronts
with small-scale, high-radius indentations can be observed in active
braid systems (Warburton et al., 2003). This distinction is not always
clear-cut. The form of terrace back-slopes is usually blurred by
colluvial input and slope degradation, but subsurface remote sensing
Fig. 2. Spacial sequencing of alluvial units: (a) stacked units produced by aggradation with the coupled removal of upper unit material, (b) stepped units in which vertical incision is
accompanied by lateral shift so as to eliminate the valley-centre assemblages from earlier units, and (c) row units produced by lateral shift to give a horizontal age sequencing across
a bedrock basal surface.
295J. Lewin, P.L. Gibbard / Geomorphology 120 (2010) 293–311

Author's personal copy
capability may allow a greater understanding of this concealed aspect
of terrace form.
The deposits of most concern in the present context are largely
channel-belt gravels, but to a disconcerting extent the gravelly
deposits of meandering, wandering and braided streams appear
similar in restricted exposures or boreholes.They are not very
different since all have pool fills, and bar surface, lateral accretion
and bar-tail deposits (cf. Bluck, 1971; Lewin, 1976; Forbes, 1983; Lunt
and Bridge, 2004). If there were a full autogenic sequence, from
bottom to top this could consist of lag deposits, scour-pool infills,
planar and trough cross-bedding, lateral accretion and bar-front
avalanche deposits, bar-tail fines, upper-level channel fills, and (above
coarse bedform level) inner accretionary bank oblique stratification
and overbank fines. But because of bar and channel migration,
because there may be several channels in the system scouring to
different depths (chute channels in the case of meanders, and a
network of channels in the case of braiding), and because of lateral
system migration, deposits have multiple truncation surfaces running
through them. Bedrock erosion is set by scour-pool depth which can
be several times average channel depth (Best and Ashworth, 1997),
particularly at channel junctions and confined valley margins. As was
suggested above, to a very large extent what get best preserved are lag
deposits, bottom scour fills, and sub-horizontal gravel sheets showing
multiple erosional truncations (Siegenthaler and Huggenberger,
1993; Lunt and Bridge, 2004) whatever the morphological style of
the channel pattern. This does not best preserve biological materials
in the way that upper alluviating levels do (e.g., higher-level cutoff
channel fills, bar-tail particulate organic detrital accumulations, or
sedimentation in floodplain ponds). It is naturally also difficult in
small exposures to differentiate between deposits produced by one of
several sinuous branches of a multiple-channel braid system and the
deposits of a singular meandering channel. Clast orientation, an inner
accretionary bank, or bar-tail deposits may help to distinguish
channel style (Bluck, 1971, 1974, 1976; Arche, 1983), but this is
often not possible where the upper levels of autogenic sequences have
been removed and there are limited lateral or down-channel
exposures.
At a smaller scale, energy levels may also be deduced from grain
sizes, sorting, bedding and clast-matrix character (Miall, 1996), whilst
a range of features like pebble clusters, grading, stratification and
mass gravity flows may be seen in section (Collinson and Thompson,
1982). Cold-climate features such as intraformational wedge casts and
involutions may also be present (Seddon and Holyoak, 1985; Martini
et al., 1993; Mol et al., 2000; Vandenberghe, 2001). These allow
interpretation of floodplain environments additional to palaeontolo-
gical evidence, with some limited bearing on the pattern and lateral
activity of the river itself. For example, nival and permafrost regimes
may give high seasonal flows with load structures and wedge casts
relating to in-channel and extra-channel processes (French, 2007).
Very high energy levels are more likely to produce straight channels
with planar bedforms and gravel sheets, rather than channel-splitting
and sinuousity-generating stalled bedforms of lesser (but still high)
energy levels (Miall, 1996; Bridge, 2003). Permafrost also leads to
bedrock shattering by ice segregation or near-surface freeze–thaw
activity leaving distinctive signatures (Murton, 1996; Murton et al.,
2006), whilst thermal erosion of ice-rich sediments can cause
exceptionally rapid bank retreat (Church and Miles, 1982; French,
2007, chapters 4 and 10.1).
Some indication of the range in cold-climate macro-scale alluvial
style may be given by contemporary analogues (Table 1). As has been
documented by Vandenberghe (2001), cold-climate rivers may be of
all types. Braid systems (generally at higher slopes and greater bed-
sediment supply rate) may be relatively straight with lateral switch-
ing. Lunt and Bridge (2004) reported that the active channel-belt of
the braiding Sagavanirktok River in Alaska (in effect an active gravel
alloformation) was some 2.4 km wide and up to 7 m thick. Meanders
may shift slowly, but over a broad migration plain. Theoretically,
channel patterns can also be sequenced down-valley (braiding–
Fig. 3. Mid-latitude Quaternary terrace elements produced during bedrock incision and aggradation. Incision and lateral planation (I) is accompanied by a transiting and co-
formational “working depth”of sediment. Later phases of aggradation (II) may preclude further broad incision, though there may be incision episodes (III) and ongoing reworking of
alluvial units. The terrace is finally abandoned following terminal deep incision (IV).
296 J. Lewin, P.L. Gibbard / Geomorphology 120 (2010) 293–311

Author's personal copy
wandering–meandering–anastomosing) with a decline in slope, grain
size and distance from bed-sediment source. But as Table 1 shows, in
practice the sequencing is more complex, depending on local slope
and tributary sediment inputs. Rivers also show different degrees of
incision; such auto-confinement may limit migration plain width in
the course of expanding it. Other rivers oscillate across broad alluvial
fans or coastal plains.
Excess sediment accumulation may lift the channel off the bedrock
floor, and theory and experiment (though generally involving linear
incision potential) suggests that a low bed-sediment transport rate
should be most effective for bedrock incision (Sklar and Dietrich, 1998;
Whipple et al., 2000; Whipple and Tucker, 2002). So which pattern is
most effective for lateral bedrock planation is less obvious than it might
seem, even though a full range ofchannel patterns may all be developed
in cold climates over permafrost (Table 1). It seems possible that whilst
the lateralmigration of meanders may be slower, itmay also be broad in
extent and, with a lower bed-sediment load over a longer timespan,
could be effective for bedrock planation. With aggradation involving a
lifting of channels off the bedrock floor, braid patterns may efficiently
transport bed materials down straight channel-belts but without
effecting so broad a bedrock incision. On the other hand, channel
switching, migration and deep-pool scour may enhance it. It should not
be unexpected for Pleistocene channel styles and bedrock planation
potential to change down-valley and to vary in effectiveness.
Finally for comparison, historically active, gravel-transporting
braided and wandering river styles do exist in Britain (Table 1), but
they are mostly small (the Spey is an exception), have steep-gradients
and are in quite narrow valleys. Actively meandering gravel-bed
rivers are much more common (e.g., Bluck, 1971, 1976; Macklin and
Lewin, 1989), but again with exceptions (e.g. the Frome in Dorset),
these are not generally found in lowland England where contempo-
rary rivers are inactively meandering. Although gravel mobility and
terracing is achieved in Holocene steep-gradient streams, there is
little evidence of post-glacial bedrock incision and planation even
there, as would be expected if cold–warm transitions were effective.
In the lowlands, virtually all larger river systems have Holocene
sediments inset above a basal Pleistocene fill.
3. Surface gradients and terrace planform remnants in
southern England
It is striking that the gravel terraces of larger rivers in southern
England for the most part have quasi-linear long profiles and very
gentle gradients (Table 2). These steepen in tributaries and head-
waters, but mid-catchment profiles on the Thames and Solent are
almost linear for 100 km and more. Gradients of less than 1 m km
−1
are general, and not very different from present-day rivers. Terrace
gradients on the meandering Warwickshire Avon (Maddy et al., 1999)
are also less than 1 m km
−1
though this is a little greater than along
the contemporary river which meanders within a meandering-valley.
Those of the lower Severn are also a little steeper (Maddy et al., 1995),
but still lower than the active Arctic river and British historically
braided or wandering rivers presented in Table 1. Extreme discharges
would be needed for gravel transport at such gradients, or possibly
high-concentration gravity flows where indicated by poorly sorted
diamictons. Gravity-flow deposits (unstratified, matrix-supported
gravels) are not commonly found in modern gravel-bed rivers (Bridge
and Lunt, 2006), but they are occasionally reported in the context of
near-bedrock Pleistocene fluvial deposits (e.g. Maddy et al., 1998;
Gibbard et al., 2008). Interpretation of the primary deposition of older
sediments can be greatly inhibited by disturbance and cryoturbation.
In England, the layout and lateral extent of preserved terrace
fragments varies considerably. Terrace sets do show down-valley
continuity, but they are seldom “paired”across valleys. Runs tend to
be on one sideof the valley only, butthese can shift acrosspresent valley
axes, as in the case of the West Knighton Gravel on the Frome near
Dorchester (Fig. 4) and the Kempton Park Gravel on the Thames, on the
right bank at Maidenhead but the left bank downstream at Iver (Fig. 5).
In the Upper Thames above Eynsham, terraces are dominantly on the
left bank, whilst the former Solent river terraces form a staircase on the
northern side of the valley, suggesting an overall unidirectional shift.
There are locations where terrace pairing does exist, particularly in up-
valley locations (e.g. upper Kennet and Blackwater on the Thames, or
the Salisbury Avon). In the Middle Thames below Windsor the inset
Kempton Park Gravel and buried Shepperton Gravel suggest an active
Table 2
Quaternary terrace dimensions in southern England.
Slope
(m/km)
Alluvial depth to contact
(m)
Bedrock contacts observed Bedrock type Max. width
(km)
Middle Thames (Gibbard, 1985, 1989)
Beaconsfield 0.39 7 m 10 7C, 3R 1.5
Gerard's Cross 0.46 5 m 20 10R, 10C 2.8
Winter Hill (lower unit) 0.66 6 m 17 12C, 4R, 1L 2.9
Black Park 0.48 5–9 m 17 12C, 1R, 5L 2.3
Boyn Hill (a) 0.39 4 m 11 6C, 2R, 3L 1.2
Boyn Hill (b) 0.27 4 m 11 2C, 7R, 2L
Lynch Hill (a) 0.51 3–7 m 11 4R, 7C 2.7
Lynch Hill (b) 0.44 4–8 m 28 4R, 24L
Taplow (a) 0.87 3–7 m 11 9C, 2R 5.2
Taplow (b) 0.49 3–8 m 29 4R, 25L
Kempton Park 0.31 7–9 m 22 22L 5.8
Shepperton(a) 0.32 4–10 m 37 3K, 4G, 30C
Shepperton (b) 0.39 8 m 36 7C, 4R, 25L
North Surrey, South Berkshire (Gibbard 1982) 0.8 18 6Ba, 5Bk, 6B, 1L
Blackwater, Lodden and Whitewater (Gibbard 1982) 1.3 12 10Ba, 2B
Solent River (Allen and Gibbard 1993)
(Below Wareham) 0.8
(Above Wareham) 1.4
Lower Severn (Maddy et al. 1995)
Wooldridge 2 m+
Spring Hill 0.6 3–4 m (7.2 max)
Bushley Green 0.4 6.9 m max
Kidderminster Station 0.5 5 m+ (7.9 max)
Holt Heath (Main) 10 m max
C, Chalk; R, Reading Beds; L, London Clay; G, Gault Clay; K, Kimmeridge Clay; Ba, Barton beds; and Bk, Bracklesham Beds. (a) and (b) refer to upper and lower reaches where there is a
gradient change.
297J. Lewin, P.L. Gibbard / Geomorphology 120 (2010) 293–311

