13
Alluvial Geoarchaeology in Ireland
Abstract
Due to its predominantly low relief and high rainfall a large proportion of the area of midland Ireland is characterised by a patchwork of lakes, raised mires and alluvial floodplains. This provides the potential for multi-proxy and multi-scale environmental reconstruction. The linkage of lakes, bogs and alluvial sequences has the potential to provide less equivocal information on key questions in Irish archaeology including; the transition from foraging to farming, regionality, ritual use of the landscape and climatic impact on prehistoric activity. This paper demonstrates the potential of this approach through several case-studies. These include studies of raised mire and alluvial data from the Little Brosna Valley in the middle Shannon Basin, the Liffey Basin in Co. Kildare and the Lee in Co. Cork. These, along with selected other studies illustrate both the relative lack of archaeological excavations in Irish floodplains and the potential that exists herein. The conclusion reached is that, despite a lack of geoarchaeological excavations, the geology and climate of Ireland has produced and preserved large areas of buried floodplain and alluvio-lacustrine landscapes and also provides an almost unique opportunity to compare alluvial geoarchaeology with other data sources pertinent to the making of the Irish landscape.
Introduction: The Geological and Geomorphological Context
Ireland is a low, eroded landscape dominated by lakes, bogs and rivers. Its highest peak, Mount Carrentuohill in the Macgillycuddy’s Reeks, reaches only 1041 m above sea level, whilst the majority of the Irish landscape lies beneath 200 m above sea level. Geologically, Ireland is old, largely composed of Palaeozoic rock cut by Mesozoic and Tertiary palaeosurfaces upon which there is a discontinuous covering of Quaternary glaciogenic sediments. The hydrogeomorphic result is that most drainage is surficial, except on the limestone outliers, and, except on the mountain blocks, rivers are of low slope with Quaternary deposits being the most erodible sediments. Many of the lowland rivers are really glacially cut channels linking staircases of lakes which themselves are small remnants of earlier, much larger lake systems. The climate, which is oceanic and dominated by persistent low intensity rainfall, provides the ideal conditions for peat accumulation. Upland areas with more than 1250 mm rain per year carry blanket peat, whilst in the lowlands, hydromorphic and topographic basins often carry raised ombrogenous bogs. The result are rivers with high discharges per unit area, but with relatively un-flashy regimes, although peat-dominated catchments have moderately flashy regimes.
Although the postglacial ecological and climatic changes will have been common to all Irish rivers, the response to such changes will have varied from catchment to catchment because of their great diversity in geomorphological properties (Glanville et al. 1997). On the basis of criteria such as rainfall and discharge, topography and stream power and the nature and availability of surface materials, it is possible to suggest there may be four broad classes of river basins in Ireland:
- Those of high rainfall and high power and low sediment availability, characterised by the western mountains and epitomised in Donegal.
- Those with moderate rainfall and moderate power, moderate to high sediment availability and characterised by upland and hilly areas of glacigenic sediments such as the Lee, Blackwater and Foyle.
- Those with moderate to low rainfall, low energy but moderate to high sediment availability characteristic of the glaciated lowlands and exemplified perhaps by the Suir, the Liffey and the Boyne.
- Those with moderate rainfall, low energy, negligible clastic sediment availability but with extensive mires and peatlands, such as the lower Shannon.
Only Type (2) and (3) can be expected to contain alluvial records of environmental change and the accessibility of such records may be best in Type (3).
A review of the Quaternary geology of Ireland has recently been provided by Coxon (2001). The key factor for Holocene alluvial development is the coverage of much of the island by unconsolidated, generally coarse sediments (sands and gravel) of glacial and glaciofluvial origin and a topography largely fashioned by the glacial processes of erosion, deposition and periglacial processes. These factors delimit the degrees of freedom of Irish fluvial systems and therefore their development since human arrival. When humans arrived in the Early Mesolithic, Ireland was largely forested and rivers would have been the main arteries of transport–indeed, they remained so until relatively recently.
A variety of field methodologies have been employed in the studies reviewed in this paper, however, most follow standard geoarchaeological methodologies based on survey, coring and palaeoenvironmental analyses. Dates are quoted as in the original published text.
The History of Alluvial Studies in Ireland
The first systematic observations on alluvial deposits can be traced back to the founding of the Geological Society of Ireland in 1831, the geological mapping by the father of Irish geology Richard Griffiths, and the activities of the Geological Survey of Britain and Ireland which had been founded in 1845. A persistent problem was the representation of both the solid and drift (Quaternary and Holocene) on one sheet (Herries-Davies 2001). Therefore Irish geology has always required an interest in the Quaternary and so it is unsurprising that Ireland has produced or attracted outstanding figures in the field of Quaternary Science such as Charlesworth, Farrington, Praeger, Jessen and G. F. Mitchell. Many of these scientists had an interest in archaeology and most particularly G. F. Mitchell (1912–1997), who combined these interests in his classic Reading of the Irish Landscape, first published in Ireland in 1986. In many ways G. F. Mitchell could be regarded as the father of Irish geoarchaeology, although in comparison to Britain, alluvial geomorphology has, until recently been an under-researched area; however, recent studies of river behaviour and alluvial stratigraphies include Croke (1991; 1994), Gallagher and Thorp (1995), Glanville et al. (1997) and Thorp and Gallagher (1999).
