Assessment of methodology

One of the main aims of the Seabed Prehistory project has been to develop survey methods currently used in the various marine industries to assess their potential for gathering archaeological data from the seabed. The various methods are assessed below.

Bathymetric survey

Bathymetric survey, which maps seabed topography, is not an appropriate tool for assessing, identifying or studying submerged prehistoric landsurfaces and their associated archaeological deposits. It provides models of the modern horizon rather than any evidence of relict palaeogeographies or buried stratigraphy.

Seismic survey

Shallow seismic survey can provide a sufficient understanding of the subsurface geological structure as to allow features such as palaeochannels to be interpreted and modelled in 3D if the correct seismic source is used. Seismic surveys can help archaeologists to understand the palaeogeography of an area and can be used to guide further work such as vibrocore surveys. However the seismic data must be collected at a line spacing appropriate for delineating features of interest.

Geotechnical survey

Vibrocore survey

Archaeological input into the vibrocore survey locations is considered central to the success of environmental reconstruction and the development of palaeogeographic models.

Grab sampling survey

Grab sampling survey methodology can be applied for archaeological purposes. The process has successfully retrieved artefacts from the upper layers of the seabed. Consequently, it will also be an effective tool for establishing negative results and determining an absence of near-surface or eroding archaeological deposits, which would be both significant and fragile, and particularly at further risk from the impacts of dredging.

The trialled grab-sampling method could be easily implemented and is complimentary to the benthic (marine ecological) survey already undertaken as part of the Environmental Statement process. It has proven to be cost effective method of undertaking empirical evaluation for archaeological assessment.

Geoarchaeological analysis

Geoarchaeological core logging and descriptions were shown to provide significant additional information to geological logs and photographs. They provide sedimentary evidence of the depositional processes involved, as well as descriptions of the sediment types.

Geoarchaeological assessment in offshore circumstances through archaeological access to vibrocores is possible and productive. The value of environmental data to palaeogeographic reconstruction, providing dynamic environmental parameters through which to assess likely human presence, should not be underestimated.

The palaeo-Arun in prehistory

The survey work undertaken in the palaeo-Arun study area produced a wealth of data which has allowed archaeologists to model the prehistoric landscape which now lies submerged in the English Channel.

The evidence used to reconstruct the prehistoric landscape came primarily from seismic and environmental data. The surface terrain was modelled directly from the seismic data. The choice of strata to represent the past Arun landscape surface was based upon a peat layer that could be easily traced within the seismic data. Where such peat was not present the surface of the substantial deposits of gravel that underlay the most recent marine sediments were used.

While sources of evidence for the vegetation came predominately from the pollen analysis, the topography of an area has a substantial bearing on the distribution of this vegetation. Reconstruction of the area’s ecology was therefore considerably influenced by evidence for the nature of both the terrain and hydrology.

In turn both river and sea will have had a substantial affect on past vegetation patterns. The degree of marine influence has a direct bearing on the expanse and nature of peat formation and the type of salt marsh environment. Similarly the extent of the river’s floodplain and degree of tidal influence will affect the vegetation and deposits at the river edge.

In terms of reconstructing these hydrological patterns the main sources of information come from foraminifera and diatoms as well as the sediments themselves. Complementary to this are also certain maritime indicator species to be seen within the pollen sequences.

Background

The phase of this landscape’s history spans a period of at most 3,000 years between 10,000-7,000 BP. However, the main source of evidence relates to a series of peat horizons dating from around 9,600 to 9,000 BP. For the purposes of the reconstruction this period was divided into three broad vegetational phases which are outlined below. The final phase, beginning around 9,100 BP, forms the basis for the 3D computer animated visualisation

No direct evidence of fauna was recovered from the survey area and consequently such evidence for the people and animals occupying southern England at this time comes from that available other archaeological sites. It is probable that some changes within the fauna would have taken place during each phase. So that within Phase I and to a lesser extent in Phase II, large herbivores of more open conditions, such as reindeer and horse would have been more common.

Within Phase III we would expect to find a greater element of woodland fauna, so that wild boar, red deer, and roe deer would all be common within the woodland on the upper gravel terraces. Of the larger herbivores those present would have consisted of species often found in small groups or solitary individuals such as European elk, and the now extinct auroch. It is likely that such animals may have come down to the floodplain, and we might also expect to find brown bears within this region, while further up the channel away from the tidal Gullreaches the european beaver might be found. There were probably few large predators within Britain at this time, wolves, fox, lynx and wild cat are all likely to have been present.

A great deal of bird life is also likely to have been present within the channel, with many common shore species present such as waders, cranes, herons, as well as gulls.

Phase I: 9629±50 BP

The channel would have been the main feature in the landscape. Seismic data only reveals the shape of the channel during its final phase, though during its earliest phase it is likely to have had several braided channels running within it; fed by spring and summer melt-water from retreating glaciers but dry in winter. This scenario is analogous with what is known of modern environments peripheral to modern retreating glaciers.

The river channel would have been fringed with various species of grasses, sedges, rushes and probably either marsh cinquefoil and/or silverweed. Areas of the channel edge which were under a greater maritime influence would have featured species such as sea plantain, sea pink, spurrey and various members of the goosefoot family.

The channel would have been set in a valley landscape with a species-rich short turf grassland, with an abundance of plants such as rock-rose, gentian, spring sandwort and low growing Cruciferae. Along with these species were more unusual species such as mountain avens and moonwort. Within the more rocky parts of this landscape we might also find various species of saxifrage.