Author's personal copy
channel-belt of up to around 4 km in width. The Taplow Gravel appears
to have covered an even wider migration floor in the vicinity of
Heathrow Airport.
By contrast, other terrace remnants are set within incising valley
meanders, as on the Warwickshire Avon (Fig. 6). Here different loops
appear to have been developed and sedimented at different stages,
with units of later ascribed date on the bends down-valley from
Pershore, whilst Maddy et al. (1999) have pointed to earlier loop
abandonment upstream. Many other smaller meandering-valley
systems with inset terraces exist in southern England which have
been less fully investigated (Tamar, Taw and Torridge in Cornwall and
Devon; the upper Avon system in Wiltshire; the Windrush and
Evenlode in Oxfordshire; and the Yare in Norfolk). Do meandering
valleys betoken meandering rivers? Holocene analogues may be
helpful. Chiverrell et al. (2010) demonstrated that over time channel
locations on the inside of a valley meander system oscillated across
depositional surfaces on the inside of major bends rather than
following a unidirectional shift from inner to outer locations within
sedimenting arcs. Active braid systems in meandering valleys may
also unify and deepen into single dominating channels against
confined valley margins, so that a constrained and concentrated
curving channel reach functions as a meandering one. Ingrowing
meandering valleys, with outer-bend curvature and inner-margin
sedimentation, are not necessarily indicative of an overall meandering
channel style as was once believed (Dury, 1964).
The size of preserved terrace fragments relative to catchment size
varies considerably. The Warwickshire Avon has broad spreads of
dissected gravel east of Stratford—upon Avon and where the tributary
Arrow joins the Avon from the north. These appear to be Wolstonian
glacial outwash materials (Old et al., 1991) which are being reworked
and funnelled through the incising lower Avon system. The Severn
was similarly affected by glacial outwash from a later-stage (Maddy
et al., 1995, 1999). The middle Thames has large quasi-rectangular
terrace blocks with straight to high-radius upper and lower margins,
suggesting a looping valley-floor at some stages but not at others
(Gibbard, 1985, 1989). The lower river has also has been greatly
modified by ice redirection following Anglian Stage glaciation (for a
summary, see Ehlers et al., 1990; Catt et al., 2006) to give a radial splay
of fluvial sediments, but the lateral extent of units is large in any case.
The gravel spreads of the unglaciated Hampshire Basin (Allen and
Gibbard, 1993; Westaway et al., 2006) are also extensive, and it seems
that the breadth of terracing in both Tertiary basins reflects relatively
easy erosion and lateral planation of the bedrocks involved. This
autogenically creates accomodation space at floodplain level for
sediment storage, the volume and spatial spread reflecting lateral
activity rather than deep vertical aggradation. This contrasts with the
narrower floors of incised Chalk valleys, as on Salisbury Plain and
where the Thames traverses the Chiltern Chalk outcrop.
Terraces have been globally matched to Marine Isotope Stages (MIS)
(Bridgland and Westaway, 2008); the Solent River seems to have more
benches than known stages. This is of course true elsewhere —on the
middle Somme and Seine (Antoine et al., 2000 Fig. 11; Antoine et al.,
2007) or even Kukla's classic sections on the Svratka where three rock-
cut benches underlie Weichselian deposits (Kukla, 1975). Antoine et al.
(2007) note some situations where there are a few large steps, some
where there are many smaller ones (c. 10 in the last 1 Ma), and one
where there is a lack of incision between two glacial–interglacial
sequences. This should not be unexpected if bedrock-cutting and
aggradations are to an extent considered independently. Lateral
planation accompanying downcutting may not reach the valley-side
every time so that successive bedrock-cutting sweeps become inset.
Alternatively, both erosional and aggradational sweeps may exceed the
lateral extent and depth of earlier ones and so eliminate some benches
or bury several. Whilst surfaces may in general match the rock steps
beneath them, they can also be fewer or greater in number. This draws
attention to the need for examination of both bedrock and surface
morphology. Possibly also the multiple terraces of the Solent relate in
part to unidirectional shifting of the alluvial system southwards, or to
accelerated incision related to the marine dismemberment of the
system. It should be stressed that no dates for the higher alluvial
deposits of the Solent have yet been obtained. In many cases incision
appears to follow a pre-existing substrate slope–fluvial sediment
contact, presumably for hydrological reasons. Such behaviour, if
repeated, gives rise to an off-stepping sequence in which pairing of
terrace remnants are the exception. Moreover, it can in general provide
a step-like series of surfaces which might be interpreted as indicating
tectonically driven uniclinal shifting (for other interpretations, see
Westaway et al., 2006).
Overall, the spatial distribution and extent of terraces suggests
laterally-mobile channel-belts, some becoming inset into developing
valley meanders (the Warwickshire Avon), others showing linear
shelves and unidirectional valley-floor shift (upper Thames, Solent
River). The spatial extent of terraces broadly reflects river size, valley
relief and especially bedrock erodibility. A reservoir of glacigenic
material was in some situations available for reworking through
fluvial systems, and elsewhere gravels from headwater sources cap
Fig. 4. Terrace gravel members in the Wareham–Dorchester area on the Rivers Frome and Piddle (from Allen and Gibbard, 1993).
298 J. Lewin, P.L. Gibbard / Geomorphology 120 (2010) 293–311

Author's personal copy
clay vale bedrock. Whilst graphic evolutionary models of terracing can
imply a general and progressive narrowing of valley floors over time,
this is not necessarily the case in practice when whole alluvial belts
shift sideways in location during incision. Where pairing does exist, as
on the Exe outside glacial limits, Brown et al. (2009) suggested that a
small volume of sediments was involved overall, and that this was
reworked from one terrace level to another as incision took place.
Consumption of upper terraces as lower ones are formed, means that
their preserved spatial extent is based both on formational and
destructional processes (Lewin and Macklin, 2003).
Reconstructing former channel styles is also not straightforward.
The gravel sediments have generally been taken to indicate braided
rivers (see below), and this may be true even of units inset within
meandering valleys. On the other hand, arcuate terrace-front
excisions could result from large-scale meandering at the bench
cutting stage, or a focusing of multiple channels into a dominating
single one at undercut bends. There can be a down-valley style
progression from braiding to meandering, with tributary-related
pattern switches associated with sediment input, as is observed on
contemporary river systems (Table 1). Which style was most effective
for bedrock incision is not fully clear; meandering incision under
lower sediment throughput favoring rock exposure followed by, or
locally interrupted by, braiding aggradation is a possibility not
hitherto recognised.
Fig. 5. Terrace gravel members in the Middle Thames (from Gibbard, 1989 following Hare 1947).
299J. Lewin, P.L. Gibbard / Geomorphology 120 (2010) 293–311

Author's personal copy
4. Basal surfaces, bottom sediments and “working depth”
alluvial units
According to Castleden (1977, 1980), the basal rock surfaces he
observed, primarily beneath the gravels of the Nene valley-floor, are
either horizontal, slightly concave, cut by channels or scalloped. He
called these “fluvioperiglacial pediments”,or“gallery pediments”
when there were sets at higher levels. These might or might not relate
to surface form. The terms are not ones generally now in use, but the
features identified are clearly the same as generally occuring beneath
cold-climate terrace sediments in England and elsewhere. Many
terraces, like the rivers they border, are developed in clay vales, but in
practice they may also be found on a considerable range in lithologies
(Table 2) which are likely to be differentially susceptible to planation
processes. Chalk in cold climates, for example, is reducable to a paste
by interstitial ice (Lautridou et al., 1986), leaving flint bands more or
less intact, as may be seen in coastal cliff sections cutting across
valleys east of Brighton.
In detail and according to lithology, terraced bedrock commonly
shows near-surface rock shattering by ice, half-detached blocks, and a
lag of coarse but only partly abraded particles in the process of
incorporation into basal sediments (cf. Antoine, 1994; Van den Berg
and van Hoof, 2001). Large residual boulders may be involved (Collins
et al., 1996). Bedrock surface relief can also be complex (Fig. 7).
Excavation of the London Clay surface beneath gravels at Stoke
Newington showed small-scale features in the form of a set of slightly
elongate scour pits c. 10–50 cm deep and 40–100 cm across (Harding
and Gibbard, 1981). At a larger scale, surface and bedrock contours at
Nightingale Estate at Hackney Downs, north London (Gibbard, 1994)
are only broadly related. Collins et al. (1996) reported a large elongate
subgravel depression (c 1.0 km×300 m) up to around 10 m deep
which they attributed either to confluence scour or subsidence
associated with Chalk dissolution. Berry (1979) reported numerous
“channel-like”features cut into London Clay beneath gravels in
central London. Some are associated with tributary junctions, whilst
an exceptional one at Battersea Power Station meandered with
excavated elongate scours on outer bends up to 20 m deep. What
appear to be large valley-margin scour features of the same order are
found in the rockhead surface beneath the Loire at Tours (Bernouf
et al., 2003). In even greater detail, GPR profiling beneath the
Devensian (Weichselian) Holme Pierrepont Sand and Gravel at the
confluence of the Trent and Soar shows a highly uneven bedrock
surface (Howard et al., 2007). It is clear that sub-gravel bedrock
surfaces can be far from planar, with deformation, enclosed scour
hollows, linear features and detailed fretting at a range of scales in
addition to the general forms noted by Castleden (1977, 1980). Many
appear to result from fluvial scour, but post-depositional solution
subsidence and diapiric activity in clays best explain some bulges and
depressions.
Exposed valley-floor bedrock is distinctively vulnerable to erosion
under periglacial conditions. Groundwater saturation allows maxi-
mum bedrock brecciation by ice segregation, whilst the surface
seasonal thawing from the thermal effect of flowing water promotes
the detachment and transport of this comminuted material. But the
process is a complex and perhaps dual one (Murton, 1996; Murton
et al., 2006); field evidence suggests both shattering by ice
segregation deep within permafrost, and annually repeated freeze–
thaw processes within the shallow active layer. In either case
incremental removal by seasonal melt discharges would be facilitated,
as would lateral bank erosion in thawing ice-rich material (Church
and Miles, 1982). Both shattering and seasonal removal may occur in
wetter valley-floor environments under relatively arid periglacial
conditions, conditions under which slope solifluction would also not
deliver a blanketing valley-floor carpet of alluvial material.
Gravel thickness on most bedrock terraces in south-east England is
not generally large (Table 2), and proportionately no greater than the
7 m reported from the active Savanirktok system in Alaska (Lunt and
Bridge, 2004). Borehole data record little more than the “working
depth”of active channel-belt systems. At some sites where detailed
investigations have been undertaken, upper levels have been shown
to have been reworked over an extended time period (e.g. Maddy
et al., 1998; Lewis et al., 2001; Gao et al., 2007; Langford et al., 2007),
but for bedrock scour to have been effective rock exposure with only
transitory alluvial burial was likely. Solifluction inputs are a major
source of river sediment in periglacial climates, and these appear to be
favoured under moist and cold conditions rather than arid ones
(Matsuoka, 2001). Conversely, relatively arid periglacial environ-
ments are still subject to nival flow regimes, with seasonal rivers
flowing over exposed and weakened bedrock being less laden with
solifluction products. On this basis, valley-floor bedrock planation
appears to have been most effective when slope sediment input was
less so, though basal sediment adjacent to the rock surface and the
surfaces themselves suggest that this was achieved in a cold climate.
Overall, we view mid-catchment terraces as essentially cold-climate
straths (Fig. 1c) with limited sediment cover, but with later reworking
within a small vertical range at some sites.
Fig. 6. Terraces of the Warwickshire Avon near Evesham (based on British Geological Survey). The Survey numbers terraces from youngest to oldest; names for gravel members
underlying these surfaces are from Maddy et al. (1999).
300 J. Lewin, P.L. Gibbard / Geomorphology 120 (2010) 293–311

Author's personal copy
5. Sedimentation
Lowland southern British river terrace sediments predominantly
consist of waterlain, sorted gravel and sand occurring in suites of
sedimentary structures that indicate deposition under broadly similar
fluvial conditions. Individual members do show minor deviations, but
facies can be discussed together and interpreted using facies models,
such as those developed by Miall (1977, 1978, 1996).
The most commonly occurring facies in lowland Britain is massive
or horizontally crudely-bedded gravel (using Miall's terminology,
facies Gm) (Fig. 8a). Parallel to the valley axis, the sedimentary bodies
are elongated in the downstream direction. This phenomenon has
frequently been observed from gravel-bed rivers (Miall, 1977, 1996;
Bryant, 1983b; Briant et al., 2008) and apparently results from
downstream extension of the sediment body. The clasts have a great
size range, their form determined by local source lithologies, but the
largest are most frequently of medium to fine gravel (1–30 cm).
Larger clasts occur and may reach over 50 cm, but these are rare. The
gravels are generally massive, but often show poorly developed
horizontal bedding (Gao et al., 2007); individual clasts are only rarely
imbricated. This probably reflects the relative paucity of blade
and rod-shaped pebbles, the irregular shape of flint and the high
Fig. 7. Contoured sub-alluvial bedrock surfaces at different scales: (a) on London Clay, TQ 339866 (from Harding and Gibbard, 1981), (b)on London Clay, TQ 345860 (from Gibbard
1994), and (c) on Tertiary sediments overlying Chalk, SU 566664 (from Collins et al., 1996).
301J. Lewin, P.L. Gibbard / Geomorphology 120 (2010) 293–311

Author's personal copy
sphericity of vein quartz and quartzite clasts (cf. Bluck, 1967). The
gravel is generally matrix-supported (Gmm) and matrix-rich, but
clast-supported gravel (Gmc) is common, particularly in the coarser
particle sizes. Interstices are filled with silt, sand and granules.
Secondary clay may also be present. Individual gravel units are
typically up to c. 90–125 cm thick but are frequently superimposed to
produce multi-storey units of greater thickness. Laterally the gravel
units may be persistent and can be traced across exposures for over
30 m. They usually rest on an erosional base and often occupy
channels cut into the underlying substrate.
Broad, shallow channel scours are frequently cut into the gravel
facies (Fig. 8b). These are filled by interbedded parallel cross-bedded
gravel and sand (facies Gp and Sp). The “channels”(possibly the
truncated lower levels of formerly deeper features) and scour features
range up to 15 m in width to 1.5 m in depth, but are often much
smaller (Maddy et al., 1998; Lewis et al., 2001; Gao et al., 2007).
Individual foresets comprise a complex of pebbly sand, sand and
gravel, the grain size showing considerable variation. Individual units,
often elongated downstream for 20–30 m, may be as much as 2–3m
in thickness. The base of these deposits is again erosional. The coarsest
clasts frequently form a “lag”at the base. Major channels are formed
by lateral channel switching and migration at high water stage or bar
dissection during falling water conditions (cf. Williams and Rust,
1969; Miall, 1977). These channels, particularly the larger examples,
show multiple infills of trough bedded gravel, trough pebbly and
scour-fill sands (facies Gt, St and Ss). Decreasing upward dip in foreset
angle is often observed. Reactivation surfaces (cf. Collinson, 1970) are
common. There is a tendency towards upward-fining channel fills, but
this trend is seldom completed, since the sequences are often
truncated by a reactivation surface and overlain by second or later-
generation basal sediments.
Trough cross-bedded pebbly sand (facies Gt and St) occurs at all
levels in the deposits normally filling channel-like features cutting
across earlier sediments and alternating with gravel facies Gm. The
“channels”are of similar proportions in cross-section to those
mentioned above, but individual units, often elongated downstream
for 20–30 m, may be as much as 2–3 m in thickness. The base of these
deposits is again erosional.
Fig. 8. (a) Irregularly bedded gravel and sands at Westmill Quarry, Hertfordshire (TL 342162), (b) a basal “channel”in gravels at Block Fen, Cambridgeshire (TL 427838), (c) ice-
wedge casts at Block Fen, Cambridgeshire (TL 427838), (d) involutions at the base of soliflucted diamicton (partially “brickearth”) at South Stifford, Essex (TQ 593794), (e) irregular
chalk–gravel interface effected by bedrock solution at Leavesdaen Green, Hertfordshire (TQ 097996), and (f) grossly disturbed fluvial and associated sediments resulting from
solutional collapse at Chalfont Lane, Hertfordshire (TQ 017917).
302 J. Lewin, P.L. Gibbard / Geomorphology 120 (2010) 293–311