Lakes, Rivers, Bogs and Archaeology
At the end of the Last Glacial period, Ireland presented a landscape of bare tundra mountains surrounded by marshy, deglaciated lowlands with moraines, eskers, outwash and meltwater complexes and often poorly integrated river systems. There were a large number of lakes of varying size, which had resulted from the impoundment of meltwater. Major lacustrine basins included Lough Boora-Ree, Lough Derg, Lough Erne and Lough Allen, Lough Gara, Sheelin, and Derravaragh in the Shannon system, as well as Lough Neagh in the north (Fig. 13.1). These large lake systems still existed when Ireland was occupied in the Early Mesolithic and would probably have been accessed by river. The shores of these lakes and rivers were of particular importance to Early Mesolithic forager-gatherer-hunters living in a largely wooded environment. Unsurprisingly, Mesolithic sites such as Mount Sandel, Co. Londonderry, detail the exploitation of freshwater resources such as salmon, trout, eels and even the seeds of the White Water Lily (Nymphae alba) (Woodman 1985). The greater aquatic dependency of the Irish Mesolithic is clearly a function of the lack of native mammals, including the Red Deer (Cervus elaphus). The distribution of sites of the Early Mesolithic type (Sandelian) shows a marked bias to major river valleys, lakes and the coast (then an inland cliff) with the densest concentrations around Lough Neagh and down the Bann Valley, the coast north of the Boyne Valley and along the Blackwater Valley in Co. Cork (Anderson 1993). There is little evidence to suggest this is a reflection of sampling bias and recent studies in Denmark have shown that differential sampling typically alters the absolute densities of sites but not relative differences across the landscape (Odgaard and Rasmussen 2000). It would be more realistic to regard the geoarchaeological context of the Irish Early Mesolithic as alluvio-lacustrine in nature.
Figure 13.1: The major rivers of Ireland and the study areas discussed in the text. The boxes are the River Lee/ Gearagh study area (Brown et al. 1995), the Barrow (Zvelebil et al. 1996), the Shannon (Hooyer 1991; Vader 1993; Aalbersberg 1994) and the Liffey (Thorp and Glanville 1999).
Sites of Later Mesolithic date, on the other hand, whilst still clustering around major valleys and lakes, appear to have been more widely dispersed. In a recent survey of the Lough Swilly region, Co. Donegal, Kimball (2000) notes a bias of Later Mesolithic sites towards estuarine and lowland floodplain ecozones. However, there remains a major visibility problem caused by the expansion of peatlands in the Middle Holocene (Littletonian period). Early landscapes are buried, such as the Neolithic field systems of Co. Mayo (Caulfield 1983), by expanding blanket bog and raised mires, and evidence of later activities entombed. This expansion of peatlands even constrained some rivers (see The Shannon System below) and thus the area of alluviation and so, although indirectly, influenced the location and local pattern of early farming. The oldest sites interstratified by alluvial sediments have been found to date from the Later Mesolithic-Early Neolithic (c. 6000–5000 cal. BC). Historical examples include the sites interstratified with diatomite in the Bann Valley excavated by Hallam Movius in the 1930s (cited in Mitchell and Ryan 2001). In the Neolithic, large lake systems linked by rivers still existed and travel may well still have been predominantly by water as illustrated by finds such as the 15 m long Lurgan logboat found in Addergoole Bog near Tuam, Co. Galway, in 1902 and now on display in the National Museum, Dublin.
Recently there has been a re-evaluation of Irish linear monuments and the recognition of cursus-type monuments (Newman 1999). Cursus monuments appear to be particularly associated with river valleys and water, with terminations close to the river or tributary stream–a factor some have implicated in the interpretation of their meaning (Barclay and Hey 1999). The most striking Later Bronze Age and Iron Age aspect of alluvial environments appears to be the ritual deposition of artefacts and occasionally skulls in rivers as well as lakes. The record may be particularly rich for Ireland due to a relative lack of channel change in the Later Holocene (as discussed later). In the Early Christian era, both lake and larger riverine islands become favoured sites for ecclesiastical establishments including hermitages, chapels, monasteries and convents. This variable association with lakes and rivers provides intellectual justification for geoarchaeological studies.
