| Geography |
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 Vale of York
For the purposes of the survey, the Vale of York was defined as an area of around 60,000 ha, 20,000 ha of which being past or present wetlands. The archaeological programme investigated the presence of archaeological material through field walking, coring, geophysical survey, excavation and the systematic analysis of aerial photographs and information held in museums and Sites and Monuments Records. This was closely integrated with the palaeoenvironmental programme determining the extent and character of past and present wetlands in the Vale, and to investigate Post-glacial (Holocene) wetland development of the major rivers in the survey area. In addition, wetland deposits from Skipwith Common, Askham Bog and the Humber foreshore were examined.Archaeological Survey
The survey revealed a near absence of Mesolithic and early Neolithic sites and finds. This is surprising, as these have been observed in all other regions thus far surveyed in the Humber wetlands. There is, however, ample evidence for human presence and activity in the late Neolithic and Bronze Age in the Vale of York. The lighter, sandier soils attracted activity throughout later prehistory, and the Bronze Age is particularly well represented on the Humber foreshore. Here, the three Ferriby boats and associated structures from the intertidal wetlands at North Ferriby have been previously identified (Wright 1990), and the current survey recorded an additional 17 prehistoric waterlogged sites on the intertidal wetlands, between Brough and North Ferriby, including trackways, fishtraps and platforms, with some evidence of late Bronze Age to early Iron Age sites also represented. These sites illustrate the biological productivity of the estuary, with fishing and waterfowling being practised extensively. The trackways are thought to have been used for the exploitation of the saltmarshes for pasture. The construction of two of these trackways investigated during the survey accords well with this concept - the hurdle structures would have been flexible and very sturdy, while the stakes would have provided stability in anchoring the structure in a fluid environment.
Prior to the survey, the Iron Age period was represented by two cemeteries of square-ditched barrows on Skipwith Common, with one barrow having been recently excavated. Late Iron Age and Roman period settlement remains at Redcliff have also been identified previously (Crowther 1987), and Iron Age activity has been recognised throughout the Vale of York from aerial photographs, especially on the sandier soils. In the Holme-on-Spalding Moor area, a high concentration of Iron Age material associated with iron working has previously been recovered (e.g. Millett and Halkon 1988), and important finds from the wetlands of the former Walling Fen area include the Hasholme boat and various settlements (Millett and McGrail 1987). However, the distribution of settlements and sites of Iron Age date in this area is not reflected in the rest of the Vale of York. Roman period sites and material have been recorded from the Vale of York at Redcliff, Brough and Faxfleet which were all important sites for communication and transport to and from York, and a fort at Stamford Bridge and a settlement at Kexby are also known from the margins of the Vale of York. The main contribution of the current survey to the archaeology of the Roman period has been the discovery of the Roman settlement at Sutton upon Derwent. This settlement, dated to the second or third century, is considered to be part of the trading network. Its location on the north bank of the River Derwent, at the confluence with Pocklington Beck, the relative wealth revealed by the small-scale excavation and its morphology make the settlement stand out from others in the Vale of York. It may have acted as a stop-over for river traffic travelling to Roman period settlements at Kexby, Stamford Bridge, Crambeck, Malton and beyond, or towards Pocklington. At less than 10 km from York, it would also be a good crossing point of the Derwent and the Pocklington Beck for travellers to Brough or to the Humber.
Archaeological finds from the Early Medieval period are particularly rare, and the majority of information for this period comes from documentary sources or finds on the periphery of the region. By the eleventh century, even Domesday records a paucity of activity in much of the Vale of York, and no evidence of Early Medieval occupation has been found to be associated with the wetlands. This is surprising, considering the continuing importance of York in the Early Medieval period. The discovery of settlements of this date on the River Ouse may be inhibited by the presence of existing settlements, especially Selby.
The archaeology of the Middle Ages after Domesday shows renewed activity in the form of monastic settlements and no less than 114 identified moated sites. For the Humberhead Levels, it has been argued that moated sites are to be found in the areas with the heaviest soils which were not attractive to earlier settlers (e.g. Fenwick 1997), and such a spatial correlation can also be noted in the Vale of York.
Melton foreshore
Walkover survey of the Humber foreshore between Brough and Ferriby revealed a wealth of archaeological sites eroding out of the peat and clay outcrops. Of these sites, Melton-25 represents the largest of the excavated structures identified during the survey. The site was located 8m from the eroding edge of a peat shelf, within the tidal range on the north shore of the Humber. The survey to date has located a number of interesting sites, but this particular site represents rare, in situ structural remains. The alluvial clays, silts and peat deposits at Melton are already known to have produced archaeological structures (Crowther 1986) directly adjacent to this site. Prior to excavation only a portion of the site was visible. It appeared as several extant roundwood boughs that sat clear of the alluvium and were visible even above the detrital overburden. The excavation showed the structure to be a ‘hurdle’ constructed from a series of small similar-sized pieces of hazel branches (rods), woven around evenly spaced, larger cross-pieces (sails). The hurdle had the appearance of being a single unbroken panel, placed flat on the ground surface to form a ‘trackway’, with a downward slant to the east of 10°. The excavated section consisted of five north-south orientated sails, with a larger number (up to fifty) of east-west orientated rods roughly woven side by side around the sails. The western end had been badly eroded, and portions of the main timbers were missing, but the remainder of the buried wood was in remarkably good condition, with the manufacturing technique easily discernible.