Author's personal copy
Of sand facies, only two types are well represented. The most
common is the scour-fill sand (facies Ss), which consists of scours,
often asymmetric in form, ranging from a few centimetres to about
50 cm and up to about 1.5 m in width. The sediment lenses often
occur within or resting upon bar-facies gravel (facies Gm) and
comprise pebbly sand. The bedding of the sand is frequently parallel
to that of the scour surface. The only other sand facies present is
horizontally bedded, often massive sand (facies Sh). These beds are
normally of very local extent and up to 20 cm in thickness.
Fine-grained sediments occur very infrequently. Only laminated
and massive fines (facies Fl, and occasionally Fm), are represented by
beds of silty clay, clayey silt and organic sediment, the latter
apparently restricted to the lower gravel members. This may reflect
a local mixed input of sediment sizes, rather than the sorting of long-
distant transport of bedload. The deposits fill channels or floodplain
depressions that have been scoured into underlying sediments or
remained as unfilled hollows. They occasionally represent the
completion of a vertically accreted fining-upward sediment sequence,
beginning as facies Sr (ripple-bedded fine to medium sand) and
ending in fine sediment. The fine facies are often grey to brown in
colour and contain interbedded sand and even gravel bands. Organic
remains, where present, almost invariably include both autochtho-
nous aquatic plant and animal fossils and also allochthonous material
washed-in during flood events from vegetated stable tracts and
neighbouring slopes (Rust, 1972).
Large-scale planar erosion or bounding surfaces (cf. Miall, 1996)
are a common feature of the gravel members, as noted by Bryant
(1983b) in the Upper Thames Valley. Erosional surfaces almost
invariably underlie medium to coarse gravel accumulations (facies
Gm and Gt) and in any one exposure may be present at various levels.
They may be very persistent and are often traceable almost
continuously across a quarry face. Where periglacial ice-wedge casts
have been found they often occur beneath such surfaces and indeed
may be truncated by them. It is noteworthy that these surfaces are
particularly well developed where the contemporaneous valley was
relatively wide, for example in the Heathrow area in the Middle
Thames.
The main mass of gravels (facies Gm) seems to represent
accumulation and migration of low-amplitude longitudinal bars. The
gravel was probably laid down in horizontal sheets, but stratification
is often obscured by accumulation of finer particles in open
interstices. Deposition of such units occurs only at peak discharges
(Hein and Walker, 1977). Interbedded sand lenses probably resulted
from flow in secondary channels over bar surfaces during low-energy
water flow.
Facies St and Sh result from small-bedform migration. The
migration of large-scale ripples or dunes either singly or in groups
gives rise to trough cross-bedded sand (facies St). Sedimentation by
sand waves, of similar proportions to dunes but with straighter crests
produces planar tabular cross-bedding (facies Sp). Such bedforms are
produced during upper flow regime flow for which there is ample
supporting experimental evidence (Friend and Moody-Stuart, 1972;
Bridge, 2003). Horizontal bedding (facies Sh) may be formed under
lower flow regime in shallow water (Harms et al., 1975)oratflood
stage (Harms and Fahnestock, 1965). In the latter case the bed is
formed by streaming of particles across the surface under upper flow
regime conditions.
During low-water periods only small-scale bedforms are produced
and these are generally restricted to the infilling of minor channels
and hollows on bar surfaces by ripple-laminated sand (facies Sr), or in
shallow water by horizontally bedded sand (facies Sh). Sand wedges
may develop during falling water stages at bar margins. Such wedges
normally show planar cross-bedding. Silt, clay and organic material
(facies Fl or Fm) accumulate in abandoned or partially-abandoned
flow channels in standing or trickling water, especially in topograph-
ically higher areas of the braid plain (Williams and Rust, 1969; Rust,
1972). In spite of their apparently low preservation potential (cf. Cant,
1976), these beds are occasionally preserved (e.g. Briant et al., 2008).
Fine drape-like laminations also occasionally occur in bar-tail facies
sediments where they probably represent deposition during the
waning phase of nival-flood events (e.g. West et al., 1999).
Braided rivers commonly only occupy part of the valley bottom at
any particular time and large areas of the braid plain may become
temporarily stable and colonised by vegetation (Williams and Rust,
1969; Rust, 1972; Miall, 1977, 1996; Rust and Koster, 1984; Reinfelds
and Nanson, 1993). During exposure these surfaces are subjected to
subaerial weathering, and ice-wedge polygons and cryoturbation may
develop in favourable localities, particularly on higher stable tracts
(Washburn, 1968; Bryant, 1983b). Renewed channel migration will
degrade such a surface, terminate subaerial weathering and result in
talik development leading to degradation of ground-ice phenomena.
Infill and truncation of ice-wedge casts together with disturbance,
truncation and burial of valley-side mass-flow deposits are found
(Fig. 8c and d),together with abandoned channel fills. Extensive
planar erosional surfaces have been attributed by Bryant (1983b) to
truncation of fluvially inactive areas, raised slightly above the active
areas of the river by large-scale channel migration. The association of
truncated infilled ice-wedge casts and other periglacial phenomena,
channel fills and earlier sediments beneath such surfaces supports this
interpretation as a normal feature of cold-climate autogenic braid
plain reworking.
In summary, actively-sedimenting lowland British cold-period
rivers were dominated by matrix-rich gravel accumulation; rivers
adopted a wandering or braided mode with a considerable degree of
reworking as indicated by multiple truncation surfaces and the
preferential preservation of lower members of autogenic depositional
suites. Sediment was supplied by slope processes, predominantly
solifluction, and bedrock scour. Potential for the latter was limited
where thick sediments accumulated. On partially abandoned or
slightly elevated braid-plain areas, where the river was less active,
shallow pools and depressions accumulated fine-grained and occa-
sionally organic material. The latter was derived from vegetation that
grew on the floodplain or immediately adjacent to it.
The channels of these river systems were unstable, shifting and
ephemeral forms, with shallow cross-sections and relatively straight
courses. However, in many instances parts of the braid plains
remained abandoned for significant lengths of time as indicated by
the development of abundant, large ice-wedge cast networks in the
sediment bodies (West, 1993). These ice-wedge systems generally
developed where the contemporaneous valleys were relatively wide,
where floodplain widths of several kilometres occurred, e.g. the
Fenland. Conversely, upstream in the same river systems where the
valley width is confined by bedrock slopes to less than 1 km (e.g. the
Great Ouse and the Nene), ice-wedge casts are relatively rare. Slopes
must have been subject to significant periglacial degradation to
provide the river systems with consistently large amounts of material
(van Huissteden et al., 2001).
The assemblage of sedimentary facies and structures present in
lowland British terrace closely approximates to the Donjek or Scott-
type depositional models proposed by Miall (1977, 1978),the
Donjeck currently incising in the middle reaches of the river (Area 2
of Rust, 1978a; facies model G
III
of Rust, 1978b), or those more
recently described on the Sagavanirktok in Alaska (Lunt and Bridge,
2004). Features suggest that the bulk of deposits accumulated in
braided-river environments (e.g. Corner, 1975; Bryant, 1983a, b;
Dawson, 1985; Gibbard, 1985, 1994; Dawson and Bryant, 1987; Allen
and Gibbard, 1993; Maddy et al., 1998; Lewis et al., 2001; Briant et al.,
2008). The consistent association with climatic indicators such as ice-
wedge casts, solifluction deposits, frozen block transport and
cryoturbation structures suggest deposition under a predominantly
periglacial regime. This is strongly supported by the facies present
which have been shown to typify nival-flood-dominated rivers in the
303J. Lewin, P.L. Gibbard / Geomorphology 120 (2010) 293–311

Author's personal copy
modern Arctic (Bryant, 1983a; Woo, 1990). Temporal fluctuations
were probably both seasonal (with high spring discharges resulting
from snow-melt with markedly reduced winter discharges, coupled
with flooding following storms in summer), and in the longer-term,
with less-marked seasonal variation and lower discharges during
minor climatic ameliorations (interstadial times: cf. below) (Church,
1974; van Huissteden et al., 2001).
6. Extra-fluvial additions
Adjacent to contemporaneous valley sides, lenticular wedge- or
tongue-shaped bodies of pebbly fine-grained (silt or clay) or “rubble
chalk”diamictons are often found both interstratified with and
overlying fluvial sediments. These diamictons vary considerably in
thickness from a few centimetres to over 2 m and have erosional bases
often showing incorporation of clasts or sediment from beneath. The
overlying fluvial sediments always show marked erosional contact
with these diamictons. The pebbly diamictons are usually massive,
with a fine-grained matrix, but the sediment is often poorly mixed
and silt laminae or sand stringers may occur. The latter rise down the
dip of the deposit, may be distorted around pebbles and may show
flow-type banding parallel to the lower bounding surface. Down-
slope clast orientation is predominant (cf. Watson, 1969; Gibbard
1985, 1994).
What was earlier called “rubble chalk”or “coombe rock”(Reid,
1887) beds comprise angular chalk fragments and flints in a putty-like
matrix of mechanically fragmented chalk. In the unmodified state this
material may be very difficult to distinguish from the upper few
metres of blocky bedrock chalk in situ. However, the material shows
considerable variability and in extreme cases, following from
incorporation of underlying material and decalcification, may be
difficult to separate from poorly sorted, massive fluvial gravel. This
sediment is well known throughout southern England, and especially
in the Middle and Lower Thames valley where it cloaks bedrock chalk
slopes (Gibbard, 1985, 1994). The deposition of this material was even
referred to a specific time period by King and Oakley (1936), but
judging from its relationship to the fluvial sequences, it has formed
repeatedly during cold periods. Slope instability can also produce
temperate diamictons which may cap earlier cold-stage materials on
steep slopes.
More generally, these pebbly clay or “rubble chalk”diamictons
were deposited by mass-flow or solifluction rather than fluvial
processes (McGregor and Green, 1983; Gibbard 1985, 1994). Their
markedly local clast content, together with the abundance of highly
angular, frost-shattered pebbles indicates a local source subject to
cold-climate weathering. The mixture of these pebbles in a non-
durable, unsorted matrix strongly suggests intermixing of material
from the immediate vicinity. Frost-dominated weathering of bedrock
under a cold-climate provides unstable slopes and down-slope flow of
water-saturated sediment would be expected (cf. Galloway, 1961),
particularly following spring-melt (cf. Hutchinson, 1991; Vanden-
berghe, 2008). Valley-side solifluction inputs are a major source of
river sediment in periglacial climates, and these appear to be favoured
under moist and cold conditions rather than arid ones (Matsuoka,
2001).
In southern Britain virtually all terrace surfaces, i.e. abandoned
channel-belt floodplains, have undergone post-depositional modifi-
cation either by later deposition, soil formation or by periglacial
processes. Acid gravels overlying calcareous bedrock lead to dissolu-
tion particularly above the water table since below it waters may be
neutralized by high CaCO
3
content, and then subsidence (Matthews
et al., 2000). The gravels of older and higher terraces in particular are
turbated as well (Fig. 8e, f). Valley-side derived surface wash, loess
and soliflucted diamicton material may bury alluvial surfaces. Where
this sedimentation is dominated by colluvium, a transverse concave
profile is commonly observed, rather than a horizontal surface. This
form is common in the Thames system (Hare, 1947; Gibbard, 1994). In
comparison to fluvial terrace depositional successions in the near-
Continent (as in the Rhine, Seine and Somme), British sequences
generally lack thick multi-phase loess accumulations. The deposits are
however, often overlain by fine-grained partially water redeposited
sediment loosely termed “brickearth”, or coarser-grained cover sand,
and this can be metres thick (Rose et al., 2000).
“Brickearth”is especially common in, although not restricted to,
the Thames Valley region (Gibbard, 1985, 1994; Gibbard and Preece,
1999). The term was applied extensively in the region to any fine-
grained deposit that was suitable for brick-making and as a result it is
very variable in character (cf. Bromehead, 1925). Throughout
southern Britain it varies from a sandy silt to remobilised bedrock
clay. Although there has been much debate over its origin, it is now
generally accepted that the most frequent type, clayey silts, were
formed as a combined loess and waterlain or colluvial deposit.
However, it is important to note that this material differs in origin
from the laminated or massive fossiliferous “brickearths”which are of
estuarine origin (cf. Hollin, 1977; Gibbard, 1994). In the Thames
system the silt component is most commonly of loessic origin
(Gibbard, 1985; Catt et al., 2006). However, the predominant bimodal
grain size distribution implies that the sediments are not solely loessic
but were modified by colluvial processes. The latter is particularly
common at valley-side situations. As Gibbard (1985, 1994) demon-
strated, because of its polygenetic origin, it is found in apparently
conflicting stratigraphical contexts, i.e. blanketing fluvial gravel units
of differing age. It was earlier thought by some writers to be
“intimately associated”(Sherlock and Noble, 1922)or“nearly
contemporaneous”(Dines and Edmunds, 1925) with the gravel on
which it rested by analogy with modern floodplain environments.
However, since the “brickearth”is generally cold-climate in origin, it
is differently sourced to modern alluvium. Clearly the term “brick-
earth”is best avoided as a lithostratigraphic term in light of the suite
of depositional or redepositional processes with which it has been
involved.
The results of analyses for heavy mineral and thermoluminescence
(TL) dating obtained by Gibbard et al. (1987) grouped around c.
17,000 BP for the main mass of silt-rich sediment in the Thames
Valley. More variable ages are found from the colluvial sediments (e.g.
over 140,000 BP at Yiewsley: Wintle, in Gibbard 1994, 1999).
Cover sand is also common, especially in the Breckland region of
Norfolk and Suffolk. Here wind action has caused recycling and
deposition of sands on fluvial sediment sequences as in the
“brickearths”. Both these sediment types mask fluvial sequences and
may often represent several subsequent cold-stage colluvial and loess
episodes. To date relatively little attention has been paid to these
accumulations in comparison to the fluvial and other sediments on
which they rest.
Of equal importance to the long-term preservation of fluvial
sediments is degradation of the exposed surfaces by periglacial slope
processes, stream erosion, gullying and bedrock cambering and
solutional subsidence (Higginbottom and Fookes, 1970; Ballantyne
and Harris, 1994). One or a combination of these processes results in
progressive degradation in terrace surfaces together with their over-
and underlying deposits with time (Gibbard, 1985, 1994). The net
result may be that, in general terms, the older the terrace remnant, the
more likely it will be that remains of its original uppermost
depositional surface will be modified, or in the older examples, lost
completely. In addition, with increased age the effect of deposition
from groundwater precipitation of oxides of iron and manganese may
lead to progressive colouration and in some instances, especially
where the aquifer is confined by overlying impermeable sediments,
and to cementation forming calcretes or ferricretes. The latter are rare
in southern England, however.
Following abandonment and incision, terrrace remnants may be
subjected to erosion and destruction especially focused at the valley-
304 J. Lewin, P.L. Gibbard / Geomorphology 120 (2010) 293–311

