Skip navigation

You are in:  Home » Research » Research projects » Controls on interglacial sedimentation

Discussion and Conclusions

The synthesis presented here proposes a clear temporal pattern of river behaviour occurs in lowland Britain in response to an interglacial climate cycle.  This response is partially and initially controlled by inheritance of coarse materials and floodplain topography from the immediately preceding cold (glacial) phase.  However, the pattern principally reflects the river’s adjustment to decreased seasonal contrast, precipitation, sediment supply and vegetational change that arise from the major climate change.  The establishment of little or no erosion, cohesive banks and vertical accretion as the dominant sedimentation mode together restrict the destruction of prior sedimentation units, and initiate a progression from large-scale lateral accretion to volumetrically smaller fine and organic sedimentation as palaeochannel fills or overbank flood deposits.   Reduction in the number of channels occupied by flowing water by progressive sedimentary infill to one or two, from the many present during the preceding cold period, encourages a progressively increasing frequency of overbank flooding and thus floodplain sedimentation.  The return to the predominantly treeless conditions of the subsequent cold period initially gives rise to increased floodplain alluviation but eventually causes incision and destruction of the fine-sediment accumulations.  Once sufficient coarse debris is available, gravel and sand emplacement follows.

fluvial phase (Fph)

channel form

dominant sediments

depositional system


sediment supply



early full-glacial

?single channel

rare - debris in transport

sediment removal, erosion of substrate

incision, gullying

?increasing coarse inorganic

patchy, herb-dominated, discontinuous ground cover




sand sheets, silts and clay drapes

floodplain splays; enlarging channel, lateral accretion

vertical accretion, deposition giving way to incision and gullying

high inorganic, especially fines

post-temperate woodland thinning progressively to shrub-, then herb-dominated communities

IV - early-glacial

Fph iv

single to multiple channel

silts and clays ± plant and animal remains

floodplain planar surface

vertical accretion and ?channel avulsion

decreasing organic, increasing inorganic

late-temperate mixed deciduous - coniferous forest

late III - IV

Fph iii

inactive meandering, anastomosing

detritus/clay/silt mud, peat, silt and clays, abundant fossils

floodplain surface becoming planar, infill of remaining depressions

channel deposition, vertical accretion and channel avulsion

high organic

full-interglacial, dense deciduous forest


Fph ii

few to multiple channels, anastomosing

detritus/clay or silt mud, peat, abundant fossils

depression and channel infill, floodplain surface sheet-like form

channel deposition, vertical accretion and channel avulsion

high organic

Full-interglacial, dense deciduous forest


Fph i

inherited multiple, stable

marl, tufa, organic mud, silt in early part, rich in fossils

channel and depression infill

deposition, initiation of cohesive bank, vertical accretion

sharply decreasing inorganic, increasing organic

early-interglacial forest, with birch, pine ± spruce

I - early II

full- and late-glacial


gravel and sand

braidplain deposition

lateral accretion and aggradation

high coarse inorganic

herb-dominated communities (shrub copses locally in late-glacial)

