Content and licensingview original scan buy a printed copy

The geology of Cardigan Bay and the Bristol Channel United Kingdom Offshore Regional Report

By D R Tappin, R A Chadwick, A A Jackson, R T R Wingfield and N J P Smith

Bibliographic reference: Tappin, D R, Chadwick, R A, Jackson, A A, Wingfield, R T R, and Smith N J P. 1994. United Kingdom offshore regional report: the geology of Cardigan Bay and the Bristol Channel. (London: HMSO for the British Geological Survey.)

British Geological Survey

The geology of Cardigan Bay and the Bristol Channel United Kingdom Offshore Regional Report

D R Tappin, R A Chadwick, A A Jackson, R T R Wingfield and N J P Smith

London: HMSO, 1994. NERC copyright 1994. Dd 292049 C20 10 /94 ISBN 0 11 884506 3

Production of this report was funded by the Department of Trade and Industry (formerly the Department of Energy) and the Natural Environment Research Council.The coastline used on many maps and diagrams in this book is based on Ordnance Survey mapping.

(Front cover) Ynys Enlli (Bardsey Island), viewed from the tip of Llŷn (Figure 1). White quartzite clasts in the Precambrian Gwna Melange can be seen on the hillside. Photo: BGS collection.

(Rear cover) Index map United Kingdom Offshore Regional Report

Foreword

Substantially surrounded by onland outcrop of Precambrian and Palaeozoic rocks, the predominantly Mesozoic sedimentary basins which lie beneath Cardigan Bay and the Bristol Channel provide a fine example of the contrast which exists between the onshore and offshore geology. Only in south Glamorgan and Somerset are Mesozoic rocks seen on land and provide a hint of the extensive offshore development of Mesozoic and Tertiary strata, which in the St George's Channel Basin comprise one of the thickest sequences on the UK Continental Shelf. It was not until the drilling of the Llanbedr (Mochras Farm) borehole in the late 1960s that Triassic, Liassic and Tertiary strata were proved in the Cardigan Bay area. The presence of these relatively young rocks in close proximity to, and faulted against, the Cambrian Harlech Dome, provided the spur for further exploration leading to the discovery of other younger offshore sequences which now prove to be so widespread offshore, both within the report area and elsewhere. In fact, in this and many other areas of northern and western Britain, the coastline approximates to a real geological boundary. To date, the younger sequences in the report area have not proved rewarding to the oil industry, although farther west, gas is being recovered from the Irish sector, notably at the Kinsale Head Gasfield. The Bristol Channel is, however, an important area of aggregate extraction.

Beginning in 1966, the British Geological Survey (BGS) has carried out a reconnaissance survey of much of the UK Continental Shelf, including the report area, and has published maps on a scale of 1:250 000 (see inside back cover). The final map of the series was published in 1992. The maps are accompanied by a series of UK Offshore Regional Reports, of which this is one, that provides a fitting culmination to a mapping programme which has extended over 25 years. This work has been largely funded by the Department of Energy (now incorporated into the Department of Trade and Industry), who during the same period also contracted BGS to curate and interpret data collected by the oil industry. Information from both sources has been incorporated into the production of this report, which is also largely funded by the Department of Trade and Industry. The report draws upon a wide range of published and unpublished literature, and provides a comprehensive reference list. But, it should be noted that there is additionally a wealth of new interpretative data in the report, notably in the chapters on Permo-Triassic, Cretaceous and Tertiary rocks, and on post-Variscan sedimentary-basin development. The deep-seismic work of the British Institutions Reflection Profiling Syndicate (BIRPS), and studies by university groups, are further sources of information.

Peter J Cook, DSc Director, British Geological Survey Kingsley Dunham Centre Keyworth Nottingham. NG12 5GG. February 1994

Acknowledgements

Chapter production by individual authors during the production of this report has been as follows:

• D R Tappin — Introduction, Precambrian to Carboniferous, Permo-Triassic, Cretaceous, Tertiary, and economic geology.
• R A Chadwick — Post-Variscan sedimentary-basin development.
• Audrey A Jackson — Jurassic.
• R T R Wingfield — Pleistocene and Holocene.
• N J P Smith — Permo-Triassic (co-author).

The UK Offshore Regional Report Series is co-ordinated by the Marine Geology Group, and edited by D Evans and A G Stevenson.

In addition to the work of the authors, the report has drawn extensively upon the knowledge and expertise of many BGS staff, not only within the marine sphere, but also from the Land Survey and specialists in the fields of sedimentology, biostratigraphy, cartography and publication. In particular, much use has been made in this report of palaeontological analyses carried out on borehole material by BGS biostratigraphers.

The following are particularly thanked for their constructive comments: T D J Cameron, R W Gallois, K Hitchen, R W O'B Knox, J B Riding, G Warrington, and members of the Wales and South-Western England Group.

Chapter 1 Introduction

The Cardigan Bay–Bristol Channel report area lies off Wales and south-west England, and comprises much of the St George's and Bristol channels, the north Celtic Sea, Cardigan Bay and the Severn Estuary (Figure 1). The area lies between 51° and 53°N, and its western boundary approximates to the limit of United Kingdom (UK) waters. To the west, beyond the territorial waters of the Republic of Ireland, lies the Irish mainland. In the south-east, the Severn Estuary and Bristol Channel are almost landlocked, for they are bordered by Wales to the north, and south-west England to the south, but they pass westward into the more-open waters of the Celtic Sea. Wales is an area of outstanding natural beauty; the high rugged mountains of North Wales, the coast of Dyfed, and the Brecon Beacons in South Wales, are all popular tourist attractions. North Devon and Somerset, on the southern boundary of the area, are by comparison with Wales relatively low lying, but have attractive, cliffed coastlines.

Shelf water depths are maintained throughout the area, with a maximum of 160 m along the axis of the St George's Channel and the Celtic Deep (Figure 1) and see (Figure 59). Towards the south-west, water depths gradually increase towards the shelfbreak, which lies beyond the report area.

A notable feature of the area is the contrast between the onshore and offshore geology (Figure 2). Onshore, the rocks are largely of Palaeozoic age, whereas offshore these older rocks are overlain by Permian, Mesozoic and Tertiary strata which are locally in excess of 7000 m thick in the St George's Channel Basin (Figure 22)). For location see (Figure 1)." data-name="images/P945202.jpg">(Figure 3). However, within the deeper parts of the basins, details of the successions remain poorly understood, as exemplified by the alternative seismic interpretations presented in (Figure 22)). For location see (Figure 1)." data-name="images/P945202.jpg">(Figure 3) and (Figure 22). Only in South Wales and Somerset are the younger beds exposed onshore. A further characteristic feature of the offshore geology is the extensive distribution of thick Quaternary sediments which, in the north Celtic Sea, locally exceed 300 m in thickness.

History of offshore research

In the north of the report area, the first offshore investigations took place in the late 1940s when Browne and Cooper (1950), using pendulums in a submarine, measured a low gravity anomaly in the St George's Channel. The anomaly was 60 mGals lower than that measured at the Welsh coast, suggesting the presence of a major sedimentary basin offshore. Further gravimeter, magnetometer, and seismic-refraction surveys indicated the presence of relatively young, low-density sediments in Tremadog Bay, downfaulted against Cambrian rocks of the Harlech Dome (Powell, 1956; Griffiths et al., 1961). Three layers were identified on the western, downthrown side of the fauh: an upper layer 600 m thick interpreted as Mesozoic or Tertiary clays, a middle layer 2000 m thick considered to be Ordovician in age, and a lower layer postulated to be Cambrian beds. To test the age of these sediments, Professor Alan Wood, of the University College of Wales (UCW) Aberystwyth, suggested drilling a borehole at Mochras, which lies seaward of the bounding fault on a sand spit. Drilled between 1967 and 1969, the Mochras borehole, officially known as the Llanbedr (Mochras Farm) borehole, proved the rocks to be largely significantly younger than predicted. Triassic, Jurassic, Tertiary and Pleistocene strata were recovered (Wood and Woodland, 1968; Woodland, 1971), confirming an earlier prediction by Professor O T Jones (1952; 1956) that Triassic rocks are present beneath Cardigan Bay. Proving these thick Mesozoic and Tertiary strata in an essentially offshore setting, in close proximity to the Lower Palaeozoic rocks of the Harlech Dome, may be regarded as a milestone in continental shelf research.

In 1966, the British Geological Survey (BGS — then called the Institute of Geological Sciences) started a reconnaissance survey of the UK Continental Shelf (Fannin, 1989). After drilling at Mochras, a deep-seismic reflection survey in the St George's Channel and Cardigan Bay basins in 1968 confirmed the presence of a significant sedimentary accumulation (Bullerwell and McQuillin, 1969). Reinterpretation of this survey forms the basis of the account presented in Chapter 3. Further geophysical surveys by Birmingham University researchers (Blundell et al., 1971) outlined the major basin lineaments, with the basin fill interpreted as Triassic, Jurassic, Tertiary, and Quaternary in age. The BGS Whitethorn drilling programme from 1970 to 1975 proved the presence of Triassic, Liassic, Middle Jurassic, Tertiary and Quaternary strata (e.g. Penn and Evans, 1976). Workers from UCW Aberystwyth further refined the structure of the Irish Sea basins (Dobson et al., 1973), and mapped late Devensian glacial deposits (Garrard and Dobson, 1974).

In the southern part of the report area, the earliest sampling was carried out by researchers from the Universities of London and Bristol. Mesozoic rocks of Triassic to Jurassic age were reported to be present in the Bristol Channel in the early 1960s (Donovan et al., 1961; Lloyd, 1963), and a research cruise in the Celtic Sea in 1966 sampled chalk south of Cork (Curry et al., 1967). Davey (1970) indicated the widespread presence of low-density sedimentary rocks, and Dixon (1968) recognised the presence of up to 100 m of Quaternary sediments beneath the Celtic Deep. An early geological map of the North Celtic Sea Basin (Blundell et al., 1971) included three units: Chalk, overlain by both Paleogene and Neogene sedimentary rocks. The latter were subsequently identified as Quaternary by Delantey and Whittington (1977).

In the 1970s, more details of the geology of the Bristol Channel were published (Banner et al., 1971; Lloyd et al., 1973; Dor6, 1976), and Tertiary sediments were proved by BGS drilling in the Stanley Bank Basin (see (Figure 21); Fletcher, 1975). In the North Celtic Sea Basin, a shallow-seismic and sampling programme carried out from UCW Aberystwyth refined the geology map of Blundell et al. (1971), providing evidence of sea-bed outcrop of Tournaisian, Permo-Triassic and Upper Cretaceous sediments (Delantey et al., 1981).

With the advent of hydrocarbon exploration activity in the area early in the 1970s, multichannel seismic-reflection data were acquired. Interpretation of these profiles, together with the drilling of deep hydrocarbon wells, resulted in the confirmation and refinement of previous interpretations of basin stratigraphy, as well as advances in the understanding of basin evolution (Naylor and Mounteney, 1975; Kamerling, 1979; Barr et al., 1981; Naylor and Shannon, 1982; Tucker and Arter, 1987; Van Hoorn, 1987). In the St George's Channel Basin, the presence of a Triassic to Upper Jurassic section was confirmed; in the South Celtic Sea and Bristol Channel basins, Wealden, greensand and chalk facies were proved for the first time.

Seven hydrocarbon wells have been drilled in the St George's Channel Basin, eight in the South Celtic Sea and Bristol Channel basins, and over seventy in the North Celtic Sea Basin outside the report area in Irish waters. A selection of simplified well logs is shown in (Figure 4). Although hydrocarbon shows have been reported, no commercial discoveries have been made in UK waters. However, in the Irish sector of the North Celtic Sea Basin, gas has been brought on stream from the Kinsale Head and Ballycotton gasfields. Hydrocarbon reservoirs are mainly in Lower Cretaceous sandstones; the source rocks are Jurassic shales.

Numerous publications on the deeper structure of the Bristol Channel have been published by Professor Brooks and his co-workers from the University Colleges of Swansea and Cardiff (e.g. Brooks and James, 1975; Brooks and Al-Saadi, 1977; Evans and Thompson, 1979; Mechie and Brooks, 1984; Brooks et al., 1988). In 1982 and 1984, as part of the British Institutions' Reflection Profiling Syndicate (BIRPS) programme, deep multichannel seismic-reflection profiles were acquired within the report area to investigate the crustal structure (Figure 1); BIRPS and ECORS, 1986; McGeary et al., 1987). The southernmost line of the Western Isles–North Channel (WINCH) series of traverses crossed the north of the area over the Kish Bank and Central Irish Sea basins. The South Western Approaches Traverses (SWAT) were run across the southern basins. Data were recorded to 15 s (seconds) two-way travel time (TWIT), and reflectors were imaged from beneath the Mohorovièiæ Discontinuity (Moho), which lies at a depth of about 30 km.

During the 1980s, the BGS completed its reconnaissance surveys across the report area, and published a series of 1:250 000 scale maps of the solid geology, Quaternary geology, sea-bed sediments, Bouguer gravity anomalies and aeromagnetic anomalies (see index map inside back cover). In many instances, this work was in collaboration with university groups, most notably UCW Aberystwyth. The maps are based upon available data; BGS alone occupied 2640 sample sites in the area, with vibrocorer and grab being used at 255, and a gravity corer and grab at 1697. Pre-Quaternary rocks were sampled at 747 sites, and 101 boreholes were sunk during the Whitethorn drilling operation (Parkin and Crosby, 1982; Warrington and Owens, 1977; Wilkinson and Halliwell, 1980).

Geological summary

Evidence of the pre-Mesozoic history of the report area is largely confined to the adjacent onshore districts. In North Wales on the Llŷn Peninsula, and in Dyfed in South Wales (Figure 2), the oldest rocks are metamorphosed volcanic and intrusive rocks of late Precambrian age.

In Early Palaeozoic times, Wales and Cardigan Bay formed part of the Welsh Basin, within which mainly marine sedimentary rocks were laid down, although deposition was accompanied by outbursts of volcanic activity. Cambrian rocks are mainly confined to North Wales, where they comprise the turbiditic sands and hemipelagic mudstones of the Harlech Dome; they also occur in north Pembrokeshire. Volcanic rocks form a significant part of the Ordovician sequence which flanks the Harlech Dome on its northern, eastern and southern margins. In South Wales, Ordovician rocks form an arcuate belt of shelf and marginal, shelly, elastic sedimentary rocks. Silurian rocks make up much of Mid Wales, where they comprise turbiditic sandstones and mudstones; these pass eastwards and south-eastwards into more-shelly, arenaceous rocks.

The uplift and subaerial exposure of North and Mid Wales during the Caledonian orogeny led to modification of the Welsh Basin, and to the deposition of alluvial sands and muds in continental environments in South Wales. This continental facies periodically extended southward into Devon and Somerset, where marine-shelf conditions predominated, and sands, muds and carbonates, interbedded with fluvial grits and sands, were laid down.

Resumption of marine conditions in South Wales during the Early Carboniferous led to the deposition of shelf carbonates. Establishment of paralic environments in Namurian times led first to the laying down of grits, sands and muds, and subsequently to the development of low-lying swamps and the accumulation of coal measures during the Westphalian. In Devon and Somerset, coeval Carboniferous sedimentation took place in a shallow-marine basin; initially, during the Visean, argillaceous deposition dominated, but in early Namurian times this was replaced by turbiditic sand and mud sedimentation. A complex of turbidite, deltaic and lacustrine sediments then accumulated in Westphalian times. During the late Carboniferous Variscan orogeny, uplift and deformation eventually led to subaerial exposure. The Variscan Front, which defines the northern limit of pervasive Variscan deformation, crosses the report area — see (Figure 13).

By the end of the Variscan orogeny, much of the report area must have been subaerially exposed. Evidence of the post-Variscan history is mainly to be found offshore. A number of sedimentary basins initiated during the Permian or early Mesozoic were formed above reactivated Caledonian and Variscan thrusts, see (Figure 13). During the Permian and Triassic periods, sedimentation was continental, with sands, muds and evaporites laid down in desert, alluvial and sabkha environments. At the beginning of the Permian, sedimentation was localised within discrete depocentres, but continued subsidence during the Triassic led to the progressive expansion of sedimentary limits, and the basins merged.

At the end of the Triassic, marine environments became established; Lower and Middle Jurassic sediments are mainly mudstones with subsidiary sandstones and limestones, although limestones are more important in South Wales. Uplift during the Mid-Jurassic led to erosion of the basin margins in the south of the area, and to emergent conditions in the St George's Channel Basin, where redbeds with minor evaporite beds accumulated. Estuarine/deltaic and marginal-marine sediments characterise the Upper Jurassic, with organic-rich shales present only in the south.

At the end of the Jurassic, progressive shallowing resulted in the spread of fluvial and deltaic conditions across the area, and to the deposition of sands, muds and carbonaceous beds of the Lower Cretaceous Wealden beds. Structural demarcation during the Early Cretaceous led to further local erosion, but with the return of widespread marine conditions during the mid-Cretaceous, glauconitic sands and clays were deposited. Continued rise in eustatic sea levels led to a blanket of chalk being widely deposited during the Late Cretaceous.

During latest Cretaceous and early Cenozoic times, minor basin inversion and regional uplift led to considerable erosion. The Lundy igneous complex was emplaced during the early Eocene. Subsequent Paleogene deposition in the north and adjacent to Lundy Island was characteristically alluvial, passing south-westward into marine conditions. There was a significant hiatus, with minor basin inversion, at the end of the Paleogene; Oligocene rocks are overlain by largely glacially related Quaternary sediments, which are thickest beneath the deep waters of St George's Channel and the Celtic Deep (Figure 1).

Chapter 2 Precambrian to Carboniferous

Precambrian to Carboniferous rocks occur extensively onshore at the periphery of the report area; outcrops of Precambrian and Lower Palaeozoic rocks in the report area are confined to Wales, although they form much of the east coast of Ireland (Figure 2). Upper Palaeozoic rocks crop out more widely in South Wales, Devon, and on the southern coast of Ireland. Offshore, such rocks occur at the margins of the Mesozoic/Cenozoic basins, mostly in continuity with the onshore outcrops (Figure 2), although their offshore extent is limited. In only one instance has the Lower Palaeozoic been proved in a borehole, and the Precambrian not at all. Upper Palaeozoic rocks have been more commonly sampled, both at, or near, the sea bed in shallow boreholes and cores, and at depth in wells.

In the north of the report area, major onshore structures may be traced offshore: the Menai Straits fault system extends into the South Irish Sea Lineament, and the Bala Fault can be mapped offshore (see (Figure 13)). The predominant northeasterly trend of these structures originated in the orogenic cycles of the late Precambrian (Cadomian) and Early Palaeozoic (Caledonian); north-westerly dipping faults probably originated as Caledonian thrusts (Brewer et al., 1983). South of the Variscan Front, which crosses the report area (see (Figure 13)), the dominant east–west structural trend originated during the Late Palaeozoic Variscan orogeny.

Precambrian

During the late Precambrian to Early Palaeozoic, the report area lay on the southern side of the Iapetus Ocean, and formed the northern margin of the Gondwana continent at a high-latitudinal location in the southern hemisphere (Gibbons, 1987). The rocks of this interval are represented by assemblages formed at destructive plate margins, with ages of 700 to 550 Ma (Shackleton, 1956; Beckinsale et al., 1984; Thorpe et al., 1984; Tucker and Pharaoh, 1991). Precambrian rocks occur only on the two western promontories of Wales, in Llŷn and Pembrokeshire; their offshore outcrop is uncertain, but limited (Figure 2).

On the Llŷn Peninsula and Bardsey Island (see Anderton et al., 1992), the Mona Complex (Greenly, 1919) is a subduction complex (Thorpe et al., 1984) that comprises a mélange of limestones, quartzites, tuffs and lavas (the Gwna Group of Matley, 1928 — see front cover), with plutonic igneous rocks ranging in composition from gabbro to granite (the Sarn Complex). The outcrop is divided by a shear zone of mylonitic schists, (Gibbons, 1983).

In Pembrokeshire, Precambrian island-arc rocks include quartzose metasediments, gneisses, spherulitic rhyolites, trachytic tuffs, breccias and gneisses. These are intruded by granites, tonalites, quartz-diorites, and pegmatite veins (Wright, 1969; Shackleton, 1975; Thorpe et al., 1984).

Lower Palaeozoic

The Early Palaeozoic was the time of formation, infilling, and destruction of the Welsh Basin. Elongated in a north-easterly direction, the basin (Figure 5) was bounded to the north-west by the Irish Sea Platform (Gibbons, 1987), and to the southeast by the Midland Platform. To the south lay the landmass of Pretannia, which during most of late Precambrian to late Silurian time formed an upstanding massif (Cope and Bassett, 1987). The basin centre occupied central Wales, and probably extended into Cardigan Bay. The sequence deposited comprises mainly graptolitic shales and turbidites, with stratigraphically limited, but thick, volcanic units. To the east and south, shallow-water, shelf sediments were deposited.

Cambrian

Cambrian rocks crop out onshore in the Harlech Dome region, where they barely extend offshore, and in Pembrokeshire, where their offshore outcrop is probably limited (Figure 2). During Cambrian times, the northern margin of the Gondwana continent formed a passive shelf (Woodcock, 1990); marine transgression early in the period led to the deposition of shallow-marine conglomerates, sandstones and siltstones, which in North Wales pass up into turbidites laid down under under more basinal conditions ((Figure 5)a; Crimes, 1970; Rushton, 1974). During the mid-Cambrian, turbidites with interbedded mudstones continued to be laid down in North Wales, while shallow-water sandstones and mudstones were deposited to the south. During the late Cambrian, basinwide sedimentation of shallow-water siltstones and mudstones took place, culminating in the development of black mudstones (Rushton, 1974; Turner, 1977).

Ordovician

Rocks of Ordovician age crop out extensively in north-west Wales, and in Dyfed (Figure 2). In both these areas the strata are presumed to extend offshore, where they can only be termed undivided basement. In North Wales, the mudstones pass upwards in continuous succession from the underlying Cambrian, but in South Wales there was uplift in latest Cambrian times, leading to an unconformity at the base of the Tremadoc (Bassett, 1980). In the Bristol Channel, Tremadoc strata have been sampled in BGS borehole BH72/60 (Figure 5)a, where they comprise cleaved, grey, micaceous mudstone (Tappin and Downie, 1978).

Subduction along the northern margin of Gondwanaland began during the late Cambrian to early Ordovician. Associated back-arc spreading led to the formation of a marginal basin and to the separation of southern Britain, including the Welsh Basin, from the Gondwana continent. The northward drift of this microcontinent, named Eastern Avalonia (Soper, 1986) or Cadomia (Ziegler, 1982), led to the opening of the Rheic Ocean to the south (Cocks and Fortey, 1982). Within the Welsh Basin, the beginning of drift was marked in late Tremadoc to early Arenig times by uplift, intense subaerial erosion, and voluminous arc volcanism (Shackleton, 1954; Fortey and Owens, 1987; Kokelaar, 1988).

Marine transgression during the Arenig marked the beginning of the main Ordovician phase of sedimentation in the Welsh Basin ((Figure 5)b; Woodcock, 1990). In North Wales, the Arenig to Caradoc sequence comprises mainly marine sandstones, siltstones and mudstones, commonly in intimate association with rhyolitic tuffs, lavas, and basaltic pillow lavas erupted under submarine conditions (Howells et al., 1983; Kokelaar et al., 1984). In South Wales, Ordovician sediments along the southern and eastern basin margins comprise shelly clastic rocks and limestones of shelf facies (George, 1970; Fortey and Owens, 1987). Turbidites testify to tectonism along the basin margins, but towards the west, black mudstones were laid down. Volcanic episodes during Arenig to Caradoc times produced mainly submarine rhyolitic and basaltic lavas, and pyroclastic rocks (Jones and Pugh, 1949; Kokelaar et al., 1985). During Caradoc times, the main phase of Ordovician deposition culminated in the basinwide sedimentation of black, graptolitic mudstones (George, 1970). BGS borehole BH73/42, in Cardigan Bay (Figure 5)b, lies on the projected line of the onshore boundary between the Ordovician and Silurian (Figure 2). It encountered an undated, hard, fractured, silty mudstone.

During late Caradoc to early Ashgill times, uplift caused by the 'soft' collision of Eastern Avalonia with Laurentia (Pickering et al., 1988) led to the creation of an unconformity which can be traced across most of the Welsh Basin (Bassett, 1980; Woodcock, 1990). In the north, upper Ashgill sediments are turbiditic sandstones and siltstones. In South Wales, shelly lithologies pass northward into basinal, graptolitic mudstones interbedded with turbiditic conglomerates, sandstones and siltstones (Figure 5)c.

Silurian

Silurian rocks form much of Mid Wales, and crop out extensively at the west coast as well as in parts of south Dyfed (Figure 2). The three facies belts established during the Ordovician continued into the early Silurian (Figure 5)c. Turbidite sedimentation dominated the basin centre (Wood and Smith, 1959), while fan and deltaic sediments were laid down on the basin margins (Kelling and Woollands, 1969). Mainly carbonate sequences to the east suggest more quiescent, shelf environments. Volcanism became localised and of within-plate character (Thorpe et al., 1989).

Turbidite deposition, sourced mainly from the south, progressively migrated eastwards during Llandovery and Wenlock times. On the basin margins, shallow-water, shelly clastics and limestones were laid down, and as sea levels rose during the late Llandovery and Wenlock, increasing areas of shelf were inundated. However, in Dyfed, alluvial and deltaic strata herald the onset of the continental conditions of the Old Red Sandstone (Allen and Williams, 1978).

Uppermost Silurian sequences show a progressive decrease in marine influence due to uplift; open-sea environments were progressively replaced by intertidal, coastal-plain and continental conditions (Allen, 1974). Pøídolí rocks are shallow-marine or intertidal mudstones and sandstones with bone beds; alluvial sandstones and conglomerates occur at the top.

Upper Palaeozoic

The uplift and infilling of the Welsh Basin towards the close of the Silurian marked the beginning of the climax of Caledonian deformation, attributed to 'hard' collision between Avalonia and Laurentia (Pickering et al., 1988). The collision resulted in the cessation of subduction along the Iapetus suture zone, and the closure of the Iapetus Ocean to the north. During the Early Devonian, after closure of Iapetus, the report area lay on the southern margin of Laurentia, to the north of the Rheic Ocean on whose southern shore lay Gondwanaland, which was moving northward. Between the converging masses of Laurentia and Gondwanaland lay the Rhenohercynian basin (Holder and Leveridge, 1986; Ziegler, 1986).

Devonian

During the Early Devonian, sediments were shed southward, as northern and central Wales, part of St George's Land, were uplifted. Deposition of continental, Old Red Sandstone type occurred on an alluvial plain in South Wales, while farther south in Devon, mainly shallow-marine conditions prevailed, with alluvial advances during regression. A fluctuating shoreline therefore lay along the Bristol Channel (Figure 6) and (Figure 7), but whether South Wales and Devon formed a single basin is uncertain (Freshney and Taylor, 1980). Gedinnian to Siegenian alluvial sediments (Dartmouth Beds) laid down in south Devon ((Figure 7); Dineley, 1966) are lithologically similar to the contemporaneous Senni Beds in Wales. In north Devon, the earliest known Devonian rocks are of Emsian age.

There is evidence for a Bristol Channel landmass during Emsian and Eifelian times (Tunbridge, 1986), with sediment being shed both to the north and south (Figure 6). During the Mid-Devonian climax of Caledonian deformation (Woodcock, 1990), sedimentation appears to have been mainly restricted to south of the Bristol Channel; in Devon there is an unbroken section which continues through the Upper Devonian. In South Wales, Upper Devonian sediments are not widespread.

Although crustal shortening across Wales, the Bristol Channel and Devon took place during the Variscan orogeny (Brooks et al., 1988; Gayer and Jones, 1989), it is probable that the relative positions of these areas have remained unchanged (Freshney and Taylor, 1980).

South Wales and adjacent offshore areas

In South Wales (Figure 6), (Figure 7) and (Figure 8), the Lower Devonian comprises a generally upward-coarsening sequence of sedimentary rocks (Williams, 1980) shed southward off the rising highlands of St George's Land. Initially, during the Gedinnian to Siegenian, mudstones with channelised sandstones and conglomerates (Red Marls) were laid down on a vast alluvial plain as distal, fluvial sediments (Allen, 1974; Williams, 1980). The succeeding strata are a more proximal, fluvial sequence of conglomerates, sandstones and siltstones, with subordinate mudstones (Senni Beds, Brownstones). Again these were mainly derived from the north (Allen, 1974), although during Emsian times some alluvial-fan conglomerates were sourced from the south (Allen, 1975; Wilson et al., 1988).

Off the coast of Pembrokeshire, numerous samples of undated red mudstones and shales, with subordinate sandstones, have been sampled at the sea bed (Figure 8). The lithologies lie along strike from onshore Early Devonian rocks, which suggests that the samples may also be of this age. Similar lithologies sampled at the sea bed nearby to the south (Figure 8) are interpreted to be of equivalent age.

Middle Devonian sediments are absent in South Wales. The Upper Devonian succession begins with the Plateau Beds, of Frasnian age, which are aeolian and fluvial sandstones that pass upwards into marginal-marine sandstones and mudstones (Barclay, 1989). Above a break in the succession, the Skrinkle Sandstones and Quartz Conglomerate Group, of late Famennian age, record a transition from continental to marine environments, as shown by the upward transition from fluvial conglomerates, sandstones and siltstones, to marginal-marine sandstones, shales and limestones (Allen, 1974).

North Devon and adjacent offshore areas

Immediately off the north Devon coast, lithologies similar to those of Devonian rocks onshore have been sampled in shallow cores ((Figure 9); Lloyd et al., 1973). The strike of strata at the sea bed, as identified on sidescan sonar records, aligns with the west-north-westerly structural trend onshore, and can be traced to the margin of the Mesozoic Bristol Channel Basin.

At the base of the Devonian sequence lie cleaved mudstones with interbedded thin sandstones; these are the Lynton Beds, deposited in an offshore, muddy, shelf environment (Goldring et al., 1967). Offshore, samples equated with the Lynton Beds are dark reddish brown, micaceous sandstones and shales. More commonly sampled offshore are noncalcareous sandstones interpreted as Hangman Grits lithologies; onshore, these rocks are mainly of fluvial origin, and derived from the north (Tunbridge, 1986). The overlying Ilfracombe Beds mark a return to offshore, muddy, shelf environments; offshore samples are mainly grey, cleaved mudstones. Northwest of Morte Point, pale grey slate and shale lithologies are interpreted as Morte Slates, which were laid down as muds in a prodeltaic or offshore environment (Webby, 1966).

To the west of Morte Bay, a sample of hard, purplish sandstone is taken to correlate with the fluvial Pickwell Down Sandstone. To the south, a green-grey, cleaved mudstone is interpreted as being from the Upcott Slates, which comprise mudstones and siltstones deposited in either backswamp or lacustrine conditions (Goldring, 1971). Neither the Baggy Beds nor the Pilton Beds have been sampled south of Baggy Point.

The Devonian–Carboniferous boundary lies within the Pilton Beds, which are mainly mudstones with subordinate thin sandstones and limestones, laid down in muddy offshore environments similar to those of the underlying Lynton and Ilfracombe beds (Edmonds et al., 1979).

To the west of the offshore extension of the Sticklepath–Lustleigh fault zone, sea-bed samples are again comparable with the Devonian rocks of north Devon (Lloyd et al., 1973; Evans and Thompson, 1979). Furthermore, on the southeastern tip of Lundy Island, exposed slates may resemble the lower parts of the Morte Slates (Edmonds et al., 1979). If this is the case, then there is either sinistral movement across the Sticklepath–Lustleigh fault zone, or a change in strike offshore. Lithological comparison of offshore samples with the onshore rocks is problematic. To the north-west of Lundy Island, a cleaved, micaceous slate with micaceous sandstone (sample 376), and a slate (sample 377), have been identified as the Pickwell Down Sandstone ((Figure 9); Lloyd et al., 1973). A sandstone was also sampled at site 262 (Evans and Thompson, 1979). If these rocks are Pickwell Down Sandstone, and the Lundy exposure is Morte Slates, then the structure west of the Sticklepath–Lustleigh fault zone is markedly different from that to the east, and the beds dip towards the north rather than to the south. On sidescan sonar records, the rocks to the north of Lundy Island follow a gentle arch where they are fractured and deformed into many minor folds (Lloyd et al., 1973). Other sea-bed samples north of Lundy Island are unassigned sandstones and slates. Borehole BH72/50 sampled 3.5 m of undated, fractured, dark-grey, silty and pyritic mudstone.

To the east of Lundy Island, grey-green slates in sea-bed cores are interpreted as offshore equivalents to the possible Morte Slates on the island. A greenish grey siltstone at site 255 (Figure 9) contains brachiopod fragments. To the south of Lundy Island, most of the sea-bed samples are grey shales and slates of uncertain affinity, although it is probable that they are of Carboniferous age towards the north Devon coast.

Low-grade metamorphic sediments have been penetrated in wells in the south-west of the report area. In well 103/21-1 (Figure 6), grey-green slate, claystone and red-brown mudstone with tuff are reported as being Devonian in age (Company log). Volcanic horizons are known to occur in the Morte Slates and Ilfracombe Beds onshore. In well 102/28-1, grey-brown and red-brown siltstone, sandstone and shale, dated as either Late Devonian or earliest Carboniferous (Company report), may be equivalent to the Pilton Beds of Devon.

Carboniferous

Tournaisian To Visean (Dinantian)

During the Late Devonian to Early Carboniferous, the prevailing regional tectonic regime changed from one of compression to one of extension, and the southern margin of St George's Land was transgressed during a major marine invasion. The passage from the Devonian was transitional; in South Wales, the Dinantian was characterised by prolonged, mainly shelf-carbonate deposition (Figure 10). Sedimentation was controlled by eustatic sea-level changes superimposed on an active tectonic framework (George et al., 1976; George, 1978; Ramsbottom, 1979; Wilson et al., 1988). The carbonates represent a variety of depositional environments (Waters and Lawrence, 1987): oolitic limestones were deposited in shallow-water shoals, whereas bioclastic limestones accumulated in more offshore conditions. Siliciclastic sediments, mainly calcareous sandstones and conglomerates, are confined to the northern margin of the basin of deposition.

George (1960) first proposed the existence of a Dinantian basin to the west of Wales beneath St George's Channel. Interpretation of deep-seismic data west of Wales suggests the existence there of a linked series of Caledonide-trending Carboniferous basins: the Kish Bank, Central Irish Sea and St George's Channel basins. Their dominant bounding faults, which dip to the north-west e.g. (Figure 11), may originally have formed as Caledonian thrusts (Brewer et al., 1983).

The Kish Bank Basin forms a half-graben, downthrown to the north, in which Dinantian strata are interpreted to be present at depth (Jenner, 1981). The Central Irish Sea Basin is fault bounded to the south by the South Irish Sea Lineament (Figure 11) and see (Figure 13), an offshore extension of the Menai Straits fault system (Brewer et al., 1983).

The presence of Dinantian strata within the basin is suggested by the presence of beds of this age along strike to the north-east on Anglesey (Figure 10), where they are are down-faulted along the Menai Straits fault system.

The St George's Channel Basin formed along the offshore extension of the Bala Fault see (Figure 13). In the west of the basin, in Irish waters, limestone lithologies, interpreted as Visean in age on the basis of microflora (Company report), have been penetrated in well 42/17-1 (Figure 10), which sampled bioclastic and partly dolomitised limestone. By correlation with this well, the Dinantian can be traced on seismic line SWAT-2 as a package that thickens south-eastwards towards the controlling St George's Fault (Figure 22)). For location see (Figure 1)." data-name="images/P945202.jpg">(Figure 3). Syndepositional faulting, similar to that seen in the St George's Channel Basin, has been identified to the north-east along the Bala Fault, on the margin of the Cheshire Basin (George, 1961).

To the south of Wales, fine-grained limestones of presumed Dinantian age have been sampled off Caldy Island ((Figure 8); Lloyd et al., 1973). Some 40 km to the west, seabed samples of undated limestones, shaly limestones, and grey shales, may also be of Dinantian age. Shelf-limestone deposition, similar to that in South Wales, may thus have extended into the Bristol. Channel. Lime mud banks, termed Waulsortian reefs, similar to those penetrated in the Cannington Park borehole, are found in the Tournaisian of southern Ireland and Pembrokeshire ((Figure 10); Freshney and Taylor, 1980).

To the south of the Bristol Channel, in Devon, the deltaic Late Devonian Baggy Beds pass up gradually into the shallow-marine Pilton Beds, within which lies the Devonian–Carboniferous boundary (Edmonds et al., 1979). The Pilton Beds comprise grey shales with subordinate thin sandstones and limestones. The remainder of the Tournaisian sequence comprises dark, pyritic shales, impure limestones, and cherts; these were laid down under euxinic, starved conditions, which persisted until the end of the Visean. Off Devon, in BGS core 224 (Figure 9), a hard, thinly bedded shale with thin limestone lenses may be a Pilton Beds equivalent. Other sea-bed samples of grey shale may also be Pilton Beds lithologies.

Namurian to Stephanian

During Namurian times, the continued convergence of Laurentia and Gondwanaland led to a change from an extensional to a compressional tectonic regime (Warr, 1989). Loading by northward-propagating thrusts from the rising Variscan mountains to the south led to uplift of St George's Land as a peripheral bulge from which clastic sediments were shed southwards ((Figure 12)a; Beaumont, 1981; Kelling, 1988; Hartley and Warr, 1990). In central South Wales, beach or barrier sands were laid down, interbedded with subordinate marine mudstones; these pass southwards into marine mudstones (Basal Grits of Kelling, 1974; Ramsbottom et al., 1978). In west Wales, the sedimentary rocks comprise delta floodplain sandstones and mudstones, lagoonal and prodelta siltstones, with mudstones and barrier-bar sandstones (George, 1970). An upper sequence (Upper Shale Group), laid down over a greater area, comprises fluvial, deltaic and littoral sandstones interbedded with basin-wide marine bands (Oguike, 1969).

In Devon, the passage from the Visean into the Namurian, represented by the Crackington Formation, is transitional. It is marked by a gradual disappearance of limestone bands, and an incoming of distal turbiditic sandstones deposited in an east–west-trending basin open only to the west, with source areas in the south and south-west (Freshney and Taylor, 1980). The basin may have extended northwards into South Wales (Freshney and Taylor, 1980).

Regional compression continued into the early Westphalian (Langsettian). In South Wales, regressive fluviodeltaic sands (Farewell Rock) sourced from the north are interbedded with marine bands (Kelling, 1974). In Devon, distal turbidite deposition continued, although in mid-Langsettian times there was an input of deltaic distributary sands (Bideford Formation) and proximal, massive, lacustrine sands (Bude Formation) derived from the north, possibly from a resurgent Bristol Channel landmass (Freshney and Taylor, 1980; Melvin, 1986; Higgs, 1991). In South Wales, during late Langsettian and Duckmantian times, there was prolonged deposition of sands and silts with thick and extensive coals on low-lying, swampy, alluvial, coastal plains subjected to infrequent marine incursions (Kelling, 1974). On the margins of the basin, fluviodeltaic sands were deposited from rivers flowing both from the north and south (Figure 12)b. In Devon, the lacustrine environments of the Bideford Formation continued; the Bristol Channel landmass was evidently a positive feature at this time (Gayer and Jones, 1989).

During the late Westphalian (Bolsovian), as uplift advanced from the south, the palaeoslope in the report area reversed towards the north (Figure 12)c. Continued thrustnappe loading led to an increase both in basin subsidence and the rate of supply of sediment in South Wales (Kelling, 1988). From the rising Variscan mountains, elastics were shed northwards to be deposited as fluvial sands of the Pennant Measures, which pass northward into paralic sediments. The youngest part of the Bude Formation is of early Bolsovian age, and it is probable that sedimentation in Devon ceased at about this time (Freshney and Taylor, 1980). Progressive uplift from the south led to the continued retreat of the basin margin across St George's Land, and during the climax of the Variscan orogeny in latest Carboniferous times (Stephanian), to the uplift and subaerial exposure of the report area.

Offshore samples

West of Wales, in the St George's Channel Basin, a conformable sequence on SWAT-2 (Figure 22)). For location see (Figure 1)." data-name="images/P945202.jpg">(Figure 3), interpreted as Silesian in age, is likely to include both Namurian and Westphalian strata. Anticipated Namurian sediments would be of a paralic facies similar to that in South Wales; Westphalian beds are probably similar to those in well 42/17-1 (Figure 12); Company log), located on the basin's western margin, where two units are recognised. The lower unit is composed of 270 m of grey beds comprising fine-grained, partly muddy, angular-grained, coaly sandstone, together with claystone and siltstone. It is dated as (?)Bolsovian to Westphalian D, and rests unconformably upon Dinantian limestone. The dominant grey colour and the presence of coal suggest a waterlogged environment on an alluvial plain or delta. The upper unit, of Westphalian D age, is composed of 655 m of red and grey beds. The red beds are dark red and purple claystone and siltstone with some thin sandstone, and, towards the base, thin coals. The grey beds are similar to those below, but with thinner sandstones and fewer coals.

On the southern margin of the basin, well 103/2-1 (Figure 12) sampled 204 m of mainly red and brown, interbedded sandstone, siltstone and mudstone with minor coals, of Westphalian D age. The red-brown sediments in this well, as in well 42/17-1, are interpreted as having been deposited under alluvial conditions with low water tables. The geophysical logs of well 103/2-1 indicate upward-fining cycles which represent channel sands. The mudstones, which are mainly red-brown with subordinate grey colours, are probably overbank deposits. Similar successions are described from the upper Westphalian in the East Irish Sea Basin (Jackson et al., 1994).

In St Bride's Bay (Figure 8), BGS borehole BH73/44 sampled a brown shale of Visan to Westphalian age. The borehole's location, offshore from the Silesian exposed in Pembrokeshire, suggests it to be of this age. West of this borehole, grey, cleaved mudstone of Namurian to Westphalian age has been sampled in a sea-bed core (sample 228). In Swansea Bay, four BGS boreholes (Figure 1) have sampled rocks presumed, on the basis of location and lithology, to be of Westphalian age: grey, fine-grained, micaceous sandstone in BH72/46; grey-brown, silty mudstone in BH72/53; dark grey, micaceous, shaly mudstone in BH72/58; and grey, shaly mudstone in BH72/59.

Off the coast of Devon, numerous sea-bed samples of grey shales and slates may well be of Late Carboniferous age (Figure 9). In Barnstaple Bay, BGS borehole BH74/42 sampled grey, fissile shale, and borehole BH72/65 recovered cleaved, quartzose, silty mudstone. Farther west, borehole BH75/05 drilled a fractured, very fine-grained greywacke.