Author's personal copy
side slope. On certain substrate lithologies, especially clay bedrock,
groundwater escape gives rise to spring-sapping at the sediment
base–substrate contact. This can lead to valley-side cambering where
blocks or masses of sediment slide downhill under gravity. Such
phenomena have been observed in the Midlands, for example
(Hutchinson, 1991). It is generally assumed that this process occurs
under a periglacial regime, since slopes are generally stabilised by a
dense vegetation cover during temperate periods.
7. Terminations
The height differences between terrace levels, representing the
degree of river incision achieved, has been taken directly as a measure
of uplift (Maddy, 1997; Maddy and Bridgland, 2000; Maddy et al.,
2000). The data used in the references cited have been plotted in
Fig. 9, omitting only projected data points, and including height data
for Sussex raised beaches for comparison. These are interglacial and
relate to high sea levels, but age–height relationships suggest an
approximately equivalent rate of relative sea-level fall. However this
area is believed to have been affected by tectonic activity on the
Portsdown Anticline (Westaway et al., 2006). Average rates of valley
incision are nonetheless broadly similar, and quite small, at around
0.07–0.14 m ka
−1
, with the exception of post-Anglian incision in the
Middle Thames (1.45 m ka
−1
) which Maddy and Bridgland (2000)
attribute to glacio-isostatic rebound. In England, and following Maddy
(1997), uplift has been advocated essentially as a driver for longer-
term terracing which is only believed to occur in areas of uplift. These
concepts have been applied globally (Bridgland and Westaway, 2008),
with the further proposal that the broad epeirogenic uplift suggested
by a linkage with river incision rates could derive from mobility and
material flow in the lower crust (Westaway et al., 2002). Presumably
ongoing erosional unloading would be necessary to trigger continuing
compensatory flow to achieve progressive uplift (rather than the
cyclical effects of direct glacial, forebulge or shelf-water loading which
would reverse in direction). Uplift timing has also been modelled (e.g.
Westaway et al., 2006, Fig. 11; Bridgland and Westaway, 2008) with
high rates equated directly to more rapid incision between some
terrace levels. A different approach has been applied on a regional
scale in England in which denudational isostasy responding to
differential erosion in scarp and vale topography has been modelled
in terms of erosional unloading (Watts et al., 2000, 2005; Lane et al.,
2008a, b). Lane et al. (2008a, b) suggest response times of b50 ka, and
a spatial scale of c. 50 km. Around 50% of Cotswold relief is attributed
to denudational isostasy arising from the preferential removal of
erodible materials from the adjacent Vales of Gloucester and Evesham
in the later Pleistocene. This approach is not dependent on, nor are the
results compatible with, uplift rates derived from terrace levels.
In the long-term, relief is required for incision, whether created by
regional tectonic activity, broad-scale epeirogenic uplift, and isostatic
or eustatic mechanisms operating over a variety of timescales. Terrace
development can be involved with and be related to uplift in areas of
crustal mobility (e.g., Bull, 1991, 2007; van Balen et al., 2000, Pazzaglia
and Brandon, 2001; Antoine et al., 2007; Peters and van Balen, 2007),
but in the short term (i.e., within an glacial–interglacial cycle), the
height difference between terraces represents also the pace and
duration of geomorphological activity episodes. At this scale, uplift
rates cannot be easily separated from the pace of geomorphological
change exploiting the potential energy of prior relief creation, nor can
they be derived directly from incision rates (cf. Pazzaglia and Brandon,
2001). Downcutting varies with process regime and is variable within
catchments, for example being greater in higher energy zones or
lower bedrock resistance. Higher rates may be expected at steeper
gradients upstream (but dependent also on bedrock and discharge-
related stream power) or at and below headward-receding knick
points or zones related to base-level change or uplift.
In cold-stage England, rivers continued to transport gravels and
erode bedrock platforms at very low gradients, being effective at
lower gradients than has been the case in a Holocene interglacial
climate and run-off regime. Profiles that were “graded”in appearance
do not appear to have precluded ongoing vertical erosion at the time
(cf. Snow and Slingerland, 1987). It is true that some authorities take
profiles and terraces to be “equilibrium reference surfaces”created “as
the lowest possible longitudinal profile for a particular tectonic
setting”by lateral erosion at the end of an uplift episode (Bull, 2007,
pp.44–45), or that they represent a balance between uplift and
incision (pp. 47–48). But these are entirely theoretical (if convenient)
assumptions. In Pleistocene lowland England, we do not see evidence
for gradients being process-terminal, but rather that they were ones
of ongoing dynamic activity, with a potential trajectory for continuing
channel erosion as well as sediment throughput —but only when
environmental conditions involving stream energy and sediment
supply rates continued to be appropriate.
Fig. 9. Height–age diagram for terrace locations in southern Britain based on data from Maddy (1997), Maddy et al. (2000) and Maddy and Bridgland (2000).
305J. Lewin, P.L. Gibbard / Geomorphology 120 (2010) 293–311

Author's personal copy
We suggest that bedrock planation and channel incision were
linked and maximised under cold climates when the volume of
catchment-wide slope degradation and sediment throughput was
fairly low. Within a glacial–interglacial cycle, this was followed both
by later cold-stage aggradation and by a set of interglacial phases
(Gibbard and Lewin, 2002). Although there was some paraglacial
reworking of sediment in cold–warm “transitions”, the termination of
fluvial activity at a particular topographic level was accomplished by
renewed bedrock incision and widening planation in the early part of
the next cold-stage. This, of course, has yet to happen to the valley
floors of the last (Devensian, Weichselian) cold-stage.
8. Deposition through an interglacial–glacial cycle: the
late Pleistocene
The fluvial evolution and sedimentary sequence through the last
cold-stage period (Devensian, Weichselian, MIS 5d-2) has been
summarised by van Huissteden et al. (2001), Gibbard and Lewin
(2002) and Briant et al. (2008). In lowland regions fine, inorganic
sedimentation initiated late in the last interglacial (Eemian, Ipswi-
chian), continued into the early Devensian (early Weichselian, or c.
MIS 5), and ultimately filled the valleys to a depth of several metres,
e.g. at Histon Road, in Cambridge (Boreham, 2002; Gibbard and Lewin,
2002). These “late interglacial–earliest glacial”sediments are poorly
fossiliferous because of their increased inorganic component that
arose as a consequence of progressively deteriorating climatic
conditions. Higher discharges and sediment loads involving increased
channel width and lateral erosion–accretion along the course of
previously quiescent channels removed much of the evidence for
previous channel activity. Infilling and widening of valley floors led
increasingly to sheet-like vertical accretion of deposits particularly
during flood events. The relatively fine-grained character of these
early-glacial deposits implies that they have a low preservation
potential and are therefore rapidly removed in all but protected places
during subsequent events.
The re-activation of valley incision and gravel transport, but
lacking the full loading of glacial inputs, resulted in the removal of
pre-existing fine alluvium. It is as yet not clear precisely when gravel
transport was re-activated. However, deposition of early Weichselian
(=late M I Substages 5d-a, c. 90 ka BP) gravel and sand sequences, the
sedimentary facies of which compare closely to braided-river regimes
and therefore are of cold-climate origin, may have begun as early as c.
100 ka BP, e.g. the Summertown–Radley Upper Member in the Upper
Thames Valley and its downstream equivalents, the Reading Town
and West Thurrock members (Gibbard, 1985, 1989, 1994, 1999), and
Deeping St. James (Briant et al., 2004). If, in fact, some of these
deposits are of earlier date, then deposition in this substage appears
even rarer. There are some interstadial sediments of early Devensian
(late MIS 5) age, such as at Wretton in Norfolk (West et al., 1974;
West, 1977) or Brimpton, Berkshire (Bryant 1983b), but sequences of
this age are rare in valley systems. Their limited occurrence probably
reflects the intensity of fluvial incision.
The Devensian–early Pleniglacial (MIS 4; c. 75–58 ka BP: Lowe and
Walker, 1997) was mainly a time of non-aggradation throughout the
whole region. Apart from the broadly dated (58–148 ka BP) Cassing-
ton site in Oxfordshire (Maddy et al., 1998), none has yet been found
to correlate with MIS 4 in southern Britain and thus to fill an apparent
absence of sites equivalent to continental early Pleniglacial inter-
stadials from Britain.
Throughout Britain, there is a notable absence of Devensian
deposits dated to older than 45 ka BP (van Huissteden et al. 2001). In
recent years this observation has prompted some workers to
postulate a possible problem arising from the dating methods used.
Until recently these dates were overwhelmingly determined using
radiocarbon, the latter being close to its dating limit at ages exceeding
40 ka BP. The question therefore is whether these numbers are real or
whether they represent an artefact of the dating method. To
investigate this problem, duplicate dating of samples using Optically
Stimulated Luminescence have been undertaken, the results provid-
ing in some cases confirmation of the
14
C determination and in others
providing dates that are considerably older (Briant, 2002; Briant and
Bateman, 2009). Although resolution of this mis-match is significant
to the discussions presented here, the detail is beyond the scope of
this work. Suffice to say that, at this stage, further investigation is
required to resolve the dilemma these datings pose. Until that
happens, the conventional
14
C ages are used here.
On the basis of dates currently available, the next major depo-
sitional phase apparently occurred during the middle Devensian
Substage (middle Pleniglacial, MIS 3). Gravel-bed braided rivers, with
abundant coarse sediment supply from drainage basin slopes by
solifluction dominated in river valleys in the higher-relief areas
underlain by harder substrates. In the British fluvial successions,
permafrost may have been continuously present. As noted by van
Huissteden et al. (2001), both ocean sediment and ice-core records
indicate that several temperature oscillations occurred during MIS 3,
yet these warm-events are seldom found in coarse fluvial sequences
in southern Britain as well as on the near-Continent.
This observation implies that for the rivers to change their form
and depositional patterns to a recognizable degree, any particular
climatic oscillation must be of significant magnitude both in duration
and temperature. For example, the fluvial response to the Upton
Warren Interstadial (the acme conventionally dated to c. 43 ka BP)
and its equivalents was a reduction in seasonal flood intensity,
possibly with flow being distributed more evenly throughout the
summer months. This presumably resulted from increased vegetation
productivity, accompanied by increased or complete surface plant
cover; these factors together would have reduced surface run-off and
increased attendant slope stability, in turn restricting the supply of
coarser debris. This initiated vertical accretion in response to flooding,
rather than substantial lateral shifting, which typifies the peaked-
flood discharges and high sediment yields of the cold, stadial period
rivers. The Upton Warren sediments regularly show this type of
evolution while still retaining characteristics of stadial-type sedimen-
tary successions, i.e. fine-grained sediments interlaminated with
coarser sands indicating that annual, possibly nival-type floods still
regularly occurred. These successions are of a type associated with a
cool continental, rather than maritime climate, where cold winters
result in marked nival-flood events in spring following snow-melt.
Interglacial-type temperate fluvial sediments from lowland areas
generally lack this regular flood-cyclicity (Gibbard and Lewin, 2002).
Thus if an “interstadial”event of lower magnitude or shorter
duration than the Upton Warren (i.e., 4–5 ka) occurred, it would be
recorded only in “normal”cold, stadial-type sediments and its
preservation potential would be no higher than that of other
stadial-type fine sediments. Therefore, the remaining minor inter-
stadials of MIS 3 must have been marked by shorter durations and/or
climatic conditions that failed to cause the river systems to cross
significant process thresholds. It appears that these warmer events
were accommodated by the rivers making minor internal adjustments
and in consequence failed to initiate a response that can be identified
in the preserved sediment sequences (van Huissteden et al., 2001;
Briant et al., 2008). The low preservation potential of these sediments
is exacerbated by subsequent MIS 2 activity with its coarse-grained
deposits. The British record is considerably more fragmentary
than those of the lower energy, sand-dominated regimes of The
Netherlands and Belgium. This reworking of the sediments is
relatively far greater in situations where the river is restricted to a
narrow valley by steep, often bedrock-controlled slopes, in contrast to
where it is unrestricted and able to increase the valley width on non-
cohesive or unresistant substrates.
The maximum period of sediment deposition in MIS 3 seems to
vary significantly, with aggradation at some sites beginning before c.
306 J. Lewin, P.L. Gibbard / Geomorphology 120 (2010) 293–311