full- and late-glacial

This pattern clearly represents a complex response behaviour typical of fluvial systems.  This concept depends on the inherent complexity of natural systems, such that when disturbed or modified, they adjust in a complexand often delayed, non-uniform fashion in an attempt to re-establish equilibrium with the prevailing conditions.   Clearly the river system changes outlined here are primarily controlled by climate change, the impact of which is seen directly through seasonality, precipitation, temperature etc. but also indirectly through vegetation, soil development, discharge and sediment supply, as noted above.  These controls result in crossing of extrinsic fluvial process thresholds.  However, adjustments within the drainage systems, e.g. avulsion, arising from exceeding of intrinisic thresholds by gradient change, local sedimentation and erosion etc., will inevitably also have occurred.   The implications of this behaviour are important because they result in changes that cannot solely be attributed to climate.  In addition, adjustment of the rivers to major climate changes demonstrably results in a ‘lag’ between a climatic event and the fluvial response.  This is clearly seen at the beginning of the Holocene when rivers apparently took at least 500 years to restore equilibrium following the climate amelioration.  Even then, poor sedimentation conditioned the pattern and location of deposition.  Pattern transformation, under conditions of relatively low, fine sediment input, from braiding towards stable meandering took extended time periods.  Similarly, the progressive increase in sediment production that resulted from the climatic deterioration towards the end of interglacials apparently saw rivers adjusting by initially increasing alluvial sedimentation at many sites.  However, during the subsequent early-glacial phase, once sediment supply decreased, degradation and incision of their courses followed.  This erosional response resulted from an intrinsic theshold being exceeded and its commencement might be expected to differ temporally from system to system, valley to valley and even reach to reach.  Such incision may have been amplified by an accompanying development of permafrost in local substrates, an effect that would have further reduced debris supply.

The pattern of fluvial response to Pleistocene climate changes presented has important stratigraphical implications because it explains why sediments laid down early in interglacials are more common in the geological record than those from later in the same events.  This is because at the beginning of an interglacial, many surface irregularities and channels are available in wide braidplain surfaces inherited from the immediately preceding cold-stage river.   These topographic complexities form a mosaic of ‘accommodation space’ (cf. Emery & Myers 1996) in which deposition can occur, protected from the main river flow.  This process could eventually lead to the complete burial of the pre-existing gravel river braidplain microrelief by fine, often organic-rich floodplain sediments, where sufficient sediment is available.   Continued deposition may result in sediments on-lapping onto valley side slopes.  The relative scarcity of late-glacial sediments reflects the restricted nature of in-valley sedimentation at that period (there was a reduction in accommodation space as floodplain surfaces were levelled-out and raised, and there was little lateral channel accretion) coupled with the subsequent destructive activity by rivers with renewed energy and sediment transport capability which removed prior material.  

The frequent absence of preserved interglacial floodplain surface deposits, contrasts markedly with the common records of channel-fill sediments, and in particular those dating from the first part of interglacial events.  This arises because the early interglacial units often occupy relatively lower positions where they are protected between less easily-eroded, often coarser, substrate materials.  The floodplain surface alluvium, on the other hand, occupies a position where it can be readily degraded and removed during lateral migration and then incision in the subsequent early-glacial cold stage; it therefore has a very low preservation potential.  In essence, this implies that the younger the sediment, within any one interglacial depositional cycle, the more susceptible it is to removal during the subsequent cold period.  Indeed this cyclic behaviour is even similar to that seen in interstadials.  The more confined the river valley, the more this effect is amplified.  Later basal and marginal erosion during emplacement of gravel and sands frequently further amplifies this effect.  It also gives rise to the common occurrence of interglacial sediment sequences in pockets intercalated between cold-climate aggradations.  

Given that the interglacial-floodplain pattern proposed here holds for at least the last 0.75 My, any exceptions must be viewed as indicating potentially important variations from the general pattern.  Of particular interest are those rare sequences which represent the latter half of interglacial events.   Precisely when rejuvenation occurs shows some variation between sequences from different events.  This is a possible artifact of the record or a consequence of climatic variability between events.  As already stated, there are relatively few known sequences that preserve evidence of rejuvenation and even fewer that record the actual subsequent interglacial-glacial boundary in continuous fluvial sediment.  Two examples will suffice:

i.    Rejuvenation of rivers seems to have been particularly marked during an early Middle Pleistocene interglacial event represented at sites from the Welsh borderland to Essex.  This Mathon – Waverley Wood – Brandon – Witham-on-the-Hill – Wivenhoe event, if not simply a grouping of sequences that reflect a similar environment rather than being stratigraphically-equivalent, represents the late to post-temperate substages of a late ‘Cromerian Complex’ interglacial.  A comparable event is also represented at Sugworth, Berkshire.   These events clearly indicate renewed channel infill, and possibly immediately preceding channel cutting, late in interglacial Substages III – IV.  Moreover, the Sugworth sequence records the deposition of a point-bar complex in a channel-like depression excavated into local bedrock; both clear exceptions to the generalised pattern described above.

