The Humber Wetlands Project

Wetland Archaeology
and Environments
Research Centre,
Dept. of Geography,
University of Hull,
Cottingham Road,
Hull
HU6 7RX

WAERC@hull.ac.uk

tel : 01482 465325
fax : 01482 466340

Origins

Wetland Heritage of the Ancholme and Lower Trent Valleys

The survey of the wetlands of the Ancholme and lower Trent valleys comprised an archaeological programme, the assessment of the state of preservation of two archaeological sites and a palaeoenvironmental programme.

Archaeological field survey

The survey has discovered 50 new archaeological sites dating from the Mesolithic to the Roman period from the Ancholme and lower Trent valleys and the floodplain of the Winterton Beck. As expected, the archaeological survey of the lower Trent valley was hampered by the extensive presence of alluvial deposits and 'warp' of Post-Medieval date, covering over 13,000 ha.

Sites and finds of Mesolithic and Early Neolithic date are closely associated with the Rivers Ancholme and Trent and the Winterton Beck. As was noted in the case of the Humberhead Levels, this suggests a form of specialised exploitation of the river resources during this period, with particular attention to fishing and fowling.

The Late Neolithic to Bronze Age period sites are also situated near the River Ancholme and the Winterton Beck, although the sites are located on the floodplain edge rather than on the riverbanks. This reflects the widespread floodplain mire development at this time, but also illustrates the increased importance of agriculture, particularly the farming of domesticated cattle and sheep, relative to hunting and fishing.

Archaeological material dating to the Iron Age or later prehistory was not found during the field survey. Sites of this date studied during the survey were all identified on the valley slopes through aerial photography, a technique which is unlikely to recognise archaeological sites which have become buried by river deposits.

The Roman period is well represented in the survey results. Sites of this date were found particularly in the lower Trent valley, but also in the Winterton Beck area and Ancholme valley. The majority of Roman period finds can be attributed to the second to early fourth centuries, with significant concentrations at Trent Falls in the lower Trent valley, Winterton Beck, the area around Bishopbridge in the upper Ancholme valley, and South Ferriby and Winteringham in the lower Ancholme valley.

The existence of Roman and Romano-British sites on floodplains in the Humber wetlands has already been noted for the Humberhead Levels (Van de Noort et al. 1997), and the survey of the Ancholme and lower Trent valleys has added more evidence for the colonisation and exploitation of the Humber wetlands. The location of a large number of Roman period sites is closely linked to major lines of communication. The site at Adlingfleet (see below) is situated alongside the River Trent and is connected to large areas of hinterland via the Rivers Trent, Don and Ouse. It would also have been ideally located as a gateway for trade with the Continent. The main concentrations of Roman period sites in the Ancholme valley are near the road which branches off Ermine Street and leads towards the Roman town of Caistor, on the Lincolnshire Wolds, and at the terminus of Ermine Street, at Winteringham.

Several themes of Early Medieval or Medieval use of the wetlands of the Ancholme and lower Trent valleys were addressed, including the exploitation of the wetlands from the higher and better drained Isle of Axholme, riverbank settlement in the lower Trent valley and monastic settlements in the Ancholme valley.

South Ferriby Roman road

Research at South Ferriby-2 (Roman road) confirmed the excellent preservation qualities of the intertidal zone. Despite the prolonged periods of daily exposure of the site on falling tides, the quality of preservation of the brushwood, which formed part of the road surface, was excellent.

This may be primarily attributed to the water-retentive nature of the fine gained sediments prevalent in the Humber estuary. Brushwood from the Roman road was radiocarbon dated to the first or early second century AD. The road ran across the lower Ancholme floodplain, from Winterton to South Ferriby, and has largely been eroded by the River Humber.

Adlingfleet Roman settlement

To the north of the excavated area, five additional concentrations of pottery were identified during the survey, suggesting a series of settlement nuclei on the bank of the River Trent. The site is ideally located for trade, and it is inconceivable that its strategic location is mere coincidence. The northernmost concentration contained predominantly sherds of imported Samian pottery, and it may well be the case that a Roman (rather than Romano-British) settlement existed here

Palaeoenvironmental survey

The palaeoenvironmental survey of the lower Trent, including its tributaries the River Eau and Old River Don, and of the Winterton Beck and River Ancholme has produced a range of new information about their development. In the case of the Ancholme, this was built upon an understanding of environmental change obtained from previous research, but such knowledge was largely or wholly absent for the other rivers.