Author's personal copy
40 ka BP. However, radiocarbon-dated organic sediments in the basal
parts of the succession may give dates as young as b30 ka BP, e.g. the
Middle Thames in the London area (Gibbard, 1985; Coope et al.,
1997), Ecton on the Nene of 34 ka BP at the very base (Castleden,
1976; Briant et al., 2008), and at 29 ka BP at Sandy in the Ivel valley
(Gao et al., 1998) and the Lea Valley Arctic Bed sites (Gibbard, 1994).
This suggests that non-aggradation may have continued in some
valleys into the early part of MIS 3. Having been initiated, deposition
also continued to various times in different valley systems, with
sedimentation continuing through the latter half of the period, e.g. at
Stanwick (27 ka BP: Briant et al., 2008), Standlake Common (29 ka BP:
Briggs et al., 1985); Brandon (29 ka BP: Coope, 1968); Great Billing
(28 ka BP: Morgan, 1973; Shotton et al., 1969); and Thrapston
(26 ka BP: Shotton et al., 1970). However, the paucity of dates for
this episode after c. 20 ka BP suggests that almost all the streams had
virtually ceased depositing sediment by this time.
The majority of modern lowland western European rivers occupy
valley bottoms that were then cut and partially infilled by gravel
sequences during the later part of the Weichselian Stage (MIS 2)
where dates of b15 ka BP are commonly found, again deposited under
gravel-bed river conditions. Britain is no exception (van Huissteden
et al., 2001). These deposits underlie, or confine, the modern
floodplains and strongly influence modern channel patterns. Where
Holocene floodplains have lower gradients than the underlying
deposits, the gravel formations tend to become emergent to form
confining terraces in the upper courses, whilst downstream they are
buried beneath fine-grained alluvium (cf. Brown, 1995). There may
therefore be some longitudinal variation in response and sedimentary
pattern in modern streams that should be considered in the following
summary of late- and post-glacial floodplain evolution.
A somewhat different regime seems to have occurred during the
Late-glacial Interstadial (13–11 ka BP: Lowe and Walker, 1997). The
climate of this period was complex with an early warm peak, followed
by a cooler later part during which regional birch forest became
established. Several authors (e.g. Rose et al., 1980; Rose, 1995; Collins
et al., 1996) have considered river activity through this period and
concluded that the climatic amelioration caused rivers to reduce their
activity as peaked discharge was reduced to a more regular flow
pattern. The rivers therefore tended to adopt a single-thread mode,
possibly actively meandering where stream energy and the local
sediment supply was sufficient (e.g. in lowland Britain). They
deposited pebbly sand, sands and silts, the latter with a high organic
component. Shallow pools developed in inherited braid-plain depres-
sions and the resulting sediments contained little inorganic material.
Rose attributed this marked change in flow style and sedimentation to
regulation of sediment supply and run-off in response to increased
vegetation cover, soil development and increased infiltration result-
ing from melting of permafrost. In spite of this increase in organic-rich
fine sediment deposition, sediments from this period have rarely been
described. If this is a consequence of non-preservation, the cause
could have been later removal by the rejuvenated, energetic and
destructive rivers during the subsequent Younger Dryas (Loch
Lomond) Stadial (e.g. Collins et al., 1996; Lewis et al., 2001). This
stabilisation–incision response parallels that noted above (van
Huissteden et al., 2001) during the Middle Devensian Upton Warren
Interstadial event.
The change from the Younger Dryas Stadial to the Holocene in
England has been repeatedly discussed (e.g. Gibbard, 1985; Brown et
al., 1994; Rose, 1995; Brown, 1996; Gibbard and Lewin, 2002; Briant
et al., 2008). It is marked throughout the region by a profound change
from gravel-dominated flow regimes to predominantly fine-grained
sedimentation, comparable to that seen in the late-glacial and Upton
Warren interstadials. This change took place in response to the abrupt
climatic amelioration, the latter having occurred in less than 50 years
according to evidence from Greenland ice cores (Björck et al., 1998).
Channels occupied during the latest late-glacial time persisted into
the early Holocene (Gibbard, 1985; Brown, 1996; Gibbard and Lewin,
2002). The initial reaction was a reduction in number of flow
channels, a process that continued by vertical accretion infilling of
secondary channels in the Holocene (Brown et al., 1994). There was
therefore a “metamorphosis”from multiple shallow gravel-bed
channels to a network of fewer, deeper and narrower channels
enclosed by cohesive banks that resulted from vertical accretion of
predominantly fine sediment on floodplain surfaces (Brown, 1995,
1996). As Brown noted, this transition is marked in the valleys by
increased channel abandonments, as indicated by basal
14
C dates in
channel fills (Brown, 1995, 1996). These abandonments are accom-
panied by the exposure of the higher areas of the pre-existing braid-
plain surface to subaerial processes, including soil development and
vegetation colonisation (cf. Gibbard, 1985; Brown, 1995, 1996;
Gibbard and Lewin, 2002). This pattern is consistent with that of
earlier interglacial sedimentation in France (Antoine et al., 2000,
2007) where such sediments overlie earlier cold-stage deposits.
9. The pattern and implications of cold-period fluvial sedimentation
Gibbard and Lewin (2002) presented a model for fluvial valley
evolution and sedimentation through an interglacial climate cycle in
lowland Britain. This was divided into four “fluvial phases”(Fph I–IV).
These were intended to be seen as process phases rather than time
divisions, that is, not as time-equivalents of the vegetation substages
of West and Turner (1968) and West (1977), but rather as broad
process styles involved in the evolution of valley fills. As such the
phases could partially overlap or grade into one another, they could
potentially vary in length from interglacial to interglacial and they
could be strongly influenced by local site, reach and valley conditions.
A similar fluvial pattern can be determined for cold (glacial–
periglacial) periods, by adding two broad cold-stage styles:
•River planation and wide strath cutting on valley floors, with
“working depth”super-incumbent deposition,
•High extra-channel sediment delivery, with some aggradation and
considerable reworking.
As Gibbard and Lewin (2002) noted in their scheme, with climate
deterioration into the glacial period, woodland disappeared and was
replaced by regional herb-dominated early-glacial grassland at the
termination of interglacial conditions. Typical cold-climate regimes
with highly-peaked flow discharges progressively ensued. These
provided the energy for stream rejuvenation, channel enlargement,
remobilisation of coarse debris, rapid removal of fines, and substantial
incision into accumulated floodplain deposits (Fph IV). This equates to
the warm–cold “transition”of other researchers (e.g.,Vandenberghe,
2008). Furthermore, gullying of valley-side and floodplain sediments,
regolith and soils would lead to the incorporation of reworked fossils,
especially pollen and spores, into these accumulations. Late intergla-
cial deposits incorporate a mélange of interglacial soil and organic
fines. The redevelopment of permafrost further ensured rapid surface
water flow as infiltration became inhibited, and with it slope erosion
and solifluction exposed regolith and substrate materials providing a
source for coarse clastics materials for rivers.
Once fully cold-climate conditions were re-established valley-floor
erosion was re-activated, with coupled bedrock incision and lateral
planation, but without the later full loading of glacial and slope-
derived inputs. Previously accumulated fines were removed en masse
and transported down-valley as both disaggregated particles and
occasionally as coherent blocks (e.g. Somersham, Cambridgeshire:
West et al., 1999) resulting from channel-margin undercutting. It is as
yet not clear precisely when gravel transport was re-activated with
respect to the interglacial–glacial transition. Inevitably this may vary
between different interglacial events. However, unlike in Continental
stratigraphies, based largely on low-energy aggradational systems
(e.g. Vandenberghe et al., 2004), there appears to be a clear delay in
307J. Lewin, P.L. Gibbard / Geomorphology 120 (2010) 293–311

Author's personal copy
timing. The preservation potential of fine sediment accumulations in
otherwise high energy gravel-bed dominated streams is particularly
low, as observed frequently in modern analogue environments
(cf. Rust and Koster, 1984), hence the fragmentary nature of the
preserved sequences. This fragmentation and constant reworking of
the sediments was relatively far greater in situations where rivers
were restricted to a narrow valley by steep, often bedrock-controlled
slopes, in contrast to where they were unrestricted and able to
increase the valley width on non-cohesive or unresistant substrates.
It is under these early cold-stage conditions that we believe bed-
rock planation was dominant, without of course much sedimentary
evidence being left locally to record it.
Later on, bedrock incision may have been swamped by the
emplacement of cold-climate gravel and sands. Coarse clastics either
cap the sequence or abut any prior fine sediment channel fills that
remain (cf. Bridgland, 1994, 2000; Bridgland and Allen, 1996). This
relationship is known from many sites, for example at Ardleigh in
Essex, Maxey in Cambridgeshire and Swanton Morley in Norfolk
(Gibbard and Lewin, 2002). It is invariably erosional and therefore
implies a removal of pre-existing sediment, potentially representing a
considerable time interval. As noted above, the deposition of
overlying coarse clastics during the last cold-stage (Devensian,
Weichselian) apparently did not occur immediately after the previous
interglacial but as much as 20–40 ka later (Gibbard, 1985, 1994; Keen
et al., 1999; van Huissteden et al., 2001; Gibbard and Lewin, 2002) and
we regard the interlude as dominantly a period of valley-floor
excavation. If additional depositional activity did occur during this
time period, little evidence for it is currently available. Throughout the
later period gravel and sand deposition predominated with slope
material recruited by periglacial slope processes providing the bulk of
the clastic material and the water for peak flows in spring. In
aggrading sequences, truncated basal elements of braid-channel
alluvial architecture are most commonly preserved. Large partially-
abandoned surface areas would have been modified by discontinuous
permafrost development, and associated landforms such as pingos
and icings (naleds) in favoured localities.
10. Conclusions
Our interpretation of the evidence takes us to the conclusion that
the mid-latitude terrace-flight fragments in southeast England are
overwhelmingly the product of cold-climate processes which early in
cold-stages produced wide bedrock incising surfaces and a “working
depth”of cold-climate sediments across valley floors in repeated
episodes of activity. In places this primary landform has hosted later
phases both of aggradation and localised incision in both cold and
warm climatic episodes. Terminating bedrock incision and terrace
insetting followed with the next cold-stage. Interglacial sites and
sediments have been widely reported from low-energy environments
on floodplain surfaces, from channel fills and inset within gravel units.
Some may relate to small stream interglacial incision, but it is not
evident that there was high energy activity capable of eroding wide
bedrock straths. Just as interglacial sequences seem to lack late-stage
sediments because of erosional removal (Gibbard and Lewin, 2002,
Fph IV), so we believe that in early cold-stages river erosion was
rampant, aggradation was largely lacking in mid-catchment locations,
and eroded materials were removed down-valley. Some up-valley
aggradation came later when enhanced solifluction took place under
different periglacial conditions. Cold climates were at times capable of
producing both low-gradient planation and gravel transport involving
both bed-plucked and long-travelled material. It is an enduring myth
that cold-climate rivers were generally inefficient, suffering only a
kind of mid-system constipation in the form of aggradation. On the
contrary, larger rivers were able to erode bedrock and transport
gravels at low gradients, later aggrading for part only of cold-stages,
with considerable reworking and truncation of sedimentation
sequences (Gibbard and Lewin, 2009). Our conclusion is that
dominant valley-wide trough incision was not brief or “transitional”
(Bridgland and Westaway, 2008; Vandenberghe, 2008), but was a
persistently-lasting and generally early cold-stage phenomenon
(Antoine et al., 2000, 2007). Where genuine aggradation did
occur (rather than a co-formational “working depth”of sediment in
transit) then basal erosion became ineffective. Multiple truncations
within terrace gravel sequences suggest that vertical oscillations
within a notably limited range (c. 10 m) was characteristic, being both
autogenically- and, where biostratigraphic evidence in particular
indicates it, allogenically-driven in relation to oscilating climate (e.g.,
Maddy et al., 1998; Briant et al., 2005; Langford et al., 2007). But as
with many erosional features, and during the Pleistocene more
generally (Porter, 1989), much of the terrestrial evidence for
extended periods of activity is now missing. For evidence of strath-
type terrace formation, it is to bedrock margins that attention should
be drawn, as this is where process evidence for the final phase of
erosion is to be found. Whilst there is broad synchroneity with cold-
stages (Bridgland and Westaway, 2008), the cutting of straths and the
covering of them by aggradation do not always leave evidence of the
same number of episodes, nor do remnant terrace fragments
produced and partially destroyed during ongoing incision sweeps
always give the same number of preserved terraces in each activity
period. The availability of GPR and related technology in recent years
should now allow a greater potential for concealed information on
bedrock erosion as well as sedimentary evidence to be obtained
(cf. Bridge and Lunt, 2006).
However two caveats are in order. First, catchment systems operate
as wholes, and erosion upstream leads to sedimentation downstream,
as the thick pile of Pleistocene sediments in depobasins bears witness
(Busschers et al., 2007). Here episodes of sedimentation may be
different from, or even the reverse of, those up-valley. Aggradational
environments may involve interglacial deposits being buried intact by
influxes of cold-climate sediments, which may themselves interdig-
itate with high sea-level interglacial estuarine deposits. The situation is
rendered more complex by tectonic activity in parts of catchment
systems (van Balen et al., 2000; Antoine et al., 2007). Thus catch-all
models of Quaternary fluvial sedimentation (in which, for example,
aggradation cold–warm sequences from down-catchment are taken to
represent bedrock terrace sedimentation up-catchment), can mislead.
Second, cold-stage climates did vary; a frequently oscillating marine
oxygen isotope signal and recognised stadials were superimposed on a
long-term cycle of extended “deterioration”and abrupt “ameliora-
tion”.Thesefluctuations were not only marked by changes of
temperature, but changes in moisture availability, discharge season-
ality, evaporation and atmospheric circulation. Phases of incision,
stability and aggradation reflect these complications (Antoine et al.,
2000; van Huissteden et al., 2001) as much as occuring in a simple
glacial–interglacial “cycle”. We stress that our dominantly cold-stage
process model does not involve inevitable or simple succession, and
styles may be variably sequenced according to climate histories.
In other circumstances terraces as a whole may be formed
differently. Active uplift may swamp steep headwaters with sediment
which is then evacuated between sediment supply episodes, thus
producing fill–cut terraces in aggradational deposits. In highly
dynamic orogenic situations there may be close uplift–incision
relationships. Low-latitude terraces can reflect a response to Quater-
nary climates that have a different cyclical impact, with wet–dry or
arid–vegetated cycles. But what we do suggest is that the terrace
evidence for large mid-catchment tracts of southern and eastern
England shows that they actually formed by coupled lateral and
vertical incision during parts of cold-climate episodes, in some
locations being overun by later-stage multiple reworking and limited
aggradation. Fine sediment aggradation, reworking, depression-filling
and narrow linear incision were more characteristic of interglacials.
On this interpretation, the widespread nature and multi-episode
308 J. Lewin, P.L. Gibbard / Geomorphology 120 (2010) 293–311