The known sedimentation in this interglacial group occurred during the period dominated by coniferous forests, and in particular during their retreat and replacement with boreal-type forest.  The explanation may be a climate deterioration, seen as lowered winter temperatures, increased storage of snow through colder winters, increased spring flooding leading to higher flood flow velocities and less effective infiltration of precipitation leading to more rapid throughput to channels.  A similar effect may have occurred if the sequence represented was immediately preceded by a cold-climate oscillation but no record of such an event is known at present.

ii.    The pattern of sequences described above includes little evidence of channel migration during the interglacials.  Indeed the evidence seems to overwhelmingly indicate channel stability.  However, the substantial channel complex beneath Little Oakley village in Essex (Fig.2), thought to represent a Cromerian Stage River Thames, clearly demonstrates channel migration through the first half of the interglacial (Bridgland et al., 1990).  This migration is seen as meandering-type behaviour, with the river depositing sandy silts in the flow channel and silts and clays in point-bar-type complexes on the inside of bends.  Such behaviour implies atypical conditions, possibly with higher than normal precipitation leading to higher discharges.  However, it may merely reflect the greater size of the river itself, in comparison to those of the later Pleistocene streams; that have been studied.

Although not fitting the preserved pattern discussed above, these records do not actually conflict with the general model presented here.  Clearly the potential variability between events, particularly the possibility of enhanced discharge episodes within interglacials must be seen as a means by which deviations from the idealised model  (Table) can occur.  For the Holocene, a number of researchers have suggested that sedimentation actually focused on a number of relatively short periods associated with enhanced river discharges.  This may equally be true of previous interglacials.  Furthermore site conditions vary considerably, some allowing active meandering for extended periods during interglacials and others very little.  The importance of sequences of alluvial styles and transformations must be stressed in general, but sequence variations must be expected at individual sites.

Seen in the context of the glacial-interglacial climate cyclicity of the later Pleistocene, these observations suggest that British lowland rivers normally adopt three major behavioural modes.  In order of their importance in the geological record they are: 1. braided or wandering gravel-dominated mode, 2. Fine-sediment-dominated stable meandering  to anastomosing mode, and 3. incision mode.  The occurrence of the individual modes will of necessity be ultimately controlled by climate, but other factors include upstream-downstream variability and changes resulting from relative sea-level change and tectonics.  What we have explored in this paper is essentially the ways in which climatic forcing is mediated by fluvial systems, responding to changes in vegetation, sediment supply and discharge, to produce distinctive patterns of interglacial fluvial sedimentation style. In the geological record.  This sequence of sedimentation styles can be used to explain, at least in part, the partial and discontinuous nature of preserved interglacial strigraphies.

Notwithstanding potential local variability, if these patterns are applied to the Late Pleistocene – Holocene (130 ky), the proportions of time that each mode operated and is preserved can be broadly estimated.  The interglacial (and interstadial) mode (2) described here represents at best only 25% of the time (i.e. including the interglacial Ipswichian and Holocene, and the interstadials of the last cold stage) ; whilst the gravel-dominated (cold-climate) mode (1) represents about 35% of the time, leaving 40% of the time during which incision (erosion) and/or non-deposition occurred.  The clear implication is that, in spite of potential exceptions, the best record that can be achieved from the lowland British fluvial sequence represents perhaps 50-60% of Pleistocene time, at most.  These figures will decrease markedly backwards in time, as a consequence of the systematic destruction of older sequences (except under favourable circumstances), such that by the Middle Pleistocene a far smaller proportion of the record will be preserved.  By the Early Pleistocene, only glimpses of fluvial palaeoenvironmental sequences will remain.

Back to interglacial rivers introduction page