Two separate samples of wood were sent for radiocarbon determination; both of the timbers were Corylus verticals, and neither sample had more than twenty year rings. The dates of 1510-1310 BC (3140±50 GU-5708) and 1440-1210 BC (3070±50 GU-5709) give a mean range of 1440-1310 cal BC, placing this trackway firmly in the Bronze Age. These dates corroborate the axemark analysis, which appears to represent bronze technology.
It appears that the hurdles were built vertically, in panels, constructed like a wattle fence (Raftery 1990), and then carried to the site and laid flat on the surface. The structure could therefore have been constructed on dry land and would have been light and easily transported to the saltmarsh environment, forming a quick and effective solution to the provision of access across more marginal zones. The evidence suggests that hazel in particular was being selectively chosen, with the small withies being used for weaving into hurdles and the more substantial pieces for hammering into the mud. It also suggests that there was an abundant source of long straight timbers in the locality, and the use of a simple woodland management system to provide a supply of these timbers cannot be ruled out.
The hurdle from the Melton track is not markedly different from other known hurdle trackways from both Britain and Ireland (e.g. Coles 1986, Raftery 1990), although these examples are more commonly found crossing raised mire or carr land. The fact that the Melton track was supported by parallel rows of flanking stakes, an uncommon feature for hurdle trackways, presumably reflects the different environmental conditions in which the trackway was constructed. Although none of the verticals was connected, it appears that they were used to keep the structure in place, with the hurdle having been attached or tied to the stakes in some way. No evidence was found for this at Melton but the theory is supported by similar discoveries in the upper Fergus estuary, Eire (O' Sullivan 1995). Here, the vertical timbers flanking the track had interwoven roundwoods forming a 'post and wattle' fence, and evidence of withies and ties was found.
Sutton upon Derwent
During field walking a large pottery scatter was located adjacent to the old course of the Derwent, to the south of Sutton Hall. Following the initial discoveries, a further program of intense field walking, GPS and geophysical survey was employed to determine the size and nature of the site.In order to establish the state of preservation over the area of the site and the adjacent wetlands, an excavation program was devised based on anomalies located by the geophysical survey. Two trenches were excavated in April 1998 revealing a number of linear features. All of the features excavated contained material which indicated that this was a large, possibly wealthy settlement of the Roman period, with some evidence of a later Medieval settlement.
The artefactual evidence recovered survives from the prehistoric through to the Post-Medieval period. The majority of the evidence comes from pottery, which supports a Roman date for the site. The pottery suggests domestic activity, but the presence of slag and fired clay suggests that some industrial activity was also taking place. The slag shows evidence for the working of iron nearby, possibly within the same complex. It is likely that this amount of material is from small-scale working for the domestic unit. The presence of vessel glass and a gold earring suggests a site of some prestigious nature. Brick and tile of Roman date is also present, and is possibly residual evidence for the presence of buildings within this unit. The presence of Medieval and Post-Medieval brick and tile help develop the ideas that buildings have survived at this location or nearby for some time after the decline of the Roman settlement.The evidence from Sutton clearly attests to the presence of an important Romano-British settlement of some size and importance. Although less than 1% of the estimated area of the site was excavated, it provided evidence of large-scale archaeological features, a huge variety of Roman pottery, tiles and glass vessels, and a gold earring. The excavation showed the potential for preservation of buried features, and for more high-value finds. Although organic deposits were not encountered during the excavation, there is potential for environmental reconstruction from buried peat horizons on the Derwent Ings and the Pocklington Canal.
The geophysical survey, field walking and excavation also provided evidence for Medieval occupation. Although this has not been properly explored during this survey, it would account for the overlapping of some of the features, and would suggest further potential for the site.
On the evidence of the excavation, the Roman occupation is relatively uncomplicated, with little evidence for multiple phasing, re-cutting or cross-cut features. Dating the site is, however, more problematic, with only one dateable sherd. A date of 230-370 AD is not unlikely and the possibility that this site was settled at some time in the second century would perhaps fit with the overall settlement pattern for the area. The date for this site could coincide with a small-scale climatic amelioration in the Vale of York, allowing Roman expansion on to the wetland margins, as seen elsewhere, for example at Riccall and Holme-on-Spalding Moor. The location of Sutton at less than 10 km south-east of Roman York (Eboracum) places it within the immediate hinterland, and combined with its location adjacent to the River Derwent and the Pocklington Beck, this suggests that the site is potentially very important in terms of Roman settlement in the region.