Author's personal copy
sequencing of Quaternary terraces reflects climatic cycles as many
have maintained (Bridgland and Westaway, 2008)—but dominated
by cold-stage planation, and cold-stage aggradation at later stages
(although the amount is not generally large on terrace surfaces).
“Transitions”, we believe, mostly saw a reworking of sediments
rather than bedrock incisions, whilst interglacial activity was greatly
restricted in landforming terms.
Acknowledgement
We thank Philip Stickler (University of Cambridge) for drafting the
figures.
References
Allen, L.G., Gibbard, P.L., 1993. Pleistocene evolution of the Solent River of Southern
England. Quaternary Science Reviews 12, 503–528.
Antoine, P., 1994. The Somme Valley terrace system (northern France); a model of
river response to Quaternary climatic variations since 800,000 BP. Terra Nova 6,
453–464.
Antoine, P., Lautridou, J.P., Laurent, M., 2000. Long-term fluvial archives in NW France:
response of the Seine and Somme rivers to tectonic move ments, climatic
fluctuations and sea-level changes. Geomorphology 33, 183–207.
Antoine, P., Limondin Louzouet, N., Chausse, C., Lautridou, J.-P., Pastre, J.-F., Auguste, P.,
Bahain, J.-J., Falgueres, C., Galehb, B., 2007. Pleistocene fluvial terraces from
northern France (Seine, Yonne, and Somme): synthesis, and new results from
interglacial deposits. Quaternary Science Reviews 26, 2701–2723.
Arche, A., 1983. Coarse-grained meander lobe deposits in the Jarama River, Madri d,
Spain. In: Collinson, J.D., Lewin, J. (Eds.), Modern and Ancient Fluvial Systems,
Special Publication 6, International Association of Sedimentologists, pp. 313–321.
Autin, W.J., 1992. Use of alloformations for definition of Holocene meander belts in the
middle Amite River, southeastern Louisiana. Geological Society of America Bulletin
104, 233–241.
Ballantyne, C.K., Harris, C., 1994. The Periglaciation of Great Britain. University Press,
Cambridge.
Bernouf, J., Carcaud, N., Garcin, M., Giot, D., Galinié, H., Rodier, X., Marlet, O., Blin, Ch.,
Taberly, Ch., 2003. Fluvial metamorphosis of the River Loire during the Holocene:
and the role of natural and anthropogenic factors: a case study from the area of
Tours, France. In: Howard, A.J., Macklin, M.G., Passmore, D.G. (Eds.), Alluvial
Archaeology in Europe. Balkema, Lisse, pp. 163–171.
Berry, F.G., 1979. Late Quaternary scour hollows and related features in central London.
Quaterly Journal of Engineering Geology 12, 9–29.
Best, J.L., Ashworth, P.J., 1997. Scour in large braided rivers and the recognition of
sequence stratigraphic boundaries. Nature 387, 275–277.
Björck, S., Walker, M.J.C., Cwynar, L.C., Johnsen, S., Knudsen, K.L., Lowe, J.J., Wohlfarth, B.,
INTIMATE Members, 1998. An event stratigraphy for the Last Termination in the
North Atlantic region based on the Greenland ice-core record: a proposal of the
INTIMATE group. Journal of Quaternary Science 13, 283–292.
Bluck, B.J., 1967. Sedimentation of beach gravels: an example from south Wales. Journal
of Sedimentary Petrology 37, 128–156.
Bluck, B.J., 1971. Sedimentation in the meandering River Endrick. Scottish Journal of
Geology 7, 93–138.
Bluck, B.J., 1974. Structure and directional properties of some valley sandur deposits in
southern Iceland. Sedimentology 21, 533–554.
Bluck, B.J., 1976. Sedimentation in some Scottish rivers of low sinuousity. Transactions
of the Royal Society of Edinburgh 69, 425–456.
Boreham, S. 2002. The Pleistocene Stratigraphy and Palaeoenvironments of the
Cambridge District. PhD thesis, Open University.
Brewer, P.A., Lewin, J., 1998. Planform cyclicity in an unstable reach: complex fluvial
response to environmental change. Earth Surface Processes and Landforms 23,
989–1008.
Briant, R.M., 2002. Fluvial responses to rapid climate change in eastern England during
the last glacial period. PhD thesis, University of Cambridge.
Briant, R.M., Bateman, M.D., 2009. Luminescence dating indicates radiocarbon age
underestimation in late Pleistocene fluvial deposits from eastern England. Journal
of Quaternary Science. doi:10.1002/jqs.1258.
Briant, R.M., Coope, G.R., Preece, R.C., Keen, D.H., Boreham, S., Griffiths, H.I., Seddon, M.B.,
Gibbard, P.L., 2004. Fluvial system response to Late Devensian (Weichselian) aridity,
Baston, Lincolnshire, England. Journal of Quaternary Science 19, 479–495.
Briant, R.M., Bateman, M.D., Coope, G.R., Gibbard, P.L., 2005. Climatic control on
Quaternary fluvial sedimentology of a Fenland basin river, England. Sedimentology
52, 1397–1423.
Briant, R.M., Gibbard, P.L., Boreham, S., Coope, G.R., Preece, R.C., 2008. Limits to
resolving catastrophic events in the Quaternary fluvial record: a case study from
theNeneValley,Northamptonshire.Geological Society of London, Special
Publications 296, 79–104.
Bridge, J.S., 2003. Rivers and Floodplains: Forms, Processes and Sedimentary Record.
Blackwell, Oxford.
Bridge, J.S., Lunt, I.A., 2006. Depositional models of braided rivers. In: Sambrook Smith,
G.H., Best, J.L., Bristow, C.S., Petts, G.E. (Eds.), Braided Rivers: Process, Deposits,
Ecology and Management. Blackwell, Oxford, pp. 11–50.
Bridgland, D.R., 1994. Quaternary of the Thames. Chapman and Hall, London.
Bridgland, D.R., 2000. River terrace systems in north-west Europe: an archive of
environmental change, uplift and early human occupation. Quaternary Science
Reviews 19, 1293–1303.
Bridgland, D.R., Allen, P., 1996. A revised model for terrace form ation and its
significance for the lower Middle Pleistocene Thames terrace aggradations of
south-east Essex. In: Turner, C. (Ed.), The Early Middle Pleistocene in Europe.
Balkema, Rotterdam, pp. 123–134.
Bridgland, D.R., Westaway, R., 2008. Climatically controlled river terrace staircases: a
worldwide Quaternary phenomenon. Geomorphology 98, 288–315.
Briggs, D.J., Coope, G.R., Gilbertson, D.D., 1985. The chronology and environmental
frameworkof early man in the Upper Thames Valley. British Archaeological Reports,
British Series 13, 1–176.
Bromehead, C.E.N., 1925. The Geology of North London. Memoir of the Geological
Survey of Great Britain.
Brown, A.G., 1995. Lateglacial–Holocene sedimentation in lowland temperate environ-
ments: floodplain metamorphosis and multiple channel systems. Paläeoklima-
forschung, Special Issue 9, 1–15.
Brown, A.G., 1996. Floodpain palaeoenvironments. In: Anderson, M.G., Walling, D.E.,
Bates, P.D. (Eds.), Floodplain Processes. Wiley, Chichester, pp. 95–138.
Brown, A.G., Keough, M.K., Rice, J., 1994. Floodplain evolution in the East Midlands,
United Kingdom: the lateglacial and Flandrian alluvial record from the Soar and
Nene valleys. Philosophical Transactions of the Royal Society 348A, 261–293.
Brown, A.G., Basell, L.S., Toms, P.S., Scrivener, R.C., 2009. Towards a budget approach to
Pleistocene terraces: preliminary studies using the River Exe in southwest England.
Proceedings of the Geologists' Association. doi:10.1016/j.pgeola.2009.08.012.
Bryan, K., Ray, L.L., 1940. Geological Antiquity of the Linden Meier Site in Colorado.
Smithsonian Miscellaneous Collections 99, Washington.
Bryant, I.D., 1983a. The utilization of Arctic river analogue studies in the interpretation
of periglacial river sediments from southern Britain. In: Gregory, K.J. (Ed.),
Background to Palaeohydrology. Wiley, Chichester, pp. 413–431.
Bryant, I.D., 1983b. Facies sequences associated with some braided river deposits of late
Pleistocene age from Southern Britain. In: Collinson, J.D., Lewin, J. (Eds.), Modern
and Ancient Fluvial Systems, Special Publication 6, International Association of
Sedimentologists, pp. 267–275.
Bull, W.B., 1991. Geomorphic Response to Climatic Change. Oxford University Press,
New York.
Bull, W.B., 2007. Tectonic Geomorphology of Mountains: A New Approach to Paleoseis-
mology. Blackwell, Malden MA.
Busschers, F.S., Kasse, C., van Balen, R.T., Vandenberghe, J., Cohen, K.M., Weerts, H.J.T.,
Wallinga, J., Johns, C., Cleveringa, P., Bunnik, F.P.M., 2007. Late Pleistocene evolution
of the Rhine–Meuse system in the southern North Sea basin: imprints of climatic
change, sea-level oscillation and glacio-isostacy. Quaternary Science Reviews 26,
3216–3248.
Cant, D.J., 1976. Selective preservation of flood stage deposits in a braided fluvial
environment. Geological Society of Canada Progress and Abstracts 1, 77.
Castleden, R., 1976. The floodplain gravels of the River Nene. Mercian Geologist 6,
33–47.
Castleden, R., 1977. Periglacial pediments in central and southern England. Catena 4,
111–121.
Castleden, R., 1980. Fluvioglacial pedimentation: a general theory of fluvial valley
development in cool temperate lands, illustrated from western and central Europe.
Catena 7, 135–152.
Catt, J.A., Gibbard, P.L., Lowe, J.J., McCarroll, D., Scourse, J.D., Walker, M.J.C., Wymer, J.J.,
2006. Quaternary: ice sheets and their legacy, In: Brenchley, P.J., Rawson, P.F.
(Eds.), The Geology of England and Wales, second ed. Geological Society, London,
pp. 429–467.
Chiverrell, R.C., Foster, G.C., Thomas, G.S.P., Marshall, P., 2010. Sediment transmission
and storage: the implications for reconstructing landform development. Earth
Surface Processes and Landforms 35, 4–15.
Church, M., 1974. Hydrology and permafrost with reference to northern North America.
Permafrost Hydrology. Environment Canada, Ottawa, pp. 7–20.
Church, M., Miles, M.J., 1982. Discussion: processes and mechanisms of river bank
erosion. In: Hey, R.D., Bathurst, J.C., Thorne, C.R. (Eds.), Gravel-Bed Rivers. Wiley,
Chichester, pp. 259–271.
Collins, P.E.F., Fenwick, I.M., Keith-Lucas, D.M., Worsley, P., 1996. Late Devensian river
and floodplain dynamics and related environmental change in northwest Europe,
with particular reference to a site at Woolhampton, Berkshire, England. Journal of
Quaternary Science 11, 357–375.
Collinson, J.D., 1970. Bedforms of the Tana River, Norway. Geografiska Annaler 52A,
31–56.
Collinson, J.D., Thompson, D.B., 1982. Sedimentary Structures. Allen and Unwin, London.
Coope, G.R., 1968. An insect fauna from Mid-Weichselian deposits at Brandon,
Warwickshire. Philosophical Transactions. Royal Society of London 254B,
425–456.
Coope, G.R., Gibbard, P.L., Hall, A.R., Preece, R.C., Robinson, J.E., Sutcliffe, A.J., 1997.
Climatic and environmental reconstructions based on fossil assemblages from
Middle Devensian (Weichselian) deposits of the River Thames at South Kensington,
Central London, UK. Quaternary Science Reviews 16, 1163–1195.
Coope, G.R., Field, M.H., Gibbard, P.L., Greenwood, M., Richards, A.E.,2002. Pal aeontologyand
biopstratigraphy of Middle Pleistocene sediment in the Mathon Member at Mathon,
Hertfordhire, England. Proceedings of the Geologists' Association 113, 237–258.
Corner, R., 1975. The Tana Valley Terraces. A Study of the Morphology and Sedi-
mentology of the Lower Tana Valley Terraces Between Mansholmen and
Maskejokka. Uppsala Universitet Naturgeografiska Institutionen Rapport No. 38,
Finmark, Norway.
309J. Lewin, P.L. Gibbard / Geomorphology 120 (2010) 293–311