Chapter 3 Post-Variscan sedimentary-basin development

After the Variscan orogeny, thick sequences of Permian to Tertiary age were deposited in the sedimentary basins situated off the coasts of Wales and south-west England (Kamerling, 1979; Barr et al., 1981; Van Hoorn, 1987; Dobson and Whittington, 1987). To the west of Wales, the Central Irish Sea, Cardigan Bay, St George's Channel, and North Celtic Sea basins show predominantly north-east to south-west structural trends (Figure 13). Farther south, the South Celtic Sea and Bristol Channel basins are oriented more nearly east–west. The basins are of extensional origin, with margins controlled by major normal faults along which movement was predominantly syndepositional. Between the basins, Palaeozoic basement massifs form the land areas of Wales and the offshore ridges of the St Tudwal's Arch and Pembroke Ridge.

The present morphology of the sedimentary basins, which is outlined by the pattern of Bouguer gravity anomalies (Figure 14), largely reflects the process of crustal extension which was instrumental in their formation. Extension-related basin subsidence was punctuated by episodes of uplift and minor compression, the consequent erosion leading to marked regional variations in the stratigraphy of the basin fill. Permo-Triassic, Jurassic, and locally Lower Cretaceous strata, form essentially synrift sedimentary sequences that are unconformably overlain by a variable postrift sequence of mid-Cretaceous to Tertiary age. Maximum sedimentary thicknesses are found in the St George's Channel Basin, where more than 10 000 m of sedimentary rocks are preserved, although up to 3000 m of these are likely to be of Carboniferous age. Other principal depocentres are the North Celtic Sea Basin containing more than 5000 m of Permian and Mesozoic rocks, the Cardigan Bay Basin with more than 4000 m (although some of this sequence may be of Carboniferous age), and the South Celtic Sea Basin in which a succession more than 4000 m thick is preserved. The most important faults within the area make up the Mochras–Tonfanau–Bala–St George's fault zone, a roughly north-easterly trending system of en-échelon, syndepositional, normal faults which in places has throws in excess of 4000 m. This fault zone forms the south-eastern margin of the Cardigan Bay and St George's Channel basins (Figure 13).

Structural and plate-tectonic setting

Prior to the onset of basin subsidence, the structural framework of the region was formed in Early to mid-Palaeozoic times. Convergence of the Avalonian, Laurentian and Gondwanan continental masses during this time led to the development of the Caledonian and Variscan foldbelts within the supercontinent of Pangaea. Thus, the basement rocks upon which the sedimentary basins developed exhibit considerable structural inhomogeneity (Figure 15).

In the northern part of the area, rocks of the Caledonian foldbelt, strongly deformed at around 390 Ma, have a predominantly north-east to south-west structural trend. Locally, important fracture zones such as the Menai Straits fault system, the Llyn shear zone, and the Bala Fault can be identified either at outcrop or on seismic-reflection data. The Caledonian basement is in places unconformably overlain by weakly deformed Carboniferous strata, which on seismic pro- files are not easily differentiated from the overlying Permo-Triassic beds.

In the southern part of the area, the Variscan foldbelt, deformed at about 300 Ma, has a roughly east–west structural grain which in places is cut by several large thrust and wrench faults. The most northerly major thrust, the Variscan Front, is thought to mark the northern limit of pervasive Variscan deformation. A third tectonic province, the relatively undeformed Midlands Microcraton, is present in south-east Wales, where it formed a foreland to both the Caledonian and Variscan foldbelts.

On deep-seismic reflection profiles, the strongly tectonised rocks of the foldbelts typically have an unreflective seismic signature (Figure 18), Section 3.) See (Figure 18) for location." data-name="images/P945215.jpg">(Figure 16), though locally, important fracture or thrust zones are imaged as gently dipping reflectors (Cheadle et al., 1987). At greater depth, the unreflective foldbelt terrains pass downwards into seismically layered lower crust. The significance of this seismic layering is uncertain, but it may be due to the presence of igneous layering, metamorphic differentiation, extensional tectonic fabrics, or anastomosing shear zones. Whatever its origin, it is thought that seismically layered lower crust behaves in a ductile manner, deforming by some form of aseismic creep (e.g. Matthews and Cheadle, 1986; Meissner and Kusznir, 1987).

An end-Carboniferous (286 Ma) plate-tectonic reconstruction of part of Pangaea (Figure 17)a shows the relative positions of the present-day circum-North Atlantic continents prior to the onset of lithospheric extension in Permian times. (Figure 17)b illustrates a plate-tectonic reconstruction for mid-Cretaceous times (105 Ma), subsequent to the major extensional phases of basin development and immediately prior to the onset of North Atlantic sea-floor spreading. Relative continental motion from Permian to Early Cretaceous times can be assessed by comparing the two reconstructions.

Except for the extreme south-west of the North Atlantic region, where sea-floor spreading commenced earlier, these relative plate motions can be explained by continental lithospheric extension within Pangaea. From Permian to Mid-Jurassic times, regional east–west continental extension took place between Europe and Canada, with widespread basin subsidence on the present-day continental shelves. By the Late Jurassic, and particularly by Early Cretaceous times, active sea-floor spreading had propagated northwards to the southern part of the North Atlantic region (Figure 17)b. Separation of Newfoundland and Iberia triggered pulses of incipient rifting west of Britain, with a complex interplay of local extensional subsidence and regional thermal uplift. The onset of true North Atlantic sea-floor spreading in mid-Cretaceous times marked the cessation of continental extension and introduced a regime of passive thermal-relaxation subsidence, locally modified by episodes of uplift and inversion.

Basin structure

The cross-sections shown in (Figure 18) illustrate the form of the post-Variscan sedimentary basins and the intervening Palaeozoic basement massifs. The basins can be divided into two groups on the basis of both their structural trends and the stratigraphy of their sedimentary fills (Figure 19), (Figure 20) and (Figure 21). North of the Variscan Front (Figure 13), the sedimentary basins have a dominantly north-east to south-west 'Caledonoid' structural trend; thick Permo-Triassic and Jurassic sequences are overlain with marked unconformity by Tertiary strata, and Cretaceous beds are generally absent. In contrast, south of the Variscan Front, the sedimentary basins have a roughly east–west, 'Variscoid' structural trend. They are infilled by Permo-Triassic, Jurassic, and litho- and chronostratigraphically variable Cretaceous sequences, with only relatively thin Tertiary strata.

Basins north of the Variscan Front

Central Irish Sea Basin

In the far north of the report area, the Central Irish Sea Basin (Figure 13) has an asymmetric form, deepening towards the east-south-east (Figure 18), Section 1. The sedimentary fill is mostly of Carboniferous and Permo-Triassic age, although thin Jurassic strata may also be preserved. The infill is thickest close to the major normal faults, linked to the Menai Straits fault system, that form the south-eastern margin of the basin where it abuts against the north-western edge of the Palaeozoic basement massif of the St Tudwal's Arch. It is likely that development of the basin-margin normal faults is related to extensional reactivation of the Menai Straits fault system and the Llŷn shear zone; the latter is imaged on deep-seismic reflection profiles as a low-angle, north-westerly dipping reflector which has been interpreted as a basement thrust (Brewer et al., 1983). Patches of Tertiary strata lie unconformably upon the Mesozoic sediments, and overlap the basin-margin faults with little evidence of faulting (Figure 18), Section 1.

Cardigan Bay Basin

South-east of the St Tudwal's Arch, the Cardigan Bay Basin forms a half-graben which deepens to the south-east (Figure 18), Section 1. Major normal faults, the Mochras, Tonfanau and Bala faults, form the eastern and south-eastern margins of the basin (Figure 13). The age of the sedimentary fill is constrained both by the findings of the Mochras borehole (Woodland, 1971) and by seismic-reflection data (Bullerwell and McQuillin, 1969; Dobson and Whittington, 1987). Sequences identified as Permo-Triassic and Jurassic are both locally more than 2000 m thick (Figure 19) and (Figure 20), although the former estimate may also include strata of Carboniferous age. Cretaceous sediments are thought to be entirely absent, with Permo-Triassic and Jurassic beds being unconformably overlain by Tertiary strata that are locally more than 500 m thick (Figure 21). The basin is notable in that basin-margin faulting continued during or after deposition of the Tertiary beds, whose easterly extent, like that of the underlying Mesozoic strata, is fault bounded.

St George's Channel Basin

The St George's Channel Basin lies to the south-west of the Cardigan Bay Basin, and is separated from it by a low-relief saddle extending southwards from the St Tudwal's Arch (Figure 13). It is one of the deepest basins on the UK Continental Shelf, with a post-Variscan sedimentary fill locally in excess of 7000 m thick, underlain by up to 3000 m of undeformed Carboniferous strata. In its eastern part, the basin has the form of an asymmetric, faulted syncline, that is deepest close to its south-east margin. Farther west, it is more clearly defined as a south-easterly deepening asymmetric graben (Figure 13) and (Figure 18), Section 2. The south-eastern margin of the basin is marked by the St George's Fault, a major sub-planar to slightly listric normal fault which forms the en-échelon westerly prolongation of the Bala Fault. It probably developed by extensional reactivation of an earlier Caledonian feature, and has a throw, much of it syndepositional, which in places exceeds 4000 m (Figure 18), Section 2.

The stratigraphy of the basin fill is not well known, for no wells or boreholes have penetrated the complete succession; consequently, interpretations of the seismic-reflection data are poorly constrained (compare (Figure 22)). For location see (Figure 1)." data-name="images/P945202.jpg">(Figure 3) and (Figure 22). Nevertheless, it is likely that considerable thicknesses of both Permo-Triassic and Jurassic strata are present. In places, the Permo-Triassic sequence is thought to be more than 4000 m thick (Figure 19), with a significant proportion of salt (Dimitropoulos and Donato, 1983; Dobson and Whittington, 1987).

There is evidence of considerable migration of salt from the basin depocentre into the St George's Fault, forming a massive, linear salt wall along the basin margin (Figure 13) and (Figure 22); the salt wall has a length of at least 50 km, although it is rarely more than 3 km wide. Dimitropoulos and Donato (1983) have suggested that steep lithostatic pressure gradients within the narrow and deep St George's Channel Basin have caused the salt to flow towards the basin margins in a highly constrained manner, rather than forming randomly distributed swells and diapirs as are found in much wider and flatter basins, such as in the southern North Sea (Cameron et al., 1992). The timing of salt migration is difficult to establish, but it appears to have begun in post-Jurassic, possibly Early Cretaceous times, and may well have continued intermittently until the present.

The Jurassic sequence is in places more than 4000 m thick. Patterns of deposition differ slightly from those of the Permo-Triassic, in that the Jurassic (and Early Cretaceous?) depositional basin was even more strongly fault controlled, as indicated by the relative thickness changes across the St George's Fault (Figure 18), Section 2 and (Figure 20). Cretaceous strata have not been proved in the basin, and may be entirely absent, but the possibility remains of Lower Cretaceous Wealden beds being preserved in the axial part of the basin.

Tertiary beds, in general, rest with marked unconformity upon Jurassic strata, and form a largely unfaulted, postextensional sequence that is locally more than 1500 m thick (Figure 21). An exception to this general lack of faulting is the St George's Fault, across which Tertiary strata are down-faulted and draped into the basin to the north (Figure 18), Section 2 and (Figure 22). Similar postextensional normal faulting has been noted in North Sea basins (Badley et al., 1984), and is at least in part due to differential compaction of underlying strata across the basin margin. Deformation of a different kind affects the Tertiary beds above a large antithetic fault (Figure 22), where they are folded into a tight syncline, and a complementary anticline from which they are now largely eroded. This is apparently a consequence of reverse movement on the antithetic fault related to minor structural inversion. There is no evidence of similar reverse movement on the more important St George's Fault on this data, but reversal of the fault has been reported elsewhere (Coward and Trudgill, 1989).

Basins south of the Variscan Front

North Celtic Sea Basin

Situated along strike, en échelon with the St George's Channel Basin, only the eastern part of this major basin falls within the report area. Lying astride the Variscan Front (Figure 13), the basin is something of a hybrid; structurally it displays the dominantly north-easterly 'Caledonoid' trend characteristic of the basins north of the Variscan Front, whereas the presence of a fairly thick Cretaceous sequence is more akin to those basins farther south.

The North Celtic Sea Basin has the form of a rather symmetrical, faulted downwarp (Figure 18), Section 3, lying for the most part in the hanging wall of the Variscan Front Thrust. It is thought that extensional reactivation of the thrust played a part in the structural evolution of the basin (Cheadle et al., 1987), but relict Caledonian structures were also significant in controlling basin morphology. The North Celtic Sea Basin differs from basins to the north-east in that its Permo-Triassic sedimentary fill is generally much thinner than the overlying Mesozoic strata (Tucker and Arter, 1987). Strongly fault controlled, the Permo-Triassic sequence is locally in excess of 3000 m thick, but is generally less than 2000 m thick (Figure 19). Jurassic and Lower Cretaceous (Wealden) strata also form part of the synextensional sequence, and are in places well over 3000 m thick (Figure 20). Synsedimentary faulting became less pronounced, and by mid-Cretaceous (Aptian) times had virtually ceased, with subsequent deposition of a postextensional 'sag' sequence of middle to Upper Cretaceous sediments. The latter are mostly Chalk, which is now best preserved over the Pembroke Ridge (Figure 18), Section 3. A patchy cover of Tertiary strata unconformably overlies the Chalk (Figure 2).

South Celtic Sea Basin

Like its northern counterpart, only the eastern and central parts of the South Celtic Sea Basin lie within the report area. In the east, the basin has the form of an elongated, faulted trough, with a dominant east-north-east to west-south-west 'Variscoid structural trend (Figure 13) and (Figure 18), Section 2. Farther west (Figure 18), Section 3, an axial normal fault, with northerly downthrow, becomes a prominent structural feature. This fault is associated with a graben on the northern flank of the basin (Van Hoorn, 1987). The sedimentary fill comprises a synextensional Permo-Triassic sequence that is strongly fault controlled, and locally over 3000 m thick (Figure 19). Jurassic and, in the west, Lower Cretaceous strata also form part of the synextensional sequence, but faulting is less significant. The maximum thickness of Jurassic and Lower Cretaceous strata locally exceeds 1500 m (Figure 20).

In Early Cretaceous times, late-Cimmerian regional uplift led to erosion of the Permo-Triassic to Lower Cretaceous sequence (Van Hoorn, 1987). Erosion was particularly severe over the basin margins, where salt diapirism may have augmented uplift. Sedimentation resumed in mid-Cretaceous times, and a postextensional sequence of mid-Cretaceous to Tertiary age unconformably overlies older strata (Figure 18), Section 4 and (Figure 21). Subsequent minor compression and basin inversion in mid-Tertiary times produced a mild upwarp of the basin depocentre (Figure 18), Section 3.) See (Figure 18) for location." data-name="images/P945215.jpg">(Figure 16), with minor reversal of the faults on the basin's southern margin.

Bristol Channel Basin

The Bristol Channel Basin, which forms the easterly en-échelon continuation of the South Celtic Sea Basin, is markedly elongate along an east–west Wariscoid' structural trend (Figure 13). In the east, it has the form of a northerly deepening asymmetric graben (Figure 18), Section 4 that is bounded to the north by the east–west-trending Central Bristol Channel normal fault zone. This important structural feature has a length of more than 190 km, and is thought to have formed by extensional reactivation of an underlying Variscan thrust (Brooks et al., 1988). The basin structure is complicated by a set of north-westerly trending wrench faults that are probably reactivated Variscan structures.

Towards its western end, the basin has a more symmetrical cross-section as it passes en échelon into the South Celtic Sea Basin (Kamerling, 1979). The basin-fill stratigraphy is thought to be similar to that of the Wessex Basin of southern England (Whittaker, 1985). A dominantly synextensional depositional sequence of Permo-Triassic to Early Cretaceous age reaches a maximum thickness of over 2000 m close to the Central Bristol Channel fault zone (Figure 19) and (Figure 20). Except for small outliers near the West Lundy (Cambeak Fault of Van Hoorn, 1987) and Sticklepath fault zones, younger sediments are largely absent (Figure 21), having been removed by erosion.

Within the Bristol Channel Basin, but in a sense structurally distinct from it, two minor basins lie adjacent to the important north-west to south-easterly trending Sticklepath and West Lundy wrench faults (Figure 13) and (Figure 21). The larger of these basins, the Stanley Bank Basin (Fletcher, 1975; Arthur, 1989), has the Sticklepath fault zone as its south-western boundary. Up to 300 m of Tertiary strata are present in this basin. Similar small basins lie along the Sticklepath Fault onshore in Devon at Petrockstow and Bovey Tracey see (Figure 49). The geometry of these basins suggests a pull-apart origin during early Tertiary, sinistral, strike-slip displacement along the fault (Holloway and Chadwick, 1986). It is likely that these minor basins all formed by sinistral reactivation of Variscan wrench faults.

Arthur (1989) has proposed that Tertiary sediments also accumulated over the area between the West Lundy and Sticklepath–Lustleigh fault zones, prior to its uplift as the Lundy Horst.

Structural evolution

The post-Variscan structural evolution of the report area was largely dominated by intraplate extensional processes which led to the formation of a complex system of sedimentary basins across north-west Europe. Crustal extension and lithospheric thinning from Permian to Early Cretaceous times led to the deposition of thick, synextensional, sedimentary sequences in fault-bounded basins. By mid-Cretaceous (Aptian) times, sea-floor spreading had propagated into the North Atlantic region, and postextensional, largely unfaulted shelf sequences were deposited unconformably upon the synextensional older strata, commonly overlapping the older, faulted, basin margins. Basin subsidence was punctuated by pulses of regional uplift and minor local inversion; the consequent erosion resulted in marked variation of preserved basin fills.

Permo-Triassic extensional subsidence

In Permo-Triassic times, east–west-directed regional crustal extension (Chadwick et al., 1990), driven by lithospheric tensional stresses, led to the development of rapidly subsiding, fault-bounded, intermontane basins such as the St George's Channel, North and South Celtic Sea, and Cardigan Bay basins. Deposition was largely limited to the basinal areas, and the intervening structural highs were severely eroded. The degree to which extensional faulting controlled basin development at this time is difficult to assess because of the uncertainty in estimating thickness variations across the basin-margin faults. For example, in the St George's Channel Basin, it is possible that thickening of Permo-Triassic beds across the St George's Fault was relatively minor (Figure 22), and certainly much less than the thickening of Jurassic and younger sediments. It may be, therefore, that the role of Permo-Triassic faulting was not as significant as commonly supposed (e.g. Ziegler, 1978).

Early Jurassic extensional subsidence

The Early Jurassic tectonic regime was characterised by continued extension and subsidence of the fault-bounded basins. The effects of regional thermal-relaxation subsidence inherited from the Permo-Triassic lithospheric thinning and the erosion of structural highs, led to a rise in relative sea level, the establishment of a marine environment, and to a progressively increased area of deposition.

Mid-Jurassic ('mid-Cimmerian') thermal uplift

In mid-Jurassic times, subsidence was interrupted, particularly over the basin margins, by a significant hiatus. This hiatus is ubiquitous over the UK Continental Shelf, and may be related to uplift resulting from some form of thermal doming (cf. Dobson and Whittington, 1987), possibly associated with incipient rifting beneath the North Sea (Ziegler, 1990).

Late Jurassic extensional subsidence

Sedimentation resumed over much of the report area in Late Jurassic times, with further extensional faulting and basin subsidence. Thermal relaxation resulted in an increasingly flexural style of subsidence, a rise of relative sea level, and a likely further increase in depositional area.

Early Cretaceous ('late-Cimmerian') structural demarcation

Early Cretaceous times saw major tectonic changes which culminated in the late-Cimmerian unconformity'. Intraplate extension, which had been dominant for so long, became severely modified by embryonic plate-margin processes as sea-floor spreading penetrated northwards from the Central Atlantic region towards Newfoundland and Iberia (Figure 17)b. Rotation of the latter imported strong north–south-orientated extension to the report area. In addition to renewed crustal extension, the area was affected by regional thermal uplift as incipient lithospheric rifts propagated northwards from the active spreading centre to the south. Severe structural demarcation thus ensued, with enhanced normal faulting, for example along the Central Bristol Channel fault zone and the St George's Fault, causing local basin depocentres, such as the North Celtic Sea Basin, to receive thick, nonmarine, Lower Cretaceous Wealden deposits. However, elsewhere there was widespread erosion of earlier sequences. Uplift was particularly severe over the basin margins; for example, over 3000 m of sediment may have been removed from parts of the southern margin of the South Celtic Sea Basin (Van Hoorn, 1987). In places, margin uplift was increased by salt intrusion, notably in the St George's Channel Basin, and by oblique-slip reactivation of the basin-margin faults (Van Hoorn, 1987).

These late-Cimmerian' movements constitute the most significant tectonic event to affect the post-Variscan sediments of the area, and had a significant influence on the present-day morphology of the sedimentary basins. It is important to stress, however, that this erosional episode was not the result of crustal compression, and that it did not constitute true basin inversion, in that basin depocentres were not preferentially uplifted; erosion was most severe over the basin flanks and the intervening highs. To quote Ziegler (1987, p.418), late-Cimmerian movements (... clearly predate the Alpine plate collision and are undoubtedly of a non-orogenic nature'. In fact, the late-Cimmerian unconformity' is a good example of an unconformity associated with the transition from extensional to postextensional (the 'rift-to-drift) phases of basin evolution.

Mid- to Late Cretaceous shelf subsidence

Mid-Cretaceous (Aptian) times saw the onset of sea-floor spreading in the North Atlantic region, and the establishment of post-extensional, regional, shelf subsidence. Middle and Upper Cretaceous sedimentary sequences were deposited unconformably upon the late-Cimmerian' erosion surface. Depositional patterns are difficult to establish because of later erosion; no Cretaceous sediments have yet been proved in the report area north of the Variscan Front. However, the lack of terrigenous clastic material in the Upper Cretaceous Chalk suggests that it was laid down over much of the region (Tucker and Arter, 1987), with considerable overlap of the earlier basin margins, and covering much of the Pembroke Ridge.

End-Cretaceous ('Laramide) uplift

Deposition of the Chalk was followed by regional erosion roughly synchronous with the end-Cretaceous 'laramide' basin inversion observed in the southern North Sea (Van Wijhe, 1987). Uplift appears to have been particularly severe north of the Variscan Front, where it is believed that a thick, middle to Upper Cretaceous sequence was removed. South of the Variscan Front, erosion was most severe over the North Celtic Sea Basin (particularly in the west) and in the Bristol Channel Basin, where no middle or Upper Cretaceous sediments are now preserved. The precise mechanisms of uplift are uncertain. In the North Celtic Sea Basin, upwarping of the basin depocentre and reversal of the basin-margin faults (Tucker and Arter, 1987) are indicative of crustal shortening and true basin inversion. Elsewhere, the uplift was more regional, and specific evidence of crustal shortening is lacking. It may be that thermal uplift played a part, related to intrusion of the Lundy Granite (Figure 13) at 55 to 50 Ma (Miller and Fitch, 1962).

Early Tertiary shelf subsidence and wrench faulting

Following the 'laramide' uplift, early Tertiary times saw a return to regional thermal-relaxation subsidence and the deposition of clastic sequences. No true extensional faulting took place, but sinistral strike-slip movements on the Sticklepath and West Lundy fault zones (Holloway and Chadwick, 1986), and on the marginal faults of the Cardigan Bay Basin, led to the development of local fault-bounded basins (Figure 21). It is thought that major uplift of the Welsh and Cornubian Palaeozoic massifs may have occurred at this time (Dobson and Whittington, 1987).

Mid-Tertiary ('Helvetic') basin inversion

Sedimentation was interrupted in Oligo-Miocene times by a pulse of crustal shortening and minor basin inversion. This corresponded to the main structural inversion of the onshore Wessex Basin (Lake and Kamer, 1987; Chadwick, 1993), and was related to the 'Helvetic' phase of Alpine collision. Basin uplift was generally rather limited, and most pronounced south of the Variscan Front. The North and South Celtic Sea basins both experienced gentle upwarping of their depocentres, and local reversal of their basin-margin faults (Figure 18), Section 3.) See (Figure 18) for location." data-name="images/P945215.jpg">(Figure 16) and (Figure 18), Sections 2 and 3). This led to a thicker preservation of the postextensional sequence on the Pembroke Ridge than in the basins (Figure 18), Section 3 and (Figure 21). North of the Variscan Front, evidence of inversion is restricted to reversal of the antithetic fault in the St George's Channel Basin (Figure 22), and local upwarping of the basin depocentre. Inversion was accompanied by dextral reactivation of the Sticklepath and West Lundy fault zones, which caused minor compressional deformation of their associated pull-apart basins (Bristow and Hughes, 1971). Subsequent to inversion, the area has probably remained relatively close to sea level up to the present.

Mechanisms of crustal and lithospheric extension

BIRPS deep-seismic reflection profiles cast considerable light upon the style of crustal extension within the report area. The amount of crustal thinning, and therefore extension, varies markedly (Figure 23). Beneath the deepest part of the St George's Channel Basin, pre-Permian 'basement' crust is only 23 km thick, and beneath parts of the Bristol Channel/South Celtic Sea Basin it is 25 km thick. By contrast, the basement massifs of the Cornubian Platform and the Pembroke Ridge are underlain by 'basement' crust 30 to 32 km thick. Taking into account the fact that these massifs have suffered considerable erosion, the latter, higher, value probably affords the better estimate of pre-extension average crustal thickness.

In the upper and middle crust, extensional thinning took place by brittle faulting, with considerable crustal attenuation beneath the sedimentary basins. In the lower crust, extension was more evenly distibuted and is thought to have occurred in a dominantly ductile manner (Cheadle et al., 1987). There is some evidence of more-pronounced lower-crustal thinning beneath the sedimentary basins of the report area (Figure 23), but this is somewhat equivocal; because of data limitations, the precise thickness of the seismically layered lower crust is not well constrained (BIRPS and ECORS, 1986).

The amount of crustal thinning which accompanied basin formation can be estimated directly from (Figure 23). A crustal thinning factor t can be defined as:

t = thickness of extended crust/thickness of pre-extension crust

Thus t is smallest beneath the sedimentary basins; t = 0.72 beneath the deepest part of St George's Channel Basin, and t = 0.78 beneath the deepest part of the Bristol Channel/South Celtic Sea Basin. It is largest beneath the basement massifs where it is assumed that t = 1. The amount of crustal thinning depends largely upon the geometry of faulting in the upper and middle crust, and thus minima of t are extremely^, localised. A more meaningful quantity is given by the regional crustal thinning, averaged along the whole cross-section. The thickness of pre-Permian 'basement' crust averaged along the cross-section (Figure 23) is 28.8 km. Thus the regional crustal thinning factor t_R is given by:

t_R = 28.8/32.0 = 0.90

For extension in the plane of section, the crustal extension factor  γ is defined as:

length of extended crustal section/length of pre-extension crustal section

If crustal cross-sectional area is conserved during extension, then ã is given by:

γ = average thickness of pre-extension crust / average thickness of extended crust = 32.0/28.8 = 1.11

Thus, the bulk crustal extension factor along the section in (Figure 23) is 1.11, or 11 per cent, and the cross-section, whose present length is 210 km, experienced some 21 km of extension from Permian to Early Cretaceous times. In detail, this extension has a markedly heterogeneous distribution, being concentrated beneath the basins, and negligible beneath the basement massifs.

Another simple geometrical relationship (though not strictly valid for isostatically balanced basins) gives an estimate of the depth at which the basin-controlling normal faults detach, or flatten-out, within the crust (Gibbs, 1983). Assuming conservation of cross-sectional area:

D =   A/∆E

where:

D is the depth of detachment

A is the cross-sectional area of the post-Variscan sedimentary basins

∆E is the amount of crustal extension along the cross-section.

Measuring from (Figure 23):

A = 389 km²

∆E = extended section length less the pre-extension section length

∆E=210 − 210m/1.11 = 20.8km

Thus D= 389/20.8 = 18.7

This computed depth to detachment lies close to the top of seismically layered lower crust (Figure 23), supporting the contention that the lower crust extends in a ductile manner, its upper surface effectively forming a zone of detachment.

The manner in which subcrustal (mantle) lithosphere deforms is not well understood, and is the subject of much current debate. The simplest model (McKenzie, 1978) proposes a lithosphere which extends by pure shear to produce a symmetrical lithospheric cross-section. The model predicts significant postextensional (thermal-relaxation) subsidence situated above the earlier faulted basins. Wernicke (1985) suggested an alternative model in which the lithosphere extends by simple shear along a low-angle, deeply penetrating shear zone which offsets thinning in the crust from thinning in the deeper lithosphere. Postextensional subsidence predicted by this model is less than that predicted by the pure-shear model, and is laterally offset from the earlier faulted basins.

The presence of postextensional sedimentary rocks within the report area (Figure 21) is certainly indicative of some form of subcrustal lithospheric thinning, but its distribution is uncertain. Pure-shear lithospheric extension taking place over several tens of millions of years would produce basins in which the proportion of postextensional sediments varies between about 20 and 50 per cent of the total basin fill, depending upon the duration of extension and the dimensions of the basin (Cochran, 1983). The proportional thickness of the postextensional sequence preserved here is rather less than this, being zero in the Bristol Channel Basin, 15 per cent in the St George's Channel Basin, and about 20 per cent of the total thickness in the Celtic Sea area (Figure 13) and (Figure 21).

This relative paucity of postextensional sedimentary rocks has led to the suggestion of depth-variant modes of litho-spheric extension, including local simple-shear (Coward and Trudgill, 1989). However, the postextensional sequence has suffered severe erosion as a consequence of external processes such as thermal uplift and basin inversion; without these processes, these sediments would be much thicker. Therefore, it is suggested that the degree of uncertainty precludes positive identification of lithospheric extension mechanisms, and that pure-shear lithospheric extension with local heterogeneities imparted by upper- and middle-crustal faulting can account for most of the observed features of basin evolution hereabouts.

Chapter 4 Permian and Triassic

During the Early Permian, the final phase of the Variscan orogeny led to the consolidation of northern Europe into the supercontinent of Pangaea. The compressional regime then gave way to extensional stress, resulting in the formation of mainly fault-controlled basins whose orientation was determined by pre-existing structures. The location of Pangaea in low northern latitudes, together with the intracontinental position of northern Europe, resulted in an arid to semiarid climate. The Permian and Triassic sediments are therefore dominantly nonmarine redbeds which lack fossils; few successions can be dated, and the systems are commonly biostratigraphically undifferentiated.

In Early Permian times, basement massifs were uplifted; coarse-grained clastic sediments accumulated along their margins, passing basinward into sandstones, siltstones, and mudstones with evaporites. After an interval of quiescence during the Late Permian, further uplift in Early Triassic times led to a second episode of coarse-clastic deposition. During the Mid- and Late Triassic, the ingress of marginal-marine conditions led to a phase of siltstone-mudstone-salt deposition prior to marine conditions being established over the area at the end of the Triassic.

Permian rocks have been proved onshore only on the southern margin of the report area in the Wessex Basin (Figure 24)a, where their outcrop does not extend offshore (BGS Bristol Channel Solid Geology sheet). Offshore, Permian strata have been tentatively identified at depth in commercial wells on the basis of geophysical logs and lithostratigraphy. Triassic strata have been proved both onshore and offshore; they are thought to comprise almost all of the Permo-Triassic mapped as cropping out at the sea bed (Figure 2). Triassic rocks have been identified in commercial wells and BGS boreholes (Figure 24)b–d, but these are located on the basin margins or on structural highs, and few penetrate acoustic basement. The lack of stratigraphical information impedes interpretation of deep-seismic reflection data, for reflectors cannot be traced from the basin margins, across intrabasinal faults, into the basin centres. Subdivision of the thicker basinal sequences can therefore only be tentative, especially north of the Variscan Front (Figure 13), where the basin infill includes Permian, Triassic, and Jurassic strata, and, in contrast to some published accounts (e.g. Coward and Trudgill, 1989), Carboniferous rocks (Chapter 2). There are therefore few reliable estimates of Permo-Triassic thickness.

The limitations on the identification of seismic sequences constrains interpretation of the evolution of offshore basins. Onshore, on the periphery of the report area, the Permo-Triassic Central Somerset Basin is much better known, and is interpreted as an extensional rift initiated during the Permian (Whittaker, 1973). Along the basin margins, Triassic sediments rest on the pre-existing topography (George, 1970; Tucker, 1977; Whittaker and Green, 1983).

Evidence from the offshore basins indicates a similar depositional and tectonic history, although the timing of the earliest phase of basin evolution is unclear because of the lack of biostratigraphical control. Localised basin formation took place during the Permian and Early Triassic ((Figure 24)a–b; Ziegler, 1981; Shannon, 1991); basin trends were inherited from north-easterly trending Caledonian basement structures north of the Variscan Front, and from east-north-easterly aligned Variscan structures to the south of the front (Figure 13). It is likely, as shown in Chapter 3, that early deposition was accompanied by contemporaneous faulting (Tucker and Arter, 1987; Van Hoorn, 1987; Petrie et al., 1989), although other workers (Brooks et al., 1988; Bois et al., 1990) have interpreted the basins as epeirogenic sags. Later basin evolution during the Triassic saw an ever-increasing area of Triassic sedimentation, although post-Triassic erosion has reduced the extent of the deposits (Figure 24)c–d. The preservation of Permo-Triassic sediments is partly due to significant post-Triassic faulting (Brooks et al., 1988).

Permian

Identification of the Permian sequence is constrained by its nonmarine nature and lack of floral and faunal control, while correlation of sedimentary units is made difficult by variations in thickness and lithology (Whittaker et al., 1985). Furthermore, deposition in western Britain took place in localised basins. Thus, whereas major deformational events and sedimentary cycles may be interrelated between basins, detailed correlation of units is not possible. Onshore, Permian rocks crop out on the western margin of the Wessex Basin (Warrington and Scrivener, 1988; 1990), and may have been cored in boreholes in the Central Somerset Basin (Whittaker et al., 1985). Offshore, Permian strata are considered to be present in well 93/6-1 in the South Celtic Sea Basin, and well 103/2-1 in the St George's Channel Basin (Figure 24)a, (Figure 25) and (Figure 26).

Permian rocks at outcrop on the western margin of the Wessex Basin rest upon the Palaeozoic basement of Cornubia ((Figure 24); Edmonds et al., 1975). To the south, near Exeter, palynomorphs from beds low in the Permo-Triassic 'New Red Sandstone' sequence indicate Late Permian ages (Warrington and Scrivener, 1988; 1990). The sequence comprises breccias, sandstones and mudstones, which were deposited on an uneven, relict topography. Volcanic rocks, including potassium-rich vesicular lavas with subordinate olivine basalts, represent postorogenic volcanic activity associated with the emplacement of the granite batholiths of southwest England.

In the Central Somerset Basin, possible Permian strata have been identified in boreholes from their position underlying the Triassic Sherwood Sandstone Group (Whittaker et al., 1985). In the Burton Row borehole (Figure 25), the rocks are purple and brown-red, micaceous siltstone and fine-grained sandstone; in the Puriton borehole they are red, marly sandstone. Permian beds have not been identified offshore in the Bristol Channel Basin (Shannon, 1991).

In the South Celtic Sea Basin in well 93/6-1 (Figure 25) and (Figure 26), undated beds resting upon Carboniferous rocks, and overlain by the Sherwood Sandstone Group, are interpreted as Permian. They comprise 30 m of red-brown marl, overlain by 40 m of interbedded, red-brown shale and sandstone. The gamma-ray and sonic-velocity logs indicate that the lower (marl) unit is a uniform succession which probably coarsens upwards. The serrated character of the gamma-ray response in the upper unit confirms its interbedded nature, and shows the sandstone beds to be up to 6 m thick.

Within the deeper parts of the South Celtic Sea Basin, the existence of strata of Permian age has been inferred from study of deep-seismic reflection profiles (Kamerling, 1979; Van Hoorn, 1987). Early, possibly Permian, sedimentation may have been penecontemporaneous with faulting, but the basin may well have been initiated as a sag formed over a southward-dipping fault of Variscan origin (Coward and Trudgill, 1989).

North of the Variscan Front, the presence of Permian strata within the Central Irish Sea Basin is speculative. Permian strata have not been proved, but by analogy with the Kish Bank and East Irish Sea basins to the north of the report area (Jackson et al., 1994), the basin probably includes a Permian sequence (Shannon, 1991), possibly with Upper Permian evaporites. Lack of stratigraphical control limits interpretation of the seismic data, but Permo-Triassic rocks that occur at the sea bed (Figure 2) may extend down to a prominent unconformity within a synformal structure (Figure 11). The underlying, seismically bedded sequence is probably of Carboniferous, or perhaps Devonian, age.

On the southern margin of the St George's Channel Basin, well 103/2-1 (Figure 25) sampled an undated, 76 m-thick succession resting upon Carboniferous (Westphalian) strata. The succession is overlain by the Triassic Sherwood Sandstone Group, and is interpreted as being of Permian age. It comprises red-brown, variably calcareous mudstone and muddy sandstone, which is assigned a Permian age because it is lithologically distinct from the overlying Sherwood Sandstone Group. This difference is reflected in the geophysical logs; the gamma-ray response is higher and more serrated, and the sonic velocity slightly lower, than in the Sherwood Sandstone Group. The boundary with the underlying Carboniferous is taken at the top of a low gamma-ray and high sonic-velocity response which marks a sharp junction with underlying dolomite. Mudstones and siltstones in the Carboniferous section have lower gamma-ray and sonic-velocity responses than the Permian sequence.

Permian has not been proved in the centre of the St George's Channel Basin, but its presence there is inferred. By analogy with basins to the south, it may be expected to be thin or locally absent (Shannon, 1991), and to lack Upper Permian evaporites. The absence of evaporites from the southern part of the report area suggests that it was relatively elevated compared with the North Sea and Irish Sea, where inundation by the Zechstein and Bakevellia seas, respectively, occurred in Late Permian times.

Permian rocks have not been proved in either the Cardigan Bay Basin or North Celtic Sea basins. The formation of these basins may have begun during either the Late Permian (Ziegler, 1981) or the Early Triassic (Shannon, 1991).

Triassic

In Britain, three major lithostratigraphical units are recognised in the Triassic (Warrington et al., 1980); in ascending order these are the Sherwood Sandstone, Mercia Mudstone and Penarth groups. All can be identified in the report area (Figure 24)b–d, and are correlatable between the individual basins. The three groups have been subdivided in onshore basins on the basis of geophysical-log character (Whittaker et al., 1985; Lou et al., 1982; Holloway et al., 1989), and many of these subdivisions have been traced offshore (Penn, 1987).

Triassic rocks crop out along the basin margins, particularly in the south of the report area (Figure 2), and have been penetrated in wells and boreholes. Dating of the Triassic, especially in the middle and upper sections, is markedly better than for the Permian. The greater abundance of miospores in the sediments may reflect some climatic amelioration as a semiarid climate gave way to more humid conditions. For the Sherwood Sandstone Group in well 103/2-1 (Figure 27), late Scythian to Ladinian ages have been obtained. The Mercia Mudstone Group has yielded mainly Carnian to Rhaetian ages in its upper parts, although in well 102/28-1 the group includes ?Scythian and Anisian deposits in its lower part (Figure 28).

Triassic depositional limits extended beyond the localised Permian basins (Figure 24), and as sedimentation became more extensive, the basins merged (Figure 19). It is evident that by Early Triassic times (Figure 24)b, if not earlier, sedimentation had extended into the North Celtic Sea and Bristol Channel basins. Uplift of adjacent areas resulted in the deposition of coarse-grained, clastic sediment of the Sherwood Sandstone Group, which is relatively thinly developed (Figure 26) and (Figure 27).

Interconnecting basinal conditions became fully established in the Mid-Triassic (Figure 24)c. A subdued hinterland is indicated by a dominance of fine-grained mudstones in the Mercia Mudstone Group, although coarser-grained, clastic deposition continued along basin margins. A significant part of the succession in many wells is formed of evaporites.

In the Upper Triassic, grey-green mudstones of the Blue Anchor Formation are the youngest beds of the Mercia Mudstone Group; these were laid down in extensive, semipermanent lakes. Marine influences in the uppermost mudstones herald the fully marine conditions of the mid- to late Rhaetian, when the Penarth Group (Figure 24)d was deposited just before the close of the Triassic Period.

The Blue Anchor Formation is absent in the North Celtic Sea Basin, and there is north-eastward thinning of the Penarth Group in the South Celtic Sea and Bristol Channel basins. These observations suggest higher elevations towards the north-west and north, and imply that the marine incursion was later in the north of the report area. This interpretation is supported by evidence from the Mochras borehole (Figure 24)d, where sediments equivalent to the lower Penarth Group are dominantly nonmarine, but the latest Triassic sediments are marine. The southerly origin of the marine transgression is confirmed by the increasingly marine nature of the upper Penarth Group sediments towards the south-west in the South Celtic Sea and Bristol Channel basins.

Sherwood Sandstone Group

Onshore in the western Wessex Basin and in the Central Somerset Basin, the relationship of the Sherwood Sandstone Group with the underlying Permian is conformable in the basin centres, but unconformable at the margins. A similar relationship may occur offshore, where the group has been sampled in the Central Irish Sea, St George's Channel, South Celtic Sea, and Bristol Channel basins (Figure 24)b, (Figure 26) and (Figure 27).

In southern England, the Sherwood Sandstone Group was subdivided into three units by Whittaker et al. (1985); these are well displayed in the Burton Row borehole (Figure 27). The units are distinguished by their geophysical-log and lithological character, and represent an overall upward-fining sequence. At the base, unit SS1 is characterised by low gamma-ray values and relatively high sonic velocities, which represent poorly bedded pebble beds and conglomerates, with sandstone interbeds. Unit SS2 shows higher gamma-ray values, which increase upwards, and sonic velocities which decrease upwards. The profiles of both logs are more serrated, than those from unit SS1. The unit comprises sandstones with minor conglomerates and siltstone interbeds. Unit SS3 continues the trend of upward increase in gamma-ray values and decrease in sonic velocity, reflecting upward-fining into sandstones and siltstones.

Onshore, in the Wessex and Central Somerset basins, pebble beds and sandstones of the Sherwood Sandstone Group were laid down by a northward-draining fluvial system that originated in the Variscan mountains to the south (Wills, 1956; Holloway et al., 1989; Smith and Edwards, 1991). It is possible that analogous river systems laid down the sequences in the offshore basins, although a general coarsening of sediment towards the north suggests a sediment source from that direction. The overall upward-fining nature of the sequences suggests a waning coarse-sediment supply; this is indicative of an initial uplift event whose effects gradually decreased.