Silver birchA vibrocore VC3, taken some 1.5km from the main channel on the south-west edge of the survey area featured peat horizons and a high proportion of foraminifera species aassociated with tidal flats and/or lagoons, as well as mid-high salt marsh. This area would have featured a salt marsh environment in the later phases of the landscape. In this early phase however, peat formation was beginning to take place and fen marshland would have been found here and closer to the channel. Within the pools in the fen would have been white and yellow water lilies while surrounding these pools would be a species rich swathe of tall herbaceous vegetation. Comprising this would be species such as kingcup, globe flower, meadow sweet, common valerian, common bisort, monks hood and dock.

The wider landscape at this time consisted of sparse woodland largely consisting of individual isolated trees with dwarf shrubs in between. This shrubland would have consisted of grasses, crowberry, dwarf juniper, cowberry/bilberry and low growing shrub willows. Those trees which did exist were dominated by silver and downy birch on the upper gravels and on the drier sandier areas of the landscape as a whole. In addition occasional pines would also have been present on the driest reaches of the upper gravels.

Phase II: 9,300 BP

By this phase it is probable that a single gently meandering channel existed in the base of the main channel. It is unlikely that at this time the channel had any great width, probably only some 25 to 50m at most.

ButtercupThe lower lying areas at this time were becoming increasingly boggy with standing pools of water with water lilies and pondweeds surrounded by dense aquatic vegetation comprising sedges, grasses, bullrushes, bur reed and common reed. Many of these species would have also been present along stretches of the river edge, while grassland with sedges, butter cups and docks will have reached across the floodplain. Fringing the edge of the floodplain would be small stands of various species of willow.

As seen both within the channel and the lower lying areas this phase witnesses the increasing development of salt marsh. It is then probable that many species associated with such vegetation, such as cord-grasses, salt marsh grass and sea aster became increasingly prevalent as the phase progressed.

The overall impression of the landscape would have been one of denser tree cover than that of Phase I. The woodland would still have been relatively open aspect, with a dense heath/grassland understorey. The trees as such would be clustered in individual stands or copses with occasional isolated individuals in between. Pine woodland would have dominated though mostly confined to the upper reaches of the gravel and would have been interspersed with both downy and silver birch. It is probable that trees of aspen would still have still been present, perhaps also along with occasional whitebeam and sloe.

Phase III: 9,100 BP

Reconstruction of Phase III woodlandThis phase as already noted is the phase in which the reconstruction is set. While the extent and density of woodland is far greater than the previous two phases it should also be noted that in general the landscape is probably less species rich than the previous two phases. By this phase the channel would have been under considerable tidal influence. To this extent during high tides it is probable that the channel was flooded to its full width and height.

This is the final vegetation phase before the landscape was eventually lost to the sea. As such it should be remembered that we are dealing with an actively changing landscape in which the woodland was continually being pushed back by salt marsh development and the encroaching sea.

The salt marsh, by this phase, as well as being well established within the depression where the vibrocore VC3 was taken, it would also have colonised both sides of the lower reaches of the channel. The vegetation within the marsh would consist of cord-grasses, salt marsh grass, seagrass, sea plantain, sea lavender and sea aster. On the mudflats within and bordering the salt marsh would have emerged the distinctive green spikes of glasswort.mbles 

Sea purslaneUpon the more sandier and gravel areas of the floodplain, plants of sea blites, marram grass, sea kale, sea bindweed sea-radish and sea-purslane would be present.

Wet, lower-lying reaches towards the salt marsh would not have supported the oaks and elm which would have existed in the woodlands on the upper reaches of the gravels. These areas would instead have featured relict vegetation stands from phase II, including occasional trees of birches, aspens, willow and other low-growing shrubs, such as sloe and dogwood. This vegetation would have come to edge of the salt marsh forming a dense swath along its edge with the occasional tree toppling in.

The main component of the vegetation during this phase is hazel which had become well and truly dominant. The abundance of hazel within various pollen diagrams at this date has been noted by a number of authors (cf. Godwin 1975). To produce such a proliferation of pollen, as noted by Rackham (2003), such hazel rather than being in the form of low growing, coppiced shrubs seen today would have formed instead substantial woodland. Whereas today hazel is recorded as growing to an average height of 6 metres, such individual trees would more probably have attained heights of between 10 and 15m. Within this dense hazel woodland the pollen spectra also hints at occasional oaks and elm, probably wych elm. Such vegetation would have occupied the drier upper reaches of the gravels and within this area it is probable that some pines survived into this phase.

BramblesGiven the density of the canopy within these woodlands ground vegetation is likely to have been sparse. What we might expect to find are brambles, dog-mercury , bracken and occasional grasses, such as fescue. Such vegetation would have been densest within clearings where trees have died and toppled. Upon the stumps we might also expect to find ferns such as polypody.

References

Rackham, O. 2003, Ancient Woodland: Its History, Vegetation and Uses in England. Castlepoint Press, Kirkcudbrightshire.

Godwin, H. 1975, The History of the British Flora: a Factual Basis for Phytogeography, Cambridge.

3D computer models.

Visualisation

The computerised reconstruction of the palaeo-Arun landscape was created by integrating data collected from geophysical and geotechnical survey to provide a basis for the interrogation of the landscape during the early Mesolithic period. Find out more about how this model was created.

Seabed Prehistory from Wessex Archaeology on Vimeo.

Digital modelling of geophysical data

A 3D model of the submerged landscapes of the study areas, detailing the palaeovalleys and other topographical features, was created using the seismic data from the geophysical surveys. The seismic data was processed using a program called Coda Geosurvey. Find out more about how this model was created.

view the model