Author's personal copy
Davis, J.L., Annan, A.P., 1989. Ground-penetrating radar for high-resolution mapping of
soil and rock stratigraphy. Geophysical Prospecting 37, 531–551.
Dawson, M.R., 1985. Environmental reconstructions of a late Devensian terrace
sequence: some preliminary findings. Earth Surface Processes and Landforms 10,
237–246.
Dawson, M.R., Bryant, I.D., 1987. Three-dimensional facies geometry in Pleistocene
outwash sediments, Worcestershire, U.K. In: Ethridge, F.G., Flores, R.M., Harvey,
M.D. (Eds.), Recent Developments in Fluvial Sedimentology. Society of Econo-
mic Paleontologists and Mineralogists Special Publication, 39. Tulsa, Oklahoma,
pp. 191–196.
Dines, H.G., Edmunds, F.H., 1925. The Geology of the Country Around Romford. Memoir
of the Geological Survey of Great Britain.
Dury, G.H., 1964. Principles of Underfit Streams. United States Geological Survey
Professional Paper, p. 452A.
Ehlers, J., Gibbard, P.L., Rose, J., 1990. Glacial Deposits of Britain and Ireland. Balkema,
Rotterdam.
Erkins, G., Dambeck, R., Volleberg, K.P., Bouman, M.T.I.J., Bos, A.A., Cohen, K.M.,
Wallinga, J., Hoek, W.Z., 2009. Fluvial terrace formation in the northern Upper
Rhine Graben during the last 20,000 years as a result of allogenic controls and
autogenic evolution. Geomorphology 103, 476–495.
Forbes, D.L., 1983. Morphology and sedimentology of a sinuous gravel-bed channel
system: lower Babbage River, Yukon coastal plain, Canada. In: Collinson, J.D., Lewin,
J. (Eds.), Modern and Ancient Fluvial Systems. Special Publication 6, International
Association of Sedimentologists, pp. 195–206.
French, H.M., 2007. The Periglacial Environment. 3rd ed. Wiley, Chichester.
Friend, P.F., Moody-Stuart, M., 1972. Sedimentation of the Wood Bay formation
(Devonian) of Spitsbergen: regional analysis of a late orogenic basin. Norsk
Polarintitut Skifter 15, 1–77.
Galloway, R.W., 1961. Solifluction in Scotland. Scottish Geographical Magazine 77,
75–87.
Gao, C., Coope, G.R., Keen, D.H., Pettit, M.E., 1998. Middle Devensian deposits of the Ivel
valley at Sandy, Bedfordshire, England. Proceedings of the Geologists' Association
109, 127–137.
Gao, C., Boreham, S., Preece, R.C., Gibbard, P.L., Briant, R.M., 2007. Fluvial reponse to
rapid climatic change during the Devensian (Weichselian) Lateglacial in the River
Great Ouse, southern England, UK. Sedimentary Geology 202, 193–210.
Gibbard, P.L., 1982. Terrace stratigraphy and drainage history of the Plateau Gravels of
north Surrey, south Berkshire and north Hampshire, England. Proceedings of the
Geologists' Association 93, 369–384.
Gibbard, P.L., 1985. Pleistocene History of the Middle Thames Valley. University Press,
Cambridge.
Gibbard, P.L., 1989. The geomorphology of a part of the Middle Thames forty years on: a
reappraisal of the work of F. Kenneth Hare. Proceedings of the Geologists' Association
100, 481–503.
Gibbard, P.L., 1994. The Pleistocene History of the Lower Thames Valley. University
Press, Cambridge.
Gibbard, P.L., 1999. The Thames Valley, its tributary, valleys and their former courses.
In: Bowen, D.Q. (Ed.), A Revised Correlation of the Quaternary Deposits in the
British Isles: Geological Society Special Report no, 23, pp. 45–58.
Gibbard, P.L., Lewin, J., 2002. Climate and related controls on interglacial sedimen tation
in lowland Britain. Sedimentary Geology 151, 187–210.
Gibbard, P.L., Lewin, J., 2009. River incision and terrace formation in the late Cenozoic of
Europe. Tectonophysics 474, 41–55.
Gibbard, P.L., Preece, R.C., 1999. Southern England. In: Bowen, D.Q. (Ed.), A Revised
Correlation of the Quaternary Deposits in the British Isles: Geological Society
Special Report no, 23, pp. 59–65.
Gibbard, P.L., Wintle, A.G., Catt, J.A., 1987. The age and origin of clayey silt brickearth,
West London, England. Journal of Quaternary Science 2, 3–9.
Gibbard, P.L., West, R.G., Pasanen, A., Wymer, J.J., Boreham, S., Cohen, K.M., Rolfe, C.,
2008. Pleistocene geology of the Palaeolithic sequence at Redhill, Thetford, Norfolk,
England. Proceedings of the Geologists' Association 119, 175–192.
Gibling, M.R., 2006. Width and thickness of fluvial channel bodies and valley fills in the
geological record: a literature compilation and classification. Journal of Sedimen-
tary Research 76, 731–770.
Hancock, G.S., Anderson, R.S., 2002. Numerical modeling of fluvial strath-terrace
formation in response to oscillating climate. Geological Society of America Bulletin
114, 1131–1142.
Harding, P., Gibbard, P.L., 1981. Excavations at Northwold Road, Stoke Newington,
North-east London, 1981. Transactions of the London and Middlesex Archaeolog-
ical Society 34, 1–18.
Hare, F.K., 1947. The geomorphology of a part of the Middle Thames. Proceedings of the
Geologists' Association 58, 294–339.
Harms, J.C., Fahnestock, R.K., 1965. Stratification, bedforms and flow phenomena (with
an example from the Rio Grande). In: Middleton, G.V. (Ed.), Primary Sedimentary
Structures and their Hydrodynamic Interpretation: Society of Econonmic Palaeon-
tologists and Mineralogists Special Publication, 12, pp. 84–115.
Harms, J.C., Southard, J.B., Spearing, D.R., Walker, R.G., 1975. Depositional environments
as interpreted from primary sedimentary structures and stratification sequences.
Society of Econonmic Palaeontologists and Mineralogists Short Course 2.
Hein, F.J., Walker, R.G., 1977. Bar evolution and development of stratification in the
gravelly, braided Kicking Horse River, British Columbia. Canadian Journal of Earth
Sciences 14, 562–570.
Higginbottom, I.E., Fookes, P.G., 1970. Engineering aspects of periglacicial features in
Britain. Quarterly Journal of Engineering Geology 3, 85–117.
Hollin, J., 1977. Thames interglacial sites, Ipswichian sea levels and Antarctic ice surges.
Boreas 6, 33–52.
Howard, A.J., Brown, A.G., Carey, C.J., Challis, K., Cooper, L.P., Kincey, M., Toms, P., 2007.
Archaeological resource modelling in temperate river valleys: a case study from the
Trent Valley, UK. Antiquity 82, 1040–1054.
Huggenberger, P., Regli, C., 2006. A sedimentological model to characterize braided river
depositsfor hydrogeological applications. In: Sambrook Smith, G.H., Best, J.L.,Bristow,
C.S., Petts, G.E. (Eds.), Braided Rivers: Process, Deposits, Ecology and Management:
International Association of Sedimentologists Special Publication, 36, pp. 51–74.
Hutchinson, J.N., 1991. Theme lecture: periglacial and slope processes. In: Forster, A.,
Culshaw, M.G., Cripps, J.C., Little, J.A., Moon, C.F. (Eds.), Quaternary Engineering
Geology: Geological Society, Engineering Special Publication, 7, pp. 283–331.
Keen, D.H., Bateman, M.D., Coope, G.R., Field, M.H., Langford, H.E., Merry, J.S., Mighall, T.
M., 1999. Sedimentology, palaeoecology and geochronology of Last Interglacial
deposits from Deeping St James, Lincolnshire, England. Journal of Quaternary
Science 14, 411–436.
King, W.B.R., Oakley, K.P., 1936. The Pleistocene succession in the lower part of the
Thames Valley. Proceedings of the Prehistorical Society 2, 52–76.
Kukla, G.J., 1975. Loess stratigraphy of Centrel Europe. In: Butzer, K.W., Isaacs, K.W.
(Eds.), After the Australopithecines: Stratigraphy, Ecology and Cultural Cange in
the Middle Pleistocene, Mouton, The Hague, pp. 99–188.
Lane, N.F., Watts, A.B., Farrant, A.R., 2008a. An analysis of Cotswold topography: insights
into the landscape response to denudational isostacy. Journal of the Geological
Society of London 165, 85–103.
Lane, N.F., Watts, A.B., Farrant, A.R., 2008b. An analysis of Cotswold topography: insights
ino the landscape response to denudational isostacy. Journal of the Geological
Society of London 165, 83–103.
Langford, H.E., Bateman, M.D., Penkman, K.E.H., Boreham, S., Briant, R.M., Coope, G.R.,
Kenn, D.H., 2007. Age-estimate evidence for Middle–Late Pleistocene aggradation
of the River Nene 1st Terrace deposits at Whittlesey, eastern England. Proceedings
of the Geologists' Association 118, 283–300.
Lautridou, J.P., Letavernier, G., Lindé, K., Etlucher, B., Ozouf, J.C., 1986. Porosity and frost
susceptibility of flints and chalk: laboratory experiments, comparison of “glacial”
and “periglacial”surface textures of flint materials, and field investigation. In:
Sieveking, G.de C., Hart, M.B. (Eds.), The Scientific Study of Flint and Chert.
Cambridge University Press, Cambridge, pp. 269–282.
Leigh, D.S., 2006. Terminal Pleistocene braided to meandering transition in rivers of the
Southeastern USA. Catena 66, 155–160.
Leopold, L.B., Miller, J.P., 1954. A post-glacial chronology for some alluvial valleys in
Wyoming. United States geological Survey, Water-Supply Paper, p. 1261.
Leopold, L.B., Wolman, M.G., Miller, J.P., 1964. Fluvial Processes in Geomorphology.
Freeman, San Francisco.
Lewin, J., 1976. Initiation of bed forms and meanders in coarse-grained sediment.
Geological Society of America Bulletin 87, 281–285.
Lewin, J., Brindle, B.J., 1977. Confined Meanders. In: Gregory, K.J. (Ed.), River Channel
Changes. Wiley, Chichester, pp. 221–233.
Lewin, J., Macklin, M.G., 2003. Preservation potential for Late Quaternary river alluvium.
Journal of Quaternary Science 18, 107–120.
Lewin, J., Weir, M.J.C., 1977. Morphology and recent history of the Lower Spey. Scottish
Geographical Magazine 39, 45–51.
Lewin, J., Davies, B.E., Wolfenden, P., 1977. Interactions between channel change and
historic mining sediment. In: Gregory, K.J. (Ed.), River Channel Changes. Wiley,
Chichester, pp. 353–367.
Lewis, S.G., Maddy, D., Scaife, R.G., 2001. The fluvial system response to abrupt climate
change during the last cold stage: the upper Pleistocene River Thames fluvial
succession at Ashton Keynes, UK. Global and Planetary Change 28, 341–359.
Lowe, J.J., Walker, M.J.C., 1997. Reconstructing Quaternary Environments. Longmans, London.
Lunt, I.A., Bridge, J.S., 2004. Evolution and deposits of a gravelly braid bar and a channel
fill, Sagavanirktok River, Alaska. Sedimentology 51, 415–432.
Macklin, M.G., Lewin, J., 1989. Sediment transfer and transformation of an alluvial
valley floor —the River South Tyne, Northumbria, UK. Earth Surface Processes and
Landforms 14, 233–246.
Macklin, M.G., Jones, A., Lewin, J., 2009. River response to rapid Holocene
environmental change: evidence and explanation in british catchments. Quater-
nary Science Reviews. doi:10.1016/j.quascirev.2009.06.010.
Maddy, D., 1997. Uplift-driven valley incision and river terrace formation in southern
England. Journal of Quaternary Science 12, 539–545.
Maddy, D., Bridgland, D.R., 2000. Accelerated uplift resulting from Anglian glacioeu-
static rebound in the Middle Thames Valley, UK?: evidence from the river terrace
record. Quaternary Science Reviews 19, 1581–1588.
Maddy, D., Green, C.P., Lewis, S.G., Bowen, D.Q., 1995. Pleistocene geology of the Lower
Severn valley. Quaternary Science Reviews 14, 209–222.
Maddy, D., Lewis, S.G., Scaife, R.G., Bowen, D.Q., Coope, G.R., Green, C.P., Hardaker, T.,
Keen, D.H., Rees-Jones, J., Parfitt, S., Scott, K., 1998. The Upper Pleistocene deposits
at Cassington, near Oxford, England. Journal of Quaternary Science 13, 205–231.
Maddy, D., Lewis, S.G., Keen, D.H., 1999. Pleistocene palaeomeanders of the River Avon,
Warwickshire. Proceedings of the Geologists' Association 110, 163–172.
Maddy, D., Bridgland, D.R., Green, C.P., 2000. Crustal uplift in southern England:
evidence from the river terrace records. Geomorphology 33, 167–181.
Martini, I.P., Kwong, J.K., Sadura, S., 1993. Sediment ice rafting and cold climate fluvial
deposits: Albany River, Ontarion, Canada. In: Marzo, M., Puigdefábregas, C. (Eds.),
Special Publication 17, International Association of Sedimentologists, pp. 63–76.
Matsuoka, N., 2001. Solifluction rates, processes and landforms: a global review. Earth
Science Reviews 55, 107–134.
Matthews, M.C., Clayton, C.R.I., Rigby-Jones, J., 2000. Locating dissolution features in the
Chalk. Quaterly Journal of Engineering Geology 33, 125–140.
McGregor, D.F.M., Green, C.P., 1983. Post-depositional modification of Pleistocene
terraces of the River Thames. Boreas 12, 23–33.
310 J. Lewin, P.L. Gibbard / Geomorphology 120 (2010) 293–311