Offshore, the Sherwood Sandstone Group comprises sandstones, siltstones and mudstones (Figure 27); it is generally finer grained than onshore, perhaps reflecting a more-distal depositional environment. A threefold subdivision, comparable to that in the Burton Row borehole, is present in well 93/6-1 in the South Celtic Sea Basin, and was also recognised by Penn (1987) in well 103/2-1 in the St George's Channel Basin, where the three units are not as well defined. Units SS2 and SS3 are identified in well 103/18-1 in the Bristol Channel Basin, and unit SS3 farther south-west in well 102/28-1.

In the Bristol Channel Basin at well 103/18-1 (Figure 27), the section below 2100 m is interpreted as Sherwood Sandstone Group by comparison with adjacent wells. Although Penn (1987) assigned a Permian age to this sequence, the geophysical-log character shows similarities with the Sherwood Sandstone Group as illustrated by Whittaker et al. (1985). From 2100 m to the terminal depth at 2197 m, red-brown sandstone beds up to 4 m thick are interbedded with red-brown mudstone. In well 103/21-1 (Figure 27), 34 m of brick-red sandstone with marl interbeds lie below the Mercia Mudstone Group, and rest upon Devonian; although undated, this section is assigned to the Sherwood Sandstone Group.

In the South Celtic Sea Basin (Figure 27), the upward-fining sequence in well 93/6-1 comprises a lower 80 m of clean sandstone (unit SS1) which the geophysical logs show to be very uniform. This unit is overlain by 58 m of harder, interbedded sandstone, marl and limestone of unit SS2, which is succeeded by unit SS3 comprising 24 m of interbedded claystone and fine-grained, argillaceous sandstone. The gamma-ray response from unit SS3 indicates upward-fining towards the overlying Mercia Mudstone Group. To the north-east, in well 102/28-1, interbedded mudstone and siltstone are interpreted as equivalent to unit SS3 on the basis of their geophysical-log character.

In the St George's Channel Basin, the succession in well 103/2-1 (Figure 27) is of late Scythian to Ladinian age, and comprises 169 m of interbedded red-brown sandstone, siltstone and clay, with quartz pebbles towards the base (Barr et al., 1981). The succession differs from those in well 93/6-1 and the Burton Row borehole, as the geophysical logs show no overall upward-fining character. The three ill-defined component units shown in (Figure 27) are those established by Penn (1987).

To the north of well 103/2-1, 99 m of fine- to medium-grained sandstone, with individual beds 3 to 15 m thick, are present at the bottom of well 106/28-1 (Figure 27). The sandstone is well sorted and friable, with either a siliceous cement or an argillaceous matrix; it is interbedded with brick-red, slightly silty or sandy mudstone with anhydrite. These beds were assigned to the Sherwood Sandstone Group by Barr et al. (1981) on the basis of their sandy nature, together with their position beneath the Mercia Mudstone Group.

On the eastern margin of the Central Irish Sea Basin in BGS borehole BH71/53 (Figure 24)b, 5 m of grey, interbedded siltstone, sandstone and conglomerate, palaeontologically dated as Scythian to Anisian (Wilkinson and Halliwell, 1980), are assigned to the Sherwood Sandstone Group. The sand is coarse grained and quartzitic, with some plant debris; the conglomerate is formed of quartz pebbles. The siltstone is micaceous, sandy, and incorporates plant debris. The location of these sediments, possibly at the base of the Sherwood Sandstone Group, invites comparison with the basal sections farther south in the Central Somerset Basin.

Mercia Mudstone Group

The Mercia Mudstone Group has been proved at wells in all offshore basins except the Central Irish Sea Basin (Figure 24)c and (Figure 28). The group crops out at the basin margins, where it has also been sampled in BGS boreholes and shallow cores. In basin centres, there is conformable passage into the Mercia Mudstone Group from the underlying Sherwood Sandstone Group. Evidence from onshore areas shows that components of the group become progressively younger towards the basin margins, where they overlap older strata. In the Central Somerset Basin, the Mercia Mudstone Group oversteps the Sherwood Sandstone Group (Figure 26), and in South Wales, upper Mercia Mudstone Group sediments overlap Silurian, Devonian and Carboniferous rocks.

In offshore wells, the base of the Mercia Mudstone Group is taken, by comparison with the Burton Row borehole (Whittaker et al., 1985), at a marked change on gamma-ray and sonic-log profiles that reflects an upward passage from interbedded sandstones, siltstones and mudstones into mudstones (Figure 28). Four lithological units are distinguishable in offshore wells: in ascending order these are: (i) red mudstones and siltstones with subordinate sandstones, gypsum and halite; (ii) halites interbedded with claystones, mudstones, sandstones, dolomite and gypsum; (iii) red, commonly silty, claystones; and (iv) grey-green mudstones with subordinate dolomite and gypsum.

In many wells these four lithological units can be related (Penn, 1987) to the six geophysical-log units (A to F) identified in wells in the Wessex Basin (Lott et al., 1982), and traced to the Burton Row borehole by Whittaker et al. (1985). The lower mudstone/siltstone unit (i) in the offshore successions equates with units A and B, the evaporite unit (ii) with unit C, the upper red claystones (iii) with units D and E, and the grey-green mudstones (iv) with unit F.

Differentiation of units A and B is not always practicable offshore, as the lithologies are uniform mudstones and siltstones. Unit C comprises evaporites, with individual halite beds up to 40 m thick. The sands and dolomites of unit D in onshore wells were not previously recognised offshore, but have been traced through a prominent geophysical-log unit (Figure 28) which may represent a possible lithostratigraphical equivalent; its log character is distinctive, but the red clay-stones that form it are apparently lithologically indistinguishable from the succeeding mudstones that are correlated with unit E. The grey mudstones of unit F, a correlative of the Blue Anchor Formation (Warrington et al., 1980), are present in the South Celtic Sea and Bristol Channel basins, but absent from the St George's Channel Basin.

A wide range of depositional environments is represented by the Mercia Mudstone Group (Warrington and Ivimey-Cook, 1992). In South Wales, there are marginal and mudstone facies (Tucker, 1977; Waters and Lawrence, 1987), the former including continental and lacustrine shore-zone facies. The continental facies comprises breccias laid down as sub-aerial screes, as well as conglomerates and sandstones interpreted as streamflood and mudflow deposits, with interbedded sandstones and mudstones which are the products of sheetfloods. There is a rapid transition from these coarse-grained sediments into the lacustrine shore-zone facies, which is composed of mudstones and algal carbonates. The mudstone facies was deposited in either a hypersaline lake or an epeiric sea (Warrington, 1970), or on a playa, desert plain, or a supratidal flat (Wills, 1976). In South Wales, the evidence indicates deposition in large, ephemeral water bodies that periodically dried up (Waters and Lawrence, 1987).

Offshore, the dominance of fine-grained claystones, mudstones and siltstones, together with the presence of thick halites, suggests basinal conditions some distance from the sources of the continental facies of South Wales. The ability to correlate subdivisions of the group between basins indicates that the depositional areas were linked (Figure 24). Lack of core offshore hampers interpretation, but the depositional environment may have been comparable with that of the mudstone facies in South Wales. By analogy with Cheshire Basin sequences (Arthurton, 1980), it can be surmised that finer-grained, clay-sized material was probably transported in dust storms, and the coarser, silt- and sand-grade sediment by water.

The environment of halite deposition is imprecisely known (Warrington and Ivimey-Cook, 1992). Warrington (1974, p.708) considered that in Britain 'the evaporites may have been precipitated from water of marine origin which was evaporated under arid conditions in an environment dominated by 'continental influences'. Triassic evaporites in the East Irish Sea Basin to the north are interpreted as having been laid down in a coastal marine sabkha with strong continental influences (Jackson, et al., 1994). In South Wales, the grey-green mudstones of the Blue Anchor Formation, at the top of the Mercia Mudstone Group, have been interpreted as mainly lacustrine in origin, with marine influences in the uppermost part of the formation (Warrington and Ivimey-Cook, 1992).

Central Somerset and Wessex Basins

The Mercia Mudstone Group crops out along the northern and southern margins of these basins, where a 455 m-thick sequence has been proved in the Burton Row borehole (Figure 26) and (Figure 28). It is divisible into the six geophysical-log units established by Lott et al. (1982). Units A and B have moderately high, poorly serrated, gamma-ray and sonic-velocity values; unit A comprises siltstone and sandstone, and unit B is marl. Unit C is characterised by low gamma-ray values and a generally high sonic-velocity response; these reflect the halite-bearing nature of the unit, with interbeds of marl and halite imparting the serrated log character. Unit D displays an overall upward increase in gamma-ray values, and a decrease in sonic velocity compared with the underlying unit C; the rocks within this unit are commonly sandy and dolomitic. The log character of unit E is similar to unit B, reflecting a reversion to the deposition of monotonous marls. Unit F displays higher sonic-velocity values and lower gamma-ray values than unit E; it is dominantly a grey marlstone with dolomite interbeds, and is equivalent to the Blue Anchor Formation.

In the Wessex Basin, the Mercia Mudstone Group comprises mainly red, silty mudstones with thin sandstones; these pass upwards into the grey-green shales and mudstones of the Blue Anchor Formation (Edmonds et al., 1975).

Bristol Channel Basin

Mercia Mudstone Group rocks have been proved in wells 103/18-1 and 103/21-1 (Figure 24), (Figure 26) and (Figure 28), and are found at outcrop along the southern and northern basin margins where they have been sampled in BGS boreholes. The offshore outcrops merge with those onshore both in South Wales and in the Central Somerset Basin. Up to 160 m of Mercia Mudstone Group are present at Cardiff, where they rest upon a pre-existing, irregular topography (George, 1970; Tucker, 1977). The thickness increases southwards into the Bristol Channel Basin, where up to 500 m may be present (Brooks et al., 1988).

In the western part of the basin, 960 m of Mercia Mudstone Group were penetrated in well 103/18-1 (Figure 28); this increase is mainly due to the development of a Mid-to Late Triassic evaporite unit which separates lower and upper clastic units. The lower unit is mainly variably calcareous, red-brown mudstone with silty interbeds; gypsum and halite are present throughout. The evaporite unit comprises halite beds up to 50 m thick, with interbedded mudstone, claystone and sporadic sandstone. The upper clastic unit is dominantly red-brown, variably calcareous claystone; it passes upwards into the grey-green mudstone of the Blue Anchor Formation (unit F of Lott et al., 1982).

The Mercia Mudstone Group in well 103/21-1 (Figure 28) comprises dominantly red-brown mudstone, claystone and sandstone; it is similar to that in well 103/18-1, but lacks a significant development of thickly bedded evaporites, although anhydrite and halite are disseminated throughout the middle of the group as thin beds. Despite the absence of thick evaporites, there is a good correlation with the Mercia Mudstone Group in well 103/18-1; the thinner section is attributed to its location on a high.

BGS offshore boreholes in the basin (Figure 24)c have sampled red-brown mudstones and marls, with gypsum both in vein and disseminated form. For the most part the rocks are barren of miospores, but they are comparable lithologically to the Mercia Mudstone Group. Furthermore, Norian to Rhaetian ages have been obtained from borehole BH72/57, and a gravity-core sample at site 676 gave a Carnian to Norian age (Wilkinson and Halliwell, 1980). Many of the samples contain reworked Carboniferous and Devonian miospores.

South Celtic Sea Basin

Wells 93/6-1 and 102/28-1 proved comparable Mercia Mudstone Group successions (Figure 24)c, (Figure 26) and (Figure 28). Basal, monotonous, brown, calcareous and noncalcareous, silty clay-stones become less homogeneous and coarser grained upwards. The succeeding evaporites comprise halite beds up to 140 m thick, with interbeds of red-brown mudstones and dolomites. The upper claystone unit is divisible into three (Penn, 1987); unit D is 30 m thick in both wells and comprises red-brown claystones with subsidiary dolomites, whereas unit E is dominantly a red-brown claystone and is 130 m thick in both wells. Unit F comprises grey to black claystones and marls with subordinate dolomite and anhydrite; it correlates with the Blue Anchor Formation. The variable lithology in unit F is reflected in the serrated geophysical-log traces which show that dolomite becomes more common in the uppermost part.

St George's Channel Basin

The Mercia Mudstone Group is present in three wells in this basin (Figure 24)c, with over 1700 m penetrated in well 103/2-1. Upper and lower clastic units, composed mainly of mudstones and claystones, are separated by an evaporite unit (Barr et al., 1981).

The lower clastic section comprises dominantly brick-red, hard, calcareous mudstones and siltstones with some pale green interbeds. There are also siltstone and sandstone beds, and scattered anhydrite nodules. A core from well 106/28-1 (Figure 28) shows vertical and horizontal fractures interpreted as desiccation structures. The geophysical logs from the two wells reflect the uniformity of the lithologies, and although Penn (1987) identified units A and B in well 106/28-1, any geophysical-log change within the section is very subtle. Evaporites (unit C) are well developed. Well 103/2-1 was sited on a salt pillow and halite over 800 m thick comprises two subdivisions (Figure 28). The lower portion, with a serrated log character, represents interbedded halite, anhydrite, mudstone, sandstone and dolomite; the upper part is predominantly halite in beds up to 100 m thick. In well 106/28-1 (Figure 28) there are 750 m of interbedded mudstone and halite, the latter forming 30 to 40 per cent of the unit. The serrated geophysical-log character recorded in this well reflects mudstone and halite interbeds, with individual halites up to 30 m thick.

The upper clastic unit in both wells consists mainly of red-brown, variably calcareous claystones which may be silty and sandy (Figure 28). There is a considerable difference in the thickness of the post-evaporite sections in the two wells; 450 m are present in well 103/2-1, and 210 m in well 106/28-1. Although there is some uncertainty in correlation, it appears that the upper part of the section in well 106/28-1 is missing or condensed. Penn (1987) subdivided the mudstone into three units comparable with those in the Wessex Basin (Lott et al., 1982), but it is probable that only correlatives of units D and E are represented. The sandstones and dolomites that usually form unit D are not present, but there is a possibly equivalent unit with a distinctive log character that has gamma-ray and sonic-velocity values lower than those in the beds both above and below it. Unit E comprises red-brown claystones that are partly silty.

Triassic sediments, interpreted as Mercia Mudstone Group, have been penetrated in well 106/24a-2B, which is located (Figure 24)c on the flank of the 50 km-long salt wall in the St George's Channel Basin. The well entered the Triassic section at 2342 m below sea level, and penetrated 20 m of reddish brown sandstone and mudstone before drilling a sequence of anhydrite and halite (unit C of Lott et al., 1982) which extends to the terminal depth at 3220 m. Halite has also been identified within the basin on deep-seismic reflection profiles (Barr et al., 1981; Dimitropoulos and Donato, 1983).

Cardigan Bay Basin

No wells in this basin have penetrated beds older than latest Triassic (Penarth Group), although the presence of older Triassic rocks is indicated on seismic profiles (Chapter 3; Dobson and Whittington, 1987). BGS borehole BH71/47, on the north-western margin of the basin (Figure 24)c, sampled red-brown mudstone with grey-green bands and gypsum, lending support to the interpretation of Triassic beds at depth in the basin.

Penarth Group and equivalent rocks

The Penarth Group (Figure 29) comprises two formations: the Westbury Formation, and the overlying Lilstock Formation which is subdivided into the Cotham and Langport members (Warrington et al., 1980). The basal beds of the Lias, below the first appearance of the ammonite Psiloceras, are Triassic in age (Warrington and Ivimey-Cook, 1990), but because this first appearance cannot be identified from well data, the Triassic–Jurassic boundary is here informally taken at a convenient and identifiable geophysical-log marker.

The group is well exposed in the western part of the Central Somerset Basin (Edmonds et al., 1975), where the Westbury Formation comprises up to 15 m of very dark grey mudstones, with thin, argillaceous limestones, and bone beds. The overlying Gotham Member of the Lilstock Formation comprises just over 1 m of pale grey, calcareous mudstones, limestones, siltstones and sandstones. The Langport Member comprises about 3 m of pale grey, lenticular or nodular limestones, and grey or grey-blue mudstones which mark the passage up into the Lias Group.

The Westbury Formation in the St Fagans borehole (Figure 29) is composed of relatively radioactive, dark grey shale that is characterised on geophysical logs by a gamma-ray peak and low sonic velocity. The Cotham Member comprises calcareous, silty mudstone which, by comparison with the underlying black shale, has lower gamma-ray and higher sonic velocity responses. The Langport Member includes hard, fine-grained limestone which gives rise to a sonic velocity peak and low gamma-ray values. This interpretation differs from that of Penn (1987).

In offshore wells, the Penarth Group is identified on the basis of geophysical-log comparison with onshore wells (Figure 29). It is thickest and best developed in the South Celtic Sea and Bristol Channel basins, and thins towards the east and north. The group is absent in wells 106/28-1 and 106/24a-2B in the St George's Channel Basin (Figure 24)d, and has not been proved in the Central Irish Sea Basin. A marginal facies of similar age to the Penarth Group has been sampled from the Cardigan Bay Basin in the Mochras borehole.

The Westbury Formation offshore (Figure 29) is composed mainly of dominantly grey to grey-green, calcareous mudstones with interbedded limestones and dolomites. Individual beds are less than 2 m thick, and give rise to rapid, short-wavelength sonic-velocity and gamma-ray responses that are particularly well seen in wells 93/6-1 and 102/28-1, and which contrast markedly with the geophysical-log profiles of the underlying Blue Anchor Formation. The formation thins towards the east and north, and fluctuations in the gamma-ray and sonic-velocity logs become less extreme as the formation thins.

Succeeding the Westbury Formation is a sequence of calcareous mudstones and grey limestones of the Lilstock Formation; the sequences in the Bristol Channel and South Celtic Sea basins most closely resemble that in the Burton Row borehole. The mudstones of the Cotham Member are identifiable from high gamma-ray values and low sonic velocities, whereas the limestones of the Langport Member give a low gamma-ray response with high sonic velocity.

Bristol Channel Basin

The Penarth Group is recognised in both wells in this basin (Figure 29). In well 103/21-1, the group is 32 m thick, and the lower boundary is taken at an upward decrease in gamma-ray values and an increase in sonic velocity. The basal Westbury Formation comprises brown to grey, calcareous shale interbedded with limestone; individual beds are 1 to 2 m thick, and give rise to a varied, short-wavelength gamma-ray and sonic-velocity profile. The overlying Cotham Member is a brown-grey mudstone represented on the geophysical logs by a high gamma-ray and a low sonic-velocity response. The Langport Member is composed of hard limestone, and is defined by low gamma-ray values and a high sonic velocity.

The Westbury Formation thins to 10 m in well 103/18-1, where its base is placed at a decrease in the gamma-ray response compared with that of the underlying Mercia Mudstone Group. The geophysical-log character of the Lilstock Formation is similar to that in well 103/21-1.

BGS borehole BH75/08 (Figure 24)d recovered interbedded grey clay and limestone of probable Rhaetian age (Wilkinson and Halliwell, 1980); these bear a marked similarity to the cuttings descriptions from the wells. Individual beds are between 0.5 and 1.5 m thick.

South Celtic Sea Basin

The Penarth Group occurs in both wells in this basin (Figure 29). Its base is taken at a change from the serrated gamma-ray and sonic-velocity curves of unit F of the Mercia Mudstone Group, to the extreme and short-wavelength alternations of the Westbury Formation. The latter are especially evident on the sonic-velocity log, and reflect interbedded calcareous mudstones, dolomites and sandstones. The overlying Lilstock Formation comprises calcareous mudstones and limestones; a tentative subdivision into the Cotham and Langport members has been made, the latter being identified as a hard limestone marked by low gamma-ray values and high sonic velocity.

St George's Channel Basin

The Penarth Group in well 103/2-1 (Figure 29) is thinner than in the South Celtic Sea Basin; the top of the Mercia Mudstone Group is taken at a distinctive sonic-log marker where the velocity decreases sharply. Above this marker, the Westbury Formation gives rise to low sonic-velocity and gamma-ray values. The overlying Lilstock Formation is marked by high velocity and variable gamma-ray values; the Cotham Member is marked by a high gamma-ray peak and the Langport Member at a gamma-ray low. Cuttings descriptions from the well show the group to comprise shale, clay-stone, siltstone and dolomite; these are mainly grey, but there are some red-brown beds. In well 106/28-1 (Figure 24)d, the group has not been identified, and a distinctive log break at the top of the Mercia Mudstone Group is interpreted as the boundary of the Triassic with the Jurassic.

Cardigan Bay Basin

The oldest Triassic rocks sampled in the Mochras borehole (Figure 24)d are of Late Triassic, possibly Norian, age (Warrington, 1971). The sediments are of a facies not observed in deposits of this age elsewhere in Britain. The sequence was not logged geophysically, but two formations are present (Harrison, 1971).

The lower, Terrigenous Formation, is dominantly red-brown calcitic sandstone. Breccio-conglomerates are sandwiched between rhythmic alternations of calcitic sandstone and siltstone, gravel bands, silty limestone and marly shale.

Cross-bedding, wedge-bedding and load casts are present. Deposition has been interpreted as occurring in a playa or sabkha environment, with the coarser-grained units being deposited from flash floods.

The overlying Carbonate Formation consists of grey-green, dolomitised carbonates with calcitic, partly haematitic sandstone, which may be fine grained or gravelly. Although there are also bedded breccio-conglomerates, this formation marks a reduction in terrigenous input, and reflects the passage from a nonmarine to a dominantly marine environment of deposition (Dobson and Whittington, 1987). The biota is poorly preserved, but indicates a late Rhaetian age for the Carbonate Formation, which is probably a correlative of the Lilstock Formation.

Chapter 5 Jurassic

The Jurassic is largely preserved in major fault-controlled basins, (Figure 20) and (Figure 30), where it generally rests conformably on marine beds of the Upper Triassic Penarth Group. The Jurassic has a cover of Cretaceous and younger sediments in the North and South Celtic Sea basins; farther north it crops out or is overlain by Tertiary and Quaternary deposits. Jurassic strata also subcrop in the Kish Bank Basin immediately to the north of the report area (Jackson et al., 1994), and may occur in the Caernarfon Bay and Central Irish Sea basins (BGS Cardigan Bay Solid Geology sheet).

Onshore, adjacent to the report area, the Lower Jurassic is preserved immediately east of the River Severn, in north Somerset, and in the Vale of Glamorgan (Figure 30). The Lower Jurassic is relatively thin north-east of the Mendip Hills, thickening north-eastwards into the Vale of Gloucester and southwards into the Glastonbury Syncline. Middle and Upper Jurassic strata are confined to the main crop from Yorkshire to Dorset (Figure 31); they are absent in Wales, and from most of north Somerset apart from isolated outliers.

During the Jurassic Period, the report area lay within palaeolatitudes 30°N to 40°N, and was drifting steadily northwards (Smith and Briden, 1977). The climate may have been broadly comparable with present subtropical climates, with a strong seasonal rainfall and temperature variation in the range 8° to 24°C. Seasonal aridity is evidenced by the presence of redbeds of Bathonian age, and minor amounts of anhydrite at several stratigraphical levels indicate local evaporitic conditions. Eustatic sea-level changes have been documented in the northern North Sea (Vail and Todd, 1981), and there may be evidence of these changes to the west of Britain (Millson, 1987a; b).

Development of the basins in which the Jurassic is found was initiated during the Permo-Triassic or earlier. Thick, largely continental, Permo-Triassic strata usually underlie the Jurassic in the major basins (Figure 19 and (Figure 20). By the end of the Triassic, peneplanation of the landscape, combined with a relative rise in sea level, led to a widespread transgression and marine deposition. This trend continued throughout the Early Jurassic, which was also a time of steady basin subsidence; a sedimentation rate over 3.5 cm/1000 years has been estimated in the depocentres, compared with less than 1.7 cm/1000 years on the highs (Millson, 1987a).

Onshore, the Lower Jurassic has been proved in west Wales at Mochras (Ivimey-Cook, 1971; O'Sullivan et al., 1971), east of Wales in the Wem Outlier, and in South Wales in the Vale of Glamorgan (Figure 30). The Mochras borehole (see (Figure 33) proved 1305 m of marine sediments above late Rhaetian marine carbonates, which in turn overlie redbeds of terrigenous origin. The succession is about three times as thick as any other section in the British Isles (Ivimey-Cook in Allen and Jackson, 1985). The succession consists of interbedded mudstone and siltstone of basinal facies; despite its proximity to the Lower Palaeozoic Welsh Massif, there is little evidence of coarse-clastic material, slump structures, or other indicators of a contemporaneous adjacent shoreline or fault scarp.

In South Wales, near Bridgend, rocks of Hettangian and earliest Sinemurian ages overstep the Penarth Group to rest on Carboniferous Limestone. Lithologies comprise basinal calcareous mudstones and argillaceous limestones, with marginal or shoreline arenaceous facies preserved around Carboniferous 'islands'. It is likely that the Early Jurassic transgression of the Welsh Massif followed a similar pattern to that of the London Platform, where the Lower Lias overlaps the Triassic to rest on Lower Palaeozoic rocks. There is internal overlap within the Lower Lias, and the entire sequence thins towards the platform (Horton et al., 1987).

In late Toarcian or early Mid-Jurassic times, there was a widespread hiatus in sedimentation; this is recorded at the basin margins, where Bajocian/Bathonian sediments rest unconformably on Lower Jurassic clays. Mid-Jurassic earth movements, caused by the onset of sea-floor spreading in the Atlantic, led to an increase in the rate of subsidence, and some erosion on the basin margins. Sedimentation in the basins is estimated at 9.8 cm/1000 years during the Bathonian.

During the Mid-Jurassic, an extensive area of lagoonal/brackish deposition covered the East Midlands (Figure 31), passing northwards into fluviodeltaic sediments. To the south, shallow-marine shelf carbonate deposition prevailed around the Cotswolds, which were occasionally emergent and subject to erosion. South of the Mendip Hills, deposition of marine muds dominated in the Bathonian. Correlation of the resultant patchwork of diachronous facies is difficult onshore, and matching these with the offshore sequences is necessarily tentative. The contemporary coastline probably enveloped a Welsh promontory, and skirted the English Midlands and London Platform. Deltaic/lagoonal/nearshore-marine facies are found in the Cardigan Bay and St George's Channel basins, although anhydrite within the sediment indicates more-prolonged emergence.

By late Mid-Jurassic times, more-uniform marine conditions were re-established. The Bathonian–Callovian boundary is marked by a widespread, transgressive horizon which is succeeded by three, well-developed, upward-coarsening cycles. These cycles represent progradational, delta-fringe sediments which were probably deposited in somewhat deeper water (Millson, 1987a). The sandy units wedge out southwards and are replaced by carbonate beds in a thinner succession on the northern flank of the Pembroke Ridge (Figure 30). South of the ridge, the mudstones are more uniform and terrigenous, and shoal influences die out southwards. The Middle Jurassic is absent over much of the south-western part of the South Celtic Sea Basin, probably due to Late Jurassic/Cretaceous erosion.

Sedimentation rates reduced during the late Mid-Jurassic and early Late Jurassic, when basin subsidence and infill were cyclic. There is evidence of uplift and erosion to the south of the Pembroke Ridge during the Kimmeridgian, but in the St George's Channel Basin there was rapid deposition at an estimated rate of 10 cm/1000 years. Upper Jurassic strata are preserved only in the St George's Channel and Bristol Channel basins (Figure 30). In early Late Jurassic times, there was a return to shallow-water lagoonal sedimentation in the St George's Channel Basin, and during periodic subaerial exposure, desiccation produced anhydritic and reddened mudstones. In the Bristol Channel Basin (Evans and Thompson, 1979), Oxfordian rocks consist of thick, fine-grained sandstones interbedded with glauconitic siltstones, lignites, and dolomitic limestones. Charophytes and a restricted, dwarfed, molluscan assemblage suggest an estuarine or deltaic environment, and the entire sequence thins northwards, indicating derivation from the south.

During Kimmeridgian times a variety of rock types were deposited in the report area, but relatively uniform depositional environments prevailed over a wide area of southern England and the North Sea (Gallois and Cox, 1976; Cameron et al., 1992). Relatively shallow, marine-shelf conditions, and stratification of the water column, provided conditions for the efficient preservation of organic matter (Miller, 1990). These organic-rich clays are the principal petroleum source rock both onshore and in the North Sea; they occupy the core of the Bristol Channel Syncline (Figure 30), and individual beds have been traced across southern and eastern England (e.g. Gallois and Cox, 1976). However, such clays have not been proved in the South Celtic Sea Basin, where Portlandian strata rest on the lower Oxfordian (Penn, 1987). North of the Pembroke Ridge, a thick sequence of Kimmeridgian is preserved in the synclinal core of the St George's Channel Basin, but the organic content is much reduced and the microfauna is dominated by freshwater elements (Mil!son, 1987a), indicating a continuation of the lagoonal/nearshore conditions of Oxfordian times, with some elastic input from deltaic sources. Portland and Purbeck beds of possible lagoonal and sabkha origin are preserved in the Bristol Channel and St George's Channel basins.

This review is based largely on the studies of Millson (1987a) and Penn (1987). The nomenclature follows that of Penn (1987), with the addition of group names where appropriate; Millson's (1987a) terminology is nevertheless shown in (Figure 32). However, few of the formations used by these authors are formally defined. The use of geophysical logs to correlate onshore and offshore sequences (Penn, 1987) is beset with difficulties because of the small number of offshore wells, the much thicker succession offshore, the variability of the Middle and Upper Jurassic facies, the separation of the major basins, their internal faulting, and the development of salt walls. Nevertheless, stratigraphical equivalence of onshore and offshore sequences is tentatively assumed. Similarly, problems with the interpretation of seismic-reflection profiles have led to differences between published interpretations.

Lower Jurassic

The base of the Jurassic, taken at the first appearance of the ammonite Psiloceras planorbis J de C Sowerby), occurs 2 to 5 m above the base of the Lias Group onshore. Thus, the lowest part of the Lias Group is Late Triassic in age (Cope et al., 1980a), but in this report is treated together with the Lower Jurassic. The Lias Group consists of medium to dark grey, fossiliferous mudstones with a variable carbonate content, interbedded with argillaceous limestones, siltstones and, more rarely, fine-grained sandstones. The group is subdivided into the Lower, Middle and Upper Lias, the middle sequence being usually more arenaceous. Onshore, the Lias Group crops out in the Vale of Glamorgan, where only the lower part of the sequence (Blue Lias) is present, and in north Somerset where it is about 51 m thick and overlain by the Inferior Oolite Group on Brent Knoll ((Figure 30); Whittaker and Green, 1983).

Offshore, the Lias Group has been proved in all boreholes which have penetrated the base of the Middle Jurassic. In the St George's Channel and Cardigan Bay basins, thicknesses range from 393 m in well 103/2-1 to 1305 m at Mochras (Figure 30), (Figure 33), and (Figure 37)." data-name="images/P945233.jpg">(Figure 34); seismic-reflection evidence suggests that the thickness at Mochras is maintained, or even exceeded, along the axes of these basins (Barr et al., 1981). To the south, drilled thicknesses range from 494 m in well 103/18-1 to 1075 m in well 102/29-1 (Figure 35), and the sequence thins north-westwards towards the Pembroke Ridge, which was an axis of uplift.

Lower Lias

In the Bristol district (Figure 30), the Lower Lias is subdivided into the Blue Lias and the overlying Lower Lias Clay (Donovan and Kellaway, 1984). The Blue Lias is Hettangian to early Sinemurian in age, and the Lower Lias Clay is late Sinemurian to mid-Pliensbachian (Figure 32).

The Blue Lias is well developed in South Wales and north Somerset; 150 m were proved in a borehole near Bridgend (Wilson et al., 1990), and it is about 275 m thick at the north Somerset coast. The Blue Lias is interpreted as a shallow-marine deposit; it consists of cyclic alternations of limestone and shale that, where fully developed, form an upward-changing succession from shale to mudstone to limestone (Whittaker and Green, 1983). The shales are finely laminated, pyrite rich, bituminous in part, and include ammonites, fish and bivalves. The abundance of sulphides, in association with an absence of burrowing organisms, indicates deposition in anaerobic conditions. The mudstones are poorly fissile, mid to dark grey, calcareous, and show trace-fossil mottling. The limestones are dark bluish grey and porcellaneous; some are lenticular or nodular, others are uniform in thickness and can be traced over long distances. The rhythmic variation in these sediments has been attributed to changes in sea level (Hallam, 1964), but Donovan and Kellaway (1984) suggested that they were deposited as alternations of laminated, carbonate-poor, and bioturbated, carbonate-rich, beds. Further concentration of carbonate took place within the carbonate-rich beds during early diagenesis.

In South Wales, the Blue Lias consists of two facies. A deeper-water facies comprises interbedded mudstones and limestones in varying proportions. This interdigitates with a marginal, shoreline facies of conglomerates, oolitic, peletal and skeletal limestones developed around the massifs (Waters and Lawrence, 1987; Wilson et al., 1990).

The overlying Lower Lias Clay is seen mainly in boreholes, where it comprises clay that is calcareous and silty in places, with scattered bands of argillaceous limestone and mudstone (Whittaker and Green, 1983; Donovan and Kellaway, 1984). It is 198 m thick in the Burton Row borehole, and 124 m thick in the Dundry borehole (Figure 30).

Penn (1987) subdivided the offshore Lower Lias into five units (LL1 to LL5), which he correlated with the onshore succession (Figure 35). The Blue Lias (LL1 of Penn, 1987) can commonly be identified by the serrated motif at its base, and varies from 215 to 230 m in thickness. It passes upwards into more-argillaceous beds (LL2 to LL5) of the Lower Lias Clay, which consist of medium to dark grey, calcareous mudstones that are in part pyritic or organic rich. In the east, a number of crude, upward-coarsening cycles are developed as mudstones grade into fine-grained, muddy sandstones. The Lower Lias Clay has a well-developed, asymmetrical, cyclic, log motif that reflects variations in carbonate content. The cycles may be about 15 m in thickness, but decrease to less than 5 m where the sequence is thinner. The frequency and thickness of the cycles are therefore of use in determining the position of the well in relation to the contemporary basin margin.

To the west of the report area in Irish waters, the lateral equivalent of the Blue Lias is the Liassic Limestone' of Robinson et al. (1981), which is mainly Hettangian in age. Microfaunas of brackish and freshwater aspect indicate proximity to a basin margin south-west of Ireland. North of the Pembroke Ridge, the Lower Lias occurs in wells 103/2-1 and 106/28-1 (Figure 37)." data-name="images/P945233.jpg">(Figure 34), probably in BGS borehole BH71/45 (Figure 36), and in the Mochras borehole (Figure 33).

At Mochras (Figure 33), the Lower Lias is 896 m thick. The Blue Lias is not well developed but may be represented by the lowest 150 m of dark grey mudstone, in which there are some thin, interbedded limestones. This unit passes up into medium to dark grey mudstone with fewer interbedded limestones; siltstone increases slightly in abundance towards its top. To the south-west in well 106/28-1 (Figure 37)." data-name="images/P945233.jpg">(Figure 34), the Blue Lias (46 m) passes up into dark grey siltstone and mudstone of the Lower Lias Clay (149 m), which gives rise to invariant sonic and gamma-ray log responses. This is overlain conformably by Middle Lias.

In well 103/2-1 immediately north of the Pembroke Ridge, the Lias is 388 m thick and Hettangian to late Pliensbachian in age; the Sinemurian is abnormally thick in comparison with that of well 106/28-1 (Figure 37)." data-name="images/P945233.jpg">(Figure 34). The succession in well 103/2-1 compares with those in wells in the South Celtic Sea Basin (Figure 35), and may similarly indicate deposition in shallower water. Some thin, interbedded limestones, which may be equivalent to the Blue Lias, occur in the lower part. These pass up into dark grey, silty mudstone, and then into grey to black siltstone that gives a distinctive, asymmetrical, sonic-velocity trace. In the upper part, particularly between about 625 and 705 m depth, a cyclic log pattern is developed.

Cores and sidescan-sonar records suggest that the Blue Lias is present offshore on the flanks of the Bristol Channel Syncline (Evans and Thompson, 1979), where it is overlain by grey shales and mudstones of the Lower Lias Clay. Farther west, the Lower Lias has been proved in the wells drilled on a north-easterly trending high which crosses the South Celtic Sea Basin (Figure 35); this high is well defined by gravity data (Figure 14). The Lower Lias in well 103/21-1 (Figure 30) is 428 m thick and passes up conformably into the Middle Lias; in well 102/28-1 it is approximately 604 m thick and overlain unconformably by Cretaceous Wealden strata.

Middle Lias

The Middle Lias is late Pliensbachian in age. In the north of the report area it is somewhat coarser grained than both the Lower and Upper Lias, whereas in the South Celtic Sea Basin it is characterised by limestone interbeds. In the Mochras borehole, the Middle Lias is 147 m thick ((Figure 33); Ivimey-Cook, 1971; O'Sullivan et al., 1971). It comprises interbedded siltstone and mudstone, with proportionally more siltstone than the Lower Lias. Thin limestones occur, and a 1 cm-thick bed of sandstone and 5 cm of conglomerate have been recorded. The beds are bioturbated in part, and in addition to ammonites, contain belemnites, crinoid debris, sparse plant fragments, bituminous mudstone, and lenses of jet. The geophysical-log signatures are finely serrated, reflecting the interbedded nature of the succession, with an overall decrease in the gamma-ray intensity, and an increase in sonic-velocity values, towards the top of the Margaritatus Zone.

In the St George's Channel Basin, the Middle Lias in well 106/28-1 ((Figure 37)." data-name="images/P945233.jpg">(Figure 34); Penn, 1987) is 110 m thick and consists predominantly of siltstone, with some interbedded mudstone, and thin sandstones at the base. These sediments pass up into 37 m of grey, fine-grained sandstone, which is overlain by mudstone and siltstone of the Upper Lias. The Middle Lias has not been identified in well 103/2-1, although uppermost Lias Group sediments lying beneath the Middle Jurassic unconformity may be Middle Lias.

To the south, there is a passage from shallow-marine to subaerial deposition. In well 103/21-1 (Figure 35), the lower part of the 44 m-thick Middle Lias consists of calcareous mudstone, which is mainly grey, but is red-brown in places. It has a variable silt and glauconite content, and some thin sandstone or siltstone interbeds. Its upper part is mainly argillaceous limestone. In well 102/29-1, the Middle Lias is 74 m thick and consists mainly of mudstone, with thin limestone in its upper part; the succession includes two thin coals.

The Middle Lias is absent in South Wales, but is represented in Avon and Somerset (Figure 30) where it is 70 m thick. It consists largely of grey mudstone, with shales and limestones in its lower part. On the northern limb of the Bristol Channel Syncline, BGS boreholes BH72/67 and BH72/68 (Figure 36), sampled mudstones and limestones assigned to the Middle Lias.

Upper Lias

The Upper Lias is Toarcian in age (Figure 32), and is over 300 m thick in the South Celtic Sea Basin. In the Mochras borehole (Figure 33), where it is 262 m thick and divisible into two units (UL1 and UL2), the uppermost Toarcian zone is incomplete, and the Upper Lias is overlain unconformably by Tertiary sediments (Ivimey-Cook, 1971). The lower unit consists of grey mudstone with sparse siltstone and limestone bands, sporadic calcareous ironstone nodules, and a few plant fragments. The gamma-ray log response is characteristically invariant, but at a depth of 812 m, there is a thinly laminated, yellowish grey shale below which the gamma-ray values are consistently higher. In the upper unit, UL2, the mudstone is interbedded with silty mudstone, which is in places rich in plant debris. Limestone within the Bifrons Zone is hard, medium grey and silty towards the base; together with cross-bedding and an increase in the abundance of bioturbation, these features may indicate temporary shallowing of the seas. The topmost beds are slightly micaceous.

To the south-west, BGS borehole BH71/46 (Figure 36) proved 5 m of slightly micaceous, thinly laminated mudstone; the borehole lies close to the base of the Middle Jurassic (Parkin and Crosby, 1982), and is probably part of the Upper Lias. Farther south-west, in the St George's Channel Basin, well 107/21-1 penetrated 291 m of grey, calcareous, slightly micaceous mudstone of Toarcian age; this shows a characteristically invariant gamma-ray and sonic-velocity log response (Figure 37)." data-name="images/P945233.jpg">(Figure 34). In well 106/28-1, the Upper Lias (172 m thick) consists of blue-grey siltstone with a gamma-ray response similar to that in well 107/21-1.

In the South Celtic Sea Basin (Figure 35), the Upper Lias in wells 103/21-1 (259 m thick), 102/29-1 (311 m) and 103/18-1 (65 m), comprises grey-brown, calcareous, interbedded, silty mudstones and siltstones. In well 103/18-1, the thinner Upper Lias contains distinctive reddish grey to yellowish orange, calcareous beds in its lower part. These beds may be equivalent to the Junction Beds, a condensed deposit onshore in Avon and Somerset which represents most of the Toarcian Stage (Cope et al., 1980a; Millson, 1987a). In the Bristol Channel Basin, the Upper Lias is about 90 m thick (Evans and Thompson, 1979) and consists of brown-grey mudstones and pale grey, silty shales.

Middle Jurassic

The Middle Jurassic comprises beds of Aalenian to Callovian age (Figure 32), which in England form the Inferior Oolite Group, the Great Oolite Group, the Kellaways Formation, and the Lower and Middle Oxford Clay (Cope et al., 1980b). Strata of Mid-Jurassic age are absent in Wales, and only Lower Jurassic rocks are recorded from the South-West Approaches (Evans, 1990), the Irish Sea, and Northern Ireland (Jackson et al., 1994). Farther afield, Middle Jurassic rocks occur in the Hebridean area (Fyfe et al., 1993) and to the west and south-west of Ireland (Trueblood and Morton, 1991; Colin et al., 1992).

The Middle Jurassic over most of central and southern England is characterised by shallow-water sediments with evidence of numerous facies changes and erosion surfaces. Correlation of these sediments is notoriously difficult, and the series thickness is much reduced from that seen offshore. The Middle Jurassic is about 1000 m thick in the deeper parts of the St George's Channel and Cardigan Bay basins, where sedimentation may have been continuous, but on the basin margins there is evidence of uplift and erosion. Over the St Tudwal's Arch (Figure 30), an estimated 650 m of Lias was removed prior to the deposition of ?Bajocian to Bathonian sediments upon upper Sinemurian and Pliensbachian beds (Penn and Evans, 1976). North of the Pembroke Ridge, sandstones of ?Bajocian age rest on Pliensbachian sediments. (Figure 37) shows that the most-deeply buried Middle Jurassic in the St George's Channel Basin lies in the south-western part, with smaller basins lying on either side of the St Tudwal's Arch. In the Cardigan Bay and St George's Channel basins, Middle Jurassic strata form the cores of synclines which plunge southwestwards beneath the Tertiary cover (Figure 30).