Recent Palaeoenvironmental and Geoarchaeological Studies
The Shannon System
The Shannon is an unusual European river being a series of channels linking lakes (Fig. 13.2). Draining much of midland Ireland, and although only 205 km in length, the Shannon is the largest river in Ireland. The longest alluvial section, the Middle Shannon, connects Lough Ree to Lough Derg (60 km) and is characterised by a floodplain of bogs and seasonally flooded lands called callows (Heery 1993). The callows are of international importance for wildfowl, harbouring several endangered species including the Corncrake as well as 17 rare plant species (Heery 1993). This ecological heritage, along with its vulnerability has been recognised by the European Union (EEC-STEP Programme; Hooyer 1991; Vader 1993). This ecological value has originated through the interaction of natural alluvial hydrological and sedimentary processes and human activity. The channel of the Middle Shannon repeatedly bifurcates around alluvial islands–probably a residual element of an earlier multiple-channel (anastomosing) form. This is unsurprising given the extremely low slope and fine calibre sediment supply (Smith and Smith 1980; Nanson and Knighton 1996). Indeed, anastomosing channels seem to have been far more common in the lowlands of north-west Europe before Late Holocene alluviation and channel engineering (Brown 1997). On the valley floor are a number of raised (ombrotrophic) mires, some of which constrain the alluvial floodplain. Both the floodplain and bogs have developed in between several esker ridges orientated across the valley, especially well-developed south of Athlone. The bogs result from the low relief, ubiquity of glacial basins and high rainfall distributed throughout the year. The stratigraphy and developmental history of the callows has been revealed in a number of ditch sections and core transects. The general stratigraphy consists of basal sand and gravels of glaciofluvial origin, grey lacustrine silts and clay capped by a sand unit, white lake marls beneath peats and, in some areas especially close to the river, superficial units of clay, silt and fine sand.
Detailed work in the Little Brosna Valley (Fig. 13.3) by Aalbersberg (1994) suggests the basal lake clays are of Late-Glacial age with the sand unit marking a sudden drop in lake level at the beginning of the Holocene. The white lake marl is highly calcareous containing abundant shells as well as ostracods, diatoms, Characeae (calcareous algae), insect and plant remains. The lakes were shallow and highly productive, although oligotrophic, due to a lack of nitrogen. The upper stratigraphy of the Little Brosna Valley, and large areas of the Shannon callows, shows hydroseral development with lakes being succeeded by reedswamp, sedge/fen, wet woodland and in some places Sphagnum dominated raised mire. Typical Brosna cross-sections reveal fine sands, silt and clay feathering out over the peat in the channel/levee zone. This stratigraphy reveals a high degree of channel stability during the period of peat growth and alluvial deposition. This alluviation is regarded by Aalbersberg (1994) as being the result of deforestation and cultivation in the catchment which contains the Slieve Bloom Mountains. In some places, such as Pollagh Bog on the Shannon, flood layers interdigitate with Sphagnum peat (Heery 1993). In other areas, the upper stratigraphy is locally variable, such as at Clonmacnoise where the peat accumulated to 3 m without evidence of alluviation, whilst only 9 km to the south at Bishops Islands the peat is covered by 1 m of silt and clay (Turbridy 1987).
The stratigraphic work by Aalbersberg (1994) in the most downstream reaches of the Brosna has provided a chronology and record of environmental conditions during mire and floodplain development. He suggests that the shallow calcareous lake terrestrialised to Cladium-Phragmites swamp with a ‘diffuse’ multiple-channel river pattern around 6200 BP (5300–5000 cal. BC). Subsequently, secondary channels silted leaving a single channel system until a second phase of anastomisis began around cal. AD 1200 and a final contraction to a largely single channel system around cal. AD 1750 (Fig. 13.4). Palaeoecological analysis of Lusmagh Bog in the Brosna Valley reveals a phase of local burning prior to the elm decline after which occur the first large-scale anthropogenic impacts on the vegetation. The pollen diagram reveals a strong increase in herbs and some cereal-type pollen grains in zone BPLMBPZ-4, c. 4300–4150 BP (3100–2500 cal. BC). There is also abundant proxy-palaeoclimate data from the Brosna/Middle Shannon area. From Lusmagh Bog itself, there are five recurrence surfaces which Aalbersberg (1994) has correlated with the continental sequence (RYV-RYI) and in nearby Annagh Bog there are four recurrence surfaces (RYIV-RYI). Studies by Barber et al. (2003) at Mongan Bog near Clonmacnoise suggest the wettest period in the records occurred c. 2650 years BP (830–800 cal. BC) and c. 1000 BP (cal. AD 990–1030) with dry periods c. 5500 BP (4360–4260 cal. BC) and 1800 BP (cal. AD 130–320). The 830–800 cal. BC date may correspond with the Iron Age/Bronze Age (Subboreal/Subatlantic) climatic deterioration associated with reduced solar activity at the start of the so-called ‘Homeric Solar Minimum’ (van Geel et al. 1996). A site to the east of the Shannon catchment in the headwaters of the River Boyne, Carbury Bog, displays an abrupt appearance of the oceanic species Sphagnum imbricatum which has been dated by AMS wiggle matching at 850 cal. BC and also coincides with the sharp rise in δ14C (van Geel et al. 1998).
Figure 13.2: Map of the Shannon catchment showing sites mentioned in the text (adapted from Hooyer 1991). The box is the study reach of the River Little Brosna (Aalbersberg 1994).
Figure 13.3: A geomorphological map of the lower Little Brosna river valley (adapted from Aalbersberg 1994).