The potential for survival of evidence for either a Roman or Medieval period crossing of the Derwent, or more likely the Pocklington Beck, has also yet to be explored. With the Derwent Ings to the south and the marshy floodplain area of The Flags to the west the area immediately adjacent to the site could be expected to possess the capacity for wet preserved archaeology and organic remains.
Palaeoenvironmental survey
The palaeoenvironmental programme focused on the main rivers in the Vale of York, with additional palynological work undertaken at Askham Bog, to the south of York, and at Skipwith Common. The rivers in the Humber wetlands were created in the early Post-glacial (Holocene) period, and the development of floodplain wetlands was a response to sea-level rise during the Holocene. The research focused primarily on the timing of this development, and has demonstrated its time-transgressive nature.The floodplain of the River Aire was investigated at Carlton Marsh and Birkin, completing the work previously undertaken in the Humberhead Levels (Lillie 1997). Wetland development in the lower Aire floodplain started at c. 5000-4000 cal BC with the expansion of fen carr environments. The floodplain of the River Wharfe was investigated to the east of Ryther and Nun Appleton at East Ings and the floodplain of the River Ouse was investigated at Faxfleet, Asselby Island, Selby and Stillingfleet. At Asselby Island, the earliest date for wetland development in the Ouse floodplain was 5240-5060 cal BC, and the maximum extent of wetland development was reached by 2180-1920 cal BC.
The floodplain of the River Derwent was investigated at Asselby, at Ellerton and on the Wheldrake Ings. South of Loftsome Bridge, the river has been diverted to the south-west to follow a canalised route, entering the River Ouse to the north of Long Drax. The course of the palaeo-Derwent follows a south-easterly route to the north of Asselby and Knedlington, underlying Howden and entering the River Ouse opposite Hook at Kilpin Pike. Palynological and radiocarbon analyses of the floodplain peats of the palaeo-Derwent at Asselby, the lowest reach investigated, suggest the onset of biogenic sedimentation from c. 5000 cal BC, with floodplain peat development continuing to post-2000 cal BC, that is from the Mesolithic to Bronze Age periods.
The River Foulness/Walling Fen area was investigated by transect coring across the Ellerker Sands. Pollen analysis suggests the earliest date for the onset of peat formation early in the Neolithic period. A birch and alder fen-carr is attested in the wetter areas of the floodplain, with hazel, oak and possibly lime on the drier fen margins. The upper temporal limits of wetland development are placed at some time in the post-Roman period.
The intertidal exposures of alluvium and peat on the Humber foreshore consist of a large expanse on either side of East Clough, including the Melton area from where many prehistoric archaeological sites were recorded, and a smaller exposure immediately south-east of North Ferriby. Radiocarbon analyses of the peat exposures at East Clough have shown that wetland development on the Humber floodplain had commenced by 2448-2041 cal BC (Long et al. 1998). Pollen analysis of a basal peat sample at East Clough reflected the presence of oak, alder and willow, with hazel growing near the sampling site and with a mosaic of intertidal vegetational zones including areas of open water, tall herb and saltmarsh communities, as well as open meadow areas.
Pollen analysis of the top of the peat sample at East Clough shows an assemblage indicative of oak and lime dominated woodland. Some limited open grassland is also attested, but no evidence for the encroachment of saltmarsh environments was recovered from this sample, and the presence of 'bog-oaks' suggests that the peat surface was drying out. A date of 3000-2000 cal BC is suggested by the pollen spectrum.
Skipwith Common is one of two isolated wetland areas in the Vale of York investigated as part of the current survey. Coring found limited remnants of peat which are fragments of deeper deposits cut-over since at least the fourteenth century. Pollen analysis of the basal peat from one of the hollows, which now remain as shallow lakes, indicates a predominantly open landscape around the sampling site, with areas of birch and sedge dominant locally. The subsequent appearance in the samples of hazel and the other forest trees (elm, oak, pine, lime and alder) represents a typical sequence of woodland spread in lowland Britain in the Late-glacial and early Holocene, c.10,000 to 7500 BP, which was also identified at Askham Bog. The top of the sequence was not recovered due to its high water content, but the presence of a lens of watery peat above early Holocene deposits is typical of re-growth in valley peats that have been cut-over (e.g. Gearey 1996). The peats from Skipwith Common therefore hold no information on the human impact on vegetation in the Bronze and Iron Age periods. As at Skipwith Common and in contrast to the river floodplains, the sequence from Askham Bog demonstrates that wetland formation began in the early Holocene, c. 9000 BP, and continued until possibly the early historic period.
The work at Skipwith Common and Askham Bog shows that isolated wetlands existed in the Late-glacial period and, in the case of Askham Bog, survived largely intact into the late Holocene. Detailed chronostratigraphic and biostratigraphic analysis of the deposits provides an overview of the vegetation history for that part of the Vale of York, and evidence for up to four phases of forest clearance, presumably representing the impact of anthropogenic activity on the environment, before the fifth and final phase of extensive forest clearance. Unfortunately, problems with the high-resolution radiocarbon dating of the upper part of the sequence prevented absolute dates for these phases of clearance from being obtained.