Middle Jurassic fluviodeltaic sediments in the St George's Channel Basin are comparable to parts of the Yorkshire Deltaic Series and the successions of the Cotswolds (Penn and Evans, 1976). Lagoonal and marginal-marine facies possibly covered the whole of central England and parts of Wales (Figure 31). Farther south, approximately to the south of the Pembroke Ridge–Mendips line, a fully marine, calcareousmudstone facies prevailed.

On the southern flank of the St George's Channel Basin, the Middle Jurassic may thicken towards the basin axis. The succession is complete in well 107/21-1, where it is estimated to be 1045 m thick (Figure 37)." data-name="images/P945233.jpg">(Figure 34); in well 107/16-1 (Figure 30), it is likely that the uppermost Lias Group was penetrated, and that the Middle Jurassic is 1165 m thick.

Inferior Oolite Group

The tripartite Inferior Oolite Group is Aalenian to Bajocian in age, and in the southern Cotswolds (Figure 30) ranges from about 12 to 46 m in thickness (Cave, 1977). The Lower Inferior Oolite in the Cotswolds is particularly variable in thickness, and comprises oolites and shell-detrital carbonates. The Middle Inferior Oolite comprises mostly sandy and silty bioclastic limestones, that are only sparsely oolitic. The Upper Inferior Oolite is made up of dominantly shell-fragmental, somewhat argillaceous limestones. All these sediments were deposited in warm, shallow seas, and nonsequences separate the three divisions; the Middle Inferior Oolite is locally absent, indicating greater uplift and erosion at that time.

Offshore, the equivalent beds are considerably thicker. In well 107/21-1, in the St George's Channel Basin, Penn (1987) subdivided the group into two units equivalent to the Lower and Middle Inferior Oolite, and the Upper Inferior Oolite (Figure 37)." data-name="images/P945233.jpg">(Figure 34). The lower unit is 124 m thick and consists of sandstone with some interbedded mudstone; the geophysical-log character and lithology suggest a deltaic origin. The sandstone is pale grey, fine grained, calcareous and poorly to well cemented, with poor to good porosity. The mudstone is soft, grey and micromicaceous, grading into siltstone in places. Gamma-ray values are low, with a cylindrical motif, whereas sonic velocities are high, with sharp peaks where the sandstone is tightly cemented by calcite (Penn, 1987). The upper unit is 133 m thick and formed of grey, calcareous siltstone grading into fine-grained sandstone. Sandstone beds near the middle, and at the top, are reflected by sharp peaks on the sonic-velocity response.

In well 107/16-1 (Figure 30), the Inferior Oolite equivalent is 230 m thick and the subdivision into two units is not clear. The base is marked by a 14 m-thick, pale grey, fine-grained sandstone with calcite cement and traces of carbonaceous material. The sandstone passes up into a medium to dark grey, calcareous siltstone with a few interbedded sandstones, at the top of which lies a 25 m-thick, pale grey sandstone with low to fair porosity.

In well 103/2-1 (Figure 37)." data-name="images/P945233.jpg">(Figure 34), the lower part of the Inferior Oolite equivalent consists of 29 m of sandstone with thin limestone beds. Above lie 90 m of limestone and sandy limestone with glauconite, and some pyrite towards the top. Deposition in a shallow-water environment is indicated.

In the South Celtic Sea Basin, the sequence is thinner than in the St George's Channel Basin, and comparable to that seen in the south Cotswolds (Cave, 1977). Penn (1987) identified the equivalent of the Inferior Oolite in well 103/21-1, where it is 49 m thick and predominantly comprises glauconitic mudstone (Figure 35). In well 103/18-1 in the western Bristol Channel Basin, a largely siltstone succession 54 m thick may indicate a position nearer the basin margin. Farther east in the Bristol Channel Basin, the equivalent of the Inferior Oolite Group may be present, but is difficult to distinguish within the uniform shales that make up the Middle Jurassic in this region (Evans and Thompson, 1979).

Great Oolite Group

The Great Oolite Group is largely Bathonian in age, but ranges from late Bajocian to earliest Callovian (Cope et al., 1980b). Onshore, adjacent to the report area, it is subdivided into the Fuller's Earth, Great Oolite/Frome Clay, Forest Marble and Cornbrash. Equivalent formations have been identified offshore (Figure 32).

Onshore, a carbonate facies dominates the sequence from the Cotswolds to the London Platform, passing northwards into a brackish sequence in Northamptonshire and Lincolnshire, and into a deltaic facies farther north in Yorkshire (Figure 31). On the London Platform, the carbonate facies of the Great Oolite Group overlaps both the Lower Jurassic and the. Inferior Oolite Group to rest directly upon Devonian rocks. A similar onlapping relationship may have existed on the Welsh Massif. In the St George's Channel Basin, deposition appears to have kept pace with rapid subsidence; the lithologies are comparable with those found onshore (Penn and Evans, 1976), but the succession is considerably thicker, with some 800 m penetrated in well 107/21-1 (Figure 37)." data-name="images/P945233.jpg">(Figure 34).

To the south of the shallow-water carbonate facies (Figure 31), somewhat deeper-water, open-shelf conditions prevailed.

Mudstone dominates this sequence, and there is little limestone; the Great Oolite Group here is largely undivided. In the Bristol Channel Basin, an estimated 330 m of grey mudstone separating Toarcian and Callovian beds is probably equivalent to both the Inferior and Great Oolite groups. In the west of the Bristol Channel Basin, the Great Oolite Group is represented by 60 m of siltstone in well 103/18-1 (Figure 35).

In the South Celtic Sea, 106 m of the Great Oolite Group (Figure 35) underlies Cretaceous rocks. The sequence comprises grey-brown siltstone with some sandy siltstone, and has been subdivided into Fuller's Earth and Frome Clay (Penn, 1987).

Fuller's Earth

The Fuller's Earth is 46 m thick at its type section near Bath ((Figure 30); Penn et al., 1979), where it is subdivided into the Lower and Upper Fuller's Earth by the Fuller's Earth Rock. On land, the Lower Fuller's Earth is subdivided into a number of well-defined sedimentary cycles (Penn et al., 1979). Each opens with a shelly bed that has a sharp erosive base, passing up into silty, calcareous mudstone capped by a band of thin, porcellaneous, burrowed, limestone. The Fuller's Earth Rock is 3 to 10 m thick and comprises hard, rubbly, shelly limestone with subordinate calcareous mudstone. The Upper Fuller's Earth is variable in lithology, consisting of mudstone with an increasing proportion of limestone towards the top. It contains the locally commercially exploited Fuller's Earth Bed, an ash-fall deposit that consists of montmorillonite-rich clay with glass shards and silt-grade feldspars. Deposition, as indicated by the fauna, took place in an inner-shelf environment (Penn et al., 1979). The Fuller's Earth thins towards the north-east, where it is replaced by a carbonate, ramp-and-shoal facies.

The Fuller's Earth reaches a maximum thickness of about 300 m in the St George's Channel Basin. Rapid subsidence in the basin coincided with a considerable input of terrigenous sediment, reflecting uplift to the north. The Fuller's Earth thins southwards, and passes towards the Pembroke Ridge into a transitional facies of limestone and mudstone. In the South Celtic Sea Basin, a 60 m-thick mudstone reflects deposition in open-marine conditions.

A Fuller's Earth equivalent is 290 m thick in well 107/21-1 (Figure 37)." data-name="images/P945233.jpg">(Figure 34), where it consists of olive to dark grey mudstone with variable pyrite and carbonate content. The mudstone is interbedded with thin, argillaceous limestone, and thin siltstones occur towards the top. Microfaunal evidence indicates a moderately deep, fully marine environment of deposition (Millson, 1987a). The Fuller's Earth Rock is 64 m thick and comprises pale green, fine- to medium-grained, tightly cemented sandstone. The Upper Fuller's Earth is a dark grey mudstone passing up into calcareous sandstone; it thins southwards across the Pembroke Ridge.

On the northern flank of the Pembroke Ridge in well 103/2-1 (Figure 37)." data-name="images/P945233.jpg">(Figure 34), blue-grey mudstone with thin limestones of the Lower Fuller's Earth (76 m) passes up into 11 m of limestone of the Fuller's Earth Rock, and thence into a sequence of interbedded limestone and mudstone akin to a transitional facies as seen onshore (Penn, 1987).

In the Cardigan Bay Basin, on the northern edge of the depositional area, the Fuller's Earth was sampled in BGS borehole BH74/22 (Figure 36). Reddish brown, bioturbated sandstone, including mud-filled cracks and upward-fining cycles, was interpreted by Penn and Evans (1976) as outer-delta sediment on the margin of a fluviatile source. In borehole BH72/38, the sandstone appears to pass into fine-grained algal limestone similar to some beds in the White Limestone (Penn and Evans, 1976), which was deposited on a nearshore shelf with local algal developments in shallower parts of the lagoonal environment (Sumbler, 1984).

The Fuller's Earth is probably present in the Bristol Channel Basin (Evans and Thompson, 1979). In the South Celtic Sea Basin, the Lower Fuller's Earth is similar in thickness to onshore sections; in well 103/21-1 it is 51 m thick (Figure 35) and passes upwards into 4 m of limestone which correlates with the Fuller's Earth Rock (Penn, 1987).

Great Oolite

Onshore, the Great Oolite is a transgressive sequence of shallow-shoal limestone that passes southwards into the more-open-marine beds of the Frome Clay, which is 30 to 35 m thick south of Bath ((Figure 30); Penn et al., 1979). The carbonate facies was identified by Penn (1987) in the St George's Channel Basin, where it is well developed in well 107/21-1 (Figure 37)." data-name="images/P945233.jpg">(Figure 34). A Frome Clay equivalent has been identified south of the Pembroke Ridge in well 103/21-1 (Figure 35).

In well 107/21-1, the Great Oolite is 183 m thick (Figure 37)." data-name="images/P945233.jpg">(Figure 34) and comprises pale grey to white limestone that is silty in its lower part, and oolitic and fossiliferous in its upper part. Interbedded mudstone is grey, soft and calcareous, grading into siltstone and fine-grained, tightly cemented sandstone. Geophysical logs show low gamma-ray values, and high sonic-velocity values for the limestones, but the log signature is deeply serrated over the middle part of the formation where mudstone beds are more common. A similar oolitic limestone occurs in well 106/24-2 (Figure 37)." data-name="images/P945233.jpg">(Figure 34), where it overlies mudstone with thin, scattered, limestone beds. In nearby well 107/16-1 (Figure 37), the base is less-clearly defined since limestone, sandstone and mudstone are interbedded in the Upper Fuller's Earth; here the Great Oolite is about 226 m thick.

In the South Celtic Sea Basin in well 103/18-1, the Great Oolite Group is represented by 60 m of siltstone overlain by the Cornbrash. In well 103/21-1 (Figure 35), the Frome Clay comprises 37 m of mudstone overlain unconformably by the Lower Cretaceous. In both wells, the Great Oolite Group contains thin interbedded limestones, with a prominent limestone at the base.

Forest Marble

Onshore, the Forest Marble is a dominantly argillaceous but highly variable sequence with significant amounts of sandstone and limestone; it is 30 to 40 m thick around Bath ((Figure 30); Penn et al., 1979). Lithological boundaries are commonly abrupt, channels occur throughout the sequence, clay galls and pebbles are common, and the fauna suggests a restricted, shallow-marine environment.

Offshore in the St George's Channel Basin, beds correlated with the Forest Marble rest on the topmost oolitic limestone of the Great Oolite, indicating a widespread change in lithofacies. At the base of the formation, upward-fining sequences are interpreted as fluviodeltaic bar sands, with some of the thinner, poorly developed sandstones representing crevasse-splay or levee deposits (Milison, 1987a). In the upper part of the formation, subaerially deposited anhydritic redbeds show a subdued geophysical-log motif; these were sampled in BGS shallow cores in Cardigan Bay ((Figure 36); Penn and Evans, 1976). Subtle upward-coarsening units may be the result of episodic marine incursions, whereas bioturbated beds, rootlet beds, and more rarely, mudcracks, suggest intervals of exposure. Reddening and the absence of organic-rich beds or coals point to a semiarid climate.

In the St George's Channel Basin in well 107/21-1 (Figure 37)." data-name="images/P945233.jpg">(Figure 34), the Forest Marble is 242 m thick and comprises rhyth mically interbedded mudstone, siltstone and fine-grained sandstone. These give rise to a well-developed asymmetrical gamma-ray motif in its lower part, indicating a predominance of sharp-based, upward-fining sequences. Millson (1987a) interpreted these geophysical-log responses as those of fluvial, channel point-bar sediments. The sandstone is hard, pale to whitish, and calcareous; it is generally marked by low gamma-ray and high sonic-velocity values, indicating tight cementation. The siltstone and mudstone are reddish brown, with traces of anhydrite in their upper parts where the geophysical-log traces are more subdued and the sedimentary cycles less obvious. Microfaunas and microfloras indicate a freshwater input (Millson, 1987a).

In the Cardigan Bay Basin, grey-brown mudstone with shell-grit in BGS borehole BH74/23 (Figure 30) and (Figure 36) is particularly carbonaceous towards the top, and is interpreted as having been deposited in a lagoon (Milison, 1987a; Penn and Evans, 1976). Boreholes BH71/50, BH75/15 and BH75/17 were all considered by Milison (1987a) to have penetrated upper Bathonian sediments. Borehole BH75/15 shows evidence of the development of soil profiles in strata that are interpreted as interfingered alluvial/deltaic sands and low-lying swamp muds, that pass upwards into lagoonal beds or shallow-marine sediments deposited in water with restricted circulation.

In the western Bristol Channel Basin in well 103/18-1 (Figure 35), the equivalent of the Forest Marble may lie within the Great Oolite Group. In the eastern Bristol Channel Basin, a Bathonian, probably late Bathonian, ostracod fauna was recorded from a single sample described as a pale grey sandy clay with much lignite (Lloyd et al., 1973). The sample is of marine origin, and represents a shallow, near-shore environment.

Cornbrash

The Cornbrash spans the Bathonian–Callovian boundary, and is 0.5 to 8 m thick in the south Cotswolds (Figure 30) and (Figure 32); Cope et al., 1980b). The Lower Cornbrash (Bathonian) is generally composed of pale, bioclastic, micritic, and mainly nodular limestones. The Upper Cornbrash (Callovian) is typically less than 1 m thick, and consists of cream to yellowish marls and sandy limestones. In places, an intraformational 'boulder bed' indicates penecontemporaneous erosion (Cave, 1977). On geophysical logs, the Cornbrash gives rise to a distinctive gamma-ray low and sonic-velocity peak (Whittaker et al., 1985); these characteristics are recognised in the St George's Channel Basin in well 107/21-1 (Figure 37)." data-name="images/P945233.jpg">(Figure 34), where the unit is 8 m thick, and in well 107/16-1 (Figure 37) where it is 4.5 m thick, but are less obvious in wells 106/24-2 (Figure 37)." data-name="images/P945233.jpg">(Figure 34) and 106/24-1 (Figure 37). In the western Bristol Channel Basin, the formation was identified in well 103/18-1 ((Figure 35); Penn, 1987), where it rests unconformably on lower to middle Bathonian beds of the Great Oolite Group.

Kellaways Formation, Lower and Middle Oxford Clay

The youngest part of the Middle Jurassic is Callovian in age. In Wiltshire and Gloucestershire (Figure 31), the Kellaways Formation is about 25 m thick and has traditionally been divided into the Kellaways Clay, a silty clay with impersistent beds of sand, and the overlying, thinner, Kellaways Sand (Cave, 1977). Offshore in the St George's Channel Basin, a similar sequence is seen in wells 107/21-1 (Figure 37)." data-name="images/P945233.jpg">(Figure 34) and 107/16-1. Geophysical logs show a well-defined funnel motif, with gamma-ray values decreasing, and sonic velocities increasing, upwards. The silty mudstones are grey to greyish green and calcareous; the overlying sandstones are pale grey, fine to medium or coarse grained, calcareous, and tightly cemented. South of the Pembroke Ridge in well 103/18-1 (Figure 35), the Kellaways Clay is 13.5 m thick, with 8.5 m of Kellaways Sand above. The sandstone has a sharp base and top, and is olive-grey, calcareous, and silty with some pyrite.

A similar gamma-ray motif to that of the Kellaways Formation is seen in both the Lower and Middle Oxford Clay, again reflecting an increase in sandstone towards the top of each cycle. Geophysical logs have been used to correlate the Oxford Clay over an area stretching from Yorkshire to Dorset (Whittaker et al., 1985). The sandy sequence in the St George's Channel Basin (Figure 37)." data-name="images/P945233.jpg">(Figure 34) is most like that in north Yorkshire, where the Lower and Middle Oxford Clay are replaced by sandstones reflecting shallowing in the seas along the basin margin. In well 107/21-1 in the St George's Channel Basin, the Lower and Middle Oxford Clay are 43 m thick, and consist of pale to dark grey, calcareous mudstone passing up into pale grey, fine-grained sandstone.

In well 103/18-1 in the western Bristol Channel Basin (Figure 35), the Lower and Middle Oxford Clay are 70 m thick, and consist predominantly of olive-grey, silty mudstone, but lack the well-defined log motif seen farther north in the St George's Channel Basin. This may indicate deposition in deeper water, and it is possible that the Pembroke Ridge had less influence on sedimentation at this time.

Farther south, the upper Middle Jurassic beds are not preserved in well 103/21-1 (Figure 35), as they were probably removed before the deposition of Early Cretaceous Greensand. To the east in the Bristol Channel Basin, the Oxford Clay equivalent is lithologically similar to that seen onshore in southern England, and the sequence, including the Upper Oxford Clay of Oxfordian age, is estimated to be 370 m thick (Evans and Thompson, 1979).

Upper Jurassic

The base of the Upper Jurassic corresponds to the base of the Upper Oxford Clay (Cope et al., 1980b), a horizon that has been correlated in onshore boreholes using geophysical logs (Whittaker et al., 1985). The regular cycles of basin subsidence and infill of the late Mid-Jurassic continued into the early part of the Late Jurassic. In the St George's Channel Basin, deposition kept pace with rapid subsidence, and a thick sequence is present (Figure 38). South of the Pembroke Ridge, only the lowest part of this succession is preserved, and a marginal-marine to deltaic sequence in the Bristol Channel suggests that uplift and erosion began earlier in the south.

The Upper Jurassic crops out in the Bristol Channel Syncline, and has been proved elsewhere in wells 107/16-1, 107/21-1, 106/24-1, 106/24-2 and 103/18-1 (Figure 30), (Figure 37)." data-name="images/P945233.jpg">(Figure 34) and (Figure 35). In well 107/21-1, Penn (1987) identified the Upper Oxford Clay, the West Walton, Ampthill Clay and the Kimmeridge Clay formations. This terminology is retained here, although there is considerable lithological variance with the typical onshore succession.

Upper Oxford Clay, and the West Walton and Ampthill Clay formations

In the St George's Channel Basin, the Upper Jurassic is best seen in well 107/21-1 (Figure 37)." data-name="images/P945233.jpg">(Figure 34), where the Upper Oxford Clay consists of 45 m of blue-grey, silty mudstone that becomes siltier towards the top. The overlying West Walton Formation equivalent comprises 66 m of soft, grey-brown and calcareous mudstone interbedded with pale grey, fine- to medium-grained, loose- to well-cemented sandstone. Beds classified as Ampthill Clay Formation are 214 m thick in well 107/21-1. They consist largely of soft, reddish brown, calcareous mudstone which is grey brown in its lower part, interbedded with loose- and well-cemented sandstones which have sharp bases and tops and some upward-fining cycles; the geophysical-log response is characteristically ragged. The sandstones are interpreted as fluviodeltaic, coastal-plain sand bodies (Millson, 1987a); anhydrite recorded in well 106/24-2 indicates some desiccation.

In much of the Bristol Channel Basin the equivalent beds are estimated to be 520 m thick (Evans and Thompson, 1979). The beds consist of grey-green, fine-grained sandstones and siltstones, in which glauconite and lignite fragments are common. The beds were deposited in a nonmarine to transitional environment; restricted and dwarfed molluscan assemblages suggest estuarine or deltaic deposition. To the west in well 103/18-1 (Figure 35), the Upper Oxford Clay is about 18 m thick, but the uppermost Upper Oxford Clay, and the West Walton and Ampthill Clay formational equivalents are missing, possibly due to faulting (Penn, 1987).

Kimmeridge Clay Formation

The thickest Kimmeridge Clay Formation in the report area is the 958 m proved in well 106/24-1 in the St George's Channel Basin (Figure 30), where it rests conformably on the Ampthill Clay Formation and is overlain unconformably by Tertiary strata. The top two zones of the Kimmeridgian are missing (Penn, 1987). This thick sequence and the thinner succession in well 106/24-2 (Figure 37)." data-name="images/P945233.jpg">(Figure 34) indicate considerable subsidence in the St George's Channel Basin during the Kimmeridgian. Similar but less pronounced subsidence was experienced in southern and eastern England (Figure 38). Onshore, the Kimmeridge Clay is divided into lower and upper subformations at a marked change in the ammonite fauna; the Upper Kimmeridge Clay contains the rhythmically bedded mudstones that include oil shales.

Penn (1987), reviewing previous classifications, suggested that the beds identified by Barr et al. (1981) as Kimmeridgian to Portlandian in well 106/24-1 (Figure 37)." data-name="images/P945233.jpg">(Figure 34) are entirely Kimmeridgian in age, and that the Portlandian stage is missing. In wells 106/24-1 and 106/24-2 (Figure 37)." data-name="images/P945233.jpg">(Figure 34), the Kimmeridge Clay Formation consists mainly of mudstones, with traces of glauconite and lignite. These are interbedded with pale grey limestones and sandstones, and vary in colour from brown and red to orange. In well 106/24-1 the sandstone is pale grey, contains traces of glauconite, is quartz bonded with calcite cement, and has a poor porosity. The beds become sandier, with more interbedded limestone, towards the contact between the Lower and Upper Kimmeridge Clay. The geophysical-log signature is moderately serrated, but shows no clearly defined pattern, although there is some evidence of upward-coarsening cycles in the lower part of the Upper Kimmeridge Clay. Anhydrite occurs in well 106/24-2.

There is no indication of the presence of oil shales in the St George's Channel Basin. The microfauna is dominated by freshwater forms, suggesting a lagoonal or overbank origin for these sediments (Millson, 1987a). Westwards in the Irish sector of the North Celtic Sea Basin, a similar thickness of beds is recorded in well 50/11-1, where anhydrite occurs throughout, indicating more-prolonged intervals of subaerial exposure and desiccation.

The Kimmeridge Clay Formation is estimated to be 330 m thick in the core of the Bristol Channel Syncline (Evans and Thompson, 1979), where the rocks are comparable with the organic-rich shales onshore. The formation has been proved in BGS borehole BH72/64 (Figure 36), where it consists of dark grey mudstone with siltstone. Lignitic fragments have been recorded from the synclinal axis in a shallow core recovered at the sea bed.

Portland and Purbeck groups

Portlandian strata subcrop beneath Tertiary in the centre of the St George's Channel Basin, and in the west of the Bristol Channel Basin, In the St George's Channel Basin in well 107/16-1 (Figure 30), the uppermost part of the Jurassic has been classified as Portlandian. These beds consist of yellow to red-brown mudstone, with a variable carbonate content and some interbedded, coarse-grained, yellow, loosely cemented sandstone. No limestone or gypsum has been recorded, and the beds are overlain by 'Tertiary to Recent' strata. Portlandian strata may also be present in the St George's Channel Basin in well 107/16-1, and have been recorded in the Bristol Channel Basin both in well 103/18-1 and BGS borehole BH72/55 (Figure 35) and (Figure 36). In well 103/18-1, the sequence rests on Oxford Clay, with most of the Oxfordian and Kimmeridgian strata missing. Kamerling (1979) interpreted the contact as a disconformity, but Penn (1987) argued that it may be a fault, as it is unknown for the Portlandian to overstep Kimmeridgian and Oxfordian strata elsewhere in southern Britain. Furthermore, the well is located on a major fault system (Van Hoorn, 1987). No similar hiatus has been recorded eastwards in the Bristol Channel, where most of the Kimmeridgian zones have been identified by Lloyd et al. (1973).

In well 103/18-1, 128 m of Portlandian beds are divided into a lower part, equivalent to the Portland Group, and an upper part equivalent to the Purbeck Group (Figure 35). The Portland Group is 45 m thick, and consists of medium grey to dusky yellow, calcareous mudstone and siltstone with pale to dark grey limestone at the top. The Purbeck Group is about 83 m thick, and consists of dark grey to dusky yellow, soft, noncalcareous silty mudstone with a 26 m-thick basal limestone that contains two gypsum beds. The Purbeck beds pass upwards into Cretaceous sandstone.

Chapter 6 Cretaceous

During the Cretaceous Period, major tectonic events took place which had a significant impact upon the geological evolution of the report area (Chapter 3). The Early Cretaceous was dominated by rifting to the south-west, which led to the formation of well-defined sedimentary basins in which mainly continental-clastic sediments, derived from the uplifted basin margins, were laid down. In contrast, from the mid-Cretaceous onward, the area suffered postextensional thermal subsidence during which an increasingly marine sequence of sediments was deposited, progressively overstepping the pre- vious basin margins.

The sedimentary sequence may be divided into four main facies associations ((Figure 39); Naylor and Shannon, 1982). Initially, the freshwater, lagoonal conditions of the Late Jurassic continued, with deposition of limestones and shales of Purbeck facies. These pass upwards into the sands, mudstones and coals of the Early Cretaceous Wealden facies, which was laid down in marginal-marine, alluvial, and deltaic environments. During the Albian, the continental/marginal environments were gradually replaced from the west by marine conditions, and an interbedded sequence of glauconitic sands and clays of Gault and Greensand facies were deposited. Finally, as marine conditions expanded and became fully established, calcareous claystones and siltstones were laid down, to be followed during the Cenomanian by chalk, which ranges into the Maastrichtian.

Cretaceous rocks are almost entirely absent from the land surrounding the report area, and offshore, Cretaceous rocks have been proved only south of the Variscan Front. The thickest and most complete sequence is preserved in the

North Celtic Sea Basin, where it reaches over 3000 m in thickness (Figure 39). Thinner sequences are present in the South Celtic Sea and Bristol Channel basins.

Lower Cretaceous

The deposition of the Purbeck and Wealden facies was mainly confined to discrete basins within an extensional structural setting similar to that in which older Mesozoic sediments were deposited. Neither Purbeck nor Wealden strata have been proved north of the Variscan Front (Figure 13), but the thickest sedimentary sequence in the St George's Channel Basin has not been drilled, and it is possible that strata of this age may be preserved in the basin centre (Chapter 3). Although Purbeck strata were identified in a well in the St George's Channel Basin by Barr et al. (1981), these are now considered to be of Kimmeridgian age (Penn, 1987).

In all three southern basins there is an Early Cretaceous unconformity of varying magnitude (Figure 39). In the centre of the North Celtic Sea Basin (Figure 40), sedimentation was continuous across the Jurassic–Cretaceous boundary, but towards the basin margins, the Valanginian progressively oversteps the Ryazanian on to Jurassic and Triassic strata (Naylor and Shannon, 1982). In the South Celtic Sea Basin, the Purbeck is absent, and Wealden strata of Aptian or older age rest with marked unconformity upon rocks which may be as young as Mid-Jurassic. In the Bristol Channel Basin, the Purbeck facies extends into the Ryazanian without break, but much of the younger Wealden is absent, and the Ryazanian sediments are overlain by Albian greensands (Kamerling, 1979). The Ryazanian Stage name is used after Rawson et al. (1978), and is taken to be the time equivalent of the Berriasian.

The Early Cretaceous unconformity is probably partly the result of uplift caused by right-lateral shear displacement of the Palaeozoic basement massifs on the basin margins during the north-east to south-west opening of the Bay of Biscay (Van Hoorn, 1987; Shannon, 1991). There is no evidence of regional basin shortening at this time. The accumulation of an almost complete Wealden section in the centre of the North Celtic Sea Basin indicates that subsidence here was greater than uplift, a situation also noted by Hillis (1988; 1991) in the Western Approaches Basin to the south.

According to Van Hoorn (1987), the uplift was initiated in the south-west of the South Celtic Sea Basin in Mid- to Late Jurassic times, and thereafter propagated north-eastwards into the Bristol Channel Basin during the Early Cretaceous. Uplift was particularly intense along the basin margins, with up to 3300 m of section eroded on the southern margin of the South Celtic Sea Basin.

The sources of the Wealden clastics in the North Celtic Sea Basin were the west and south-west including Ireland, Wales and Cornubia (Allen, 1981). However, in the Bristol Channel and South Celtic Sea basins, the source was in Wales and Cornubia (Tappin and Morton, unpublished). There is no evidence to suggest derivation from the St George's Channel/Cardigan Bay area. Provenance studies from the South Celtic Sea and Bristol Channel basins (Tappin and Morton, unpublished) indicate that the depositional area over Cornubia increased during Ryazanian to Aptian times. The heavy-mineral assemblages in the lowermost Wealden of the Bristol Channel Basin were derived mainly from the Cornubian granites, the Culm (Carboniferous) and the Old Red Sandstone, with secondary sources from New Red Sandstone and Jurassic sediments. The Cornubian granites were the only source for the younger, Aptian sandstones of the South Celtic Sea Basin.

In the South Celtic Sea and Bristol Channel basins, deposition was continental/marginal until Aptian times, but in the west and south of the North Celtic Sea Basin, marine incursions during the Barremian presaged the marine conditions of the mid- and Late Cretaceous. The advance of marine conditions from the west is attributed to collapse of the western high as the plate-tectonic regime changed from 'rift to drift' (Allen, 1981), that is syn- to postextensional (Chapter 3).

Mild regional warping during the Aptian, the Austrian event of Ziegler (1978), led to the formation of a disconformity with the overlying Albian. There was also an Aptian regression (Figure 39), which briefly interrupted the overall sea-level rise; by the mid- to late Albian, shallow-marine conditions were established across much of the report area (Colin et al., 1981; Naylor and Shannon, 1982; Ainsworth et al., 1987). Sedimentation extended beyond the previous limits of Wealden deposition and on to the Palaeozoic highs. The deposition of thin Gault and Greensand facies sediments was accompanied by minor flexuring (Kamerling, 1979; Naylor and Shannon, 1982).

Purbeck facies

The Purbeck facies in the North Celtic Sea Basin comprises a relatively monotonous sequence of grey marls with sandstone intervals and thin limestone stringers. The Cinder Bed, taken as the lithological base of the Cretaceous in England (Rawson et al., 1978), is absent, but the ostracod zonation established by Anderson (1973) for The Weald in south-east England has been used with only slight modification to identify the Jurassic–Cretaceous boundary (Colin et al., 1981). A notable feature of the ostracod faunas is that both their abundance and diversity decrease dramatically across the boundary. The uppermost Purbeck beds are of early Ryazanian age.

Purbeck beds have been penetrated in well 103/18-1 (Figure 41) in the Bristol Channel Basin (Penn, 1987), where a succession similar to that of southern England is present. The lowermost Purbeck in this well is of Jurassic age, and represented by correlatives of the Gypsum Beds, Broad Oak Calcareous Member, Plant and Bone Beds, and the lower Arenaceous Beds. At the top of the Purbeck are 21 m of sandstone and limestone of the upper part of the Arenaceous Beds and Grey Limestones, within which lies the Cinder Bed.

Wealden facies

North Celtic Sea Basin

The Purbeck beds pass up into a mainly nonmarine, Wealden sequence of sandstones, shales and coals. These are of Ryazanian to early Albian age, and are up to 2000 m thick ((Figure 39); Ainsworth et al., 1981; Naylor and Shannon, 1982). In England, the Purbeck–Wealden boundary is in the middle of the Cypridea Setina Zone of Anderson (1973), which in the North Celtic Sea Basin cannot be identified because of sparse biotas (Colin et al., 1981). However, the Wealden is both less calcareous and less fossiliferous than the Purbeck, and has common sphaerosiderite, which is lacking in the later facies.

The Ryazanian to Hauterivian sediments are mainly fluvial sandstones and alluvial shales laid down in brackish and fresh waters (Naylor and Shannon, 1982). Sediment sources were generally to the west, although both the Irish and Cornubian massifs contributed sediment (Allen, 1981). During the Hauterivian, finer-grained, muddy sediments of Weald Clay facies were laid down in the western part of the basin in tropical, coastal lagoons and on floodplains in deltaic environments (Ainsworth et al., 1987). This argillaceous sequence extended at least as far east as the Kinsale Head Gasfield, which is located in the basin centre. Together with brackish-marine influences in the southern part of the basin, these marginal environments heralded the basin-wide marine conditions of the Albian (Ainsworth et al., 1987). In the eastern part of the basin, nonmarine, coarse-grained, clastic deposition continued uninterrupted, with source areas in the Irish and Welsh massifs, and possibly Cornubia (Allen, 1981; Naylor and Shannon, 1982). The presence of abundant sphaerosiderite nodules in both the clays and sands indicates that abundant plant fragments were laid down in fresh or brackish water in floodplains and lagoons (Colin et al., 1981).

The deposition of argillaceous sediments continued in the west until the early Barremian, when, as the marine influence increased (Ainsworth et al., 1987), a complex pattern of fluvial, alluvial and deltaic sedimentation took place in the basin centre. This led to the development of sandstones, siltstones, mudstones and minor coals (Naylor and Shannon, 1982).

The continued eastward advance of marine conditions during the Aptian is best illustrated by the disappearance of sphaerosiderite in well sections ((Figure 40); Colin et al., 1981); in well 57/6-1, sphaerosiderite is found only at the base of the Aptian, but to the north-east in well 47/30-1, it occurs throughout the interval. In the southern part of the basin, Aptian sediments are grey and carbonaceous; they were deposited in a high-energy, shallow-marine, sublittoral environment which microbiotas indicate to have been warm, shallow, and carbonate rich (Ainsworth et al., 1987). In the basin centre around the Kinsale Head Gasfield, sandstones, siltstones, mudstones, root-zone claystones, ironstones and minor coals record a lower delta-plain environment (Colley et al., 1981). The sediments show a high degree of cyclicity, with upward-coarsening cycles common. The sandstone units are normally 3 to 10 m thick, but may range up to 25 m in major distributary bodies of limited areal extent. The argillaceous sediments are levee and overbank deposits, whereas coals, palaeosols with roots, iron- and pyrite-rich claystones, and carbonaceous shales, are interdistributary swamp and marsh deposits.

South Celtic Sea Basin

The Wealden in the South Celtic Sea Basin is up to 500 m thick and rests unconformably upon pre-Cretaceous strata, the youngest of which are the Bathonian sediments penetrated in well 93/2-1 (Figure 41). Wealden beds crop out at the sea bed in the east (Figure 2), where on the south side of the basin they overlie strata of Mid-Jurassic age with marked unconformity. On the north side of the basin, the relationship is paraconformable, indicating uplift along the south side of the basin during the Late Jurassic to Early Cretaceous (Van Hoorn, 1987).

The 280 m-thick Wealden section penetrated in well 93/2-1 is the thickest recorded in the basin, and the uppermost beds have yielded an Aptian palynomorph assemblage (Company report). Two units are recognised: the lower is 150 m thick and comprises mainly very fine- to medium-grained, poorly sorted, muddy sandstone and siltstone with mudstone. It is probably representative of an interdistributary environment, with clays and silty sands deposited on the floodplain. The upper unit is 130 m thick and comprises clean, fine- to coarse-grained, moderately sorted, quartzose sandstone in beds up to 18 m thick, with minor siltstone and brown to grey mudstone. Lignite occurs throughout the unit, which is mainly noncalcareous except for the uppermost 15 m. Both units are interpreted as having been deposited in a nonmarine, alluvial environment, which palynomorph assemblages indicate to have been deltaic (Company report).

The thick sandstones in the upper unit are probably major distributary sand bodies.

In wells 102/28-1 and 102/29-1, 15 m of calcareous, pyritic and carbonaceous sandstones of Aptian age were sampled at the base of the Wealden (Figure 41). The sediments are similar to the uppermost Wealden in well 93/2-1, and unconformably overlie the Jurassic. The sandstone in well 102/28-1 is overlain by 12 m of noncalcareous, grey, silty and carbonaceous claystone of early Aptian age; it includes a rich, reworked assemblage of Hauterivian and Barremian ostracods and miospores, as well as Jurassic foraminifera. A similar, but calcareous, 14 m-thick claystone with minor sandstone interbeds is present in well 102/29-1.

Bristol Channel Basin

Wealden beds in the Bristol Channel Basin occupy the axis of an east–west-trending syncline formed during Early Cretaceous deformation (Figure 2) and (Figure 42), and have been proved in well 103/18-1 (Figure 41) and several BGS bore-holes (Figure 43). In well 103/18-1, there are 195 m of interbedded, noncalcareous, mottled, red-brown and grey clay-stone and siltstone, with grey argillaceous sandstone (Kemerling, 1979; Penn, 1987).

The BGS boreholes, located in the eastern part of the basin (Figure 43), sampled interbedded sandstones and claystones deposited in alluvial environments (Tappin and Morton, unpublished). A number of specific environments have been identified:

1.     Floodplain with high water table. In BGS boreholes BH73/59 and BH74/44, the sediments are sandstones which fine upwards into green clays with plant matter. The presence of siderite and pyrite in the claystones indicates high water tables. Crevasse splays are represented by structureless sandstones 1 to 2 m thick that are very fine to fine grained, muddy, poorly sorted, and noncalcareous with coarse-grained carbonaceous fragments and siderite nodules. The clays have rootlet traces, but are little affected by pedogenesis. Some of the very fine- and fine-grained sandstones are finely laminated and show ripple-drift lamination. The lamination is commonly disturbed by slumping, faulting and bioturbation. Freshwater bivalves and gastropods are present, and these laminated sediments are interpreted as having been deposited in more-permanent standing bodies of water.

2.     Floodplain with fluctuating water table. In borehole BH73/40, fine-grained, muddy sand, silt and clay, similar to those described above, are interpreted as crevasse-splay and overbank deposits. However, the clay is variegated, green, purple, brown, ochre and red, indicating deposition in a variably oxidising environment affected by a fluctuating water table. The presence of burrows indicates that standing-water conditions prevailed at times.

3.     Distributary-channel fills. In borehole BH74/44, a sand body some 4 m thick is interpreted as a distributary-channel fill. The sand is fine to coarse grained, with moderate sorting. Sedimentary structures include planar cross-bedding, ripple-drift lamination, and channelling. Carbonaceous debris is common. The sediments were deposited under a high-energy flow regime, perhaps as a point bar.

Clay minerals in the sediments (Tappin and Morton, unpublished) are dominated by kaolinite, chlorite and mica. In sections interpreted as being near the Purbeck–Wealden transition, smectite is present, and kaolinite is absent. The discrimination and mutual exclusion of different clay minerals is interpreted to be the result of weathering processes in the hinterland. Similar clay relationships have been identified in The Weald (Sladen and Batten, 1984), where they have been attributed to changes in weathering of the source rocks as local climate changed during uplift and erosion in the source areas. When source areas were low lying, the rainfall was low and semiarid conditions prevailed, so that alkaline and poorly leached sources resulted in a low proportion of kaolinite in the clays. When the hinterland was uplifted, rainfall increased, resulting in acidic and well-leached sediments with an increased proportion of kaolinite.

Gault and Greensand facies

North Celtic Sea Basin

Local subsidence in the North Celtic Sea Basin was significantly greater than elsewhere in the report area during the mid-Cretaceous, with up to 140 m of Gault and Greensand facies of Albian age being deposited in the basin centre (Figure 39). The only described sequence is at the Kinsale Head Gasfield ((Figure 45)." data-name="images/P945243.jpg">(Figure 44); Colley et al., 1981), which lies outside the report area, but is considered representative of the basin as a whole.

Nonmarine Wealden environments continued into the early Albian at the gasfield, when they rapidly became transitional in character (Figure 45)." data-name="images/P945243.jpg">(Figure 44). Uppermost Wealden sediments display evidence of small-scale vertical burrowing in the siltstones and sandstones, which suggests deposition in brackish to marine waters. Overlying these transitional beds is a marine sequence of interbedded sand and clay, interpreted as a shoreline deposit. At the base of this marine sequence is a 4 m thick, very fine- to fine-grained, rippled and cross-stratified sandstone (B sand) with carbonaceous fragments; The top of this B sand is clay rich, and contains silt and clay laminae.

Overlying the B sand is a sequence of glauconitic, silty claystone deposited on a shallow-marine shelf. This passes upwards through clay-rich siltstone and fine-grained sandstone into a clay-free sandstone up to 50 m thick, termed the A sand, which forms the main reservoir in the gasfield. The transition reflects a shallowing of the sea, probably due to local flexuring (Naylor and Shannon, 1982). The A sand is mainly very fine to fine grained, massive, glauconitic and fossiliferous; bioturbation has destroyed any sedimentary structures. The sandstone was laid down as a stacked series of offshore, shallow-marine sand bars which accumulated during three main phases of deposition. There are numerous subunits separated by impermeable beds, either of calcified sandstones or laminated clays. The A sand is overlain by a further sequence of grey, soft, fossiliferous, bioturbated, and in part silty, claystone of Albian age. This is similar to the underlying clay unit, but more homogeneous.

South Celtic Sea and Bristol. Channel Basins

Gault and Greensand facies are exposed at the sea bed in the eastern part of the South Celtic Sea Basin, and have been penetrated in all commercial wells (Figure 41). They comprise shallow-water clastic sediments and limestones of (?)mid-Albian to early Cenomanian age (Weighell, 1980). In well 93/2-1, 20 m of calcareous and glauconitic siltstone and mudstone of (?)mid-Albian age are overlain by 25 m of (?)late Albian to early Cenomanian calcareous and glauconitic sandstone and siltstone with limestone (Weighell et al., 1981). Similar glauconitic mudstones, siltstones, and variably cemented sands, together with several prominent limestones, are found in wells 102/28-1 and 102/29-1. In well 103/21-1, there are 57 m of undated greensand which give a geophysical-log response comparable to the greensand sequences in the Quadrant 102 wells.