The landscape archaeology of the Shannon Valley has, until recently, received less attention than its palaeoecology. Prehistoric structures, particularly barrows, occur on the alluvial callows, and occasionally on present and past alluvial islands, as at Long Island near Clonown, a location that may have had special ceremonial significance (Brown 2004). In relation to prehistoric domestic archaeology, the potential of the area is illustrated by the excavations by the Irish Wetland Unit at Clonfinlough, Co. Offaly (Maloney et al. 1993; O’ Sullivan 1998). Clonfinlough was an enclosed farmstead built in wetlands on the south side of a small lake. It lies just to the east of the River Shannon, from which it is separated by an esker ridge upon which there is evidence of Bronze Age activity (O’Sullivan 1998). The site is palisaded with an associated trackway and was constructed between 917 and 899 BC (dendrochronology date). The site has excellent organic preservation and has yielded two boat paddles presumably for use on the Shannon.
Recent excavations of the Clonmacnoise Bridge (O’Sullivan and Boland 2000) have dated its construction via dendrochronology to AD 840, making it the oldest recorded bridge in Ireland. Associated with the bridge were 11 dugout boats, axes, a copper-alloy basin and many other industrial and agricultural artefacts. The bridge was constructed of 50 oak posts fitted with base-plates to prevent further subsidence into the clay riverbed. Further analysis may reveal environmental details but it would appear that the location, width and depth of the river were similar in the ninth century as it is today. The historical stability of the planform of the Shannon (and probably other Irish lowland rivers) is probably due to the combination of low slope and therefore energy and high resistance from clay and peat riverbanks. The archaeological result is a continuity of landscape affording high potential for studies of alluvial landscape change due to a high probability of long-term preservation of artefacts and environmental records.
Figure 13.4: Sumary didagram of the developmental phases and palaeoecological information from Lusmagh Bog. Little Brosna Valley (adapted from Aalbersberg 1994).
Figure 13.5: A map of the Liffey catchment with sites mentioned in the text.
The River Liffey
The River Liffey, sourced in the north-west Wicklow Mountains and, flowing west, north and then east to Dublin, passes through at least four distinct reaches–in the mountains, the now flooded Blessington Basin, through the intensively farmed lowlands of Co. Kildare underlain by a complex of deglaciation tills and outwash sands and gravels, and finally the strongly urbanised reach from Leixlip to the sea (Fig. 13.5). In all reaches, save those above 300 m, settlement and agriculture have been almost continuous since Neolithic times, but of the river behaviour very little is known save for the Dublin city and the mountain reaches. The following account relates entirely to the mountain reach where the effects of environmental controls and human activities can, in part, be analysed.
The 100 km2 upper River Liffey catchment, on the north-western side of the Wicklow Mountains, is defined by 600–850 m encircling ridges. Long valley side slopes sweep smoothly down to narrow valley floors which lie at 220–350 m. The underlying granite rocks have provided the glacial till smearing the lower slopes and valley floors. Along the valley floors, the Post Glacial rivers have eroded a narrow pan-shaped trough countersunk 2–5 m into tills, moraines and outwash trains. This trough contains Holocene alluvial sediments which have been sourced mainly, but not entirely, from the glaciofluvial deposits. There are three sub-basins to the catchment wherein the valley floors are sufficiently wide to contain alluvial floodplains–the Coronation Plantation, the Ballydonnell and the Athdown Basins. Between these, the valley floor passes through gorge-like narrowings.
Today, blanket peat covers the ridges and valley sides but in the lower part of the valley, where slopes are gentler and the till thicker, farmland and coniferous forests are extensive. Here and in the adjacent Blessington Basin there is some indirect archaeological evidence for Neolithic and Bronze Age settlement and more direct evidence for settlement from the Late Iron Age and onwards. Above 300 m in the upper basins, however, the harsh climate and small extent of mineral soils have always limited farming to several small areas.
Figure 13.6: A schematic cross-section of the terraces of the River Liffey after Glanville et al. (1997).
Pollen analysis (Glanville 1999) of blanket peats and peaty soils within the Holocene alluvial terraces show that the mountain catchment, between c. 7000 and 2000 cal. BC, was dominated by closed forests (initially oak, elm, pine then alder and birch) and that blanket peat did not begin extending until after 2000 cal. BC, although it had formed in wetter valley floor locations by 5000 cal. BC and on high plateaux by 7000 cal. BC. After 2000 cal. BC, the vegetation began changing as woodlands were opened up for agriculture, grasslands expanded, Ericaceae appeared and blanket peat began spreading downslope and thickening. Between 500 cal. BC and cal. AD 1000, a vegetation pattern not unlike the present day, but without the conifer plantations, had developed–small areas of birch, alder and oak woodlands and scrub, large areas of grasslands and pastoral agriculture with limited cereals, and extensive blanket peat on the upper slopes and on valley floor alluvia. ‘Intensive’ agriculture appears to have been located on the more freely draining tills and outwash deposits on the lower slopes in the Athdown Basin and on the limited alluvial terraces in the upper parts of the catchment.