Author's personal copy
Miall, A.D., 1977. A review of the braided river depositional environment. Earth Science
Reviews 13, 1–62.
Miall, A.D., 1978. Lithofacies types and vertical profile models in braided river deposits:
a summary. In: Miall, A.D. (Ed.), Fluvial Sedimentology. Canadian Society of
Petroleum Geologists Memoir 5, pp. 597–604.
Miall, A.D., 1996. The Geology of Fluvial Deposits. Springer, Berlin.
Mol, J., Vandenberghe, J., Kasse, C., 2000. River response to variations of periglacial
climate in mid-latitude Europe. Geomorphology 33, 131–148.
Morgan,A.V., 1973. The Pleistocene geology of thearea north and west of Wolverhampton,
Staffordshire, England. Philosophical Transactions. Royal Society of London 265B,
233–297.
Murton, J.B., 1996. Near-surface brecciation of Chalk, Isle of Thanet, south-east England:
a comparison with ice-rich brecciated bedrocks in Canada and Spitsbergen.
Permafrost and Periglacial Processes 7, 153–164.
Murton, J.B., Peterson, R., Ozouf, J.-C., 2006. Bedrock fracture by ice segregation in cold
regions. Science 314, 1127–1129.
Old, R.A., Hamblin, R.J.O., Ambrose, K., Warrington, G., 1991. Geology of the Country
Around Redditch. British Geological Survey, London.
Panin, A.V., Sidorchuk, A.Yu., Chernov, A.V., 1999. Historical background to floodplain
morphology: examples from the East European Plain. In: Marriott, S.B., Alexander, J.
(Eds.), Floodplains: Interdisciplinary Approaches: Geological Society of London
Special Publication 163, London, pp. 217–229.
Passmore, D.G., Macklin, M.G., 2000. Late Holocene channel and floodplain develop-
ment in a wandering gravel-bed river: the River South Tyne at Lambley, northern
England. Earth Surface Processes and Landforms 25, 1237–1256.
Pazzaglia, F.J., Brandon, M.T., 2001. A fluvial record of long-term steady-state uplift and
erosion across the Cascadia Forearc High, western Washington State. American
Journal of Science 301, 385–431.
Peters, G., van Balen, R.T., 2007. Pleistocene tectonics inferred from fluvial terraces of
the northern Upper Rhine Graben, Germany. Tectonophysics 430, 41–65.
Porter, S.C., 1989. Some geological implications of average Quaternary glacial
conditions. Quaternary Research 32, 245–261.
Reid, C., 1887. On the origin of dry chalk valley and of Coombe Rock. Quarterly Journal of
the Geological Society 43, 364–373.
Reinfelds, I., Nanson, G., 1993. Formation of braided river floodplains, Waimakariri
River, New Zealand. Sedimentology 40, 1113–1127.
Rose, J., 1995. Lateglacial and early Holocene river activity in lowland Brit ain.
Paläoklimaforchung 14, 51–74.
Rose, J., Turner, C., Coope, G.R., Bryan, M.D., 1980. Channel change in a lowland river
catchment over the last 13,000 years. In: Cullingford, R.A., Davidson, D.A., Lewin, J.
(Eds.), Timescales in Geomorphology. Wiley, Chichester, pp. 159–175.
Rose, J., Lee, J.A., Kemp, R.A., Harding, P.A., 2000. Palaeoclimate, sedimentation and soil
development during the last glacial stage (Devensian), Heathrow Airport, London,
UK. Quaternary Science Reviews 19, 827–847.
Rust, B.R., 1972. Structure and process in a braided river. Sedimentology 18, 221–245.
Rust, B.R., 1978a. A classification of alluvial channel systems. In: Miall, A.D. (Ed.), Fluvial
Sedimentology: Canadian Society of Petroleum Geology, Memoir, 5, pp. 187–198.
Rust, B.R., 1978b. Depositional models for braided alluvium. In: Miall, A.D. (Ed.), Fluvial
Sedimentology: Canadian Society of Petroleum Geology, Memoir, 5, pp. 605–625.
Rust, B.R., Koster, E.H., 1984. Coarse alluvial deposits. In: Walker, R.G. (Ed.), Facies
Models: Geological Society of Canada Reprint Series, 1, pp. 53–69.
Schirmer, W., 1995. Valley bottoms in the late Quaternary. Zeitschrift für Geomorpho-
logie Suppl.-Bd. 100, 27–51.
Seddon, M.B., Holyoak, D.Y., 1985. Evidence of sustained regional permafrost during
deposition of fossiliferous Late Pleistocene river sediments at Stanton Harcourt
(Oxfordshire, England). Proceedings of the Geologists' Association 96, 53–71.
Sherlock, R.L., Noble, A.H., 1922. The Geology of the Country Around Beaconsfield.
Memoir of the Geological Society of Great Britain.
Shotton, F.W., Blundell, D.J., Williams, R.E.G., 1969. Birmingham University radiocarbon
dates III. Radiocarbon 11, 263–270.
Shotton, F.W., Blundell, D.J., Williams, R.E.G., 1970. Birmingham University radiocarbon
dates IV. Radiocarbon 12, 385–399.
Sidorchuk, A., Panin, A., Borisova, O., 2003. The late glacial and Holocene palaeohy-
drology of Northern Eurasia. In: Gregory, K.J., Benito, G. (Eds.), Palaeohydrology:
Understanding Global Change. Wiley, Chichester, pp. 61–76.
Siegenthaler, C., Huggenberger, P., 1993. Pleistocene Rhine gravel: deposits of a braided
river system with dominant pool preservation. In: Best, J.L., Bristow, C.S. (Eds.),
Braided Rivers: Geological Society of London Special Publication, 75, pp. 147–162.
Sklar, L., Dietrich, W.E., 1998. River longitudinal profiles and bedrock incision models:
stream power and the influence of sediment supply. In: Tinkler, K.J., Wohl, E. (Eds.),
Rivers Over Rock: Fluvial Processes in Bedrock Channels: Geophysical Monograph
107, American Geophysical Union, pp. 237–260.
Snow, R.S., Slingerland, R.L., 1987. Mathematical modeling of graded river profiles.
Journal of Geology 95, 15–33.
Van Balen, R.T., Houtgast, R.F., Van der Wateren, F.M., Vandenberghe, J., Bogaart, P.W.,
2000. Sediment budget and tectonic evolution of the Meuse catchment in the
Ardennes and the Roer Valley Rift System. Global and Planetary Change 27, 113–129.
Van den Berg, M., van Hoof, T., 2001. The Maas terrace sequence at Maastricht, SE
Netherlands: evidence for 200 m of late Neogene and Quaternary surface uplift. In:
Maddy, D., Macklin, M.G., Woodward, J. (Eds.), River Basin Sediment Systems:
Archives of Environmental Change. Balkema, Rotterdam, pp. 45–86.
Van Huissteden, J., Gibbard, P.L., Briant, R.M., 2001. Periglacial fluvial systems in
northwest Europe during marine isotope stages 4 and 3. Quaternary International
79, 75–88.
Vandenberghe, J., 2001. A typology of Pleistocene cold-base rivers. Quaternary
International 79, 111–121.
Vandenberghe, J., 2008. The fluvial cycle at cold–warm–cold transitions in lowland
regions: a refinement of theory. Geomorphology 98, 275–284.
Vandenberghe, J., van Overmeeren, R.A., 1999. Ground penetrating radar images of
selected fluvial deposits in The Netherlands. Sedimentary Geology 128, 245–270.
Vandenberghe,J.,Lowe,J.,Coope,G.R.,Litt,T.,Zöller,L.,2004.Climaticand
environmental variability in the Europe sector during the last interglacial–glacial
cycle. In: Battarbee, R., Gasse, F., Stickley, C. (Eds.), Past Climate Variability through
Europe and Africa. Kluwer, Dordrecht, pp. 393–416.
Warburton, J., Davies, T.R.H., Mandl, M.G., 2003. A meso-scale field investigation of
channel change and floodplain characteristics in an upland braided gravel-bed
river, New Zealand. In: Best, J.L., Bristow, C.S. (Eds.), Braided Rivers: Geological
Society of London Special Publication, 75, pp. 241–255.
Washburn, A.L., 1968. Periglacial Processes and Environments. Arnold, London.
Watson, E., 1969. The slope deposits of the Nant Iago valley near Cader Idris, Wales.
Biuletyn Peryglacjalny 18, 95–113.
Watts, A.B., McKerrow, W.S., Fielding, E., 2000. Lithospheric flexure, uplift, and
landscape evolution in south-central England. Journal of the Geological Society of
London 157, 1169–1177.
Watts, A.B., McKerrow, W.S., Richards, K., 2005. Localized Quaternary uplift of south-
central England. Journal of the Geological Society of London 162, 13–24.
Werritty, A., Ferguson, R.I., 1980. Pattern changes in a Scottish braided river over 1, 30
and 200 years. In: Cullingford, R.A., Davidson, D.A., Lewin, J. (Eds.), Timescales in
Geomorphology. Wiley, Chichester, pp. 53–68.
West, R.G., 1977. Pleistocene Geology and Biology. Longmans Green, London.
West, R.G., 1993. Devensian thermal contraction networks and cracks at Somersham,
Cambridgeshire, UK. Permafrost and Periglacial Processes 4, 277–300.
West, R.G., Turner, C., 1968. The subdivision and zonation of interglacial periods.
Eiszeitalter und Gegenwart 19, 93–101.
West, R.G., Dickson, C.A., Catt, J.A., Weir, A.H., Sparks, B.W., 1974. Late Pleistocene
deposits at Wretton, Norfolk. II. Devensian deposits. Philosophical Transactions.
Royal Society of London 267B, 337–420.
West, R.G., Andrew, R., Catt, J.A., Hart, C.P., Hollin, J.T., Knudsen, K.-L., Miller, G.F.,
Penney, D.N., Pettit, M.E., Preece, R.C., Switsur, V.R., Whiteman, C.A., Zhou, L.P.,
1999. Late and Middle Pleistocene deposits at Somersham, Cambridgeshire, UK: a
model for reconstructing fluvial/estuarine depositional environments. Quaternary
Science Reviews 18, 1247–1314.
Westaway, R., Maddy, D., Bridgland, D., 2002. Flow in the lower continental crust as a
mechanism for the Quaternary uplift of southeast England: constraints from the
Thames terrace record. Quaternary Science Reviews 21, 559–603.
Westaway, R., Bridgland, D., White, M., 2006. The Quaternary uplift history of central
southern England: evidence from the terraces of the Solent River system and
nearby raised beaches. Quaternary Science Reviews 25, 2212–2250.
Whipple, K.X., Tucker, G.E., 2002. Implications of sediment-flux-dependent river incision
models for landscape evolution. Journal of Geophysical Research 107, 3.1–3.20.
Whipple, K.X., Hancock, G.S., Anderson, R.S., 2000. Processes of incision into bedrock:
implications for the dynamics of landscape evolution. Geological Society of America
Bulletin 112, 490–503.
Williams, P.F., Rust, B.R., 1969. The sedimentology of a braided river. Journal of
Sedimentary Petrology 39, 649–679.
Woo, M.-K., 1990. Permafrost hydrology. In: Prowse, T.D., Ommanney, C.S.L. (Eds.),
National Hydrological Research Institute Science Report, 1. Environment Canada,
Saskatoon, pp. 63–75.
311J. Lewin, P.L. Gibbard / Geomorphology 120 (2010) 293–311