The Gault and Greensand in the Bristol Channel Basin were probably penetrated in well 103/18-1 (Figure 41), where, overlying the Wealden, a 13 m-thick unit is interpreted from the gamma-ray log as a sandstone that becomes cleaner upwards (Penn, 1987). According to Penn (1987), a thin Gault unit may lie at the base of the sandstone.

Upper Cretaceous

During the Late Cretaceous, a combination of high sea levels (Figure 39) and regional subsidence led to the widespread deposition of chalk, a limestone biomicrite composed mainly of algal debris. The chalk blanket was extensive, for deposition took place beyond previous Cretaceous limits, and all but the most persistent highs were inundated (Hancock, 1975; Tyson and Funnell, 1987). Subsidence was more regional in character than during the Early Cretaceous. Although thicker sedimentary sequences continued to accumulate within the preexisting basins, rather than on the intrabasinal highs, this was probably due to the effects of differential compaction, rather than to any underlying tectonic cause. Sedimentation was generally uniform within the basinal sequences, but there was minor deformation during the Cenomanian and at the Coniacian–Santonian boundary.

Offshore, the Upper Cretaceous is now confined to the south of the report area, where it is found either at the sea bed or overlain by a cover of Tertiary and/or Quaternary sediments (Figure 2). The thickest Upper Cretaceous section of 1200 m lies in the North Celtic Sea Basin (Figure 39), (Figure 40) and (Figure 45), despite the removal of Campanian and Maastrichtian Chalk along Late Cretaceous to early Tertiary inversion axes (Tucker and Arter, 1987). Cenomanian to Santonian Chalk is of a remarkably constant thickness (Tucker and Arter, 1987).

In the South Celtic Sea and Bristol Channel basins, there are up to 1000 m of Upper Cretaceous sediment (Figure 39) and (Figure 46). Thickness variations along the southern basin margin are due to syndepositional halokinesis triggered during the Early Cretaceous deformational phase. Thinning towards the east is due to subsequent uplift and erosion  (Figure 47). On the eastern margin of the basin, westerly dipping seismic reflectors, which sea-bed sampling have shown to be chalks and calcareous claystones of Cenomanian to Coniacian age, are truncated at the sea bed (BGS Lundy Solid Geology sheet).

The original depositional limit of the chalk is not well defined; the absence of a marginal facies in the Coniacian to Maastrichtian sections hinders interpretation of the location of basin margins. Lack of significant clastic material within the chalk suggests that land was distant (Tucker and Arter, 1987), or that elevations were low (Tyson and Funnel!, 1987). Hancock (1975) argued that the lack of detritus was due to nonseasonal climatic conditions that produced little or no erosion of the surrounding land.

Although now preserved only in the south of the report area (Chapter 3), it is probable that Late Cretaceous sediments were laid down within all basins (Hancock, 1975; Tyson and Funnell, 1987). Cenomanian to Turonian green-sands have been sampled on the northern and north-eastern margins of the North Celtic Sea Basin (Weighell et al., 1981; Delantey et al., 1981), passing southwards into the deeper-water chalk facies. Late Cretaceous deposition probably extended northwards into the St George's Channel Basin, which may have formed a shallow gulf (Dobson and Whittington, 1987). Towards the east, Late Cretaceous sedimentation extended into the Bristol Channel Basin, probably connecting with the chalk basins of southern England (Weighell et al., 1981).

The highlands of the massifs surrounding the basins would have been at least partly transgressed during the Turonian and Campanian highstands (Hancock, 1975; Tyson and Funnel!, 1987). Marginal Cenomanian glauconitic limestones and sandstones, and Turonian to Coniacian chalk, are found in south Devon (Hart, 1982). Provenance analysis of the Cenomanian sandstones suggests that during their deposition, the Cornubian granites, with the exception of Dartmoor, were exposed (Hancock, 1969). Turonian to mid-Campanian microfaunas occur as residues in the Tertiary flint gravels in Devon (Hart, 1982), and Senonian (Turonian to Campanian) chalk is present in a small outlier preserved by karstic subsidence in southern Ireland (Walsh, 1966).

The Chalk in the report area mainly comprises a hard lower sequence of Cenomanian to Coniacian age, and a soft upper sequence of Santonian to Maastrichtian age (Weighell et al., 1981). Throughout, there are traces of glauconite, pyrite, and quartz sand; flint is common. At the base of the Chalk there are greensands and marls; on the basin margins, the former are as young as Turonian. The overall downward increase in hardness, reflected by an increase in sonic velocity (Figure 46), is due to postdepositional cementation, early lithification, and burial diagenesis. Using geophysical logs, major lithostratigraphical units may be traced across the report area, and correlated with the onshore Chalk sequences in southern England (Figure 46).

In the South Celtic Sea and Bristol Channel basins, the Upper Cretaceous succession is divided lithostratigraphically into the Lower, Middle, and Upper Chalk by comparison with the Winterborne Kingston borehole in the Hampshire Basin ((Figure 46); Wood et al., 1982; Lott, 1982). The Lower, Middle and Upper Chalk divisions are dated respectively as Cenomanian, early and mid-Turonian, and late Turonian to Maastrichtian (Rawson et al., 1978). Deposition of the Lower and Middle Chalk took place during a sea-level rise which culminated in the late Turonian highstand (Figure 39). During the deposition of the Upper Chalk, relative sea level remained generally high until after a mid-Campanian highstand, when it gradually fell during Maastrichtian times.

Lower Chalk

Early Cenomanian deposits are characterised by lateral and vertical facies changes associated with variations in water depths. Initially, in the South Celtic Sea and Bristol Channel basins, and on the northern margins of the North Celtic Sea Basin, there was a continuation of the deposition of shallow-water greensands (Weighell et al., 1981). In the deeper waters of the central North Celtic Sea Basin, coeval sediments are calcareous siltstones and claystones (Colley et al., 1981). With a rapid expansion of deeper-water conditions into the South Celtic Sea and Bristol Channel basins, the greensands were succeeded by calcareous silts and clays (Figure 46).

In the South Celtic Sea and Bristol Channel basins, the lowermost Cenomanian comprises calcareous, glauconitic greensands which pass up from the underlying Albian. This passage is similar to that observed in many areas of southern England, where the Upper Greensand is considered to be of early Cenomanian age (Rawson et al., 1978). Within the greensands, faunas are poor, and identification of the Albian–Cenomanian boundary is problematic (Weighell et al., 1981). The thickness of the early Cenomanian greensands is 10 to 15 m in wells 93/2-1 and 102/28-1, and 27 m in well 102/29-1 (Figure 46).

Overlying the greensand in all wells are 3 to 4 m of hard, glauconitic limestone or marl (Figure 46), which is characterised by high gamma-ray values and high sonic velocity. The limestone, which also occurs at Winterborne Kingston, is interpreted as a lateral equivalent of the Glauconitic Marl of southern England, where it is considered to be a diachronous unit of early to mid-Cenomanian age (Lott, 1982). The Glauconitic Marl is succeeded by 30 to 50 m of calcareous, glauconitic claystones and siltstones, which are interpreted as lateral equivalents of the Chalk Marl. The serrated nature of the gamma-ray and sonic velocity response is attributed in the Winterborne Kingston borehole to rhythmic bedding in more- and less-argillaceous chalk (Lott, 1982). According to Weighell et al. (1981), two lithologies are present: in wells 102/29-1 and 103/21-1 (Figure 41) and (Figure 46), the sediments are fine-grained marls, whereas in well 102/28-1 they comprise 29 m of marl overlain by 14 m of calcareous siltstone, and in well 93/2-1 they are 28 m of calcareous siltstone. An equivalent marl unit is interpreted to be present in the uppermost part of well 103/18-1 (Figure 41), which penetrated the feather edge of the Upper Cretaceous in the eastern part of the basin.

In the centre of the North Celtic Sea Basin, lower Cenomanian claystones and siltstones are present at the Kinsale Head Gasfield ((Figure 45)." data-name="images/P945243.jpg">(Figure 44); Colley et al., 1981) and elsewhere in the basin. In all instances, they overlie either Gault or Greensand facies. At the boundary, there is a green clay distinguished by a log response of high gamma-ray values and high resistivity/sonic-velocity, similar to that of the Glauconitic Marl, but dated as uppermost Albian. The lower Cenomanian sediments comprise some 30 m of calcareous and glauconitic siltstones and clays, which are correlated with the Chalk Marl; their interbedded nature is reflected in a serrated geophysical-log response. Towards the northern basin margin, they pass into greensands which overstep northwards on to older rocks of Permo-Triassic and Palaeozoic age (Delantey et al., 1981).

In mid-Cenomanian times, a transgressive pulse led to a further increase in water depths, and to the initiation of chalk sedimentation, which was synchronous in the basin centres (Weighell et al., 1981). In the South Celtic Sea and Bristol Channel basins, there is a sharp boundary between the Chalk and the underlying siltstones and claystones ((Figure 46); Weighell et al., 1981). This is reflected by a sharp decrease in gamma-ray values and an increase in sonic velocity, a junction that is more clearly marked than that in southern England. The upper Cenomanian Chalk is white and hard, with minor glauconite and quartz sand. The change in lithology is accompanied by an influx of planktonic foraminifera (Weighell et al., 1981), indicating deeper-water conditions. Upper Cenomanian Chalk thickness increases towards the south-west from 5 m in well 103/21-1 to 20 m in well 93/2-1 (Figure 41) and (Figure 46), indicating basin subsidence in this direction. This unit is probably a lateral equivalent of the coeval Grey Chalk of southern England, although the colour is white rather than grey, and it is less argillaceous. Gamma-ray values are lower and the curves are serrated, indicating interbedded argillaceous and less-argillaceous chalk.

The upper Cenomanian at the Kinsale Head Gasfield in the North Celtic Sea Basin comprises 12 m of uniform, hard, white to pale grey chalk with common glauconite (Figure 45)." data-name="images/P945243.jpg">(Figure 44). The geophysical-log character is similar to that in the wells in the South Celtic Sea and Bristol Channel basins. The upward passage from the lower Cenomanian silty claystones is transitional through thinly bedded, chalky limestones and siltstones (Colley et al., 1981), although elsewhere in the basin, a sharp contact suggests a hiatus between the two units.

In the uppermost Cenomanian there is a thin unit of grey marl which may be correlated with the darker Plenus Marl of the Anglo-Paris Basin (Jeffries, 1963; Weighell et al., 1981). There is a sharp increase in gamma-ray values and a decrease in sonic velocity at the base of the unit, followed by a gradual upward decrease in radioactivity and an increase in sonic velocity, giving a characteristic funnel-shaped log pattern (Figure 45)." data-name="images/P945243.jpg">(Figure 44) and (Figure 46). In the South Celtic Sea and Bristol Channel basins, the unit is less than 5 m thick. In the North Celtic Sea Basin it has been dated as Turonian, and is up to 12 m thick. Onshore in the UK, and in the North Sea, the Plenus Marl is variously dated from late Cenomanian to Turonian (King et al., 1989; Deegan and Scull, 1977; Crittenden et al., 1991). Deegan and Scull (1977) postulated that the shales are of volcanic origin because of their high montmorillonite content; a more-widely held view is that the unit correlates with a global occurrence of anoxic sediments at the Cenomanian–Turonian boundary, and that its sedimentation is related to a worldwide transgressive event (Schlanger et al., 1987). An alternative model equates the unit to a glacial lowering of sea level (Jeans et al., 1991).

Middle Chalk

In southern England, the Middle Chalk comprises a series of prominent hardgrounds and marls which are well marked on geophysical logs (Wood et al., 1982; Whittaker et al., 1985). The base of the Middle Chalk is taken at the top of the Plenus Marl, where there is an increase in sonic velocity reflecting the passage into the nodular chalk of the Melbourne Rock equivalent. The top of the Middle Chalk is taken at a geophysical-log horizon interpreted as the base of the Chalk Rock equivalent. In the Winterborne Kingston borehole, the Middle Chalk may be subdivided into three main geophysical-log units. There is a good geophysical-log correlation between the Winterborne Kingston borehole and offshore wells (Figure 46), an interpretation which is supported by palaeontological evidence (Company reports; Williamson, 1979).

The lowermost unit, with the Melbourne Rock equivalent at its base, is characterised by high sonic-velocity and low gamma-ray values, which reflect mainly hard, nodular chalks and interbedded marl seams. Offshore, in the South Celtic Sea and Bristol Channel basins in wells 93/2-1, 102/28-1 and 102/29-1, the unit is 20 to 25 m thick (Figure 46). It is up to 40 m thick in the North Celtic Sea Basin.

The Melbourne Rock and overlying nodular chalks pass upwards into a unit with higher gamma-ray values and a lower sonic velocity. This is interpreted as a softer, more argillaceous chalk, the uppermost part of which is a lateral equivalent of the Glynde Marl (Whittaker et al., 1985). The unit is particularly well developed in wells 102/28-1 and 102/29-1, where it is 30 m thick (Figure 46). In the North Celtic Sea Basin the unit is up to 100 m thick, although not as well defined on geophysical logs.

Above the Glynde Marl, a prominent, high sonic-velocity response identified in all wells in the South Celtic Sea and Bristol Channel basins is interpreted as a lateral equivalent of the Spurious Chalk Rock of southern England (Wood et al., 1982). It marks the base of the uppermost Middle Chalk unit, and its high velocities reflect chalk hardgrounds. In the North Celtic Sea Basin, the unit is identified but not well defined.

Upper Chalk

Lithostratigraphical subdivision of the Upper Chalk is less clearly defined than in the Lower and Middle Chalk because geophysical-log markers are uncommon. The base of the Upper Chalk in southern England (Rawson et al., 1978) is taken at the base of the Chalk Rock hardground development. In the Winterborne Kingston borehole (Wood et al., 1982; Whittaker et al., 1985), the Chalk Rock is marked by a sonic-velocity peak a little above the Spurious Chalk Rock (Figure 46). In the South Celtic Sea and Bristol Channel basins, a correlative horizon has been tentatively chosen in wells 93/2-1, 102/28-1 and 102/29-1; above it, the top of the Turonian has been palaeontologically identified. In the North Celtic Sea Basin, the base of the Upper Chalk is not well defined on geophysical logs.

Above the base of the Upper Chalk there is a gradual increase in chalk softness up-section; this is reflected by a decrease in sonic velocity (Figure 46), although its serrated character indicates interbedded hardgrounds and flint layers. In the South Celtic Sea and Bristol Channel basins, there are several sonic-velocity log breaks, one of which lies within the upper Coniacian to lower Santonian interval. Although dating of the Coniacian and Santonian is constrained by lack of diagnostic faunas (Williamson, 1979), the breaks in wells 93/2-1 and 102/28-1 are probably coeval. They are significant because they mark a distinct upward change to softer chalk (Williamson, 1979; Weighell et al., 1981). Additionally, there is a coincident change in foraminiferal faunas (Williamson, 1979; Weighell et al., 1981); the

Turonian to Coniacian interval is characterised by abundant, keeled, planktonic foraminifera such as Praeglobotruncana spp. and Globotruncana spp., whereas the Santonian to Maastrichtian sediments are dominated by calcareous benthonic species such as Stensioeina spp., Gavelinella spp. and Bolivinoides spp., with significant numbers of arenaceous benthonic forms. The lithological and foraminiferal changes at the Coniacian–Santonian boundary may be traced into the North Celtic Sea Basin, where they are associated with either a minor unconformity (Delantey et al., 1981) or a minor stratigraphical hiatus (Weighell et al., 1981). In the Winterborne Kingston borehole there is a similar log break in the lowermost Coniacian (Figure 46), although whether the breaks are coeval is not certain.

The faunal changes at the Coniacian–Santonian boundary indicate a shallowing of water depths, and changes in the temperature and salinity of the water mass (Williamson, 1979; Weighell et al., 1981). Although the upward decrease in hardness within the chalk is due in part to compaction (Weighell et al., 1981), the changes to the water mass may have produced a subtle alteration in the composition of the ooze, which modified diagenesis. The change at the Coniacian–Santonian boundary postdates the late Turonian sea-level rise, and suggests local deformation.

Comparison of the thickness of the Upper Chalk in the Quadrant 102 wells with that in well 93/2-1 (Figure 46) indicates that differential basin subsidence, initiated in Turonian times, accelerated rapidly during the Coniacian and continued throughout the Santonian, Campanian and Maastrichtian. Above the base of the Santonian, geophysical-log units cannot be correlated between the wells. In well 102/28-1, the Santonian and Campanian Chalk, notably near the top of the Santonian, gives an irregular, serrated, sonic-velocity response indicative of interbedded marls, hardgrounds and flint layers. In well 93/2-1, most of the Campanian sequence comprises soft chalk with hardgrounds and flint horizons, but in the uppermost Campanian and Maastrichtian, the sonic-velocity log is more serrated, indicating interbedded softer chalk and hardgrounds/flint horizons. There are further log breaks at 375 m and 450 m depth.

Chapter 7 Paleogene and Neogene

At the end of the Cretaceous Period, a eustatic sea-level fall (Haq et al., 1987) led to much of the basinal areas being sub-aerially exposed, and at the same time, tectonic uplift of the basins led to an erosional regime. Succeeding deposition took place in a mainly nonmarine environment, although marine conditions prevailed in the extreme south-west of the report area. Extensively distributed Paleogene sediments have been proved within the report area in the St George's Channel Basin, as well as in the South and North Celtic Sea basins. Smaller outliers are preserved in half-grabens in the Cardigan Bay and Stanley Bank basins (Figure 48), and are thought to occur in the Central Irish Sea Basin. Neogene sediments are believed to be present only in the Mochras borehole, where the largely Oligocene succession is Lower Miocene at the top (Herbert-Smith, 1979).

North of the Variscan Front, end-Cretaceous uplift and erosion may have removed both middle and Upper Cretaceous sediments (Chapter 3), although it is not established whether or not such sediments were originally deposited here (Chapter 6). Little is known about the Central Irish Sea Basin (Figure 48), but in the St George's Channel and Cardigan Bay basins, there is a marked unconformity between Jurassic and Tertiary sediments; this may reflect the deformation which took place at the end of the Cretaceous (Tucker and Arter, 1987). Uplift in the North Celtic Sea Basin took the form of true basin inversion; pre-existing growth faults were reversed, and the Cretaceous depocentre was uplifted by up to 1000 m along the inversion axes (Tucker and Arter, 1987). Uplift in the St George's Channel Basin has also been attributed to basin inversion, with movement along the St George's and associated faults (Chapter 3).

The inversion occurred in the South Celtic Sea Basin (Figure 18), Section 3.) See (Figure 18) for location." data-name="images/P945215.jpg">(Figure 16) and (Figure 18), but to the south of the Variscan Front (Figure 48), the uplift was generally of a more regional nature (Van Hoorn, 1987). Specific causes have not been identified, but it may have been due to a thermal event related to the emplacement of the Lundy igneous complex in the early Eocene (Miller and Fitch, 1962; Dodson and Long, 1962; Harland et al., 1989). The complex is the only known representative of Tertiary igneous activity within the report area. Assuming that Cretaceous and Tertiary strata were once present in both the South Celtic Sea and Bristol Channel basins, the sea-bed exposure towards the east within these basins of successively older strata indicates that the magnitude of uplift increased in this direction.

The dating of the uplift is poorly constrained. The most reliable dates are from outside the report area in the North Celtic Sea Basin (Tucker and Arter, 1987), where the uplift is post-Maastrichtian to pre-mid-Eocene in age. In the Cardigan Bay Basin, the oldest Tertiary sediments are of possible early Oligocene age, and were deposited penecontemporaneously with the final phase of uplift (Dobson and Whittington, 1987). In the Stanley Bank Basin, it is likely, by comparison with the sequences farther south in the Tertiary basins of Devon ((Figure 49); Edwards and Freshney, 1982), that sedimentation was initiated during the early Eocene, but that uplift had probably ceased by Oligocene times. The absence of Cretaceous rocks from the St George's Channel Basin precludes dating of the initiation of uplift, but the oldest Tertiary strata are mid-Eocene in age.

The morphology of the main Tertiary basins is controlled by pre-existing structure; the north-easterly trend of the St George's Channel Basin is Caledonoid, and the east-northeasterly trend of the South Celtic. Sea Basin is Variscoid (Chapter 3). The relationship between the Tertiary and Mesozoic depocentres varies; in the St George's Channel Basin they are mainly coincident, but in the South Celtic Sea Basin they are laterally offset (compare (Figure 20) and (Figure 48).

Sedimentation over much of the report area, particularly in the north and east, took place in a nonmarine environment, and fluvial and lacustrine conditions prevailed. The Tertiary sediments are best known from the Mochras borehole (Figure 50) and (Figure 51); Woodland, 1971), at the faulted margin of the Cardigan Bay Basin, and from the onshore basins in Devon (Edwards and Freshney, 1982). The Devon basins lie along the Sticklepath–Lustleigh fault zone, which also bounds the Stanley Bank Basin (Figure 48) and (Figure 49).

The sediments are poorly known offshore, as they have been sampled in only a limited number of wells and bore-holes; this constrains the interpretation of the basin sections, a problem compounded by the limited biostratigraphical dating of the dominantly nonmarine sediments. In spite of these constraints, it is apparent that in some basins the Tertiary may be subdivided into two sequences: a lower, coarse-clastic section; and an upper, finer-grained section within which lignite is commonly present.

The lower sequence of fanglomerates and rudites is well developed in the Mochras borehole and Devon, and similar lithologies are recorded in the cuttings from well 106/28-1 in the St George's Channel Basin see (Figure 55). These coarse-elastic sediments were deposited during uplift of the Welsh and Cornubian massifs (Dobson and 'Whittington, 1987; Edwards and Freshney, 1982). Dating indicates this deposition to have taken place during Eocene to Oligocene times.

The upper, finer-grained part of the Tertiary sequence is composed of sands, silts, clays and lignites; these were deposited in fluvial and floodplain environments. In the Mochras borehole, the sediments are of possibly early Oligocene to early Miocene age (Herbert-Smith, 1979), and in the St George's Channel Basin, Eocene to Oligocene ages are recorded. In Devon, the upper sequence is dated as Eocene to Oligocene. In the Stanley Bank Basin it is Oligocene in age, although it is likely by comparison with the basins in Devon that Eocene sediments may be present.

Through Tertiary times, the climate changed from tropical/subtropical to temperate as Britain drifted northwards (Wilkinson et al., 1980). The coarse-grained fluvial deposits laid down in largely oxidising environments during the Eocene and possibly the early Oligocene show the development of silcrete, suggesting that conditions may initially have been semiarid. The later units of the Oligocene, many of which were de posited under reducing conditions in a swampy environment and include abundant lignite, indicate that by this time the climate was wetter. The generally hot climatic conditions led to the development of deep weathering profiles, and subsequent erosion led to the deposition of kaolinite-dominated clays typically developed in Devon and at Mochras (Bristow, 1968; O'Sullivan, 1979). In the Oligocene section of the Bovey Basin in Devon, the in-situ flora from the floodplain deposits is dominantly subtropical in aspect (G C Wilkinson in Edwards and Freshney, 1982). However, the lignite was derived from upland areas, such as the Dartmoor Massif, where cooler conditions prevailed. The description of the Tertiary palynological assemblages in the Mochras borehole suggests strong affinities with those of the Bovey Basin (Herbert-Smith, 1979).

Whereas the Paleogene sediments in the north and east of the report area indicate a dominantly nonmarine environment of deposition, marine conditions prevailed in the southwest, as evidenced in the South Celtic Sea Basin by the common presence of glauconite. Periodically, this marine influence penetrated farther north and east, for sediments sampled in BGS boreholes marginal to the St George's Channel Basin are glauconitic, and sediments within the Stanley Bank Basin contain marine microplankton.

During Oligocene to Miocene times, a second Tertiary phase of deformation took place. It was of limited extent, and mainly affected the basins south of the Variscan Front, where uplift of about 350 m took place, perhaps associated with basin inversion (Chapter 3). Reactivation of pre-existing faults, notably dextral movement along the Sticklepath–Lustleigh fault zone, resulted in compressional deformation of the Tertiary sediments adjacent to the faults (Holloway and Chadwick, 1986).

Where thick Tertiary deposits accumulated, such as in the St George's Channel Basin, sediment loading triggered salt movement (Tucker and Arter, 1987), notably along the St George's Fault where a salt piercement structure was emplaced (Figure 22) and (Figure 48). In the South Celtic Sea Basin, salt movement was triggered by faulting (Kamerling, 1979).

Basins North of the Variscan Front

Central Irish Sea Basin

The presence of Tertiary strata within the Central Irish Sea Basin (Figure 48) is implied from seismic interpretation (Dobson and Whittington, 1979), but no samples have as yet been acquired. The sequence lies with marked unconformity upon Triassic and Lower Jurassic strata. The seismic data show the Tertiary to be characterised by generally flat-lying, locally folded, weak but persistent reflectors. Disruption of the reflectors was attributed to salt movement.

Cardigan Bay Basin

The Cardigan Bay Basin Tertiary sequence lies in an eastward-dipping half-graben that terminates against the Mochras and Tonfanau faults (Figure 48). The deep-seismic line illustrated in (Figure 51) shows a strongly unconformable relationship between the Tertiary and underlying Mesozoic strata. There are up to 600 m of Tertiary strata in the basin, of which 524 m were sampled in the Mochras borehole ((Figure 50); Woodland, 1971; Dobson and Whittington, 1987).

Mochras Borehole

Within the Middle Oligocene to Lower Miocene (Herbert-Smith, 1979) succession in the Mochras borehole, three lithostratigraphical units were erected by O'Sullivan (1979). The boundaries between the three units (Figure 50) are lithologically diffuse, but are well defined on the geophysical well logs, where breaks occur at 405 and 280 m depth. From sedimentological analysis, Dobson and Whittington (1987) identified three fine-grained, and one coarse-grained, facies. All represent deposition in nonmarine, alluvial environments (Figure 52) and (Figure 53).

The lowest lithostratigraphical division in the borehole is termed the Basal Red Unit, because of its predominant colour. It is composed of numerous, coarse-grained, cobble conglomerates (facies 4 of Dobson and Whittington, 1987) in intimate association with upward-fining sequences of sands, silts, silty clays and clays (facies 3) between 1 and 2 m thick. The conglomerates are interpreted as subaerial mass-flow deposits emplaced on alluvial fans proximal to active fault scarps. At least ten phases of ?early Oligocene fault movement have been identified from the deposition of successive conglomerate units penetrated in the borehole. The conglomerate beds are up to 14 m thick, and are either clast or matrix supported; the clast-supported types show a crude imbrication. Clasts are up to boulder size, and comprise only Lower Palaeozoic rocks that were strongly oxidised and argillised prior to erosion. The transportation of these soft clasts was attributed by Dobson and Whittington (1987, p.347) to '... deposition as subaerial mass flow deposits transported in low strength variable viscosity debris flows and surging debris flows with occasional intersurge flows'. The two facies in the Basal Red Unit contain mainly oxidised forms of iron, notably haematite, which is present as 'iron shot' and hardpan. There is a lack of siderite and significant organic carbon, indicating that the sediments were deposited in a well-drained, oxidising environment.

The Transitional Series overlies the Basal Red Unit. The lithological boundary is itself transitional, and is marked by a change in dominant colour from red to grey. On geophysical logs (Figure 50), the boundary is taken at a break to a less-irregular gamma-ray and sonic-velocity character. The series is characterised by upward-fining cycles of silt and clay, mainly the finer sections of facies 2, which have a limited component of carbonaceous material and lignite. The depositional environment is interpreted as a floodplain.

The uppermost subdivision is the Lignite and Clay Unit. The passage from the Transitional Series is marked by an increase in the proportion of lignite. On geophysical logs, the boundary is taken at an increase in the irregularity of the gamma-ray and sonic-velocity response (Figure 50). The sediments are olive-green in colour, and formed of upward-fining cycles of facies 1 and 2. Each cycle typically comprises sharp-based, thin, partly coarse-grained sand, interpreted as a crevasse splay, that passes upwards through coarse- and fine-grained silts into clays laid down in a humid, water-logged swamp or well-drained plain supporting thick vegetation. Upward-coarsening cycles at the bases of sands are interpreted as levee deposits. The sediments are usually structureless, bioturbated and organic rich, with plant debris scattered throughout the upper parts of the cycles. Well-developed lignite occurs at the top of facies 1. Spherulitic siderite is common, and illite is associated with plant material in a seatearth association. The general environment (Dobson and Whittington, 1987, p. 345) was one of '... brief episodes of sedimentation separated by long periods of pedogenesis when sediment supply was largely absent'.

Major-element geochemistry shows no variation within the section cored in the borehole, and the sediments are interpreted as being derived from a deeply weathered hinterland (Dobson and Whittington, 1987). The source rocks were acidic, and suffered desilication during weathering, resulting in aluminium and titanium enrichment. The presence of kaolinite and illite indicates intense leaching of the source rocks; the presence of gibbsite is indicative of tropical conditions.

In contrast, the mineralogy shows a marked variability attributed to differences in diagenesis due to changing environmental conditions (Dobson and Whittington, 1987). Variability is particularly evident in the iron mineralogy. In the Basal Red Unit, the presence of haematite, together with the absence of organic matter, indicates oxidising conditions associated with a well-drained environment, though not so well drained as to allow the formation of calcrete. Higher in the section, the presence of siderite indicates waterlogged conditions. In facies 1, which is the most common, plant material and rootlets indicate perennial swamp conditions, and grey soils are characteristic. In facies 2, swamp conditions again prevailed, but there was seasonal variation in the height of the water table, as indicated by the sporadic presence of haematite.

Other strata

Sediments similar to those sampled at the Mochras borehole were penetrated in a borehole at Tonfanau, another location on the eastern margin of the Cardigan Bay Basin (Figure 48). Here, 71 m of variegated, crudely graded units of conglomeratic-breccia, sand, silt and clay with lignite fragments were drilled (Institute of Geological Sciences, 1971). The sediments have lithological similarity with those at Mochras, have a comparable structural position, and are dated as Oligocene (Wilkinson et al., 1980).

On the western margin of the Cardigan Bay Basin, two BGS boreholes (Figure 48) recovered deposits which are undated, but lithologically similar to Tertiary sediments elsewhere in the basin. Borehole BH70/08 penetrated 9.5 m of pale grey clay and muddy sand with carbonaceous fragments and lignite beds. Borehole BH74/25 penetrated 51 m of grey to brown, fine-grained sand with lignite.

St George's Channel Basin

The thickest Tertiary sequence in the report area is within the St George's Channel Basin (Figure 48). The basin is divided by the St George's Fault, along the south-western portion of which a salt wall has been emplaced (Dimitropoulos and Donato, 1983). North-west of the fault, over 1500 m of Tertiary are preserved in a faulted syncline which becomes a half graben south-westwards. The axis of this north-western part of the basin is coincident with that of the underlying Mesozoic basin, although on profile SWAT-2 (Figure 54) the Tertiary rests unconformably upon the underlying beds, which are of late Mesozoic age. On SWAT-2, several unconformities have been identified within the Tertiary sequence; these have been tentatively correlated with those to the southeast of the St George's Fault (Figure 55). South-east of the St George's Fault there are over 1000 m of Tertiary strata; unconformities within this sequence have been correlated with depositional breaks in well 106/28-1 (Figure 55). In the north-east, the sequence rests with marked unconformity upon older beds (Dimitropoulos and Donato, 1983), a relationship that becomes less pronounced south-westwards in the vicinity of well 106/28-1 (Figure 48) and (Figure 55).

Four wells have been drilled in the basin (Figure 48); the overall lithological character of the Tertiary sediments within these wells shows a marked similarity to those sampled in the Mochras borehole. They are predominantly sands, silts and clays, with common lignite and woody fragments. A similar floodplain environment of deposition is envisaged, with channel sands and overbank deposits. Conglomerates have been identified in the lower section of well 106/28-1 (Figure 55); their absence in other wells may be more apparent than real, as their component clasts may be soft and argillised, as in the Basal Red Unit in the Mochras borehole.

The Paleogene successions are dated as Eocene in well 107/21-1, and Eocene to Oligocene in well 106/28-1. It is not possible to correlate stratigraphically between the wells because of the nonmarine character of the sediments, but there is a similarity in log character between well 106/24-1 (Figure 55), which is undated, and the Mochras borehole (Figure 50). Well 107/16-1, located on the north-eastern basin margin, penetrated about 100 m of undated Tertiary sand and clay. In the three other wells, the Tertiary section may be subdivided into a number of units (Figure 55).

Well 107/21-1

This well (Figure 55) is located towards the eastern margin of the basin, where three, mainly sandy units of mid- to late-Eocene age were penetrated (Barr et al., 1981).

Unit 1 at the base comprises 142 m of grey to brown, non-calcareous and calcareous clay, with fine- to medium-grained, partly coarse-grained sandstone and some lignite beds. The serrated geophysical logs indicate numerous upward-coarsening cycles (possible crevasse splays), and less-common upward-fining cycles (channel sands). A fluvial environment is indicated for these sediments, with the lignite and clay perhaps deposited as overbank accumulations.

Unit 2 is made up of 130 m of mainly medium- to coarse-grained sandstone with brown to black lignite and brick-red, noncalcareous claystone. The sandstone is interpreted as having a fluvial origin, with periodic invasions by swamps leading to the deposition of lignite and clay. The red colour in the clay suggests frequent subaerial exposure during which oxidising conditions were established.

Unit 3 at the top is composed mainly of very fine- to fine-grained sandstone with subordinate grey claystone and lignite. The finer-grained nature of this unit, as reflected in the higher gamma-ray values, together with the grey colouration of the clay indicating waterlogged conditions, suggests an overbank environment subjected to frequent flooding.

Well 106/24-1

Well 106/24-1 (Figure 55) is located in the centre of the St George's Channel Basin, and penetrated a much thicker, undated Tertiary sequence which has been subdivided into three units comparable in lithological and geophysical-log response with the Tertiary of the Mochras borehole (Figure 50). The basal unit (1) is 239 m thick, and shows a serrated log character that reflects its variable lithology of grey, silty and sandy clay with lignite, together with sandstone and siltstone. Upward-fining and upward-coarsening cycles indicate fluvial environments, and the grey colouration of the clays implies waterlogged, reducing conditions. Carbonate rocks within the section are described as 'dolomites' on the lithologs, but it is likely, by comparison with the cored section in the Mochras borehole, that these are iron carbonates, probably siderite, supporting the interpretation of high water tables. The periodic establishment of vegetated surfaces is indicated by the lignite.

Unit 2 forms a section 161 m thick that is relatively uniform on the geophysical logs. It is composed of mainly grey, slightly silty clay, with some lignite. Poorly drained, overbank environments are suggested, with sedimentation probably taking place in standing bodies of water on the alluvial plain. The uncommon presence of lignite beds suggests the infrequent establishment of swampy conditions.

The top unit (3) is composed of mainly grey, but partly brown, clay, lignitic clay and lignite, laid down in waterlogged conditions with frequent development of vegetated surfaces. The serrated log character, especially of the gamma-ray curve, reflects the interbedded nature of the unit; the low gamma-ray values register the lignite beds.

Well 106/28-1

Well 106/28-1 (Figure 55) is located to the south of the salt wall that divides the St George's Channel Basin; it sampled 930 m of late Eocene to early Oligocene sediments (Barr et al., 1981), and may be subdivided into five units that can be correlated with the seismic profile of SWAT-2 (Figure 54).

At the base is a 234 m-thick unit (1) with a very serrated log character representing an interbedded succession, mainly of sandstone and siltstone, with subordinate claystone and lignite. Conglomerate is described from the lower part of the unit. The sandstone is fine to coarse grained, and contains pyrite and siderite; carbonate rock described as 'dolomite' on the logs is again likely to be iron carbonate. The logs suggest both upward-fining and upward-coarsening cycles within the sediments, which are interpreted as having been deposited mainly under fluvial conditions. The grey to brown colours in the clay suggest a fluctuating water table on the flood-plain; the uncommon presence of lignite indicates infrequent colonisation by plants.

Above lies the 76 m-thick unit 2, which has low gamma-ray values and moderate sonic velocities of comparatively uniform character. It comprises fine- to medium-grained, pyritic sandstone and siltstone of fluvial origin. Unit 3 is 43 m thick, and displays a serrated geophysical-log response representing interbedded lignite, up to 20 m thick, and brown siltstone with mica, pyrite and carbonaceous debris. The sediments suggest deposition in fluvial environments, probably on an alluvial plain subject to flooding, on which there was episodic plant colonisation. Unit 4 is a 66 m-thick fine- to coarse-grained channel sandstone with some granule-size components.

At the top lie 312 m of mixed lithologies that form unit 5, which comprises sandstone, lignite, siltstone and red-brown claystone. The siltstone in the lower part of the unit is described as 'dolomitised', but is again interpreted as siderite. The sandstone is fine to very coarse grained and pyritic, probably carbonaceous, with lignite interbeds. One lignite bed is logged at almost 100 m in thickness. The lithologies suggest alluvial environments with well-vegetated surfaces, the varicoloured clay reflecting a fluctuating water table.

Teifi Basin

Between the St George's Channel Basin and the Welsh mainland lies a small basin, here termed the Teifi Basin, in which Tertiary sediments are preserved (Figure 48). BGS borehole BH74/21 recovered 3 m of white, lignitic clay with pink mottling; this has been dated as Tertiary (Warrington and Owens, 1977). In BGS borehole BH72/66, 10.6 m of late Eocene to mid-Oligocene, dark green and pale brown, laminated clay with plant fragments was cored (Warrington and Owens, 1977). Common glauconite in the sediment indicates a marine influence, the most northerly marine incursion encountered in the Paleogene of the report area.

Basins south of the Variscan Front

South of the Variscan Front, the Tertiary depocentres are not coincident with those of the Mesozoic basins, and Tertiary sediments commonly transgress over the Mesozoic basin margins on to Palaeozoic basement (Tucker and Arter, 1987; Van Hoorn, 1987). Sediment thicknesses in the southern basins are not as great as those to the north (Figure 48).

South Celtic Sea Basin

Tertiary sediments, dated from a sparse microfloral assemblage as mid-Eocene to late Oligocene, can exceed 500 m in thickness in the South Celtic Sea Basin (Figure 48). In BGS borehole BH89/10 (Figure 56), 77 m of vivid green, finely laminated sandstone with interbeds of silty clay and dark brown to black lignite were penetrated. The sand is micaceous, glauconitic and lignitic throughout. The lignite occurs both as individual woody fragments and as detrital clasts, mostly of sand-sized grains forming beds and lenses. Near the base of the borehole succession, distinctive, thin, brown, carbonaceous, silica-cemented sandstone occurs beneath the main lignite development; it may represent seatearth or ganister. Lignite is most abundant towards the base of the sequence, which is interpreted as having been deposited in a marginal-marine environment under low-energy conditions. The upward passage from abundant lignites into more sandy, glauconitic sediments suggests increasingly marine conditions, although the persistent presence of lignite indicates a close proximity to a well-vegetated landmass. Confirmation of marine conditions is provided by the presence of Rhaxellatype sponge spicules.

On the northern flank of the South Celtic Sea Basin, there is a small sub-basin with probably less than 200 m of unsampled Tertiary strata (Figure 48). The sediments dip to the north-west, where the basin terminates against a fault.

Bristol Channel Marginal Basin

To the south-east of the South Celtic Sea Basin lies a basin, herein named the Bristol Channel Marginal Basin, where up to 300 m of Tertiary sediments are preserved (Figure 48). In the east, the basin has a downfaulted northern margin, but to the west, Tertiary sediments rest unconformably upon Mesozoic strata. On seismic records, the sedimentary sequence is similar to that recorded in the Stanley Bank Basin (see below), and two units are identified: a lower one of discontinuous reflectors interpreted as sandy (?fluvial) in nature, and an upper sequence of continuous, parallel reflectors, which sampling at the sea bed has shown to comprise clays and lignites.

On the eastern margin of the basin, two small Tertiary sub-basins are preserved along the West Lundy fault zone (BGS Lundy Solid Geology sheet; Arthur, 1989). The beds within each basin are in part faulted and in part downwarped. BGS borehole BH75/20, in the southern sub-basin, penetrated 4 m of undated, brown and white clay.

Tertiary basins along the Sticklepath–Lustleigh fault zone

A number of Tertiary accumulations occur along the Sticklepath–Lustleigh fault zone, notably onshore at the Bovey and Petrockstow basins in Devon, and offshore in the Stanley Bank Basin (Figure 49). Farther to the north-west, and perhaps lying along the same fault zone, are possible Tertiary deposits at Flimston in Pembrokeshire (Figure 48), where white, and red-and-white mottled, and 'nearly black clays' have been described (Dixon, 1921).

There is an intimate genetic relationship between faulting and basin formation along the Sticklepath–Lustleigh fault zone, the basins having formed by its reactivation in early Eocene times, and the consequent downwarping along left-stepping, en-échelon fractures (Edwards and Freshney, 1982; Holloway and Chadwick, 1986: Arthur, 1989). Early Tertiary sinistral movement along the fault zone onshore is considered to be of the order of 6 km (Holloway and Chadwick, 1986), but evidence from the Lundy area suggested to Arthur (1989) a considerably greater displacement of 28 to 40 km. Arthur (1989) also proposed that the fault movement and basin formation offshore were associated with the emplacement, during the early Eocene, of the Lundy Granite. The Lundy Granite (Brooks and Thompson, 1973) forms part of the Lundy Horst, which is bounded to the east by the Sticklepath–Lustleigh fault zone, and to the west by the north-westerly trending West Lundy fault zone (Arthur, 1989), the Cambeak Fault of Van Hoorn (1987), along which Tertiary sediments are also preserved (Figure 48). It was proposed by Arthur (1989) that Tertiary sediments accumulated in a graben formed between these two fault systems. Subsequent uplift led to the formation of the Lundy Horst (Figure 48), from which Tertiary sediments were eroded, leaving remnant basins both to the east and west.

Basins in Devon

In Devon (Edwards and Freshney, 1982), there are restricted occurrences of flint-bearing gravels of Paleocene and Eocene age in the Haldon Hills, at Cadham, and at Orleigh Court (Figure 49). Some are derived from the in-situ weathering of Chalk, but the younger gravels were probably deposited by rivers flowing over plains at the foot of upland areas stretching eastwards from Dartmoor towards Dorset. There are larger basins at Bovey and Petrockstow, where the gravels are overlain by thick, younger, Eocene to Oligocene clays, lignites and sands.

Within the Bovey and Petrockstow basins, deposition of the lower, gravelly, braided-stream, fluvial deposits is attributed to uplift and erosion of the Dartmoor Massif in late Paleocene to early Oligocene times. A substantial thickness of sediments accumulated in these basins. Initially, deposition was confined to particularly narrow valleys and linear basins along the Sticklepath–Lustleigh fault zone; latterly, more extensive fluviolacustrine deposition took place. Sediment thickness in the Bovey Basin is up to 1100 m, of which the lower 700 m was deposited during active faulting in a central trough (Edwards, 1976). In the Petrockstow Basin, the sediments are up to 780 m thick, and again it appears that the lower sediments were deposited during active, fault-induced subsidence (Freshney, 1970).

