Following consultation with representatives from the marine aggregate industry, preliminary prospecting took place in the area of the palaeo-Arun river within the Owers Bank region of the English Channel, c. 18km south of Littlehampton in Sussex. The prospecting stage of the survey involved collecting six lines of sub-bottom profiling data using a boomer seismic source.

The study area was chosen after studying the data from these survey lines and identifying a large palaeochannel. It was approximately 20m deep and 300m wide and was likely to have been a major geographic feature in the palaeolandsurface, implying a high archaeological potential. The palaeochannel feature was targeted for the study because these features are intimately related both to the presence of aggregate resources and to the potential presence and survival of archaeological material. These areas are very suitable for testing methodology because they represent challenging and varied geomorphology. They also have the best potential for the survival of organic material which is ideal for palaeoenvironmental reconstruction.

Survey work took place between 1-18 July 2003 and 27-29 September 2003. Twenty vibrocores, 108 seabed grab samples and 245km of seismic survey data were collected over a 3.5km by 1km area, but primarily in a central 1km² around the palaeochannel feature.

The wealth of data retrieved allowed archaeologists to reconstruct the palaeo-Arun prehistoric environment. As part of Round Two of the Seabed Prehistory project this reconstructed environment was recreated as a 3D computer animated visualisation.

Geophysical survey

3D model of submerged palaeo-Arun valleyFrom a preliminary review of this line of data it was possible to determine that the survey area covered a wide valley landform, approximately 2.5-3km wide, with the axis of the valley orientated NW-SE. The higher topography in the SE corner of the survey area was a bedrock feature. On the sloping valley side of this highland was an area of bright reflectors, which were interpreted as being peat horizons. The wide valley landform contained a sediment-filled palaeochannel at the centre of the survey area.

This palaeochannel was orientated E-W, approximately 200-300m wide and ranged from 14-18m deep. The sediments filling the channel were interpreted as being a fine grained unit composed of sand/silt and a coarse-grained gravel deposit which was situated on the southern bank of the palaeochannel.

The dimensions of the gravel deposit vary along the length of the channel. These gravels must have been deposited in a high-energy environment because a large amount of energy would be required to transport sedimentary particles of this size. The other fine-grained sand and silt deposits will have been laid down when the channel was in a low energy environment, which allowed these small sedimentary particles to drop out of suspension. The fine-grained sand and silts onlap the gravel deposit and were therefore deposited at a later stage.

The bedrock underlying the palaeochannel can be seen beneath the fine-grained channel in-fill and at a higher level on either side of the palaeochannel, indicating that the channel is incised in the bedrock. A small palaeochannel feature was observed on selected lines of seismic data only. This palaeochannel was cut into the bedrock underneath the main palaeochannel being studied.

Geotechnical survey

Vibrocore survey

Deploying the vibrocorerOn 27-28 September 2003, a 6m hydraulic vibrocorer was used to acquire twenty vibrocores from ten locations within the study area. Each core was cut into 1m lengths, capped and labelled and taken back to WA for comprehensive logging. The second cores from each location were recovered using black vibrocore liners and kept separately in a darkened container to prevent exposure to light which would have rendered them unusable for Optically Stimulated Luminescence (OSL) dating.

Grab sampling survey

Flints recovered during grab samplingOn 27-29 September 2003, a total of 156 grabs were taken. From this, 108 samples, some of them combined, were collected. The initial visual examination on board the vessel identified four flints of potential anthropogenic origin. Very occasionally, a sample appeared to contain elements of peat or decaying organic remains within ‘humic’ clay. Where a large amount of this material was present within the sample, it was retained as a whole sample and processed accordingly in the laboratory.

The laboratory processing of samples was conducted through a nest of sieves of mesh sizes 9.6mm (classed as 10mm for processing purposes), 4mm and 1mm. Slag, clinker and flint were all retrieved. Slag and clinker were both readily identifiable; the flints, however, being relatively small, were harder to categorically define. Non-diagnostic fired clay of unknown function and date also occurs. Apart from the presence of two bird bones, only fossilised bone was encountered.

119 flints were retrieved from 50 of the 108 samples. Of those flints recovered, 22 came from the 10mm fraction, four of which were identified aboard the survey vessel, and 97 from the 4mm fraction. Of the 22 flints coming from the larger mesh twelve have elements indicative of human activity. Five of the identified flints have cortex visible, generally on one side of the flint (dorsal). Cortex is the outer surface of a piece of flint and so its presence on these five flakes indicates that these may be primary or secondary flakes from the initial stages of shaping the flint tool. None of these flints appear to be of a diagnostic type, though one has minor blade-like characteristics and another looks like a possible piercer.

Geoarchaeological analysis

The sedimentary deposits analysed, especially the palaeochannel fills, are substantial and appear to have accrued over a relatively short period of time given their size. The geoarchaeological core log descriptions highlight what appear to be either flood couplets or patterns of deposition produced by spring tides or tidal surges, in some palaeochannel fill deposits. It would potentially be possible to further clarify the rate of deposition if these features prove to deposition events produced by spring tides or tidal surges. These depositional events can be interpreted more precisely, by quantifying the accumulative seasonal tides, in a similar manner to counting tree rings, to produce a high resolution estimate of the rate of deposition.

There are two Late Devensian and early post-Devensian meltwater ‘pulses’ identified in the regional sea level curves, one at 14,500BP and one at 10,500-11,000BP, which may have been the source of this high volume sediment regime.

Peat horizons visible in vibrocore VC3

Environmental analysis

The initial assessment of the vibrocore samples from the palaeochannel sequences and the upper part of the wider palaeovalley form sequence all produced Boreal pollen assemblages. These contained a strong woodland component with fewer herbs, dominated by pine, oak, elm and hazel. They also indicated the presence of saline brackish water or marine conditions in the locale.

This relative uniformity of the vegetational assemblage suggested all the depositional environments were broadly contemporary. However, significant fluctuations in proportions of species in the various cores implied significant changes in vegetation and environment. For example, there was a discernible difference in the pollen and foraminifera assemblage of the lower part of one of the cores. The pollen assemblage in this core generally represented low and limited herbaceous diversity, dominated by arboreal and shrub elements, however, grasses were ‘extremely important’ in the lower part of the core. The foraminifera in particular suggested a range of environments from marine brackish near the base to salt marsh in the upper sediments.

As a result, fuller analysis of samples from cores from the wider valley edge deposits, palaeochannel and from the peat horizon in the wider valley floor deposits was undertaken.