Stout (1994) has summarised what little is known of the archaeology in the north-west Wicklow Mountains. Whilst no evidence has been found in the upper part of the mountain catchment for pre-Medieval settlement, settlement of Neolithic and later date is suggested from the Athdown area and especially from the Blessington Basin reach of the Liffey. The most direct evidence for settlement in the mountain catchment, however, comprises ring forts of supposed Early Christian age in the Athdown Basin.
In the valley floor widenings of the mountain reach there are two Holocene alluvial terraces flanking the current floodplain (Fig. 13.6; Glanville et al. 1997). These have been extensively dated using radiocarbon analysis and dendrochronology and by pollen spectra correlations with dates on pollen ‘master’ curves. The oldest terrace spans 5800–100 cal. BC whilst the younger is dated from c. 100 cal. BC–cal. AD 1600. Both terraces comprise several layers of sandy alluvium separated by palaeosols and the floodplains appear to have been constructed by episodic overbank flows and vertical accretion. Construction of the first floodplain was terminated by one or several major flood events c. 100 cal. BC when parts of the floodplain were stripped down to the underlying coarse gravel sheet that is present under all Holocene alluvia in the valley floors and when the river bed was scoured into this gravel by c. 5 m. Floodplain rebuilding over the wide, exposed gravel bars followed. A similar event or group of flow events terminated the second floodplain between c. cal. AD 1450 and 1800. Floodplains built post c. cal. AD 1600 comprise numerous thin laminae of coarse sand and peat laid down on a quasi-annual basis. This sedimentation style is associated with the onset of blanket peat gullying on the interfluves during the ‘Little Ice Age’ and of accelerated peat fuel cutting after c. cal. AD 1750 (Thorp and Gallagher 1999).
Rates of sediment accumulation on the floodplains began to increase after c. 2000 cal. BC at an exponential rate, stabilising somewhat between cal. AD 400 and 1100 after which the rate dramatically increased. The initial increase in sedimentation rate seems roughly coeval with both the beginnings of vegetation modification by people and with climatically induced soil paludification and blanket peat expansion. Undoubtedly, the several sedimentation events that constructed the two floodplains were caused by rainfall-run-off events but the sediment yield may reflect, at least in part, an increasing access by run-off to soil and sediments induced by growing agricultural activity, both spatially and intensively, from the Bronze Age onwards. The two incision events, on the other hand, occurred when specific flood discharges exceeded the sedimentation–erosion threshold, a threshold which had itself been shifted by the increasing floodplain height. This was such that the next overtopping flow was so powerful as to have induced massive channel erosion rather than additional overbank sedimentation. The floods of ‘Hurricane Charlie’ in 1986 crossed analogous thresholds and produced similar effects to those of c. 100 cal. BC and c. cal. AD 1600–1800.
Post cal. AD 1750, settlement and agricultural activity in the mountain reach of the Liffey has concentrated on the free draining gravel-dominated tills and outwash sediments veneering the lower hillside slopes and it may be reasonable to assume that previous farming generations made similar resource judgements; the terraces are generally devoid of signs of prehistoric settlements. However, in an exposure of the Mid-Holocene terrace at Kilbride, a circular hearth was seen with indications that it was used to fire crack blocks of vein quartz to make crude stone tools. Samples of the charcoal gave dates of 4043 and 3951 cal. BC, indicating a Neolithic age. The most extensive area of floodplain terrace surface is from the mid-Holocene in the Athdown Basin. However, soil paludification and hydromorphic peat development commenced on this surface c. 700 cal. BC and thereafter put this area beyond intensive use. In the upper part of the catchment the small terraces may have been mainly used for grazing although in the nineteenth century the mid-Holocene terrace was cultivated by the three farms which then existed there.
It would appear that prehistoric settlement in the mountain Liffey catchment was not on the floodplains, which were of limited spatial extent and often wet, but on the adjacent well-drained valley footslopes. However, in the higher parts of the catchment in the areas marginal to settlement, the small floodplain terraces were perhaps the only sites for habitation. Enhanced sediment yields from cultivated and habited lower valley sides seem to have contributed to floodplain formation after c. 2000 cal. BC. It is difficult, however, to judge a discernable distinct effect upon run-off and river discharge of colonisation although it would be reasonable to assume run-off coefficients were enhanced by the reductions in vegetation bulk mass. There is some suggestion in the alluvial stratigraphies that the increase in sedimentation rate after c. 2000 cal. BC was caused mainly by increased frequency of overbank sedimentation events. This is certainly the case after c. cal. AD 500. The latter relates primarily to the development of peat erosion systems and to increased storm frequencies during the Little Ice Age but it is not clear what the relative roles of natural climate moistening and human created vegetation modifications were between those two periods.