In the Bovey Basin, the upper section (the middle and upper Bovey Formation) is formed of sands, silts and clays, with some thick lignite beds. These sediments are interpreted as having been deposited on river floodplains and in short-lived lakes. Whereas some plant material preserved in the sediments is derived from in-situ vegetation such as palms, heathers, ferns and swamp species, the majority of the lignite was introduced from a sequoia forest growing on upland areas to the north and west (G C Wilkinson in Edwards and Freshney, 1982).

In the Petrockstow Basin (Freshney, 1970) there is a marked cyclicity of sedimentation, with upward-fining sequences of sands through silts to clays and seatearths. These are typical of alluvial conditions, but some broader cyclicity is attributed to a tectonic control on subsidence and sedimentation. Fining of the sediments in a north-westerly direction is attributed to river flow away from the Dartmoor Massif towards the Bristol Channel.

Stanley Bank Basin

The Stanley Bank Basin (Figure 57), first identified by Fletcher (1975), contains up to 335 m of sediment (Brooks and James, 1975; Kamerling, 1979) of mid-Oligocene age (Boulter and Craig, 1979). It is a half-graben developed along the extension of the Sticklepath–Lustleigh fault zone, which forms the western margin of the basin. The sediments sampled from the basin are markedly similar to the upper parts of the successions in the Petrockstow and Bovey basins.

Seismic records show the basin to be infilled by three units (Davies, 1987). At sea bed in the south and east, there is an unsampled basal unit of hummocky clinoforms which pro-grade towards the centre of the basin (facies B in (Figure 57)); these are interpreted to represent a braided-channel system (Davies, 1987). They pass upwards into the major part of the succession, which is formed of uniform, parallel reflectors (facies A) which dip towards the faulted western margin of the basin. This parallel-bedded sequence has been penetrated by 6 BGS boreholes, which sampled mainly clays and lignites (Figure 57). The clays are red, grey and brown, in part mottled, silty and sideritic. Rootlets and plant fragments are found, and there are interbedded sandstones. The lignites are up to 5 m thick. The sequence is interpreted as a distal-floodplain deposit laid down in a back swamp or shallow lagoon (Boulter and Craig, 1979: Davies, 1987). A marine incursion is indicated by the presence of microplankton in borehole BH73/36 in the north of the basin. Drainage of the floodplain is suggested to have been to the north-west (Davies, 1987), continuing the direction indicated at Petrockstow (Freshney, 1970). At the top of the succession, the third unit recognised by Davies (1987) is now thought to be of Pleistocene age (see Chapter 8).

Lundy Igneous Complex

The only igneous activity of proven Tertiary age identified within the report area is represented by the Lundy igneous complex ((Figure 48) and (Figure 57); Dollar, 1942; Dangerfield, 1982). The exposed part of the intrusion is formed mainly of coarse- and fine-grained granite, comprising feldspar and quartz megacrysts set in a groundmass of potassium feldspar, quartz, plagioclase and mica. The granite is intruded by over 230 dykes (Edmonds et al., 1979), and further dykes can be identified from magnetic anomalies offshore (BGS Lundy Aeromagnetic Anomaly sheet). On Lundy Island, most of the dykes are of mafic composition: olivine-dolerite, analcime-bearing olivine-dolerite, olivine-free dolerite and quartz dolerite. There are also dykes of intermediate composition represented by trachyte and trachyandesite. The ages of the dykes (50 to 54 Ma) and the granite (50 to 55 Ma) are indistinguishable; they were emplaced during the early Eocene (Dodson and Long, 1962; Miller and Fitch, 1962; Musset et al., 1976; Harland et al., 1989).

Gravity and magnetic evidence indicate that the Lundy Granite does not extend far offshore (Cornwell, 1971), and that it lies at the eastern margin of a much larger basic intrusion, between 2.5 and 4 km thick, which lies at shallow depth (Brooks and Thompson, 1973). The horst block, within which the igneous complex lies, is bordered to the east and west by north-westerly trending faults (Figure 48). The presence of the igneous complex at this location may be associated with the intersection of these faults with east–westtrending, reactivated, Hercynian structures (Kamerling, 1979; Dangerfield, 1982; Arthur, 1989).

Chapter 8 Pleistocene and Holocene

Within the report area, deposits of Neogene age are believed to be present only at Mochras; there are no deposits of mid-Miocene to early Pleistocene age, for during this time there was uplift and erosion (Dobson and Whittington, 1987), but later Pleistocene and Holocene sediments are abundant. Quaternary strata may be recognised on seismic profiles both by their unconformable relationship to pre-Quaternary strata, and by their commonly subhorizontal stratification. However, even in the thickest sequences, major erosion surfaces are apparent, and nowhere do the deposits represent a continuous depositional sequence. Borehole correlations indicate that the Quaternary deposits are unlithified, with acoustic velocities that average 1.70 km/s, but range from some 1.65 km/s or less for the uppermost loose and soft sediments, to some 1.85 km/s for the underlying, more-consolidated units.

Using these acoustic velocities, Quaternary thicknesses in the report area are estimated to range up to some 375 m (Figure 58). The thickest deposits are concentrated in the 25 to 80 km-wide St George's Channel and Celtic Deep troughs; these are roughly coextensive with the deepest water in the St George's Channel and the Celtic Deep (Figure 59). The bathymetric troughs form part of the Celtic Trough, which extends northwards through the western Irish Sea and the North Channel. Deposits unconformably overlying Paleogene strata of the Stanley Bank Basin north of Lundy Island have similar seismic character to Quaternary deposits found farther west, and are not considered to be Neogene as suggested by Davies (1987).

The St George's Channel Trough is flanked to the west and east respectively by the Irish and Welsh platforms, which are low-relief shelves of variable width, with water depths only locally greater than 60 m. The Celtic Deep Trough is less well defined in the deeper waters of the north Celtic Sea, although lesser depths characterise the Nymphe Bank and Lundy platforms to the north-west and south-east respectively (Figure 58) and (Figure 59). To the south-west lies the Haig Fras Platform. A loosely connected series of enclosed, infilled depressions, containing Quaternary deposits up to 175 m thick (Delantey, 1980), underlies the north-western part of the Nymphe Bank; this North-West Nymphe Bank Trough has no apparent bathymetric expression.

Quaternary deposits on the platforms are generally less than 50 m thick, and are absent off many headlands and over very considerable areas of the Bristol Channel and its approaches. However, on the platforms, up to 250 m of sediment locally form the fills of major incisions, which are elongate, enclosed depressions less than 5 km wide and up to 20 km long (Figure 60) and (Figure 61). Some of the major incisions, both on the platforms and in the troughs, are incompletely filled, and are marked by elongate, enclosed, bathymetric deeps (Figure 59).

Subdivision of the sequence is based on seismic interpretation, supported by the borehole data. Although about 100 BGS boreholes have been drilled in the report area, only 5 have sampled the Quaternary sequence of the troughs. Many of the units either occur only in the deposits of the Celtic Trough, or are attenuated and difficult to correlate across the platforms. Four seismic facies are recognised:

1. Tabular stratified. These are stratiform units, typically tens of metres thick and laterally extensive for tens of kilometres, that are internally distinguished by closely spaced, horizon tal reflectors.

2.       Tabular unstratified. With an overall geometry similar to the tabular-stratified type, but lacking internal continuous reflectors, although there may be discontinuous, dipping reflectors.

3.       Lenticular infills. Channel or depression infill deposits that can be hundreds of metres in thickness. Acoustic signatures include: chaotic with ill-defined channels and dipping reflectors; discontinuous reflectors; draped reflectors; cross-stratified and horizontal reflectors; and transparent, perhaps with very faint reflectors.

4.       Lenticular upstanding. Bank or sediment-wave deposits which are largely confined to sea-bed features.

The Quaternary deposits of the report area have been divided into six formations (Hession, 1988; BGS Quaternary Geology sheets), which in order of decreasing age are: the Bardsey Loom, Caernarfon Bay, St George's Channel, Cardigan Bay, Western Irish Sea, and Surface Sands formations (Figure 61). These formations are seen on seismic profiles largely to overlie one another in the St George's Channel and Celtic Deep troughs, although nowhere are all formations present. Considerable lateral transition, including interdigitation, occurs from the Caernarfon Bay Formation into the St George's Channel Formation. Subdivision of each formation into facies and informal members may be made, although it was not attempted for the Bardsey Loom and St George's Channel formations on BGS Quaternary Geology map sheets.

The cold and warm stages of the Quaternary are referred to by different names in Ireland, Britain and north-west Europe (Figure 61), although it is assumed that the time intervals referred to are broadly equivalent (Mitchell et al., 1973; West, 1977; Synge, 1981; Warren, 1985; Bowen et al., 1986; McCabe, 1987). Even the stage classification of deposits within Britain and Ireland remains controversial (Bowen et al., 1986; Scourse et al., 1992). Onshore sections are not generally continuous with those offshore, making correlation extremely difficult, and north-west European stage names are used here. Their application to the offshore sequence must however remain tentative, for two problems bedevil the biostratigraphical dating of Quaternary deposits. Firstly, the time-span is too short for most species to have undergone appreciable evolutionary changes. Consequently assemblages reflect environmental variation. Secondly, in thick offshore Pleistocene sequences, the proportion of interglacial deposits preserved is generally about 1 per cent of the section, although interglacial stages spanned some 10 per cent of Pleistocene time (Jensen and Knudsen, 1988). The bulk of the deposits were laid down in cold environments, and are barren or have a poor and restricted fauna and flora, particularly the pre-Western Irish Sea Formation units. Biostratigraphical analyses on boreholes and cores in this area have been carried out mainly by R Harland and D M Gregory.

By analogy with the threefold division of the North Sea Quaternary sequence, where major erosion surfaces are dated as late Elsterian, late Saalian, and late Weichselian (Cameron et al., 1987), Wingfield (1989) proposed that a crude event stratigraphy can be applied to the offshore Quaternary sequences of Britain. This framework, when applied to the report area, produces the correlations shown in (Figure 61), although four major cycles occur here, the earliest of which is thought to have started in pre-Elsterian times. The eroded channels of the earliest cycle are however not deep enough to be classified as major incisions, which are arbitrarily defined as those cut deeper than 100 m.

In south-west England, tills are recorded at scattered points along the coast between Fremington, on the shores of Barnstaple Bay, and the Scilly Isles ((Figure 59); Stephens, 1973; Scourse, 1991). Otherwise, the scattered Pleistocene deposits are mostly periglacial and cold-water marine, except for raised-beach gravels tentatively classified as Hoxnian or Ipswichian (Stephens, 1973). These scattered coastal exposures indicate that the extreme limits of one or more ice sheets lay southwards of the report area. Till on the northern Scilly Isles has been dated at 20 000 ± 2000 years BP (Scourse, 1991). Farther offshore to the south-west as far as 49°30'N, rare patches of glacial diamicton are attributed to local grounding by sea ice during the substantially lower sea levels of late Weichselian times (Pantin and Evans, 1984; Evans, 1990; Scourse et al., 1992). The three generations of major incisions within the report area extend to a southern limit of about 51°N (Figure 60) and (Figure 61), and the earlier fourth generation of lesser incisions to 50°40'N. These were believed by Wingfield (1989) to indicate that lowland or tidewater ice fronts were established there for hundreds to a few thousands of years on at least four occasions. From these long-term ice fronts, relatively short-duration surge advances of up to 100 km to the south may well have occurred.

Bardsey Loom Formation

These oldest strata of the Quaternary section are confined to the St George's Channel and Celtic Sea troughs, and to the Haig Fras Platform. In the St George's Channel Trough, the deposits consist of tabular-stratified beds up to 50 m thick, which occupy shallow rockhead basins some 3 to 10 km in width (Figure 62). No samples have been obtained from these sediments, but their acoustic signature of discontinuous, concave-up, cuspate reflectors, with onlap on the basin margins, may be tentatively interpreted as rudimentary bedding within shallow channels, possibly in a fluviatile or shallow-marine environment.

In the Celtic Deep Trough, and in particular to the southwest on the Haig Fras Platform, the Bardsey Loom Formation is better preserved and comprises two members. A lower, lenticular-infill member up to about 75 m thick is confined to isolated incisions, whereas an upper, tabular-stratified member occurs more widely and is some 20 m thick. In BGS borehole BH89/10 (Figure 63), the upper member was proved as 22 m of clay, sand, pebbly sand and gravel, with layers of peat. The deposits have a sparse microbiota indicative of a cold environment, and given their position at the base of the Quaternary sequence below the base of the pre-Caernarfon Bay Formation incision level, they are tentatively ascribed a pre-Elsterian age. This contention is supported by an amino-acid ratio determination which suggests an Elsterian age for sediments well above the stratigraphical level of the Bardsey Loom Formation in BGS borehole BH89/10.

Caernarfon Bay Formation

This formation consists of four informal members (Figure 61): the Lower Unstratified, Bedded, Incision Infill, and Upper Unstratified members. The base of the Incision Infill member is marked by a major erosion surface, which is termed the first generation of major incisions.

Lower Unstratified member

This unit of tabular-unstratified facies deposits is up to 70 m thick, and occurs in the Celtic Deep Trough and on the Haig Fras Platform. On seismic profiles, the member lacks internal reflectors, but may exhibit discontinuous reflectors and hyperbolic point sources. It either lies upon the Bardsey Loom Formation with apparent conformity, or overlies a subhorizontal unconformity above pre-Quaternary rocks. The member has been proved in BGS boreholes BH81/08A and BH89/10 (Figure 63) as olive-grey till comprising hard diamicton of matrix-supported, gravelly, muddy sand with broken shells and abundant chalk and lignite fragments. Sea-bed cores have proved similar till on the Haig Fras Platform. The acoustic signature of this member is therefore thought to represent unsorted deposits laid down in subglacial or ice-proximal conditions. Since the incision surface above this member is considered to have been formed during late Elsterian glaciation, these sediments are surmised to be of earlier Elsterian age.

Bedded member

This member comprises tabular-stratified deposits that are up to 35 m thick in the St George's Channel Trough, up to 70 m thick in the Celtic Deep Trough, and less than 20 m thick on the Haig Fras Platform. Reflectors in the member vary from closely to widely spaced, and may have either strong or weak acoustic signatures. In the St George's Channel Trough, the few and restricted (less than 10 km wide) preserved remnants of the member either conformably overlie the Lower Unstratified member of the formation, or grade laterally into it (Figure 62). In the south, the member is more widely preserved, and overlies the Lower Unstratified member conformably. Only one BGS borehole, BH81/08A (Figure 63), has penetrated this facies; recovery was very poor, but the indications were of sand with occasional clay beds, and scattered pebbles with shell debris. The depositional environment is unknown, but the member immediately predates the supposed late Elsterian cutting of the erosion surface which truncates it.

Incision Infill member

The Incision Infill member forms lenticular infills more than 200 m thick within boat-shaped depressions that down cut through the Bedded member, the Lower Unstratified member, older Quaternary strata, or down some 150 m into pre-Quaternary rocks (Figure 62). The facies entirely fills the eroded depressions, except that the conformably overlying tabular unit, the Upper Unstratified member, may show a tendency to thicken across such infills. The boat-shaped depressions are termed the Intra-Caernarfon Bay Formation major incisions (Figure 60). The major incisions of this generation in the report area are up to some 40 km long and 8 km wide, and are confined to the troughs, unless the in-filled incisions north of Lundy Island are of this generation. This facies is unsampled in the report area, but boreholes farther north in the Irish Sea (Jackson et al., 1994) penetrated up to 77 m of these deposits. In ascending order the infill comprises: diamictons of stiff clay with stones, muds with clasts from pebble to boulder size, sands, muds and clays. Such infill deposits are typical of major incisions (Holmes, 1977; Long and Stoker, 1986a, b; Wingfield, 1990), and are either penecontemporaneous with the creation of the incisions or postdate it. The sediments are probably largely late Elsterian in age, but some later kettle-hole infill may be Holsteinian or even earliest Saalian.

Upper Unstratified member

The Upper Unstratified member is well developed throughout the Celtic Deep and St George's Channel troughs, as well as in central parts of Cardigan Bay on the Welsh Platform. It forms a massive, tabular-unstratified unit up to 90 m thick, and is the thickest widespread unit in the Quaternary sequence. On acoustic profiles it is distinguished by its virtual lack of internal reflectors. It either succeeds the Incision Infill member conformably, or overlies the Lower Unstratified member or older strata across a planar, subhorizontal, erosion surface. Where the Upper Unstratified member overlies the Lower Unstratified member, the two are distinguishable only because of the intervening erosion surface, which forms a strong reflector traceable for many kilometres.

In BGS borehole BH73/41 in the north, the member is composed of till, possibly subglacial lodgement till. In borehole BH89/10 in the south, it comprises sandy or muddy diamicton with dropstones (Figure 63). The latter has a sparse, cold-water flora, and is probably an ice-sheet-proximal, glaciomarine deposit laid down in early Saalian times. In the Celtic Deep Trough, deposits of this facies both underlie, and pass laterally into, the tabular-stratified deposits of the St George's Channel Formation (Figure 61) and (Figure 62).

St George's Channel Formation

This formation occurs almost exclusively within the Celtic Trough (Figure 64), and comprises tabular-stratified deposits of variable thickness; it is locally absent in the trough due to postdepositional erosion. The acoustic signature is of closely spaced, subhorizontal reflectors which are laterally continuous over many kilometres, with disconformities indicated by oversteps. In the Celtic Deep Trough, the unit attains 125 m in thickness. In the south, the deposits can be equally divided into lower and upper members separated by a widely developed, low-angle unconformity. Both members show facies changes towards the margins of the trough. The horizontally disposed reflectors of the central parts pass laterally into more-coarsely stratified, prograding wedges up to 75 m thick. These pass laterally into tabular-unstratified deposits that form the upper parts of the generally underlying, Upper Unstratified member of the Caernarfon Bay Formation (Figure 61) and (Figure 62).

Three boreholes through the St George's Channel Formation (BH71/56, BH73/41 and BH89/10; (Figure 63)) all proved muds with minor shell debris and sporadic pebbles. Although these have been described as interglacial deposits (Jasin, 1976; Garrard, 1977; Delantey and Whittington, 1977; Whittington, 1980; Hession, 1988), indications in the microbiota are overwhelmingly of boreal and cold waters, suggesting arctic-like depositional conditions. Only one interval, less than 1.5 m thick, in the mid-parts of sections from boreholes BH71/56 and BH73/41, shows slight signs of cool-temperate conditions. An amino-acid ratio determination on this formation (Figure 63) suggests a date of c.370 ka BP (Knudsen and Sejrup, 1988); this indicates a Holsteinian age, but the general framework provided by erosion surfaces has been taken to suggest a Saalian age.

Cardigan Bay Formation

This formation compares with the Caernarfon Bay Formation in that it has upper and lower tabular-unstratified members (Figure 61), with a middle member which largely comprises the lenticular infills of major incisions. However, there is no equivalent of the Caernarfon Bay Formation's Bedded member. All three members occur in sequence only in the troughs (Figure 62), although the Upper Till member is widely represented on the platforms, as, to a lesser extent, is the Bedded and Ina member. Only in the north are the Upper and Lower Till units formed of till, for they pass southward into thinner sands and gravels.

Lower Till member

This tabular-unstratified member lies upon an erosion surface with a gentle topographic variation of some 15 m (Figure 62). The member ranges in thickness from 90 m in the north, to less that 5 m at the southern limit of the Celtic Deep Trough in borehole BH89/10 (Figure 63). Internally, it may exhibit discontinuous, dipping, and cross-cutting reflectors, which produce a chaotic acoustic character that contrasts strongly with the close, continuous, parallel reflectors of both the underlying St George's Channel Formation and the overlying Bedded and Infill member.

Only two borehole sections of the member are available. BGS borehole BH71/54 in Caernarfon Bay on the Welsh Platform, just outside the report area, proved 7 m of very stiff clay with abundant pebbles (Jackson et al., 1994). Borehole BH89/10, in the Celtic Deep Trough (Figure 63), found 5 m of sand with lithic granules, sporadic pebbles, and shell clasts. These few data may suggest that the facies consists of subglacial lodgement till, passing southwards into ice-proximal glaciomarine sands. The member is thought to have been deposited by Saalian ice prior to the cutting of the second generation of major incisions.

Bedded and Infill member

Seismically, this member has two distinct subdivisions, which do not necessarily occur together: a lenticular infill lower portion, and a tabular-stratified upper part. The sequence of subunits in this member has an overall similarity to that in the Bardsey Loom Formation, although it attains over double its thickness. In detail, its lower part is similar to the Incision Infill member of the Caernarfon Bay Formation, and is up to 230 m thick in an incision 20 km south of Carnsore Point (Figure 60). The lower part of the Bedded and Infill member occupies erosional downcuts of various sizes, the largest of which (greater than 75 m deep) are shown on Figure 60. Seismic profiles across these lenticular infills show a comparable range of acoustic characters to those seen in the Incision Infill member of the Caernarfon Bay Formation. Its upper portion is similar to the tabular-stratified St George's Channel Formation, and is up to 20 m thick on the Welsh Platform, but attains 60 m thickness in the trough at about 53°N (Figure 62). This upper, tabular-stratified facies is extensively preserved in the trough, where it is seismically imaged as having parallel reflectors whose spacing reduces up the section.

Where the Bedded and Infill member crops out, a few seabed cores have recovered sand. In boreholes (Figure 63), the lower part of the Bedded and Infill member has been found to consist of sands with subordinate muds, whereas the upper part comprises fine-grained silty sands and sandy clays. Both parts have yielded sparse microbiota indicating arctic-like or boreal conditions of deposition. These cold conditions, combined with the stratigraphical position of the member, may indicate a late Saalian to early Weichselian age, with Eemian sediments either absent of poorly developed. An amino-acid ratio determination from this member in BGS borehole BH89/10 indicates an age of some 120 ka BP (Knudsen and Sejrup, 1988), which is Eemian or earliest Weichselian (Figure 63).

Eemian sea levels were similar to the present level or possibly no more than 7 m higher (Hails, 1983; Coleman and Roberts, 1988; Chen et al., 1991), and climatic conditions for the c.12 ka duration of the interglacial were similar to, or slightly warmer than, the climate in the Holocene. The relatively higher Eemian sea level may have formed the Courtmacsherry Raised Beach along the Irish coast (Synge, 1985).

Upper Till member

This member forms most of the sea bed of St George's Channel and Cardigan Bay (BGS Quaternary Geology sheets). It is a tabular-unstratified unit, generally some 30 m thick, that has been sampled by numerous cores, and in bore-holes (Figure 63), as a stiff to hard diamicton of clay with varying amounts of sand, gravel, shell, cobbles and boulders.

South of 52°N, the Upper Till member thins progressively^, as far as 51º20′N, where it appears to wedge out between tabular-stratified deposits (Figure 62). However, sea-bed cores have proved some scattered samples of the member across the Lundy Platform. In the Bristol Channel to the east of 5°W, 49 per cent of some 1600 BGS samples proved pre-Quaternary bedrock, but only 37 found the Upper Till member. Similarly, of 65 boreholes there, only 4 found the member, to a maximum thickness of 8 m in borehole BH73/60 (Figure 63). On the Haig Fras Platform, scattered samples of till near the sea bed are interpreted from the seismic profiles to represent the Lower Unstratified member of the older Caernarfon Bay Formation.

The Upper Till member is interpreted to be a subglacial lodgement till, a product of the late Weichselian glaciation. Due to subsequent erosion, this till is preserved only as patches offshore to the south of about 51°20'N. However, the occurrence of till proven to date from 20 000 ± 2000 years BP on the northern Scilly Isles (Scourse, 1991) suggests that Weichselian ice may have advanced temporarily some 130 km southwards of the limit of extensively preserved till.

The onshore Midlandian (Weichselian) subglacial tills of Ireland may be matched with the Upper Till member of the Cardigan Bay Formation. Along the east coast of Ireland from County Down to County Wexford, the Midlandian till complex has been variously described as the 'Irish Sea Drifts', the 'shelly drifts', or the 'Irish Sea Till' (Eyles and McCabe, 1989). These deposits are the subject of continuing controversy as to whether they consist dominantly of subglacial and flow tills, or comprise subordinate subglacial till only in the lower parts, overlain by subaerial sandur deposits and a suite of glaciomarine, ice-proximal gravels, through prodeltaic sands, to distal muds (Eyles and McCabe, 1989).

Western Irish Sea Formation

This formation comprises both localised incision infill deposits up to 200 m thick, and overlying, more-widespread, tabular-stratified deposits that are generally less than 10 m thick. Both parts overlie a marked erosion surface, which is locally deeply incised to define major incisions (Figure 60), (Figure 61) and (Figure 62). The deposits display marked facies changes both vertically and laterally; these changes are attributed to sedimentary passages from proximal to distal settings in relation to sediment supply. Five facies are recognised, which from oldest to youngest are: Chaotic, Prograded, Mud, Sarnau, and Codling Bank facies. The Mud facies is a boreal or glaciomarine deposit with a restricted fossil assemblage; the other facies are either barren or have restricted and apparently largely reworked fossil assemblages. Eyles and McCabe (1989) have recognised deltaic deposits on land around the St George's Channel at up to 120 m above present sea level; these late Devensian deposits were laid down in proximal-glaciomarine conditions, and may correlate with the Western Irish Sea Formation offshore.

Chaotic facies

This facies either has an amorphous seismic signature lacking reflectors, or consists of cross-cutting and irregular reflectors. It is confined to the basal parts of major incision infllls, as in Tremadog Bay (Figure 62) where it has been proved in a number of boreholes as sands, gravels and cobbles, with rafts of till and disturbed clay (Figure 63). In both the Celtic Deep Trough and in the Bristol Channel, the seismic resolution is seldom fine enough to distinguish the components of the formation where it is only a few metres thick, but numerous seabed cores have proved Chaotic facies of sandy gravel unconformably overlying older Quaternary deposits or bedrock. To the north, in the Irish Sea (Jackson et al., 1994), samples from this facies have included a sparse microbiota indicative of arctic-like conditions, and it is probable that it was laid down in ice-proximal, glaciomarine or glaciolacustrine conditions during the Weichselian.

Prograded facies

These sediments comprise prograding wedges that are proved in boreholes (Figure 63) to comprise sands which form the dominant parts of the major-incision infills of the third generation (Figure 61) and (Figure 62). Similar sands are proved in seabed cores of the Western Irish Sea Formation's thin, tabular deposits, whose distribution is not mapped as they are commonly too thin to be seismically resolved. This situation contrasts with that in the western Irish Sea (Jackson et al., 1994), where the Western Irish Sea Formation is widely and thickly developed, and both lateral facies changes and direction of progradation can be readily mapped. The facies is thought to be prodeltaic and glaciomarine, becoming increasingly ice distal with time as Weichselian ice retreated.

Mud facies

This facies has a near-transparent acoustic signature with parallel, subhorizontal reflectors that form tabular-stratified units. The deposits are black to greenish grey, shelly, sulphide- or glauconite-rich silts (Figure 63) with a cold-water, marine microbiota. Scattered pebbles and very sparse cobbles in the silts are thought to be dropstones from floating ice. The Mud facies may pass down into deposits of the Prograded facies, but some seismic profiles show a lateral passage. The Mud facies is interpreted as a distal, glaciomarine deposit that became increasingly temperate-marine with time as the climate warmed at the end of the Weichselian.

Sarnau facies

In Cardigan Bay, three low, smooth-topped ridges project seaward for up to 15 km at the sea bed (see (Figure 67); BGS Cardigan Bay Sea Bed Sediments sheet). These ridges (Welsh – sarnau, singular sam, meaning causeway) are covered by gravel, cobbles and boulders, and are formed of clastsupported, clayey diamictons. Garrard and Dobson (1974) inferred that the sarnau are late glacial median moraines of piedmont glaciers extending from the valleys in the adjacent Cambrian Mountains. The Mochras borehole (Figure 63), drilled landward of Sam Badrig (see (Figure 67)), penetrated 50 m of sandy diamicton thought to be of this facies; it overlies fine-grained strata possibly representing the Prograded facies.

Although the Sarnau facies is difficult to resolve on seismic profiles due to both the shallow water depths and the coarse-grained nature of the deposits, it is interpreted that Sarn Badrig in part overlies a base-Western Irish Sea Formation major incision and its infilling deposits of both Chaotic and Prograded facies. A morainic origin is therefore questionable, and the sarnau may be remnants of late-glacial sandur. St Patrick's Causeway, running south from the south coast of County Wexford to the Saltees (see (Figure 67)), forms a sarn of similar lithology (Hession, 1988). The Sarnau facies is similar to the more extensive Codling Bank facies.

Codling Bank facies

This facies is restricted to the region of the Codling Bank off the County Wicklow coast (mostly out of the report area north of 53°N), where the apparently dissected remnants of an earlier continuous mantle occur (Jackson et al., 1994). The facies forms the upstanding features in this region, including the shoals comprising the Codling Bank itself. The Codling Bank facies is some 15 m thick, and overlies deposits representing both the Prograded and the Chaotic facies of the Western Irish Sea Formation, which form the infill of two major incisions. Warren and Keary (1988) record that dredging operations on the Codling Bank obtained large quantities of ballast almost entirely in the cobble to boulder sizes. Both the Codling Bank and the Sarnau facies bear comparison with the uppermost late-glacial facies described onshore in Leinster as the Screen Hills braid-plain deposits (Eyles and McCabe, 1989), or the succeeding diamicton they attributed to subaqueous ice-rafting. Alternatively, a comparison might be made with the south Icelandic sandar, which Maizels (1989) identified as dominantly formed by jiikulhlaup floods.

Surface Sands Formation

Deposits of this formation are either absent or extremely thin, generally less than 2 m thick, and commonly less than 0.5 m over the greater part of the report area. However, in the nearshore and intertidal areas, there are localised thicknesses of 20 m (Adams and Haynes, 1965; Haynes and Dobson, 1969; Allen, 1990), with up to 60 m recorded as the youngest infill facies of some base-Western Irish Sea Formation major incisions (Figure 60). The base of the Surface Sands Formation is everywhere an unconformity; this erosion surface has been proved in many sea-bed cores. Where the Prograded facies of the Western Irish Sea Formation passes up without apparent break into present-day sand waves (Hession, 1988), it is inferred that a surface of reworking underlies these upstanding lenticular features.

The Surface Sands Formation has three morphological divisions: the Sea Bed Depression, SL1 and SL2 members. All are dominantly sandy, and the first two are in their upper parts the products of present-day marine processes, but each may pass down into deposits laid down in shallower-water or subaerial conditions. In intertidal areas, muds comprise an important, though subordinate, part of the Surface Sands Formation.

Both the SL1 member and the upper parts of the Sea Bed Depression member have rich faunas and floras indicative of temperate, marine conditions. Depth indications either verge upwards towards, or are similar to, those of the present localities, and the fossil assemblages are essentially those of the present sea-bed communities.

Sea Bed Depression member

This member is locally developed as the fill, either partial or complete, of hollows cross-cutting the bulk of the infill in base-Western Irish Sea Formation major incisions (Figure 60). It may be up to 60 m thick. Wingfield (1990) postulated that these hollows are large kettle holes, some of which have an incomplete postglacial sediment infill, and form enclosed bathymetric deeps (Figure 62). The Sea Bed Depression member is mostly identified on profiles that cross present-day enclosed deeps. In Tremadog Bay, BGS borehole BH70/08 proved 27 m of this member, comprising sandy silt with shell debris and a rich, temperate-marine microbiota, beneath the SL1 member (Figure 62) and (Figure 63).

The SL I and SL2 members

Following Pantin's (1977; 1978) work in the east Irish Sea, the more extensively developed parts of the Surface Sands Formation are divided into two members: a lower SL (sedimentary layer) 2, and an upper SL1.

The SL2 member is diachronous, and comprises the deposits laid down across a surface of erosion formed before, and during, the early stages of the late-glacial and postglacial marine transgression. It has a fauna and flora indicative of temperate climatic conditions, although the fossils indicate shallower water depths than at present. Basal lag gravels and reworked deposits represent shallow-marine, and locally, beach deposits; these have been found in hundreds of sea-bed cores above this erosion surface across the Western Irish Sea Formation deposits. This member includes estuarine silts with peats in the Severn wetlands, now reclaimed and undergoing erosion at the shoreline (Allen, 1990). Farther offshore in deeper waters, it includes sands in relict and moribund tidal sand ridges (Kenyon et al., 1981; see (Figure 67)). A few of these tidal sand ridges were reported by Pantin and Evans (1984) to have on their surfaces patches of till dropped from floating ice, suggesting that these parts of SL2 are contemporaneous with the Western Irish Sea or Cardigan Bay formation deposits.

SL1 lies disconformably above an erosion surface across the SL2 member or older strata. It represents presently mobile sediments, which include transitory and thin deposits, lenticular sand bedforms, and active tidal sand ridges up to 40 m thick. Nearer shore and inshore, it passes into intertidal sands, mudflats and saltmarshes (Haynes and Dobson, 1969; Allen, 1990).

The present sea floor and its evolution

Eyles and McCabe (1989) have recognised late Weichselian and Holocene events in the report area; these were the result of changes in sea level produced by the interaction of glacioeustasy, glacioisostasy, and forebulge development and rebound associated with the growth and wastage of the ice cap centred in Scotland. There was major sea-level fall (Figure 65)c associated with the growth of the late Weichselian ice sheet and its southerly spread to the Celtic Sea, and concurrent till deposition (Upper Till facies of the Cardigan Bay Formation).

With glacioisostatic depression, relative sea level rose to tidewater margins over 70 m above present sea level (Figure 65)d–f, and proximal to distal glaciomarine deposits (upper Western Irish Sea Formation) were laid down. Due to final ice-sheet melting and isostatic rebound, sea level fell to some 120 m below present level in the north Celtic Sea after 11 000 years BP (J D Scourse, written communication, 1992; Figure 65)g, and to lesser levels in Cardigan Bay before 9000 years BP (Figure 65)h, forming a temporary Holocene (Littletonian) land bridge to Ireland. This produced a surface of erosion across the glaciomarine deposits, and there was subsequent deposition of shallow-water deposits (the SL2 member of the Surface Sands Formation). Sea level then rose eustatically until the Holocene transgression attained its present tidal level at about 5000 years BP (Figure 65). This produced the present sea-bed sediments (the SL1 member of the Surface Sands Formation), with incomplete infill of offshore kettle holes (Sea Bed Depression member).

Physiography

The present sea floor is very much flatter than the surrounding land, even when compared to relatively flat land, and may be divided into areas of net erosion and net deposition. In areas of net erosion, relict sediments of Pleistocene age, or pre-Quaternary rocks, form the sea bed. Relict bedforms also occur, many of which formed at times of lower sea level. The following five bathymetric zones occur in the report area (Figure 59):

1.       Coastal embayments. These comprise embayments and estuaries, but exclude the large, open bays or bights comprising the Bristol Channel and Cardigan Bay. The embayments are up to about 10 m deep, and include up to some 50 per cent intertidal ground, incorporating drying banks. The embayments are the drowned valleys of the Holocene sea-level rise (Eyles and McCabe, 1989).

2.       Inner shelf platforms. These vary in width from 5 km off Llŷn, to greater than 200 km in the Bristol Channel, and have gentle gradients of 1:100 to 1:2000. Steeper gradients only occur over sandbanks, at open coasts, about rocky prominences, and in enclosed deeps. Water depths on the inner shelves are generally 10 to 60 m, but south of 51°40′N at the Celtic Sea Platform, the sea bed slopes gently to some 110 m on the Haig Fras Platform (Figure 58) and (Figure 59). During early and mid-Weichselian times, sea levels fluctuated between 15 and 75 m below present level, with 5 to 7 eustatic regressions and transgressions (Denton and Hughes, 1981). Climatic conditions varied from cool temperate to low arctic. Coastal processes would have been similar to those now operating in northern Canada (Taylor and McCann, 1983), and during this 90 ka span, they effectively bevelled the inner shelf between present water depths of 15 and 75 m. This process, which must also have operated over similar depth envelopes during at least two previous glacial stages, formed the low-gradient platforms and removed most earlier Pleistocene deposits, except those preserved below base level within incisions.

3.       The Celtic Trough. This is an extensive, broad trough of subdued slopes (1:50 or less), part of which runs south-south-westwards through St George's Channel to the southern end of the Celtic Deep. It is 60 km wide with shoulder depths of 60 m at 53°N, 50 km wide with shoulder depths of 80 m at 52°N, and some 30 km wide with shoulder depths of 100 m at its southern end. The depths in the trough are generally some 110 m, but extend to a maximum of almost 160 m. The combined water depth and Quaternary thickness of sediment in the Celtic Trough range up to about 500 m. Great troughs of comparable dimensions are not uncommon on northern glaciated continental shelves (Summerhoff, 1973), but there is no agreed explanation of their origin (Eisma et al., 1979; Fader et al., 1982; White, 1988; Salge and Wong, 1988 a, b; Boulton et al., 1988). The formation of the Celtic Trough in the Celtic Deep area could postdate the Bardsey Loom Formation and the Lower Unstratified facies of the Caernarfon Bay Formation, which may have been structurely downwarped as it formed (Figure 62), but most later units show a tendency to thicken into the trough and onlap its margins. Thus, an Elsterian or earlier, mid-Pleistocene age is likely, since when it has acted as an area of net deposition with intermittent erosive episodes. The erosion took place during the major falls of sea level produced by glacial maxima, and during the subsequent sea-level rises (events 1 and 2 of Eyles and McCabe (1989) in the case of the late Weichschian). There is no indication that the Celtic Trough's position is controlled by the underlying geology.

4.       Rocky prominences. Rocky prominences are generally restricted zones of rough and rugged topography with outcrops of rock. In coastal embayments they occur as rocky headlands, islets, or shoals; on the inner shelf they occur as islets, islands and shoals of very variable size ranging from the few hundreds of metres across, as in the case of the Tuskar Rock off Carnsore Point, or The Smalls, to the 10 km length of Lundy Island. However, most extensive areas of sea-bed rock outcrop, such as over much of the Bristol Channel, are of sedimentary rocks, and therefore of subdued topography. Rocky prominences in less than 25 m water depths were formed during the Holocene transgression, during its stabilisation at around the present sea level since about 5000 years BP (event 4 of Eyles and McCabe, 1989), and continue to form today.

5.       Enclosed deeps. These are areas less than 5 km wide, up to 30 km long, and 10 to 50 m deeper than the surrounding sea floor. Such deeps occur both on the inner-shelf platform and in the Celtic Trough. All the deeps appear to have similar dimensions, smooth sides and floors, and have rather irregular shapes in plan, with side gradients less than 1:10. These formed as kettle holes in the early Holocene (event 3 of Eyles and McCabe, 1989), and have remained partly unfilled due to slow deposition rates.

Each of these zones has different controls on sedimentary processes due to water depth, gradient, and distance from sediment supply. Deeper waters reduce both wave- and current-induced bottom stresses. Gradients modify the effects of both tidal currents and waves on the sea bed; the deepest waters occur in enclosed deeps, and the relatively high gradients out of these deeps effectively produce sumps which act as sediment traps due to weak bottom currents. The effect of waves towards both open and embayed coasts is also strongly controlled by the gradient of the sea bed; the larger-amplitude waves dissipate their force well offshore against gently shoaling gradients, as into Cardigan Bay, whereas about Pembrokeshire and Cornwall, steeper gradients allow the transfer of wave energies closer to the shore.

Sediments derived from coastal erosion and rivers are carried alongshore to collect in embayments, or pass offshore into the general marine circulation. Distance has a controlling effect on the ultimate destinations of these sediments in that the lengths of marine sediment transport paths vary with grain size; the coarser-grained components (pebbles and larger clasts) are moved little or remain as a lag, whereas the finest components (silt and clay sized) may be carried out of the report area. It is the middle size fractions which assume prominence; the finer-grained gravels, the sands, and the silts remain in circulation, largely in transient bedforms, although a comparatively minor proportion is abstracted into sediment sinks or estuaries.

Sea-bed sediments

In the report area, BGS have taken some 2900 sea-bed samples and other organisations another 6000; over half of these were taken east of 5°W in the Bristol Channel and Severn Estuary. The majority of the samples were obtained with grabs, dredges or short cores to less than 0.15 m penetration, although many BGS sample stations were cored up to 6 m depths (BGS Sea Bed Sediment sheets). A majority of the cores taken sampled more than one layer of sediment, as did grabs in some instances. Samples reveal two contrasting sea floors. The first type is a continuous mantle deposited in a regime that has attained hydraulic equilibrium. The second type forms a discontinuous layer which has not attained equilibrium with the hydraulic regime, and overlies sediments of a different regime. In neither case is the regime of deposition necessarily that pertaining at present. Pantin and Evans (1984) suggested that the sea-bed sediments (Figure 66) occur as three units: layer A — the mobile sediment; layer B — a gravelly, lag deposit; and layer C — Pleistocene sediments or solid rock. Only layer A, and to a limited extent layer B, are actively involved in the present hydraulic regime.

Layer B

Where layer A is absent or broken, the sea bed is mantled by a discontinuous, rudaceous deposit which generally forms a pebbly coquina or shelly gravel. This lag deposit, termed layer B, is generally 0.1 to 0.2 m thick, and comprises poorly sorted, sandy, shelly gravel, and coarse-grained sand. Across substrates including very coarse-grained components, such as the Codling Bank facies of the Western Irish Sea Formation or the Upper Till member of the Cardigan Bay Formation, layer B may form an armoured pavement of cobbles. Across fine-grained substrates, such as the Mud facies of the Western Irish Sea Formation, layer B may be absent or be represented by a shell lag with sand grains.

Where gravels at sea bed are several metres thick, layer B can be considered to form the active depth of reworking by the winnowing of fines. Layer B represents both mobile and immobile sediment; it is mobile where it forms the sea bed and undergoes only winnowing for long periods, with infrequent reworking, largely in situ, by wave action during storms. It is immobile when buried by layer A. A few cores indicate that layer B may bifurcate to interdigitate with layer A, and that pebbles and shells from layer B may be incorporated within layer A. Where layer A forms merely the upper, mobile part of comparatively thick, late Holocene, SL1 deposits, a relict layer B has formed on the basal disconformity.

Layer A

Across areas of gravelly sediment, largely in the St George's and Bristol channels (Figure 66), cover of this layer is commonly patchy, and less than 0.3 m thick in sand or gravel bedforms, with exposure of layers B or C. Many grab samples here are a mixture of two, or even three, layers.

