Skip navigation

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

Devensian (Weichselian) Late-glacial - Holocene (Flandrian) fluvial sequence as an analogue

The majority of modern lowland British rivers occupy valleys that were cut and partially infilled by gravel sequences during the Devensian (=Weichselian) Stage.  These deposits underlie, or confine the modern floodplains and strongly influence modern channel patterns.  Because Holocene floodplains frequently 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.  This summary is not intended to provided an exhaustive synthesis of Holocene river activity, which has been recently reviewed elsewhere, but it is included for comparison with the interglacial fluvial sequences described above.

The 'typical' lowland British extra-glacial river during the cold phases of the Devensian was actively accumulating sediment along its course; the sedimentation being dominated by gravel deposition with rivers adopting a braided or wandering mode.  Sediment was supplied by slope processes, predominantly solifluction, together with bank and channel floor erosion.  In partially abandoned or slightly elevated channel areas of the braidplain, which were only rarely inundated by  flowing water, shallow pools and depressions accumulating organic material were present.  Where parts of the braidplains remained abandoned for significant lengths of time, abundant ice wedge cast networks developed in the sediment bodies.  Vegetation on the floodplain or adjacent valley sides was treeless herb-rich grassland .

Modern analogues for active braided river systems suggest that lateral channel-shift by bank erosion is rapid and spatially extensive, such that the surface layer of entire alluvial valley floors may be reworked in a matter of decades.  Earlier material may be only fortuitously preserved in ‘islands’ or valley-margin locations, or where aggradation is in progress. Even then, where rates of lateral erosion / accretion are high, it is only the lower parts of alluvial sequences (such as deep scour-pool fills) that get to be preserved in truncated and stacked sequences.  Altogether a braided channel style is likely to be characterised by:
(i)    a laterally-extensive (possibly valley-wide) layer of channel accretionary deposits to scour pool depths produced in a relatively short time.
(ii)    rapid destruction of earlier alluvial or other materials within the channel migration belt.

In contrast, a somewhat different regime seems to have occurred during the warm phase of the Late-glacial (Windermere) Interstadial (13-11 Ka).  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.  Rose has considered river activity through this period in lowland Britain 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. Gipping valley).  They deposited pebbly sand, sands and silts, the latter with a high organic component.  Shallow pools developed in inherited braidplain depressions 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 resulting 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 of the subsequent Younger Dryas (Loch Lomond Stadial).  A parallel stabilisation/ incision response has been noted to the Middle Devensian Upton Warren (=Hengelo) Interstadial.

The change from the Younger Dryas (Loch Lomond)Stadial to the Holocene 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 Interstadial.  This change took place in response to the abrupt climatic amelioration, the latter having occurred in less than 50 years.  Such a change would have been too rapid for streams to adjust their channels and so the channels occupied during the latest Late-glacial would have persisted into the early Holocene.  The initial reaction was a reduction in number of flow channels, a process that continued by siltation of secondary channels throughout the Holocene.  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.  As this author notes, this transition is marked in the Nene and Soar valleys by increased channel abandonments, as indicated by basal C14 dates in channel fills.  These abandonments are accompanied by the exposure of the higher areas of the pre-existing braidplain surface to subaerial processes, including soil development and vegetational colonisation.  

Two things are important in determining the preserved alluvial record of the early Holocene.  First, anastomosing or more stable meandering alluvial styles are intrinsically different in type and rate of alluviation.  Lateral accretion rates are low and the zone occupied by migrating channels is restricted.  Floodplain zones liable to overbank mineral sedimentation and organic growth for extensive periods, albeit at low sedimentation rates, are much more in evidence.  Thus for a given valley floor, such sedimentation styles are:
(i)    spatially differentiated, with active lateral accretion on only a fraction of the floodplain (in contrast to braided environments) with the potential for long-term preservation of vertically-accreted channel fills and floodplain fines on much of the floodplain surface; and,.
(ii)    less destructive of prior alluvial units.

A second factor concerns the effects of the sequential change from braiding to meandering / anastomosing.  The prior braidplain surface with a complex relief of bars and interlacing multiple channels directly conditions the location of lower energy channels depositing sediment.  Alluviation in what might otherwise be a simpler channel and planar floodplain system is topographically steered, producing a distinctive transitional pattern type.
Once the transition had occurred, it is uncertain whether more than a single channel operated at any one time.  Brown considers that the lack of channel migration, combined with the low width/depth channel ratio and the preservation of floodplain archaeological sites, indicates the presence of multiple reaches of sinuous channels of comparable proportions to those found in modern analogue localities (e.g. Khopersk National Park, Russia – personal observation).  Brown quotes an example from southern Ireland where a wooded anastomosing channel system, comparable to those expected to have occurred throughout southern Britain before human interference with channel systems, is found.  It is characterised by a stable interconnecting channel pattern, small diamond-shaped islands, adjacent channels with different bed- and flow conditions, much dead-water at medium and high flow, floodplain scour often related to vegetation, debris dams and the silting-up infill of abandoned channels. Where shifts in flow channel take place, they principally occur by avulsion into neighbouring depressions.  This behaviour is triggered by local channel deposition, resulting in locally-increased gradient or blockage of individual reaches by sediment.  Avulsion typically occurs where an intrinisic threshold has been exceeded.  Clearly the complexities of these systems in part reflect the inherited topography on which they formed and which they were incompetent to destroy.