The Lee Valley and the Gearagh
The River Lee in south-west Ireland drains the Boggerach, Derrynasaggart and Shehy Mountains of west Cork (Fig. 13.7). Its source is a glacial lake at Gougane Barra and it flows 70 km to the sea at Cork through a series of reservoirs. Until 1954, a large tract of the floodplain upstream of Macroom was covered by a dense pattern of hundreds of wooded islands divided by small channels. This area of anastomosing river, called the Gearagh, was of international biological importance. A small surviving fragment has recently received attention from a bio-geomorphological perspective as it is one of the few wooded anastomosing reaches left in the British Isles and Ireland (Harwood and Brown 1993; Brown et al. 1995). More recent studies have also produced evidence of human activity in and around the floodplain. The upper Lee catchment is rich in standing archaeology with a dense cluster of wedge tombs (O’Brien 1993), stone rows and stone circles (Walsh 1993). The rows and circles are particularly common on the north side of the catchment and especially along the Foherish tributary. A stone circle sits on a spur of terrace gravels overlooking the Gearagh. Some palaeoenvironmental data is now available for this area. During the Late- Glacial, the River Lee bifurcated around a bedrock knoll in this reach leaving two tracts of outwash gravels. Only the most northerly of these tracts became the Holocene floodplain, whilst the southerly tract developed into raised mire. Pollen analysis of Annahala Bog, only 1 km from the Gearagh, reveals little or no signal of the alluvial woodland, probably due to the dominance of the mire and regional pollen signal. Stratigraphic investigations within the Gearagh have shown that the islands are typically composed of a felted organic-rich peat upon which lie the superficial sandy silt-clays (Fig. 13.8). Dating from two islands indicate that island formation began around the thirteenth–fourteenth centuries AD and may have been caused by the confinement of the multi-channel river by a southerly flood embankment. More work is needed in order to define the relationship between channel alterations, land use change and the geomorphological development of the River Lee.
Figure 13.7: A map of the upper Lee catchment showing archaeological sites and the location of the Gearagh.
Figure 13.8: Stratigraphic sections across islands in the Gearagh.
Other Alluvial-related Studies
Interdisciplinary studies in the Barrow Valley, south-east Ireland, have been aimed at reevaluating the so-called Riverford Culture (Zvelebil et al. 1996). Stratigraphic studies, particularly in the Athy Basin (Fig. 13.9) have revealed two periods of widespread alluviation, the first in the Early Holocene c. 8000–6000 cal. BC and the second after c. 2000 cal. BC. The dominant controls on the later period are not clear although based on earlier radiocarbon dating of an oak, Mitchell (1986) had speculated that Bronze Age farming could be implicated. However, as Zvelebil et al. (1996) point out c. 2000 cal. BC is also a period of climatic deterioration in Ireland and Britain. Adjacent studies of raised mire records could be used to explore these factors further. The study also has a strong taphonomic aspect revealing how the distribution of Mesolithic and Neolithic flints is biased by the sedimentation history of the basin.
Another illustration of the alluvial geoarchaeological potential in Ireland can be taken from the prehistoric monumental complex in the Boyne Valley. The Neolithic monuments of Newgrange, Knowth and Dowth are some of the most researched sites in Ireland. Studies have also expanded to include the surrounding landscape (e.g. satellite tombs). Cooney (2000) uses the indications from on-site pollen and seed studies and the character of the local area (topography, soils etc.) to produce a hypothetical land use map. There are also monuments on the valley floor (e.g. satellite tombs and a henge) and both the topography and aerial photographs suggest a palaeochannel to the north of these sites and the present channel. The use of palaeochannels for deriving local land use and land management data has been neglected in Ireland almost certainly because of the wealth of potential sites from raised mires. However, the palaeo-data that maybe derived from palaeochannels, although of local origin, is almost certainly more sensitive to clearance, management and farming on the floodplain, terraces and surrounding slopes and hence more valuable at the spatial scale of landscape archaeology (Brown 1999). The Irish Discovery Programme’s current Lake Settlement Project will hopefully go some way towards exploiting this alluvial potential as many lake sites are located at the junction of rivers with lakes, or in or by lakes which exist within alluvial landscapes (see Whitehouse, this volume).
An example of such an area is Lough Derravarragh, Co. Westmeath, where excavations and exposures at Clonava 1 have revealed Neolithic chert-knapping (Mitchell 1972; Woodman 1978). The landscape:
Figure 13.9: An annotated cross-section of the Athy Basin of the Barrow Valley (adapted from Zvelebil et al. 1996).
‘was a stretch of fen beside the lake, which had been dissected by numerous channels as the river Inny entered Lough Derravarragh’ (O’Sullivan 1998, 49).
Mesolithic scatters have also been found at Clonava which, during the Mesolithic, was a large island surrounded by a large area of wetlands. Indeed there appears to be a bias to the location of Mesolithic finds towards lake islands, promontories and lake-river junctions. Whilst these locations offer both good visibility and a variety of resources they are also the most accessible by boat. The use of islands is particularly significant and geoarchaeology has considerable potential in relation to the identification, delimitation and characterisation of past alluvial, alluvio-lacustrine and lacustrine islands. This is not simply a contextual matter as it has been argued that the location of some types of archaeological sites on islands, particularly Early Christian sites, may have far more than functional significance (Brown 2004).
Most upland rivers in Ireland are small and apart from the Liffey have received little geomorphic attention. However, some work has been done on the alluvial fans of the Macgillyguddy’s Reeks in south-west Ireland by Anderson et al. (2000). Radiocarbon dating shows episodic Holocene aggradation and incision with aggradation clustering into two periods, cal. AD 230–790 and cal. AD 1510–present. The authors correlate these phases with enhanced valley alluviation in northern England implying an overall climatic control. However, they also report that pollen analysis of peats interbedded in the fans indicate that land use change (probably overgrazing) may have been important in reducing the threshold of slope stability and so increasing aggradation during intense rainstorms.