The thin, gravelly, layer A deposits pass into thicker sands both within St George's Channel, where they pass into 'trains' of sand bedforms, and in the inner part of Cardigan Bay where, some 25 km off the coast, they merge into a sand carpet (Figure 66). Similar passages take place south-westwards into a broad belt of sand 70 km wide in the Celtic Sea (which passes farther south-west back to gravels) and into the bays off the Bristol Channel, including the Severn Estuary (Stride and Belderson, 1990). The sands forming the mobile layer A sediments are up to 40 m thick in giant sand waves (Wingfield, 1987; James and Wingfield, 1987) and in tidal sand ridges (Figure 67). In some cores, these sands are proved to overlie thin gravels similar to layer B.

In the Celtic Deep, Rosslare Bay, Tremadog Bay, the Muddy Hollow, and the Trawling Grounds of Cardigan Bay, and in many Bristol Channel bays, the sands pass progressively into muddy sands and then sandy muds (Figure 66). The largest areas of mud (less than 10 per cent sand) are mapped in the central part of Tremadog Bay and in Bridgewater Bay; considerable volumes of gas are recorded in these muds (Taylor-Smith, 1987; James and Wingfield, 1987). In the estuaries, fine-grained, mobile sediments occur both as river-channel muds, as mud in suspension for the majority of the tidal cycles, and as intertidal and salting mudflat deposits (Haynes and Dobson, 1969; Collins, 1987; Allen, 1990). These muddy sediments are over 10 m thick in places, and only their upper parts should be considered as mobile.

Oceanography

The principal controls that interact to create the presently operating hydraulic regime are climate, tidal currents, and bathymetry. Other factors of lesser importance include the increasing effects of man-made works, particularly in embayments, and the dumping of dredgings, sewage and industrial wastes (Abdullah and Royle, 1974; Barrie, 1980).

Climate

The present climate is cool temperate in a zone of variable westerly winds with frequent gales and storms; winds of force 8 and above are recorded on 35 to 45 days a year (Hydrographic Department, 1960). Both the St George's Channel and the Bristol Channel open westwards into the prevailing winds, providing fetches into the open ocean which ensure abundant wave action (Figure 68)b. The effect of waves is predominantly confined to a limited zone of wave attack between just above high water, and a few metres below low water, leading to erosion of headlands and islands. It is only within this zone that effective erosion of lithified deposits occurs in the marine environment. Less-consolidated deposits are affected by wave motions only to modest (c.15 m) depths, except during storms. However, the predominant effect of wave action in deeper waters is to produce cyclic wave loading, leading to the consolidation of the clay fraction in the sediment.

The debris from coastal erosion is carried into embayments that are protected from wave attack, and disposed to form bars and spits so as to smooth out the irregularities of the coast. That this process takes thousands of years to attain equilibrium is shown in the report area by the presence of numerous offshore islets, prominent headlands, and open bays. Yet the present wave regime has operated under similar climatic conditions, with a virtually unchanged sea level, over the last 4 to 6 ka (Tooley, 1985).

Tidal currents

Tidal ranges on spring tides in the report area vary from less than 2 m about the Arklow Bank to 14 m at Avonmouth in the Severn Estuary. Wide areas of intertidal ground are restricted to the Severn Estuary and Bridgewater Bay; only relatively small estuaries occur on the coasts of Leinster and west Wales. Farther offshore, the principal tidal effects result from the horizontal motions of the water mass set up by the tidal streams (Figure 68)a. The twice-daily reversing tidal streams produce bottom stresses so as to create bed-load partings in the St George's and Bristol channels, from which net sand-transport paths converge in the Celtic Deep area of fine-grained sediment deposition ((Figure 68)c and d; Stride and Belderson, 1990). Transient sediment movements are by the suspension of fines (silt and clay), and by the bottom traction of coarser material (sand and fine-grained gravel).

Comparison of (Figure 68)c and (Figure 66) suggests that the distribution of the sea-bed sediments, basically showing the mobile-sediment distribution (layer A) or the lack of it (layer B at sea bed), matches well with that of bottom stresses. Areas of net erosion comprise the shoals fronting open coasts where wave action is most effective, the areas of maximum bottom stress, and the areas of strong tidal streams through the St George's and Bristol channels.

Areas of net erosion are sediment sources due to sea-bed action by winnowing, which leaves gravel-, cobble- and boulder-enriched lags. The sediment provided by sea-bed erosion in these source areas is mostly sand sized; at present, mud is added almost exclusively from river input (Evans, 1982). However, the annual addition of river mud, even in the Severn Estuary, is reckoned to be only 5 per cent of the total mud volume in circulation (Kirby and Parker, 1980; Collins, 1987).

Active bedforms

Active bedforms (Figure 67), are dominantly moulded from sand, although some gravelly and muddy bedforms also occur. Stride (1982) described the passage along the sediment transport path in areas of low sediment supply as ranging from furrows and waves in gravel, through isolated, uncommon sand ribbons and sand streaks parallel to the tidal current, into transverse, horned, barchan-type, large sand waves, passing into extensive sand patches with small sand waves (here termed megaripples). Bedform trains of the low-supply sequence occur in the eastern St George's Channel north and south from the bed-load parting, westwards along the southern side of the Bristol Channel, and into the Celtic Deep from the south-west.

Stride (1982) also identified a different sequence of bed-forms in settings with abundant sediment supply. In this case he found that the gravelly sea floor has sand ribbons and elongate patches, changing down-path to a continuous carpet of sand moulded into transverse sand waves, and tidal sand ridges subparallel to the current; farther down-path there are continuous sand-wave fields. Bedform trains of the abundant-sediment supply sequence occur in the western St George's Channel off the Irish coast, in Cardigan Bay as the inshore sand carpet, westwards from Swansea Bay in the northern Bristol Channel, and south-westwards past Lundy Island (Figure 67). They also occur into the Severn Estuary and Bridgewater Bay, and from the Carnsore Point area southwestwards into the fine-grained sediments of the Celtic Deep. Sandbanks as finger-like splays parallel to tidal channels form drying banks in estuaries, notably that of the Severn.

The offshore muddy areas (Figure 66) are generally featureless except for artifacts such as trawl and anchor scars, and obstacle or wreck marks (Stride, 1982). No fluid-escape structures have been described. Evans (1982) noted that the muds of Bridgewater Bay and the Severn Estuary exhibit smooth and featureless surfaces in areas of net deposition, whereas areas of net erosion show a furrowed surface with a multitude of closely spaced, anastomosing channels up to some 1.5 m deep. The inshore drying mudflats pass landward into salting stabilised by vegetation (Haynes and Dobson, 1969), and are now extensively reclaimed at several localities.

Relict bedforms

Whereas active bedforms are attributed to present-day sedimentary processes, there is also a suite of relict bedforms (Figure 67), considered to have formed subaerially or in shallower water depths than those in which they occur at present. It should also be noted that marine features from times of relatively higher sea levels have been recognised on land (Eyles and McCabe, 1989).

Sarnau are low ridges that occur normal to the coastline in Cardigan Bay; it has been suggested that they are median moraines truncated seawards by Irish Sea ice moving south-south-eastwards from Llŷn (Hession, 1988).

Anastomosing nets of channels in St George's Channel, individually up to 200 m wide and up to a few kilometres long, are observed on sidescan sonar records by the contrasting responses of channel-fill and interchannel parts of a flat sea floor. They are interpreted to have been formed by braided rivers on subaerial sandar.

Wingfield (1987) described extensive development of polygonal textures on sidescan sonar records across sea-bed outcrops of till in St George's Channel; the features are relict, periglacial, ice-wedge polygons 15 to 80 m in diameter. Ice-wedge casts, involutions and rubble fields are periglacial features preserved across the bedrock floor of the outer Severn Estuary (Allen, 1990).

Tidal sand ridges which have least water depths of 20 m to over 60 m above their crests (Figure 67) are moribund, having formed during sea levels lower than at present (Kenyon et al., 1981). Scalps of peats exposed below high tide along the shores of the Severn Estuary record freshwater marsh deposits formed at a time of lower sea level during the Holocene sea-level rise since 7000 years BP.

Chapter 9 Economic geology

The most important economic resource identified in the report area is sand and gravel in the Bristol Channel and the Severn Estuary, for despite exploration since the early 1970s, no significant hydrocarbon discoveries have been made. Onshore, the most significant economic product from around the report area has been coal from South Wales, but extraction has not extended offshore. Feasibility studies for a barrage across the Severn Estuary have indicated the economic viability of such a project.

Sand and gravel

Dredging of sand and gravel within the report area is restricted to the Bristol Channel, where there are two main areas of extraction: the outer Severn Estuary, and off Porthcawl (Figure 69); Nunny and Chillingworth, 1986). Reserves are estimated at 40 million tons, and there is at present an annual production of approximately 2 million tons. The main ports used for offloading are in South Wales. The main licence areas are around the Holm islands in the south-western part of the estuary, where linear tidal sand ridges and sand-wave fields constitute a large volume of sand (Davies, 1980). In the dredging areas of Cardiff Grounds, Culver Sand and Holm Sand (a composite term including One Fathom Bank and Mackenzie Shoal), the sediment is medium- to coarse-grained sand, and gravel. One Fathom Bank is formed of a wide variety of sand and gravel grades, which are put to different uses: well-sorted, medium-grained sand is used as building sand; poorly sorted, coarser-grained sand is used for concrete; and gravel is utilised as fill (Davies, 1980). Off Porthcawl, the main dredging area is the linear tidal sand ridge of Nash Sand, which is formed of uniform, medium-grained sand.

Oil and gas

The generation and entrapment of hydrocarbons depends upon a combination of factors; these include the sedimentation of an organic-rich source rock, the presence of a reservoir in which hydrocarbons can accumulate, a cap rock to prevent hydrocarbon escape, and a series of tectonic events which allow the hydrocarbons to be generated after burial and to migrate into a sealed reservoir.

A comprehensive grid of commercial deep-seismic reflection data has been acquired within the report area, and 15 wells have been drilled in the UK sector (Figure 1) without commercial success. Farther west, over 70 wells have been drilled in the Irish sector, where there have been several commercial gas finds.

North Celtic Sea Basin

In the North Celtic Sea Basin, beyond the limit of the report area, the gasfield at Kinsale Head (Figure 45) has been in production since 1979. Gas reserves in the field are estimated to be 1 X 10¹² cubic feet with 47 square miles of closure at the top of the main reservoir sand — the A sand (Figure 45)." data-name="images/P945243.jpg">(Figure 44). The reservoir consists of channel, beach and shallow-marine sandstones of Aptian/Albian age which have been formed into a simple, elongate, east-north-easterly trending anticline created during Paleogene inversion (Colley et al., 1981). Gault Clay forms the cap rock for gas that is very dry; it is almost-pure methane with a very light isotopic composition, characteristics which suggest a complex origin that is still poorly understood. According to Colley et al. (1981), paraffinic crudes of uncertain origin migrated into the Wealden beds, and were subject to freshwater flushing and biodegradation to produce large volumes of isotopically light methane. This gas subsequently mixed with thermally generated methane derived from Jurassic strata, and migrated into the reservoir.

Since the discovery of the Kinsale Head Gasfield in 1971, much exploration in the North Celtic Sea Basin has concentrated on identifying similar domal structures, but with limited success. Only small discoveries, such as the Ballycotton and Seven Heads fields (Figure 45), have been made (Shannon, 1991). The limited success of exploration in the basin is attributed to the poor quality of the early seismic data, on which the chalk has masked a complex and poorly understood pre-Cretaceous structural history (Shannon, 1991). Improved seismic processing has now led to a better understanding of the basin; source rocks are known to occur in the Lower and Upper Jurassic, and potential reservoirs exist in the Permo-Triassic, Upper Jurassic and Lower to middle Cretaceous. Further exploration is anticipated to be successful, although the reservoirs may be small and structurally complex (Shannon, 1991).

South Celtic Sea and Bristol Channel basins

In the South Celtic Sea and Bristol Channel basins, the seven wells drilled have all proved to be dry. Aptian/Albian sandstones, similar to those in the Kinsale Head Gasfield, are only thinly developed, although Wealden sandstones up to 18 m thick have been penetrated in well 93/2-1 (Figure 45). Lower Jurassic sandstones up to 5 m thick in well 103/18-1 (Figure 42) have proved to be water bearing (Kamerling, 1979). Perhaps the best potential reservoirs are in the Sherwood Sandstone Group, of Triassic age, which has sandstone-dominated units over 50 m thick in well 93/6-1 (Figure 27).

The general absence of organic-rich source rocks within the two basins is a problem; the drilling results indicate that they are only patchily developed and of poor quality. According to Kamerling (1979), the early Cenomanian claystones immediately underlying the chalk offer the best potential, but have not reached thermal maturity. Jurassic, organic-rich clay-stones, which are major source rocks in the northern North Celtic Sea and North Sea basins, are not widely developed.

Silesian strata are important source rocks for gas in the southern North Sea and Irish Sea basins. In Devon, Carboniferous sandstones and mudstones contain vitrinite organic matter, usually of woody stem or bark which were carried into the prodelta and deeper-marine environments from the shallower, coal-rich environments to the north in Wales. Shales from north Devon have a total organic carbon content of between 0.49 and 1.34 per cent, and contain a dominance of gas-prone kerogen (Cornford et al., 1987). Studies of vitrinite reflectance indicate that the beds were buried to depths of between 4.5 and 7 km, and that some low-maturity shales could have generated gas on further burial. However, no cap rock has been identified, and any gas generated would have escaped during Variscan deformation.

Offshore, the few wells which penetrate the Devonian/Carboniferous indicate the pervasive presence of low-grade, nonprospective, metamorphic sediments similar to those onshore (Kamerling, 1979). The tectonic history of the two basins may also be unfavourable for hydrocarbon accumulation. Their evolution has been similar to that of the North Celtic Sea Basin, but phases of deformation during the Mid- to Late Jurassic and Early Cretaceous may have been untimely in relation to hydrocarbon generation (Kamerling, 1979).

St George's Channel Basin

North of the Variscan Front in the St George's Channel Basin, there are 3000 m of Upper Palaeozoic strata, and 6000 m of Mesozoic and Tertiary sedimentary rocks. Prospects may be better than in the South Celtic Sea and Bristol Channel basins, as oil shows have been encountered. Furthermore, thicker sections have not yet been penetrated, so that the presence of economic accumulations may remain to be proven.

Drilling objectives have included sandstones in the Upper Lias (Bridport Sand equivalent) and Middle Jurassic, with middle Cretaceous greensands, equivalent to those in the Kinsale Head Gasfield, as secondary objectives. Hydrocarbon shows have been reported in the Upper Lias as a poor gas show in well 103/2-1, from the Middle Jurassic in wells 106/24-1 and 106/24a-2B, and in the Oxfordian as a minor oil show in well 107/21-1 (Figure 30).

Source rocks are present in the basin, but are commonly marginal and immature (Barr et al., 1981). In well 107/21-1 (Figure 30), the drilled section is immature for hydrocarbon generation: above 1500 m, Tertiary and Callovian to Kimmeridgian beds are lean but gas prone. Below 1500 m, in the Lower and Middle Jurassic beds, the source potential improves to become fair to good for the generation of gas and condensate. To reach the oil window, further burial of 1000 m is necessary; such depths are attained off-structure.

In well 106/24-1 (Figure 30), source-rock analysis indicates the sediments to be marginally prospective. The Tertiary strata are thermally immature, and the source rocks lean and nonprospective. The Kimmeridgian is thermally mature and marginally prospective for heavy oil, whereas the Bathonian to lower Kimmeridgian section is thermally mature and marginally prospective for gas condensate. Source-rock analysis of the Liassic section in the Mochras borehole (Figure 33) indicates that the shales are immature, and have a low organic content (Barr et al., 1981).

The source-rock potential of the Upper Palaeozoic is likely to be better than in the basins to the south of the Variscan Front, as the rocks are unmetamorphosed. To the north-east of the report area in the East Irish Sea Basin, a widespread phase of oil and gas accumulation took place during the late Mesozoic (Bushell, 1986). Oil-prone source rocks include Dinantian limestones, Dinantian and Namurian shales, and Westphalian oil shales, cannel coals and marine bands, whereas gas-prone source rocks comprise Westphalian coals and argillaceous beds, and Namurian shales (Johnson, 1981; Lawrence et al., 1987). Permo-Triassic salt may provide an effective cap rock in the St George's Channel Basin, and it is possible that there may be prospects at depth within the basin, similar to those of the Morecambe and Hamilton gas-fields, and Lennox and Douglas oilfields, in the East Irish Sea Basin.

Severn barrage

Several schemes to build a barrage across the Severn Estuary have been proposed over the past 150 years (Clarke, 1982). Fuljames, probably stimulated by an idea of Telford's, first proposed a barrage in 1849 to provide a road and rail link across the river, and to protect against flooding. The first feasibility study on the construction of a barrage to generate electricity was completed in 1933; it concluded that it was economically justified as long as storage capacity was provided. Reappraisal of a barrage in 1943 endorsed the economic feasibility, but concluded that a storage scheme was not necessary. In 1975, a further review concluded that a barrage scheme would not produce electricity as cheaply as either coal or nuclear energy.

In 1977, a recommendation from the Select Committee on Science and Technology initiated an exhaustive study of the barrage scheme, and its report, published in 1981, identified two prospective barrage sites (Figure 69); an 'outer' location between Breaksea Point and Warren Point, and a preferred 'inner' location between Lavernock Point and Brean Down. The barrage scheme was considered to be technically feasible, and would have a generating capacity of 7200 MW with an annual energy output of 12.9 TWh, at a cost (at 1981 prices) of £5660 million (Clarke, 1982). Included in the analysis was a geological survey of the estuary (Evans, 1982); of most significance to the location of the barrage sites was the absence of a superficial sediment cover so as to allow for least-cost foundation construction. Sediment cover along the line of the inner site is less than 1 m thick, except across sandbanks where it is less than 10 m. Underlying the superficial sediments are a variety of lithologies, including Carboniferous limestone in the vicinity of the Holm islands, Triassic mudstones, and Liassic limestones and mudstones. The presence of a deep-water channel between the Holm islands influenced the decision to plan the location of caissons with turbines at this central position (Figure 69).

A follow-up study, commissioned in 1983 by the Secretary of State, confirmed much of the earlier work (Clare, 1987). It was carried out by the Severn Tidal Power Group, an association of interested companies.

References

Most of the references listed below are held in the Library of the British Geological Survey at Keyworth, Nottingham. Copies of the references can be purchased subject to the current copyright legislation.

References to BGS offshore maps are not given here, but an index map showing their distribution in the report area is presented inside the back cover.

ABDULLAH, M I, and ROYLE, L G. 1974. A study of dissolved and particulate trace elements in the Bristol Channel. Journal of the United Kingdom Marine Biological Association, Vol. 54, 581–597.

ADAMS, T D, and HAYNES, J. 1965. Foraminifera in Holocene marsh cycles at Borth, Cardiganshire (Wales). Palaeontology, Vol. 8, 27–38.

AINSWORTH, N R, and five others. 1987. Biostratigraphy of the Lower Cretaceous, Jurassic and uppermost Triassic of the North Celtic Sea and Fastnet Basins. 611–622 in Petroleum geology of North West Europe. BROOKS, J, and GLENNIE, K W (editors). (London: Graham and Trotman.)

ALLEN, J R L. 1974. The Devonian rocks of Wales and the Welsh Borderlands. 47–84 in The Upper Palaeozoic and post-Palaeozoic rocks of Wales. OWEN, T R (editor). (Cardiff: University of Wales Press.)

ALLEN, J R L. 1975. Source rocks of the Lower Old Red Sandstone: the Llanishen Conglomerate of the Cardiff area, South Wales. Proceedings of the Geologists' Association, Vol. 86, 63–76.

ALLEN, J R L. 1990. The Severn Estuary in southwest Britain: its retreat under marine transgression, and fine-sediment regime. Sedimentary Geology, Vol. 66, 13–28.

ALLEN, J R L. and WILLIAMS, B P J. 1978. The sequence of the earlier Lower Old Red Sandstone (Siluro-Devonian), north of Milford Haven, southwest Dyfed (Wales). Geological Journal, Vol. 13, 113–136.

ALLEN, P. 1981. Pursuit of Wealden models. Journal of the Geological Society of London, Vol. 138, 375–405.

ALLEN, P M, and JACKSON, A A. 1985. Geology of the country around Harlech. Memoir of the British Geological Survey, Sheet 135 and part of Sheet 149 (England and Wales).

ANDERSON, F W. 1973. The Jurassic-Cretaceous transition: the non-marine ostracod faunas. 101–110 in The Boreal Lower Cretaceous. CASEY, R, and RAWSON, P F (editors). Geological Journal Special Issue, No. 5.

ANDERTON, R, GIBBONS, W, and NICHOLSON, P G. 1992. Precambrian. 5–12 in Atlas of palaeogeography and lithofacies. COPE, J C W, INGHAM, J K, and RAWSON, P F (editors). Memoir of the Geological Society of London, No. 13.

ARTHUR, M J. 1989. The Cenozoic evolution of the Lundy pull-apart basin into the Lundy rhomb horst. Geological Magazine, Vol. 126, 187–198.

ARTHURTON, R S. 1980. Rhythmic sedimentary sequences in the Triassic Keuper Marl (Mercia Mudstone Group) of Cheshire, northwest England. Geological Journal, Vol. 15, 43–58.

BADLEY, M E, EGEBERG, T, and NIPEN, O. 1984. Development of rift basins illustrated by the structural evolution of the Oseberg Feature, Block 30/6, offshore Norway. Journal of the Geological Society of London, Vol. 141, 639–650.

BANNER, F T, BROOKS, M, and WILLIAMS, E. 1971. The geology of the approaches to Barry, Glamorgan. Proceedings of the Geologists' Association, Vol. 82, 231–247.

BARCLAY, W J. 1989. Geology of the South Wales Coalfield, Part II, the country around Abergavenny. Memoir of the British Geological Survey, Sheet 232 (England and Wales).

BARR, K W, COLTER, V S, and YOUNG, R. 1981. The geology of the Cardigan Bay-St George's Channel Basin. 432–443 in Petroleum geology of the continental shelf of North-West Europe. ILLING, L V, and HOBSON, G D (editors). (London: Heyden and Son.)

BARRIE, J V. 1980. Heavy mineral distributions in bottom sediments of the Bristol Channel, UK. Estuarine and Coastal Marine Science, Vol. 11, 369–381.

BASSETT, M G. 1980. The Caledonides of Wales, the Welsh Borderland and Central England. 34–48 in United Kingdom: introduction to general geology and guides to excursions 002, 055, 093, 151. OWEN, T R (editor). Proceedings of the 26th International Congress, Paris, 1980.

BEAUMONT, C. 1981. Foreland basins. Geophysical Journal of the Royal Astronomical Society, Vol. 65, 291–329.

BECKINSALE, R D, EVANS, J A, THORPE, R S, GIBBONS, W, and HARMON, R S. 1984. Rb-Sr whole rock isochron ages, ä¹⁸O values and geochemical data for the Sarn Igneous Complex and the Parwyd gneisses of the Mona Complex of Lfyn, North Wales. Journal of the Geological Society of London, Vol. 141, 701–709.

BIRPS and ECORS. 1986. Deep seismic reflection profiling between England, France and Ireland. Journal of the Geological Society of London, Vol. 143, 45–52.

BLUNDELL, D J, DAVEY, F J, and GRAVES, L J. 1971. Geophysical surveys over the South Irish Sea and Nymphe Bank. Journal of the Geological Society of London, Vol. 127, 229–371.

BOIS, C, GARIEL, O, and PINET, B. 1990. ECORS deep seismic surveys across Palaeozoic and Mesozoic basins in France and adjacent areas: a review. 381–401 in The potential of deep seismic profiling for hydrocarbon exploration. PINET, B, and Bois, C (editors). (Paris: Editions Technip.)

BOULTER, M C, and CRAIG, D L. 1979. A middle Oligocene pollen and spore assemblage from the Bristol Channel. Review of Palaeobotany and Palynology, Vol. 28, 163–182.

BOULTON, G S, THORS, K, and JARVIS, J. 1988. Dispersal of glacially derived sediments over part of the continental shelf of south Iceland and the geometry of the resultant sediment bodies. Marine Geology, Vol. 83, 193–223.

BOWEN, D Q, ROSE, J, MCCABE, A M, and SUTHERLAND, D G. 1986. Correlation of Quaternary glaciations in England, Ireland, Scotland and Wales. 299–340 in Quaternary glaciations in the Northern Hemisphere. BOWEN, D Q, and RICHMOND, G M (editors). Quaternary Science Review, Vol. 5, 299–340.

BRADSHAW, M J, and seven others. 1992. Jurassic. 107–129 in Atlas of palaeogeography and lithofacies. COPE, J C W, INGHAM, J K, and RAWSON, P F (editors). Memoir of the Geological Society of London, No. 13.

BREWER, J A, and five others. 1983. BIRPS deep seismic reflection studies of the British Caledonides. Nature, London, Vol. 305, 206–210.

BRISTOW, C M. 1968. The derivation of the Tertiary sediments in the Petrockstow Basin, north Devon. Proceedings of the Ussher Society, Vol. 2, 29–35.

BRISTOW, C M. and HUGHES, D E. 1971. A Tertiary thrust fault on the southern margin of the Bovey Basin. Geological Magazine, Vol. 108, 61–68.

BROOKS, M, and AL-SAADI, R H. 1977. Seismic refraction studies of geological structure in the inner part of the Bristol Channel. Journal of the Geological Society of London, Vol. 133, 433–445.

BROOKS, M, and JAMES, D G. 1975. The geological results of seismic refraction surveys in the Bristol Channel, 1970–1973. Journal of the Geological Society of London, Vol. 131, 163–182.

BROOKS, M, and THOMPSON, M S. 1973. The geological interpretation of a gravity survey of the Bristol Channel. Journal of the Geological Society of London, Vol. 129, 245–274.

BROOKS, M, TRAYNER, P M, and TRIMBLE, T J. 1988. Mesozoic reactivation of Variscan thrusting in the Bristol Channel area, UK. Journal of the Geological Society of London, Vol. 145, 439–444.

BROWNE, B C, and COOPER, R I B. 1950. The British submarine gravity surveys of 1938 and 1946. Philosophical Transactions of the Royal Society of London, Vol. A242A, 243–310.

BULLERWELL, W, and MCQUILLIN, R. 1969. Preliminary report on a seismic reflection survey in the southern Irish Sea, July 1968. Report of the Institute of Geological Sciences, No. 69/2.

BUSHELL, T P. 1986. Reservoir geology of the Morecambe Field. 189–208 in Habitat of Palaeozoic gas in NW Europe. BROOKS, J, GOFF, J C, AND VAN HOORN, B (editors). Special Publication of the Geological Society of London, No. 23.

CAMERON, T D J, STOKER, M S, and LONG, D. 1987. The history of Quaternary sedimentation in the UK sector of the North Sea Basin. Journal of the Geological Society of London, Vol. 144, 43–58.

CAMERON, T D J, and six others. 1992. United Kingdom offihore regional report: the geology of the southern North Sea. (London: HMSO for British Geological Survey.)

CAVE, R. 1977. Geology of the Malmesbury district. Memoir of the Geological Survey of Great Britain, Sheet 251 (England and Wales).

CHADWICK, R A. 1993. Aspects of basin inversion in southern Britain. Journal of the Geological Society of London, Vol. 150. 311–322.

CHADWICK, R A. LIVERMORE, R A, and PENN, I E. 1990. Continental extension in southern Britain and its relationship to opening of the North Atlantic Ocean. 411–424 in Extensional tectonics and stratigraphy of the North Atlantic margins. TANKARD, A J, and BALKEMA, H R (editors). Memoir of the American Association of Petroleum Geologists, No. 46.

CHEADLE, M J, MCGEARY, S, WARNER, M R, and MATTHEWS, D H. 1987. Extensional structures on the western UK Continental Shelf: a review of evidence from deep seismic profiling. 445–465 in Continental extensional tectonics. COWARD, M P, DEWEY, J F, and HANCOCK, P L (editors). Special Publication of the Geological Society of London, No. 28.

CHEN, J H, CURRAN, H A, WHITE, B, and WASSERBURG, G J. 1991. Precise chronology of the last interglacial period: ²³⁴U^_230Th data from fossil coral reefs in the Bahamas. Bulletin of the Geological Society of America, Vol. 103, 82–97.

CLARE, R. 1987. An introduction to the work of the Severn Tidal Power Group 1983–1985. 3–16 in Tidal Power: proceedings of the symposium organised by the Institution of Civil Engineers, London, 30–31 October 1986. (London: Thomas Telford.)

CLARKE, F J P. 1982. Severn Barrage schemes from 1849 onwards. 3–8 in Severn Barrage: proceedings of a symposium organised by the Institution of Civil Engineers, London, 8–9 October 1981. (London: Thomas Telford.)

COCHRAN, J R. 1983. Effects of finite rifting times on the development of sedimentary basins. Earth and Planetary Science Letters, Vol. 66, 289–302.

COCKS, L R M, and FORTEY, R A. 1982. Faunal evidence for oceanic separations in the Palaeozoic of Britain. Journal of the Geological Society of London, Vol. 139, 467–478.

COLEMAN, J M, and ROBERTS, H H. 1988. Sedimentary development of the Louisiana continental shelf related to sea level changes. Geo-Marine Letters, Vol. 8, 63–119.

COLIN, J-P, IOANNIDES, N S, and VINING, B. 1992. Mesozoic stratigraphy of the Goban Spur, offshore south-west Ireland. Marine and Petroleum Geology, Vol. 9, 527–541.

COLIN, J-P, LEHMANN, R A, and MORGAN, B E. 1981. Cretaceous and Late Jurassic biostratigraphy of the North Celtic Sea Basin, offshore southern Ireland. 122–155 in Microfissils from Recent and fossil shelf seas NEALE, J W, and BRASIER, M D (editors). (Chichester: Ellis Horwood Ltd.)

COLLEY, M G, McWILLIAMS, A S F, and MYERS, R C. 1981. Geology of the Kinsale Head Gas Field, Celtic Sea, Ireland. 504–510 in Petroleum geology of the continental shelf of North-West Europe. ILLING, L V, and HOBSON, G D (editors). (London: Heyden and Son.)

COLLINS, M B. 1987. Sediment transport in the Bristol Channel: a review. Proceedings of the Geologists' Association, Vol. 98, 367–383.

COPE, J C W, and BASSETT, M G. 1987. Sediment sources and Palaeozoic history of the Bristol Channel area. Proceedings of the Geologists' Association, Vol. 98, 315–330.

COPE, J C W,Gerry, T A, HOWARTH, M K, MORTON, N, and TORRENS, H S. 1980a. A correlation of Jurassic rocks in the British Isles. Part One: Introduction and Lower Jurassic. Special Report of the Geological Society of London, No. 14.

COPE, J C W,INGHAM, J K, and RAWSON, P F (editors). 1992. Atlas of palaeogeography and lithofacies. Memoir of the Geological Society of London, No. 13.

COPE, J C W,and five others. 1980b. A correlation of Jurassic rocks in the British Isles. Part Two: Middle and Upper Jurassic. Special Report of the Geological Society of London, No. 15.

CORNFORD, C, YARNELL, L, and MURCHISON, D G. 1987. Initial vitrinite reflectance results from the Carboniferous of north Devon and north Cornwall. Proceedings of the Ussher Society, Vol. 16, 468–473.

CORNWELL, J D. 1971. Geophysics of the Bristol Channel area. Proceedings of the Geological Society of London, No. 1664, 286–289.

COWARD, P M, and TRUDGILL, B. 1989. Basin development and basement structure of the Celtic Sea basins (SW Britain). Bulletin of the Geological Society of France, Vol. 8, 423–436.

CRIMES, T P. 1970. A facies analysis of the Cambrian of Wales. Palaeogeography, Palaeoclimatology, Palaeoecology, Vol. 7, 113–170.

CRITTENDEN, S, COLE, J, and HARLOW, C. 1991. The Early to "Middle" Cretaceous lithostratigraphy of the central North Sea (UK sector). Journal of Petroleum Geology, Vol. 14, 387–416.

CURRY, D, GRAY, F, HAMILTON, D, and SMITH, A J. 1967. Upper Chalk from the sea bed, south of Cork, Eire. Proceedings of the Geological Society of London, No. 1640, 134–136.

DANGERFIELD, J. 1982. The Tertiary igneous complex of Lundy. 238–248 in The geology of Devon. DURRANCE, E M, and LAMING, D J C (editors). (Exeter: University of Exeter.)

DAVEY, F J. 1970. Bouguer anomaly map of the North Celtic Sea and entrance to the Bristol Channel. Geophysical Journal of the Royal Astronomical Society, Vol. 22, 277–282.

DAVIES, C M. 1980. Evidence for the formation and age of a commercial sand deposit in the Bristol Channel. Estuarine and Coastal Marine Science, Vol. 11, 83–89.

DAVIES, C M.1987. Seismic stratigraphical sequences in the Lundy Tertiary basin, Bristol Channel. Proceedings of the Geologists' Association, Vol. 98, 355–366.

DEEGAN, C E, AND SCULL, B J. 1977. A standard lithostratigraphic nomenclature for the central and northern North Sea. Report of the Institute of Geological Sciences, No. 77/25; Norwegian Petroleum Directorate Bulletin, No. 1.

DELANTEY, L J. 1980. The geology of the north Celtic Sea. Unpublished PhD thesis, University of Wales.

DELANTEY, L J. and WHITTINGTON, R J. 1977. A re-assessment of the "Neogene" deposits of the south Irish Sea and Nymphe Bank. Marine Geology, Vol. 24, M23-M30.

DELANTEY, L J. and WHITTINGTON, R J.  and DOBSON, M R. 1981. The geology of the North Celtic sea west of 7° longitude. Proceedings of the Royal Irish Academy, Vol. 81B, 37–51.

DENTON, G H, and HUGHES, T J. 1981. The last great ice sheets. (New York: Wiley.)

DIMITROPOULOS, K, and DONATO, J A. 1983. The gravity anomaly of the St George's Channel Basin, southern Irish Sea - a possible explanation in terms of salt migration. Journal of the Geological Society of London, Vol. 140, 239–244.

DINELEY, D L. 1966. The Dartmouth Beds of Bigbury Bay, south Devon. Quarterly Journal of the Geological Society of London, Vol. 122, 187–217.

DIXON, E E L. 1921. Geology of the South Wales Coalfield. XIII: the country around Pembroke and Tenby. Memoir of the Geological Survey of the United Kingdom.

DIXON, J. 1968. Pleistocene episodes of the north east Celtic Sea trough. Proceedings of the Ussher Society, Vol. 2, 52.

DOBSON, M R, EVANS, W E, and WHITTINGTON, R J. 1973. The geology of the south Irish Sea. Report of the Institute of Geological Sciences, No. 73/11.

DOBSON, M R and WHITTINGTON, R J.1979. -The geology of the Kish Bank Basin. Journal of the Geological Society of London, Vol. 136, 243–249.

DOBSON, M R and WHITTINGTON, R J 1987. The geology of Cardigan Bay. Proceedings of the Geologists' Association, Vol. 98, 331–353.

DODSON, M H, and LONG, L E. 1962. Age of the Lundy granite, Bristol Channel. Nature, London, Vol. 195, 975–976.

DOLLAR, A T J. 1942. The Lundy complex: its petrology and tectonics. Quarterly Journal of the Geological Society of London, Vol. 97, 39–77.

DONOVAN, D T, and KELLAWAY, G A. 1984. Geology of the Bristol district: the Lower Jurassic rocks. Memoir for 1:63 360 Bristol geological special sheet (England and Wales).

DONOVAN, D T, SAVAGE, R J G, STRIDE, A H, and STUBBS, A R. 1961. Geology of the floor of the Bristol Channel. Nature, London, Vol. 189, 51–52.

DORP., A G. 1976. Preliminary geological interpretation of the Bristol Channel approaches. Journal of the Geological Society of London, Vol. 132, 453–459.

DRAPER, L. 1973. Extreme wave conditions in British and adjacent waters. Proceedings, 13th Coastal Engineering Conference, Vancouver, Canada. 157–165.

EDMONDS, E A, McKEOWN, M C, and WILLIAMS, M. 1975. British regional geology: south-west England (4th edition). (London: HMSO for Institute of Geological Sciences.)

EDMONDS, E A, WILLIAMS, B J, and TAYLOR, R T. 1979. Geology of Bideford and Lundy Island. Memoir of the Geological Survey of Great Britain. Sheet 292 with 275, 276, 291 and part of 308 (England and Wales).

EDWARDS, R A. 1976. Tertiary sediments and structure of the Bovey Basin, south Devon. Proceedings of the Geologists' Association, Vol. 87, 1–26.

EDWARDS, R A. and FRESHNEY, E C. 1982. The Tertiary sedimentary rocks. 204–237 in The geology of Devon. DURRANCE, E M, and LAMING, D J C (editors). (Exeter: University of Exeter.)

EISMA, D, JANSEN, J H F, and VAN WEERING, Tj C E. 1979. Sea-floor morphology and recent sediment movement in the North Sea. 217–231 in The Quaternary history of the North Sea. OELE, E, SCHÜTTENHELM, R T E, and WIGGERS, A J (editors). Universitatis Upsaliensis: Symposia Universitatis Upsaliensis, Annum Quingentesimum Celebrantis, Vol. 2, (Uppsala: University of Uppsala.)

EVANS, C D R. 1982. The geology and superficial sediments of the inner Bristol Channel and Severn Estuary. 35–42 in Severn Barrage: proceedings of a symposium organised by the Institution of Civil Engineers, London, 8–9 October 1981. (London: Thomas Telford.)

EVANS, C D R. 1990. United Kingdom offshore regional report: the geology of the western English Channel and its south-western approaches. (London: HMSO for British Geological Survey.)

EVANS, D J, and THOMPSON, M S. 1979. The geology of the central Bristol Channel and the Lundy area, South Western Approaches, British Isles. Proceedings of the Geologists' Association, Vol. 90, 1–14.

EYLES, N, and MCCABE, A M. 1989. The late Devensian (<22 000 BP) Irish Sea Basin: the sedimentary record of a collapsed ice sheet margin. Quaternary Science Reviews, Vol. 8, 307–351.

FADER, G B, KING, L H, and JOSENHANS, H W. 1982. Surficial geology of the Laurentian Channel and the western Grand Banks of Newfoundland. Marine Sciences, Paper 21; Geological Survey of Canada, Paper 81–22, 1–25.

FANNIN, N G T. 1989. Offshore investigations 1966–87. British Geological Survey Technical Report, WB/89/02.

FLETCHER, B N. 1975. A new Tertiary basin east of Lundy Island. Journal of the Geological Society of London, Vol. 131, 223–225.

FORTEY, R A, and OWENS, R M. 1987. Early Ordovician (Arenig) stratigraphy and faunas of the Carmarthen district, south-west Wales. Bulletin of the British Museum of Natural History (Geology), Vol. 30, 225–294.

FRESHNEY, E C. 1970. Cyclical sedimentation in the Petrockstow Basin. Proceedings of the Ussher Society, Vol. 2, 179–189.

FRESHNEY, E C. and TAYLOR, E A. 1980. The Variscides of southwest Britain. 49–57 in United Kingdom: introduction to general geology and guides to excursions 002, 055, 093 and 151. OWEN, T R (editor). Proceedings of the 26th International Congress, Paris, 1980.

FYFE, J A, LONG, D, and EVANS, D. 1993. United Kingdom offihore regional report: the geology of the Malin-Hebrides sea area. (London: HMSO for British Geological Survey.)

GALLOIS, R W, and COX, B M. 1976. The stratigraphy of the Lower Kimmeridge Clay of eastern England. Proceedings of the Yorkshire Geological Society, Vol. 41, 13–26.

GARRARD, R A. 1977. The sediments of the south Irish Sea and Nymphe Bank area of the Celtic Sea. 69–92 in The Quaternary history of the Irish Sea. KIDSON, C, and TOOLEY, M J (editors). Special Issue of the Geological Journal No. 7.

GARRARD, R A. and DOBSON, M R. 1974. The nature and maximum extent of glacial sediments off the west coast of Wales. Marine Geology, Vol. 16, 31–44.

GAYER, R, and JONES, J. 1989. The Variscan foreland in South Wales. Proceedings of the Ussher Society, Vol. 9, 177–179.

GEORGE, T N. 1960. Lower Carboniferous rocks in County Wexford. Quarterly Journal of the Geological Society of London, Vol. 116, 349–364.

GEORGE, T N. 1961. British regional geology: North Wales. (London: HMSO for Geological Survey and Museum.)

GEORGE, T N. 1970. British regional geology: South Wales. (London: HMSO for Institute of Geological Sciences.)

GEORGE, T N. 1978. Eustasy and tectonics: sedimentary rhythms and stratigraphical units in British Dinantian correlation. Proceedings of the Yorkshire Geological Society, Vol. 42, 229–262.

GEORGE, T N. and six others. 1976. A correlation of Dinantian rocks in the British Isles. Special Report of the Geological Society of London, No. 7.

GIBBONS, W. 1983. Stratigraphy, subduction and strike-slip faulting in the Mona Complex of North Wales - a review. Proceedings of the Geologists' Association, Vol. 94, 147–163.

GIBBONS, W. 1987. Menai Straits fault system: An early Caledonian terrane boundary in North Wales. Geology, Vol. 15, 744–747.

GIBBS, A. 1983. Balanced cross-section construction from seismic sections in areas of extensional tectonics. Journal of Structural Geology, Vol. 5, No. 2, 153–160.

GOLDRING, R. 1971. Shallow water sedimentation as illustrated in the upper Devonian Baggy Beds. Memoir of the Geological Society of London, No. 5.

GOLDRING, R, HOUSE, M R, SELWOOD, E B, SIMPSON, S, and ST LAMBERT, R. 1967. Devonian of Southern Britain. In International Symposium on the Devonian System. Vol. 1, 1–14. OSWALD, D H (editor). (Calgary: Alberta Society of Petroleum Geologists.)

GREENLY, E. 1919. The geology of Anglesey. Memoir of the Geological Survey of Great Britain.

GRIFFITHS, D H, KING, R F, and WILSON, C D V. 1961. Geophysical investigations in Tremadoc Bay, North Wales. Quarterly Journal of the Geological Society of London, Vol. 117, 171–191.

HAILS, J R. 1983. Coastal processes, relict shorelines and changes in sea level on selected mid- and low-latitude coasts. 29–52 in Shorelines and isostasy. SMITH, D E, and DAWSON, A G (editors). Special Publication of the Institute of British Geographers, No. 16.