The Ancholme Valley

The palaeoenvironmental survey of the Ancholme valley consisted mainly of transect coring at four locations - at Bishopbridge near the confluence of the Rivers Ancholme and Rase, and at Redbourne, Brigg and South Ferriby. Determination of the sediment stratigraphy was combined with pollen and diatom analysis, and also with the radiocarbon dating of 25 samples.

As far as the development of wetlands in the Ancholme valley is concerned, the identification of fully marine diatoms in the sand near the base of the deep palaeochannel at Brigg suggests that sometime before c. 5000 BC the Ancholme valley north of the Brigg constriction was an inlet of the Humber estuary. The rise in sea-level after that date resulted in higher water tables, initiating peat formation in the valley bottom. The vegetation changed from a fen-carr towards a reed swamp as sea-level continued to rise. This resulted ultimately in the change from a peat-forming mire towards a minerogenic mire and the onset of mineral sediment accumulation during later prehistory.

Sometime during the Roman period there was a return to drier conditions. The pollen and archaeological evidence at South Ferriby suggest that the environment dried out sufficiently to allow nearby settlement, probably on the floodplain itself. This may be indicative of falling sea-level or lower amplitude tidal regimes during the first to fourth century AD, and is paralleled in the lower Trent valley and the Humberhead Levels (Van de Noort et al. 1997). In the late or post-Roman period, a second marine transgression took place, resulting in a return to wetter conditions.

The lower Trent Valley

The palaeoenvironmental survey of the lower Trent valley included transect coring across the Trent floodplain at three locations - at Stockwith, between Amcotts and Flixborough and at Trent Falls. Transects were also taken across the floodplains of the River Eau at Scotter and the Old River Don at Adlingfleet. The coring showed that initially (during the Late-Devensian and earlier Holocene) the palaeo-rivers Trent, Don and Eau incised their channels deeply, responding to the much lower sea-level at that time. Subsequently, peat developed within the palaeochannels, apparently in response to waterlogging and impeded runoff as sea-level continued to rise. At Trent Falls, this phase of peat development is dated to c. 6000-3700 BC. To the south of Trent Falls, the channel infill sequence indicates a fining-upwards of the sediments, reflecting the reduced ability of the rivers to transport coarser material as sea-level rose towards OD.

Within the Trent valley, the development of the floodplain wetlands is characterised by a mosaic of plant communities occurring in an essentially alder fen-carr dominated environment throughout much of the Holocene, up to the more recent historical period. Peat was developing on the sand dunes around Scotter and the blown sand deposits that flank the rivers of the region after c. 6000-3990 BC, and rapid channel infilling resulted in the expansion of the floodplain mire after c. 3900-1310 BC, a situation that is mirrored throughout the region to the west of the Lincoln Edge.

Elsewhere in the vicinity of Trent Falls a previously unidentified palaeochannel has been located on the western margins of the alluvial 'island' on which the Roman period site at Adlingfleet (discussed above) is situated. Throughout the lower part of the Trent valley between Flixborough and Trent Falls, diatom evidence has indicated the presence of open areas of shallow freshwater on the floodplain, and the incidence of brackish/marine incursion at the shallowest sampling points.

Winterton Beck

The Winterton Beck, situated at the northern end of the Lincoln Edge, has produced a mid-late Mesolithic (c. 5300-4000 BC) to later historical sequence of wetland development. The mid-Holocene wetlands are characterised by alder dominated fen-carr environments with high concentrations of reeds locally. Within the palaeochannel, grass/sedge dominated wetland with a brackish water component indicative of tidal incursions has been recorded for the more recent part of the stratigraphy. At this point the contained assemblage is dated to the later Holocene, at c. AD 1000 to the present.

Additional work

Global Positioning System (GPS) equipment was used extensively during the survey of the Ancholme and lower Trent valleys. This made it possible to obtain absolute height data rapidly for all boreholes.

The GPS can also be used for earthwork survey, and its application in this field was tested on the earthworks at South Ferriby, which represent the Medieval and Post-Medieval harbour in the eastern palaeochannel of the River Ancholme. The survey was undertaken with points mostly measured at 3 m intervals, but using a higher resolution where necessary. A digital terrain model was then generated in an ArcInfo GIS (Geographical Information System).

A central channel running from south to north is blocked at its southern end, and is confined at its northern end by the stone piers of a bridge, which is documented. Various channels and mounds, forming the port, are discernible to the east of the main channel.