Conclusions: The Archaeology of River Valleys and Geoarchaeological Potential
The alluvial geoarchaeology of Ireland has yet to be seriously evaluated. However, the potential is high in the lowlands with a dominance of vertical sedimentation and bog development over erosion and fluvial reworking. The low-energy environments of most Irish floodplains have almost certainly entombed abundant evidence of alluvial landscape change, which awaits excavation. There are, however, obvious problems concerning high water-tables, but there are also unique opportunities through the combination and linkage of alluvial, lacustrine and mire based records of environmental change and the human creation of the Irish landscape. This endeavour need not be seen as entirely empirically driven as serious archaeological questions may be tackled including the location and nature of early prehistoric activity and mobility, the role of water in later prehistory and the relative role of human activity and climate in both peat growth and erosion, and in alluviation.
Acknowledgements
The authors must thank several researchers for assisting in the collation of information including M. Macklin, A. O’Sullivan, K. Barber and P. Hughes. We must also thank S. Roullard for drawing several of the diagrams.
References
Aalbersberg, G. 1994. The Little Brosna River Valley: Quaternary Geology and Palaeo-ecology. Unpublished M.Sc. Dissertation, Free University, Amsterdam, 2 vols.
Anderson, E. 1993. The Mesolithic: fishing for answers, pp. 16–24 in Twohig, E. S. and Ronayne, M. (eds.), Past Perceptions: The Prehistoric Archaeology of South-West Ireland. Cork: Cork University Press.
Anderson, E., Harrison, S., Passmore, D. G. and Mighall, T. M. 2000. Holocene alluvial-fan development in the Macgillycuddy’s Reeks, southwest Ireland. Geological Society of America Bulletin 112, 1834–49.
Barber, K. E., Chambers, F. M. and Maddy, D. 2003. Holocene palaeoclimates from peat stratigraphy: macrofossil proxy-climate records from three oceanic raised bogs in England and Ireland. Quaternary Science Reviews 22, 521–39.
Barclay, A. and Hey, G. 1999. Cattle, cursus monuments and the river: the development of ritual and domestic landscapes in the Upper Thames Valley, pp. 67–76 in Barclay, A. and Harding, J. (eds.), Pathways and Ceremonies: The Cursus Monuments of Britain and Ireland. (Neolithic Studies Group Seminar Paper No. 4). Oxford: Oxbow.
Brown, A. G. 1997. Alluvial Geoarchaeology. Cambridge: Cambridge University Press.
Brown, A. G. 1999. Characterising prehistoric lowland environments using local pollen assemblages. Quaternary Proceedings 7, 585–94.
Brown, A. G. 2004. Divisions of floodplain space and sites on riverine ‘islands’: functional, ritual, social, or liminal places? Journal of Wetland Archaeology 3, 3–16.
Brown, A. G., Stone, P. and Harwood, K. 1995. The Biogeomorphology of a Wooded Anastomosing River: The Gearagh on the River Lee in County Cork, Ireland (Occasional Papers in Geography No. 32). Leicester: University of Leicester.
Caulfield, S. 1983. The Neolithic settlement of north Connaught, pp. 195–215 in Reeves-Smith, T. and Hamond, F. (eds.), Landscape Archaeology in Ireland (BAR British Series 116). Oxford: British Archaeological Reports.
Cooney, G. 2000. Landscapes of Neolithic Ireland. London: Routledge.
Coxon, P. 2001. Cenozoic: Tertiary and Quaternary (until 10,000 years before present), pp. 387–428 in Holland, C. H. (ed.), The Geology of Ireland. Dunedin: Dunedin Academic Press.
Croke, J. C. 1991. Floodplain variability in the Glenmalur valley, South East Leinster, Ireland. Unpublished Ph.D. Thesis, University College Dublin, National University of Ireland.
Croke, J. C. 1994. Floodplain change in the Glenmalure valley, southeast Leinster. Irish Geography 27, 122–34.
Gallagher, C. and Thorp, M. B. 1995. The fluvial concentrations of heavy minerals in the Slieve Bloom Mountains, Central Ireland. Irish Geography 28, 14–34.
Glanville, W. P. 1999. Holocene River Behaviour and Environmental Change in the upper river Liffey Catchment, Co. Wicklow, Ireland. Unpublished Ph.D. thesis, University College Dublin, National University of Ireland.
Glanville, W. P., Thorp, M. B. and Gallagher, C. 1997. River change during the late Quaternary in the Upper Liffey, pp. 133–41, in Sweeney, J. (ed.), Global Change and the Irish Environment. Dublin: Royal Irish Academy.
Harwood, K. and Brown, A. G. 1993. Changing in-channel and overbank flood velocity distributions and the morphology of forested multiple channel (anastomosing) systems. Earth Surface Processes and Landforms 18, 741–48.