The increase in floodplain surface and channel stability was greatly enhanced from the early Holocene by establishment of woodland both in the catchment in general, but particularly on floodplain surfaces. Riparian vegetation reinforced bank resistance. For example, in the River Stour valley peat that began at 9.8 ky and continued until recently, indicates a succession from Betula to Salix to Alnus-dominated forest during Substage Fl I.  Indeed, Alnus remains the dominant tree on lowland British floodplains to this day.  As Brown observes, such dense woodland would certainly retard channel migration, leading instead to the stabilisation of multiple small channels that shifted by avulsion.  In addition, these small channels were highly susceptible to damming by vegetational debris and by beaver activity, creating ponds.

The overall pattern of (early) Holocene fluvial sedimentation is therefore of an anastomosing pattern of stable channels confined between cohesive banks with a predominantly forested floodplain and relict braidplain surfaces upon which the dominant if patchy sedimentation was by vertical accretion of fine, overwhelmingly organic sediment, either as in situ peat growth, tufa or inwashed plant detritus. This anastomosing channel system has been widely interpreted from Holocene sequences throughout the region, including the Nene, the Welland, the Trent and the Thames.  This behaviour has given rise to the occurrence of sediment ‘packages’ of differing ages within the interglacial but occurring at the same height in the sequence.

It is important to note that the deduced style of ‘forced anastomosis’ , dependent on the topographic control of the prior landsurface, has modern analogues, but these contrast with other anastomosing patterns produced under conditions of aggradation and avulsion.  Whilst there appears to have been a switch to dominantly fine and organic sedimentation in the Holocene, rates of sedimentation were not necessarily large.  They were also spatially restricted so that whilst some sites may have inherited depression infilling, other topographically higher parts of inherited braidplains may have been very little affected by early Holocene alluviation at all.

In the later Holocene human interference in the landscape, principally through forest clearance accelerated the delivery of fine sediment to predominantly single channel systems.  Clearance involved greatly increased surface flow, inorganic sediment input, increased flooding and therefore accelerated floodplain surface (overbank) deposition, principally of fine sediment.  Where sufficient stream energy and sediment were available and cohesive banks were lacking or destroyed, multiple-channel braided reaches and active channel migration continued in some areas, particularly in the Middle Trent valley.  However, the latter is atypical of lowland Britain.

Precisely when alluviation occurred has been the subject of considerable discussion centred upon whether depositional episodes and interspersed incision are principally climatically or anthropogenically-driven.  Attempts to determine this based on dated sequences have been collated for the whole country .  Their plots suggest that the early Holocene (Substage Fl I) was a period of relative stability with slow sedimentation.  This is followed by an apparent hiatus of around 1000 years at c. 6.0 ky BP (early Substage Fl II).  This may be interpreted as a period of erosion resulting from minor climatic deterioration  or one of low sedimentation.   Alternatively, a record of relatively small amounts of alluviation once present may have been removed by subsequent erosion. Recorded alluviation throughout the region began again shortly after 5.5 ky BP in the climatic optimum (the second half of Substage Fl II: sensu this article).  Major fine inorganic sediments were laid down by increased overbank flooding, normally attributed to the anthropogenically-induced surface erosion following the Neolithic forest clearance.  This clearance commenced at approximately 3-4 ky BP and continued virtually to the present (Fl III).

Macklin & Lewin and Macklin argue that the anthropogenic effects have served mostly to amplify or blur the impact of natural climate changes, whilst others consider that humans have been the controlling force for changes, particularly in the later Holocene.  A third group  conclude that a combination of the two have influenced fluvial systems.

These observations throw into question the validity of using Holocene fluvial sequences as analogues for older interglacials.  The solution is not straightforward.  If climate is the principal driving force behind alluviation, then the Holocene can serve as an analogue until anthropogenic effects become strong, i.e. about 3-4 ky ago.  This implies that the first half of the Holocene sequence is a valid analogue for the critical period of substages Fl I-early II.  By contrast, the modern period (substage Fl III) has an environment so influenced by humans that it might be regarded as offering little direct comparative value.  However, as will be shown below, even this period has a role in our understanding of late interglacial events, for which no other analogue is available.  

Perhaps the greatest hurdle to the direct comparison of Holocene sequences with those of preceding events is the period‘s general atypicality.  Humans have not only directly, but also indirectly influenced drainage systems throughout the region .  For example, they are probably responsible for the absence of the substantial large vertebrate populations so characteristic of earlier interglacials.  Both the direct and indirect effects of herbivorous mammals have been repeatedly noted in interglacial fluvial sequences, through the occurrence of their remains, their disturbance of the sediments, and their destruction and modification of local floodplain environments.  Their potential effects on floodplains and channels can be observed today both in pastoral areas in the region or in African game reserves, for example.  The impact of herds of elephants, hippopotamuses, bison, horses and deer on river floodplains in Britain can only be interpreted from the sediments themselves but was undoubtedly locally, if not regionally, significant.

Next page