Heery, S. 1993. The Shannon Floodlands; A Natural History. Newtownlynch: Tír Eolas.
Herries-Davies, G. L. 2001. History of Irish geology, pp. 493–504 in Holland, C. H. (ed.), The Geology of Ireland. Dunedin: Dunedin Academic Press.
Hooyer, A. 1991. Introduction to the Middle Shannon Catchment, Ireland (EEC-STEP Project Report 11/1). Amsterdam: Amsterdam Free University.
Kimball, M. J. 2000. Variation and context: ecology and social evolution in Ireland’s later Mesolithic, pp. 31–48 in Desmond, A., Johnson, G., McCarthy, M., Sheehan J. and Twohig, E. S. (eds.), New Agendas in Irish Prehistory. Bray: Wordwell.
Maloney, A., Jennings, D., Keane, M., MacDermott, C. 1993. Excavations at Clonfinlough, Co Offaly (Irish Archaeological Unit Transactions No. 2). Dublin: Irish Archaeological Unit.
Mitchell, G. F. 1972. Some ultimate Larnian sites at Lake Derryvarragh, Co. Westmeath. Journal of the Royal Society of Antiquaries of Ireland 102, 160–73.
Mitchell, G. F. 1986. The Irish Landscape. Dublin: Country House.
Mitchell, G. F. and Ryan, M. 2001. Reading the Irish Landscape. Dublin: Town House.
Nanson, G. C. and Knighton, A. D. 1996. Anabranching rivers: their cause, character and classification. Earth Surface Processes and Landforms 21, 217–39.
Newman, C. 1999. Notes on four cursus-like monuments in County Meath, Ireland, pp. 141–47 in Barclay A. and Harding J. (eds.), Pathways and Ceremonies: the Cursus Monuments of Britain and Ireland (Neolithic Studies Group Seminar Paper No. 4). Oxford: Oxbow.
Odgaard, B. V. and Rasmussen, P. 2000. Origin and temporal development of macro-scale vegetation patterns in the cultural landscape of Denmark. Journal of Ecology 88, 733–48.
O’Brien, W. 1993. Aspects of wedge tomb chronology, pp. 63–74 in Twohig, E. S. and Ronayne, M. (eds.), Past Perceptions: The Prehistoric Archaeology of South-West Ireland. Cork: Cork University Press.
O’Sullivan, A. 1998. The Archaeology of Lake Settlement in Ireland (Discovery Programme Monograph 4). Dublin: Royal Irish Academy.
O’Sullivan, A. and Boland, D. 2000. The Clonmacnoise Bridge: An Early Medieval River Crossing in County Offaly (Archaeology Ireland Heritage Guide 11). Bray: Wordwell.
Smith, D. G. and Smith, N. D. 1980. Sedimentation in anastomosing river systems: examples from alluvial valleys near Banff, Alberta. Journal of Sedimentary Petrology 50, 157–64.
Stout, G. S. 1994. Wicklow’s Prehistoric Landscape, pp. 1–40 in Hannigan, K. and Nolan, W. (eds.), Wicklow: History and Society. Dublin: Geography Publications.
Thorp, M. B. and Gallagher C. 1999. Dating recent alluvial sediments in the Wicklow Mountains. Irish Geography 32, 112–25.
Turbridy, M. 1987. Clonmacnoise Heritage Zone Project. A Portfolio of Management Plans. (Unpublished final Report to EEC). Dublin: Trinity College.
van Geel, B., Buurman, J. and Waterbolk, H. T. 1996. Archaeological and palaeoecological indications of an abrupt climate change in the Netherlands, and evidence for climatological teleconnections around 2650 BP. Journal of Quaternary Science 11, 451–60.
van Geel, B., van der Plicht, J., Kilian, M. R., Klaver, E. R., Kouwenberg, J. H. M., Rensses, H., Reynaud-Farrera, I. and Waterbolk, H. T. 1998. The sharp rise of δ14C c. 800 cal BC: possible causes, related climatic teleconnections and the impact on human environments. Radiocarbon 40, 535–50.
Vader, L. 1993. Hydrological Characterisation of the Little Brosna Floodplain, Ireland (Report EEC-STEP Project). Amsterdam: Amsterdam Free University.
Walsh, P. 1993. In circle and row: Bronze Age ceremonial monuments, pp. 101–13 in Twohig, E. S. and Ronayne, M. (eds.), Past Perceptions: The Prehistoric Archaeology of South-West Ireland. Cork: Cork University Press.
Woodman, P. C. 1978. The Mesolithic in Ireland: Hunter-gatherers in an Insular Environment (BAR International Series 58). Oxford: British Archaeological Reports.
Woodman, P.C. 1985. Excavations at Mount Sandel 1973–77 (Northern Ireland Archaeological Monograph 2). Belfast: HMSO.
Zvelebil, M., Macklin, M. G., Passmore, D. G. and Ramsden, P. 1996. Alluvial archaeology in the Barrow Valley, southeast Ireland: the ‘Riverford’ culture revisited. Journal of Irish Archaeology 7, 13–40.