HALLAM, A. 1964. Origin of the limestone-shale rhythm in the Blue Lias of England: a composite theory. Journal of Geology, Vol. 72, 157–169.

HANCOCK, J M. 1969. Transgression of the Cretaceous sea in south-west England. Proceedings of the Ussher Society, Vol. 2, 61–83.

HANCOCK, J M. 1975. The petrology of the Chalk. Proceedings of the Geologists' Association, Vol. 86, 499–535.

HAQ, B U, HARDENBOL, J, and VAIL, P R. 1987. Chronology of fluctuating sea levels since the Triassic. Science, New York, Vol. 235, 1156–1167.

HAQ, B U, HARDENBOL, J, and VAIL, P R. 1988. Mesozoic and Cainozoic chronostratigraphy and eustatic cycles. 71–108 in Sea-level changes: an integrated approach. WILGUS, C K, and five others (editors). Society of Economic Palaeontologists and Mineralogists Special Publication, No. 42.

HARLAND, W B, and five others. 1989. A geologic timescale 1989 (Cambridge: Cambridge University Press.)

HARRISON, R K. 1971. The petrology of the Upper Triassic rocks in the Llanbedr (Mochras Farm) Borehole. 37–72 in The Llanbedr (Mochras Farm) Borehole. WOODLAND, A W (editor). Report of the Institute of Geological Sciences, No. 71/18.

HART, M B. 1982. The marine rocks of the Mesozoic. 181–203 in The geology of Devon. DURRANCE, E M, and LAMING, D J C (editors). (Exeter: University of Exeter.)

HARTLEY, A J, and WARR, L N. 1990. Upper Carboniferous foreland basin in SW Britain. Proceedings of the Ussher Society, Vol. 7, 212–216.

HAYNES, J, and DOBSON, M R. 1969. Physiography, foraminifera and sedimentation in the Dovey Estuary (Wales). Geological Journal, Vol. 6, 217–256.

HERBERT-SMITH, M. 1979. The age of the Tertiary deposits of the Llanbedr (Mochras Farm) borehole as determined from palynological studies. Report of the Institute of Geological Sciences, No. 78/24, 15–29.

HESSION, M A I. 1988. Quaternary geology of the South Irish Sea. Unpublished PhD thesis, University of Wales.

HIGGS, R. 1991. The Bude Formation (Lower Westphalian), SW England: siliciclastic shelf sedimentation in a large equatorial lake. Sedimentology, Vol. 38, 445–469.

HILLIS R R. 1988. The geology and tectonic evolution of the Western Approaches Trough. Unpublished PhD thesis, University of Edinburgh.

HILLIS R R. 1991. Chalk porosity and Tertiary uplift, Western Approaches Trough, SW UK and NW French continental shelves. Journal of the Geological Society of London, Vol. 669–680.

HOLDER, M T, and LEVERIDGE, B E. 1986. Correlation of the Rhenohercynian Variscides. Journal of the Geological Society of London, Vol. 141, 61–70.

HOLLOWAY, S, and CHADWICK, R A. 1986. The Sticklepath–Lustleigh fault zone: Tertiary sinistral reactivation of a Variscan dextral strike-slip fault. Journal of the Geological Society of London, Vol. 143, 447–452.

HOLLOWAY, S, MILODOWSKI, A E, STRONG, G E, and WARRINGTON, G. 1989. The Sherwood Sandstone Group (Triassic) of the Wessex Basin, southern England. Proceedings of the Geologists' Association, Vol. 100, 383–394.

HOLMES, R. 1977. Quaternary deposits of the central North Sea, 5. The Quaternary geology of the UK sector of the North Sea between 56° and 58°N. Report of the Institute of Geological Sciences, No. 77/14.

HORTON, A, POOLE, E G, WILLIAMS, B J, KLING, V C, and HOBSON, G D. 1987. Geology of the country around Chipping Norton. Memoir of the British Geological Survey, Sheet 218 (England and Wales).

HORTON, A, and six others. 1993. Thame. England and Wales Sheet 237. Solid and Drift Geology. 1:50 000. (Southampton: Ordnance Survey for British Geological Survey.)

HOWELLS, M F, FRANCIS, E H, ADDISON, R, and REEDMAN, A J. 1983. The lithostratigraphical subdivision of the Ordovician underlying the Snowdon and Crafnant Volcanic Groups, North Wales. Report of the Institute of Geological Sciences, No. 83/1.

HYDROGRAPHIC DEPARTMENT. 1960. West coast of England pilot (10th edition). (Taunton: Hydrographer of the Navy.)

INSTITUTE OF GEOLOGICAL SCIENCES. 1971. Annual report. (London: Institute of Geological Sciences.)

IVIMEY-COOK, H C. 1971. Stratigraphical palaeontology of the Lower Jurassic of the Llanbedr (Mochras Farm) Borehole. 87–92 in The Llanbedr (Mochras Farm) Borehole. WOODLAND, A W (editor). Report of the Institute of Geological Sciences, No. 71/18.

JACKSON, D I, and five others. 1994. United Kingdom offihore regional report: the geology of the Irish Sea. (London: HMSO for British Geological Survey.)

JAMES, J W C, and WINGFIELD, R T R. 1987. Aspects of sea bed sediments in the southern Irish Sea. Proceedings of the Geologists' Association, Vol. 98, 404–406.

JASIN, B. 1976. Quaternary microfauna in 5 BGS boreholes in Cardigan Bay. Unpublished MSc thesis, University of Wales.

JEFFRIES, R P S. 1963. The stratigraphy of the Actinocamax plenus subzone (Turonian) in the Anglo-Paris Basin. Proceedings of the Geologists' Association, Vol. 74, 1–33.

JEANS, C V, LONG, D, HALL, M A, BLAND, D J, and CORNEORD, C. 1991. The geochemistry of the Plenus Marl at Dover, England: evidence of fluctuating oceanographic conditions and of glacial control during the development of the Cenomanian-Turonian anomaly. Geological Magazine, Vol. 128, 603–632.

JENNER, J K. 1981. The structure and stratigraphy of the Kish Bank Basin. 426–431 in Petroleum geology of the continental shelf of North-West Europe. ILLING, L V, and HOBSON, G D (editors). (London: Heyden and Son.)

JENSEN, K A, and KNUDSEN, K L. 1988. Quaternary foraminiferal stratigraphy in boring 81/29 from the central North Sea. Boreas, Vol. 17, 273–287.

JOHNSON, E W. 1981. A tunnel section through a prograding Namurian (Arnsbergian E_2a) delta, in the western Bowland Fells, north Lancashire. Geological Journal, Vol. 16, 93–110.

JONES, O T. 1952. The drainage system of Wales and adjacent regions. Quarterly Journal of the Geological Society of London, Vol. 107, 201–225.

JONES, O T. 1956. The geological evolution of Wales and the adjacent regions. Quarterly Journal of the Geological Society of London, Vol. 111, 323–352.

JONES, O T. and PUGH, W J. 1949. An early Ordovician shoreline in Radnorshire, near Builth Wells. Quarterly Journal of the Geological Society of London, Vol. 105, 65–99.

KAMERLING, P. 1979. The geology and hydrocarbon habitat of the Bristol Channel Basin. Journal of Petroleum Geology, Vol. 2, 75–93.

KELLING, G. 1974. Upper Carboniferous sedimentation in South Wales. 185–224 in The Upper Palaeozoic and post-Palaeozoic rocks of Wales. OWEN, T R (editor). (Cardiff: University of Wales Press.)

KELLING, G. 1988. Silesian sedimentation and tectonics in the South Wales Basin: a brief review. 38–42 in Sedimentation in a synorogenic basin complex: the Upper Carboniferous of Northwest Europe. (London and Glasgow: Blackie.)

KELLING, G. and WOODLANDS, M A. 1969. The stratigraphy and sedimentation of the Llandoverian rocks of the Rhayader district. 255–282 in The Precambrian and Lower Palaeozoic rocks of Wales. WOOD, A (editor). (Cardiff: University of Wales Press.)

KENYON, N H, BELDERSON, R H, STRIDE, A H, and JOHNSON, M A. 1981. Offshore tidal sandbanks as indicators of net sand transport and as potential deposits. Special Publication of the International Association of Sedimentologists, Vol. 55, 257–268.

KING, C, BAILEY, H W, BURTON, C A, and KING, A D. 1989. The Cretaceous of the North Sea. 372–417 in Stratigraphical atlas of fossil foraminifera. (2nd edition). JENKINS, D E, and MURRAY, J W (editors). British Micropalological Society Series. (Chichester: Ellis Harwood Ltd.)

KIRBY, R, and PARKER, W R. 1980. Settled mud deposits in Bridgwater Bay, Bristol Channel. Report of the Institute of Oceanographic Sciences, No. 107.

KNUDSEN, K L, and SEJRUP, H P. 1988. Amino-acid geochronology of selected interglacial sites in the North Sea area. Boreas, Vol. 17. 347–354.

KOKELAAR, B P. 1988. Tectonic controls of Ordovician arc and marginal basin volcanism in Wales. Journal of the Geological Society of London, Vol. 145, 759–775.

KOKELAAR, B P. BEVINS, R E, JAMES, D E, and ROACH, R A. 1985. Submarine silicic volcanism and associated sedimentary and tectonic processes, Ramsay Island, SW Wales. Journal of the Geological Society of London, Vol. 142, 591–613.

KOKELAAR, B P. HOWELLS, M F, BEVINS, R E, ROACH, R A, and DUNKLEY, P N. 1984. The Ordovician marginal basin of Wales. 245–269 in Marginal basin geology. KOKELAAR, B P, and HOWELLS, M F (editors). Special Publication of the Geological Society of London, No. 16.

LAKE, S D, and KARNER, G D. 1987. The structure and evolution of the Wessex Basin, southern England: an example of inversion tectonics. Tectonophysics, Vol. 137, 347–378.

LAWRENCE, S R, COSTER, P W, and IRELAND, R J. 1987. Structural development and petroleum potential of the northern flanks of the Bowland Basin (Carboniferous), north-west England. 225–233 in Petroleum geology of North West Europe. BROOKS, J, and GLENNIE, K W (editors). (London: Graham and Trotman.)

LLOYD, A J. 1963. Upper Jurassic rocks beneath the Bristol Channel. Nature, London, Vol. 198, 375–376.

LLOYD, A J. SAVAGE, R J G, STRIDE, A H, and DONOVAN, D T. 1973. The geology of the Bristol Channel floor. Philosophical Transactions of the Royal Society of London, Vol. A274, 595–626.

LONG, D, and STOKER, M S. 1986a. Channels in the North Sea: the nature of a hazard. 339–351 in Advances in Underwater Technology. Ocean Sciences and Off hone Engineering, No. 6. (London: Graham and Trotman.)

LONG, D, and STOKER, M S. 1986b. Valley asymmetry: evidence for periglacial activity in the central North Sea. Earth Surface Processes and Landforms, Vol. 11, 525–532.

LOTT, G K. 1982. The sedimentology of the Lower Chalk (Middle-Upper Cenomanian) of the Winterborne Kingston borehole, Dorset. 28–34 in The Winterborne Kingston borehole, Dorset, England. RHYS, G H, LOTT, G K, and CALVERT, M A (editors). Report of the Institute of Geological Sciences, No. 81/3.

LOTT, G K.  SOBEY, R A, WARRINGTON, G, and WHITTAKER, A. 1982. The Mercia Mudstone Group (Triassic) in the western Wessex Basin. Proceedings of the Ussher Society, Vol. 5, 340–346.

MAIZELS, J. 1989. Sedimentology, palaeoflow dynamics and flood history of jokulhlaup deposits: palaeohydrology of Holocene sediment sequences in southern Iceland sandur deposits. Journal of Sedimentary Petrology, Vol. 59, 204–223.

MATTHEWS, D H, and CHEADLE, M J. 1986. Deep reflections from the Caledonides and Variscides west of Britain and comparison with the Himalayas. 21–30 in Reflection seismology: a global perspective. BARAZANGI, M, and BROWN, L (editors). American Geophysical Union Geodynamics Series, Vol. 13.

MCCABE, A M. 1987. Quaternary deposits and glacial stratigraphy in Ireland. Quarternary Science Reviews, Vol. 6, 259–299.

MCGEARY, S, WARNER, M R, CHEADLE, M J, and BLUNDELL, D J. 1987. Crustal structure of the continental shelf around Britain derived from BIRPS deep seismic profiling. 33–42 in Petroleum geology of North West Europe. BROOKS, J, and GLENNIE, K W (editors). (London: Graham and Trotman.)

MCKENZIE, D P. 1978. Some remarks on the development of sedimentary basins. Earth and Planetary Science Letters, Vol. 40, 25–32.

MECHIE, J, and BROOKS, M. 1984. A seismic study of deep geological structure in the Bristol Channel area. Geophysical Journal of the Royal Astronomical Society, Vol. 78, 661–689.

MEISSNER, R, and KUSZNIR, N J. 1987. Crustal viscosity and the reflectivity of the lower crust. Annals Geophysicae, Vol. 5B, 365–374.

MELVIN, J. 1986. Upper Carboniferous fine-grained turbiditic sandstones from southwest England: a model for growth in an ancient, delta-fed subsea fan. Journal of Sedimentary Petrology, Vol. 56, 19–34.

MILLER, J A, and FITCH, F J. 1962. Age of the Lundy Island granites. Nature, London, Vol. 195, 553–555.

MILLER, R G. 1990. A paleoceanographic approach to the Kirnmeridge Clay Formation. 13–26 in Deposition of organic facies. HUC, A Y (editor). (Tulsa: American Association of Petroleum Geologists.)

MILLSON, J A. 1987a. The sedimentology of the Celtic Sea Jurassic. Unpublished PhD thesis, University of Wales.

MILLSON, J A.1987b. The Jurassic evolution of the Celtic Sea basins. 599–610 in Petroleum geology of North West Europe. BROOKS, J, and GLENNIE, K W (editors). (London: Graham and Trotman.)

MITCHELL, G F, PENNY, L F, SHOTTON, F W, and WEST, R G. 1973. A correlation of Quaternary deposits in the British Isles. Special Report of the Geological Society of London, No. 4.

MUSSETT, A E, DAGLEY, P, and ECKFORD, M. 1976. The British Tertiary igneous province: palaeomagnetism and ages of dykes, Lundy Island, Bristol Channel. Geophysical Journal of the Royal Astronomical Society, Vol. 46, 595–603.

NAYLOR, D, and MOUNTENEY, S N. 1975. Geology of the North West European Continental Shelf Vol 1: The West British Shelf (London: Graham and Trotman.)

NAYLOR, D,and SHANNON, P M. 1982. Geology of offshore Ireland and west Britain. (London: Graham and Trotman.)

NUNNY, R S, and CHILLINGWORTH, P C H. 1986. Marine dredging for sand and gravel. Report of the Department of the Environment Minerals Division. (London: HMSO.)

OGUIKE, R O. 1969. Sedimentation of the Middle Shales (Upper Namurian) of the South Wales Coalfield. Unpublished PhD thesis, University of Wales.

O'SULLIVAN, K N. 1979. The sedimentology, geochemistry and conditions of deposition of the Tertiary rocks of the Llanbedr (Mochras Farm) borehole. Report of the Institute of Geological Sciences, No. 78/24, 1–13.

O'SULLIVAN, K N. IVIMEY-COOK, H C, LEWIS, B J, and HARRISON, R K. 1971. Log of the Llanbedr (Mochras Farm) borehole. 11–13 in The Llanbedr (Mochras Farm) borehole. WOODLAND, A W (editor). Report of the Institute of Geological Sciences, No. 71/18.

PANTIN, H M. 1977. Quaternary sediments from the northern Irish Sea. 27–54 in The Quaternary history of the Irish Sea. KIDSON, C, and TOOLEY, M J (editors). Special Issue of the Geological Journal, No. 7.

PANTIN, H M. 1978. Quaternary sediments from the north-east Irish Sea, Isle of Man to Cumbria. Bulletin of the Geological Survey of Great Britain, No. 64.

PANTIN, H M. and EVANS, C D R. 1984. The Quaternary history of the central and south-western Celtic Sea. Marine Geology, Vol. 57, 259–293.

PARKIN, M, and CROSBY, A. 1982. Geological results of boreholes drilled on the southern United Kingdom Continental Shelf by the Institute of Geological Sciences 1969–1982. Internal Report of the Marine Geology Unit, Institute of Geological Sciences, No. 82/2.

PENN, I E. 1987. Geophysical logs in the stratigraphy of Wales and adjacent areas. Proceedings of the Geologists' Association, Vol. 98, 275–314.

PENN, I E. and EVANS, C D R. 1976. The Middle Jurassic (mainly Bathonian) of Cardigan Bay and its palaeogeographical significance. Report of the Institute of Geological Sciences, No. 76/6.

PENN, I E. MERRIMAN, R J, and WYATT, R J. 1979. The Bathonian strata of the Bath-Frome area. Report of the Institute of Geological Sciences, No. 78/22.

PETRIE, S H, BROWN, J R, GRANGER, P J, and LOVELL, J P B. 1989. Mesozoic history of the Celtic Sea Basins. 433–444 in Extensional tectonics and stratigraphy of the North Atlantic margins. TANKARD, A J, and BALKWILL, H R (editors). Memoir of the Association of American Petroleum Geologists, No. 46.

PICKERING, K T, BASSETT, M G, and SIVETER, D J. 1988. Late Ordovician-early Silurian destruction of the Iapetus Ocean: Newfoundland, British Isles and Scandanavia - a discussion. Transactions of the Royal Society of Edinburgh, Vol. 79, 361–382.

PINGREE, R D, and GRIFFITHS, D K. 1979. Sand transport paths around the British Isles resulting from M₂ and M₄ tidal interactions. Journal of the United Kingdom Marine Biological Association, Vol. 58, 211–226.

POWELL, D W. 1956. Gravity and magnetic anomalies in North Wales. Quarterly Journal of the Geological Society of London, Vol. 111, 375–397.

RAMSBOTTOM, W H C. 1979. Rates of transgression and regression in the Carboniferous of NW Europe. Journal of the Geological Society of London, Vol. 136, 147–153.

RAMSBOTTOM, W H C. and six others. 1978. A correlation of Silesian rocks in the British Isles. Special Report of the Geological Society of London, No. 10.

RAWSON, P F, and seven others. 1978. A correlation of Cretaceous rocks in the British Isles. Special Report of the Geological Society of London, No. 9.

ROBINSON, K W, SHANNON, P M, and YOUNG, D G G. 1981. The Fastnet Basin: an integrated analysis. 444–454 in Petroleum geology of the continental shelf of North-West Europe. ILLING, L V, and HOBSON, G D (editors). (London: Heyden and Son.)

RUSHTON, A W A. 1974. The Cambrian of Wales and England. 43–122 in Cambrian of the British Isles, Norway and Spitzbergen (Lower Palaeozoic rocks of the World, Vol. 2). HOLLAND, C H (editor). (London: John Wiley and Son.)

SAGER, G, and SAMMLER, R. 1968. Atlas der gezeitenstrame für die Nordsee, den Kanal, und die Irische See. (Rostock: Seehydrographischer Dienst der Deutschen Demokratischen Republik.) [In German.]

SALGE, U, and WONG, H K. 1988a. Seismic stratigraphy and Quaternary sedimentation in the Skagerrak (north-eastern North Sea). Marine Geology, Vol. 81, 159–174.

SALGE, U, and WONG, H K. 1988b. The Skagerrak: a depo-environment for Recent sediments in the North Sea. 367–380 in Biogeochemistry and distribution of suspended matter in the North Sea and implications for fisheries biology. Mitt Geologisch-Paliiontologisches Institut und Museum, University of Hamburg, Heft 65.

SCHLANGER, S O, ARTHUR, M A, JENKYNS, H C, and SCHOLLE, P A. 1987. The Cenomanian-Turonian Oceanic Anoxic Event. I Stratigraphy and distribution of organic carbon-rich beds and the marine a^, ä¹³C excursion. 371–399 in Marine petroleum source rocks. BROOKS, J A, and FLEET, A J (editors). Special Publication of the Geological Society of London, No. 26.

SCOURSE, J D. 1991. Late Pleistocene stratigraphy and palaeobotany of the Isles of Scilly. Philosophical Transactions of the Royal Society of London, Vol. B334, 405–448.

SCOURSE, J D. and five others. 1992. New evidence on the age and significance of the Gortian temperate stage: A preliminary report on the Cork Harbour site. Proceedings of the Royal Irish Academy, Vol. 92B, 21–43.

SHACKLETON, R M. 1954. The structural evolution of North Wales. Liverpool and Manchester Geological Journal, Vol. 1, 261–296.

SHACKLETON, R M. 1956. The structure and succession of Anglesey and the Lleyn Peninsula. British Association for the Advancement of Science, Vol. 11, 106–108.

SHACKLETON, R M. 1975. Precambrian rocks of Wales. 76–82 in A correlation of the Precambrian rocks in the British Isles. HARRIS, A L, and five others (editors). Special Report of the Geological Society of London, No. 6.

SHANNON, P M. 1991. The development of Irish offshore sedimentary basins. Journal of the Geological Society of London, Vol. 148, 181–190.

SLADEN, C P, and BATTEN, D J. 1984. Source-area environments of Late Jurassic and early Cretaceous sediments in southeast England. Proceedings of the Geologists' Association, Vol. 95, 149–163.

SMITH, A G, and BRIDEN, J C. 1977. Mesozoic and Cenozoic palaeocontinental maps. (Cambridge: Cambridge University Press.)

SMITH, S A, and EDWARDS, R A. 1991. Regional sedimentological variations in Lower Triassic fluvial conglomerates (Budleigh Salterton Pebble Beds), southwest England: some implications for palaeogeography and basin evolution. Geological Journal Vol. 26, 65–83.

SOPER, N J. 1986. The New Granite problem: a geotectonic view. Geological Magazine, Vol. 123, 227–236.

STEPHENS, N. 1973. South-west England. 36–45 in A correlation of the Quaternary deposits in the British Isles. MITCHELL , G F, PENNY, L F, SHOTTON, F W, and WEST R G (editors). Special Report ofthe Geological Society of London, No. 4.

STRIDE, A H (editor). 1982. Offshore tidal sands-processes and deposits. (London: Chapman and Hall.)

STRIDE, A H and BEIDERSON, R H. 1990. A reassessment of sand transport paths in the Bristol Channel and their regional significance. Marine Geology, Vol. 92, 227–236.

SUMBLER, M G. 1984. The stratigraphy of the Bathonian White Limestone and Forest Marble formations of Oxfordshire. Proceedings of the Geologists' Association, Vol. 95, 51–64.

SUMMERHOFF, G VON. 1973. Formenschatz und morphologische Gliederung des sudostgronlandischen Shelfgebietes und Kontinentalabhanges. "Meteor" Forschungsergebnisse. Vol. C15, 1–54. [In German.]

SYNGE, F M. 1981. Quaternary glaciation and changes of sea level in the south of Ireland. Geologie en Mijnbouw, Vol. 60, 305–315.

SYNGE, F M. 1985. Coastal evolution. 115–131 in The Quaternary geology of Ireland EDWARDS, K J, and WARREN, W P (editors). (London: Academic Press.)

TAPPIN, D R, and DOWNIE, C. 1978. New Tremadoc strata at outcrop in the Bristol Channel. Journal of the Geological Society of London, Vol. 135, 321.

TAYLOR, R B, and MCCANN, S B. 1983. Coastal depositional landforms in northern Canada. 53–76 in Shorelines and isostasy. SMITH, D E, and DAWSON, A G (editors). Special Publication of the Institute of British Geographers, No. 16.

TAYLOR-SMITH, D. 1987. Geotechnical studies in Tremadog Bay. Proceedings of the Geologists'Association, Vol. 98, 385–396.

THORPE, R S, and five others. 1984. Crustal growth and late Precambrian-early Palaeozoic plate tectonic evolution of England and Wales. Journal of the Geological Society of London, Vol. 141, 521–536.

THORPE, R S, LEAF, P T, BEVINS, R E, and HUGHES, D J. 1989. Late orogenic alkaline/subalkaline Silurian volcanism of the Skomer Volcanic Group in the Caledonides of south Wales. Journal of the Geological Society of London, Vol. 146, 125–132.

TOOLEY, M J. 1985. Sea-level changes and coastal morphology in north-west England. 94–121 in The geomorphology of north-west England. JOHNSON, R H (editor). (Manchester: Manchester University Press.)

TRUEBLOOD, S, and MORTON, N. 1991. Comparative sequence stratigraphy and structural styles of the Slyne Trough and Hebrides Basin. Journal of the Geological Society of London, Vol. 148, 197–201.

TUCKER, M E. 1977. The marginal Triassic deposits of South Wales: continental facies and palaeogeography. Geological Journal, Vol. 12, 169–188.

TUCKER, R D, and PHARAOH, T C. 1991. U-Pb zircon ages for Late Precambrian igneous rocks in southern Britain. Journal of the Geological Society of London, Vol. 148, 435–444.

TUCKER, R M, and ARTER, G. 1987. The tectonic evolution of the North Celtic Sea and Cardigan Bay basins with special reference to tectonic inversion. Tectonophysics, Vol. 137, 291–307.

TUNBRIDGE, I. 1986. Mid-Devonian tectonics and sedimentation in the Bristol Channel area. Journal of the Geological Society of London, Vol. 143, 107–115.

TURNER, P. 1977. Notes on the depositional environment of the Lingula Flags in Dyfed, South Wales. Proceedings of the Yorkshire Geological Society, Vol. 41, 199–202.

TYSON, R V, and FUNNELL, B M. 1987. European Cretaceous shorelines, stage by stage. Palaeogeography, Pakeoclimatology, Palaeoecology, Vol. 59, 69–91.

VAIL, P R, and TODD, R G. 1981. Northern North Sea Jurassic unconformities, chronostratigraphy and sea-level changes from seismic stratigraphy. 216–235 in The geology of the continental shelf of North-West Europe. KLING, L V, and HOBSON, G D (editors). (London: Heyden and Son.)

VAN HOORN, B. 1987. The South Celtic Sea/Bristol Channel Basin: origin, deformation and inversion history. Tectonophysics, Vol. 137, 309–334.

VAN WIJHE, D H. 1987. Structural evolution of inverted basins in the Dutch offshore. Tectonophysics, Vol. 137, 171–219.

WALSH, P T. 1966. Cretaceous outliers in south-west Ireland and their implications for Cretaceous palaeogeography. Quarterly Journal of the Geological Society of London, Vol. 122, 63–84.

WARR, L N. 1989. The structural evolution of the Davidstow Anticline and its relationship to the Southern Culm Overfold, north Cornwall. Proceedings of the Ussher Society, Vol. 8, 67–72.

WARREN, W P. 1985. Stratigraphy. 39–65 in The Quaternary history of Ireland EDWARDS, K J, and WARREN, W P (editors). (London: Academic Press.)

WARREN, W P. and KEARY, R. 1988. The sand and gravel resources of the Irish Sea Basin. 66–79 in The Irish Sea: a resource at risk. SWEENEY, J C (editor). Special Publication of the Geographic Society of Ireland No. 3.

WARRINGTON, G. 1970. The stratigraphy and palaeontology of the 'Keuper Series' of the central Midlands of England. Quarterly Journal of the Geological Society of London, Vol. 126, 183–223.

WARRINGTON, G. 1971. Palynology of the Upper Triassic strata in the Llanbedr (Mochras Farm) borehole. 73–86 in The Llanbedr (Mochras Farm) borehole. WOODLAND, A W (editor). Report of the Institute of Geological Sciences, No. 71/18.

WARRINGTON, G. 1974. Les evaporites du Trias Britannique. Bulletin Societe Geologique de France, Series 7, Vol. 14, No. 6, 708–723.

WARRINGTON, G. and eight others. 1980. A correlation of Triassic rocks in the British Isles. Special Report of the Geological Society of London, No. 13.

WARRINGTON, G. and IVIMEY-COOK, H C. 1990. Biostratigraphy of the Late Triassic and Early Jurassic: a review of type sections in southern Britain. Cahiers Universite Catholique de Lyon, Series Scientifique, Vol. 3, 207–213.

WARRINGTON, G. and IVIMEY-COOK, H C. 1992. Triassic. 97–106 in Atlas of palaeogeography and lithofacies. COPE, J C W, INCHAM, J K, and RAWSON, P F (editors). Memoir of the Geological Society of London, No. 13.

WARRINGTON, G. and OWENS, B. 1977. Micropalaeontological biostratigraphy of offshore samples from southwest Britain. Report of the Institute of Geological Sciences, No. 77/7.

WARRINGTON, G. and SCRIVENER, R C. 1988. Late Permian fossils from Devon: regional geological implications. Proceedings of the Ussher Society, Vol. 7, 95–96.

WARRINGTON, G. and SCRIVENER, R C.1990. The Permian of Devon, England. Review of Palaeobotany and Palynology, Vol. 66, 263–272.

WATERS, R A, and LAWRENCE, D J D. 1987. Geology of the South Wales Coalfield, Part 3, the country around Cardiff (3rd edition). Memoir of the British Geological Survey, Sheet 263 (England and Wales).

WEBBY, B D. 1966. Middle-upper Devonian palaeogeography of north Devon and west Somerset, England. Palaeogeography, Palaeoclimatology, Palaeoecology, Vol. 2, 27–46.

WEIGHELL, A J. 1980. The Upper Cretaceous of the west British continental shelf. Unpublished PhD thesis, University of Wales.

WEIGHELL, A J. DOBSON, M R, WHITTINGTON, R J, and DELANTEY, L J. 1981. Upper Cretaceous geology of the Celtic Sea. 727–741 in North Atlantic Borderlands. Canadian Society of Petroleum Geologists Memoir, No. 7.

WERNICKE, B. 1985. Uniform-sense normal simple shear of the continental lithosphere. Canadian Journal of Earth Sciences, Vol. 22, 108–125.

WEST, R G. 1977. Pleistocene geology and biology. (London: Longmans.)

WHITE, W A. 1988. More on deep glacial erosion by continental ice sheets and their tongues of distributary ice. Quaternary Research, Vol. 30, 137–150.

WHITTAKER, A. 1973. The Central Somerset Basin. Proceedings of the Ussher Society, Vol. 2, 585–592.

WHITTAKER, A. (editor). 1985. Atlas of onshore sedimentary basins in England and Wales: post-Carboniferous tectonics and stratigraphy. (Glasgow: Blackie.)

WHITTAKER, A. and GREEN, G W. 1983. Geology of the country around Weston-super-Mare. Memoir of the British Geological Survey, Sheet 269, with parts of 263 and 295 (England and Wales).

WHITTAKER, A. HOLLIDAY, D W, and PENN, I E. 1985. Geophysical logs in British stratigraphy. Special Report of the Geological Society of London, No. 18.

WHITTINGTON, R J. 1980. Geophysical studies in the South Irish Sea. Unpublished PhD thesis, University of Wales.

WILKINSON, G C, BAZLEY, R A B, and BOULTER, M C. 1980. The geology and palynology of the Oligocene Lough Neagh Clays, Northern Ireland. Journal of the Geological Society of London, Vol. 137, 65–75.

WILKINSON, I P, and HALLIWELL , G P (compilers). 1980. Offshore micropalaeontological biostratigraphy of southern and western Britain. Report of the Institute of Geological Sciences, No. 79/9.

WILLIAMS, B P J. 1980. The Devonian (Old Red Sandstone) rocks of the Variscan Foreland. 57–63 in United Kingdom: introduction to general geology and guides to excursions 002, 055, 093, 151. OWEN, T R (editor). Proceedings of the 26th International Congress, Paris, 1980.

WILLIAMSON, M A. 1979. Cretaceous foraminifera from the Celtic Sea (BP borehole 93/2–1). Unpublished MSc thesis, University of Wales.

WILLS, L J. 1956. Concealed coalfields. (London and Glasgow: Blackie.)

WILLS, L J. 1976. The Trias of Worcestershire and Warwickshire. Report of the Institute of Geological Sciences, No. 76/2.

WILSON, D, DAVIES, J R, FLETCHER, C J N, and SMITH, M. 1990. Geology of the South Wales Coalfield, part VI, the country around Bridgend. Memoir of the British Geological Survey, sheets 261 and 262 (England and Wales).

WILSON, D, DAVIES, J R, SMITH, M, and WATERS, R A. 1988. Structural controls on Upper Palaeozoic sedimentation in south-east Wales. Journal of the Geological Society of London, Vol. 145, 901–914.

WINGFIELD, R T R. 1987. Giant sand waves and relict periglacial features on the sea bed west of Anglesey. Proceedings of the Geologists' Association, Vol. 98, 400–404.

WINGFIELD, R T R. 1989. Glacial incisions indicating Middle and Upper Pleistocene ice limits off Britain. Terra Nova, Vol. 1, 538–548.

WINGFIELD, R T R. 1990. The origin of major incisions within the Pleistocene deposits of the North Sea. Marine Geology, Vol. 91, 31–52.

WOOD, A, and SMITH, A J. 1959. The sedimentation and sedimentary history of the Aberystwyth Grits (Upper Llandoverian). Quarterly Journal of the Geological Society of London, Vol. 114, 163–195.

WOOD, A, and WOODLAND, A W. 1968. Borehole at Mochras, west of Llanbedr, Meirionethsh ire. Nature, London, Vol. 219, 1352–1354.

WOOD, C J, BIGG, P J, and MEDD, A W. 1982. The biostratigraphy of the Upper Cretaceous (Chalk) of the Winterborne Kingston borehole, Dorset. 19–27 in The Winterborne Kingston borehole, Dorset, England. RHYS, G H, LOTT, G K, and CALVERT, M A (editors). Report of the Institute of Geological Sciences, No. 81/3.

WOODCOCK, N H. 1990. Sequence stratigraphy of the Palaeozoic Welsh Basin. Journal of the Geological Society of London, Vol. 147, 537–547.

WOODLAND, A W (editor). 1971. The Llanbedr (Mochras Farm) borehole. Report of the Institute of Geological Sciences, No. 71/18.

WRIGHT, A E. 1969. Precambrian rocks of England, Wales and south-east Ireland. 93–109 in North Atlantic geology and continental drift: a symposium. KAY, M (editor). Memoir of the American Association of Petroleum Geologists, Vol. 12.

ZIEGLER, P A. 1978. North-western Europe: tectonics and basin development. Geologie en Mijnbouw, Vol. 57, 589–626.

ZIEGLER, P A. 1981. Evolution of sedimentary basins in North-West Europe. 3–39 in Petroleum geology of the continental shelf of North-West Europe. ILLING, L V, and HOBSON, G D (editors). (London: Heyden and Son.)

ZIEGLER, P A. 1982. Geological atlas of western and central Europe. (Amsterdam: Elsevier for Shell Internationale Petroleum Maatschappij BV.)

ZIEGLER, P A. 1986. Geodynamic model for the Palaeozoic crustal consolidation of western and central Europe. Tectonophysics, Vol. 126, 303–328.

ZIEGLER, P A. 1987. Late Cretaceous and Cenozoic intra-plate compressional deformations in the Alpine foreland — a geodynamic model. Tectonophysics, Vol. 137, 389–420.

ZIEGLER, P A. 1990. Tectonic and palaeogeographic development of the North Sea rift system. 1–36 in Tectonic evolution of North Sea rifts. BLUNDELL, D J, and GIBBS, A D (editors). (Oxford: Clarendon Press.)

The diagrams in this report were produced by the Cartographic Production Group of BGS, Keyworth, under the supervision of R J Parnaby. Cartography was co-ordinated by S Bray and R J Demaine, reprography by J E Kmieciak and S C Wilkinson.

Book production was supervised by M B Simmons; typesetting was by A R Hutchinson and page make-up by J Norman.

Figures

(Front cover) Ynys Enlli (Bardsey Island), viewed from the tip of Llŷn (Figure 1). White quartzite clasts in the Precambrian Gwna Melange can be seen on the hillside. Photo: BGS collection.

(Rear cover) Index map United Kingdom Offshore Regional Report

(Figure 1) The location of the report area, showing simplified bathymetry, licence blocks, released wells, BGS shallow boreholes, and the locations of BIRPS profiles.

(Figure 2) Pre-Pleistocene geology of the report area and adjacent districts.

(Figure 22)). For location see (Figure 1)." data-name="images/P945202.jpg">(Figure 3) Part of line SWAT-2 across the St George's Channel Basin, with an interpretation showing pre-Permian reflectors and little or no Permian sediment. An alternative interpretation is possible in view of the lack of stratigraphical control (see (Figure 22)). For location see (Figure 1).

(Figure 4) Simplified logs of selected wells from the report area.

(Figure 5) Generalised palaeogeographic sketches of the Early Palaeozoic. Based on Cope et al. (1992).

(Figure 6) Generalised Devonian palaeogeography. Modified after Freshney and Taylor (1980).

(Figure 7) Schematic relationships of the continental Old Red Sandstone rocks of Wales with the marine Devonian rocks of south-west England. After Allen (1974).

(Figure 8) Geological map of part of south-west Wales and the adjacent offshore area.

(Figure 9) Geological map of part of north-west Devon and the adjacent offshore area.

(Figure 10) Generalised Dinantian palaeogeography. Modified after Freshney and Taylor (1980).

(Figure 11) Seismic profile across the Central Irish Sea Basin. For location see (Figure 1).

(Figure 12) Generalised palaeogeographic sketches of the Silesian. Modified after Freshney and Taylor (1980).

(Figure 13) Principal structural elements of the report area, with structure contours on the base of the sedimentary basins. The sedimentary cover comprises Permian and younger rocks, and includes Carboniferous strata in the St George's Channel and Cardigan Bay basins. Based upon BGS and BIRPS seismic-reflection data, and published information.

(Figure 14) Bouguer gravity anomaly map of the report area and adjacent areas.

(Figure 15) Palaeozoic structural provinces. BASIN STRUCTURE

(Figure 18), Section 3.) See (Figure 18) for location." data-name="images/P945215.jpg">(Figure 16) BIRPS deep-seismic reflection profile SWAT-4 across the South Celtic Sea Basin. Note the gentle regional upwarp of the post-extensional sequence due to mid-Tertiary basin inversion (see also (Figure 18), Section 3.) See (Figure 18) for location.

(Figure 17) Plate-tectonic reconstructions of the North Atlantic region. Modified after Chadwick et al. (1990). a. End-Carboniferous (286 Ma), prior to the onset of continental lithospheric extension. b. Mid-Cretaceous (105 Ma), immediately prior to the onset of North Atlantic sea-floor spreading. Blue area indicates approximate extent of 'Central Atlantic' sea-floor spreading.

(Figure 18) Geological cross-sections at both true and expanded (X2) vertical scales. BGS and BIRPS seismic data, and published information.

(Figure 19) Thickness of preserved Permo-Triassic sedimentary rocks (with Carboniferous beds in the St George's Channel and Cardigan Bay basins). Based upon BGS and BIRPS seismic-reflection data, and published information.

(Figure 20) Thickness of preserved Jurassic and Lower Cretaceous (Ryazanian to Barremian) sedimentary rocks. Based upon BGS and BIRPS seismic-reflection data, and published information. For explanation of structural abbreviations, see (Figure 19).

(Figure 21) Thickness of preserved middle to Upper Cretaceous (Aptian to Maastrichtian) and Tertiary sedimentary rocks. Based upon BGS and BIRPS seismic-reflection data, and published information. For explanation of structural abbreviations, see (Figure 19).

(Figure 22) BIRPS seismic-reflection profile SWAT-2 through the St George's Channel Basin. Arrowed, convergent strata in the Permo-Triassic sequence may be indicative of salt migration from the basin depocentre into the salt wall. For location see (Figure 18).

(Figure 23) a) Line drawing of BIRPS deep-seismic lines SWAT-2 and SWAT-3. b) True-scale crustal cross-section based upon a depth conversion of these lines. The section coincides with that of (Figure 18), Section 2, but prolonged slightly to the north and south.

(Figure 24) The generalised distribution of known and inferred strata of Permian and Triassic age, either at outcrop or at depth. Faults not shown.

(Figure 25) Wells depicting Permian lithologies, with geophysical logs. For locations see (Figure 24)a.

(Figure 26) Cross-section joining wells in the South Celtic Sea and Central Somerset basins. Well locations arc shown on (Figure 24).

(Figure 27) Wells depicting Sherwood Sandstone Group lithostratigraphy, with geophysical logs. For locations see (Figure 24)b.

(Figure 28) Wells depicting Mercia Mudstone Group lirhostratigraphy, with geophysical logs. Lithostratigraphical units are based on the Burton Row borehole (after Whittaker et al., 1985), and geophysical units after Lott et al. (1982). Locations are shown on (Figure 24)c. Note the change of scale between wells 103/2-1 and 93/6-1.

(Figure 29) Penarth Group lithostratigraphy, and geophysical-log correlation. For locations of wells see (Figure 24)d.

(Figure 30) Surface distribution of Jurassic strata in and around the report area.

(Figure 31) Generalised Mid-Jurassic palaeogeography. After Penn and Evans (1976). For greater detail see Bradshaw et al. (1992)

(Figure 32) Jurassic stratigraphical nomenclature for the report area.

(Figure 33) Log of the Lower Jurassic succession in the Mochras borehole. Based on Ivimey-Cook (1971). For location see (Figure 30).

(Figure 37)." data-name="images/P945233.jpg">(Figure 34) Logs of wells from the St George's Channel Basin. For locations see (Figure 37).

(Figure 35) Logs of wells from the South Celtic Sea and Bristol Channel basins. For locations see (Figure 30).

(Figure 36) Jurassic strata recovered in BGS boreholes located in the report area. After Parkin and Crosby (1982). For locations see (Figure 30).

(Figure 37) Generalised depth to the base of Bathonian strata in the: St George's Channel and Cardigan Bay basins.

(Figure 38) Cumulative thickness of the Kimmeridgian in well 106/24-1 and parts of England. After Penn (1987).

(Figure 39) Generalised Cretaceous lithologies in the North and South Celtic Sea basins and Bristol Channel Basin, showing approximate maximum sequence thicknesses. The Cretaceous global sea-level curve is after Haq et al. (1988).

(Figure 40) North–south section across the southern part of the North Celtic Sea Basin in Irish waters. After Colin et al. (1981).

(Figure 41) Lower Cretaceous successions in wells in the South Celtic Sea and Bristol Channel basins.

(Figure 42) Sea-bed outcrop of Wealden strata in the Bristol Channel Basin.

(Figure 43) Lithologies of Wealden facies in BGS boreholes in the eastern Bristol Channel Basin. For locations see (Figure 42).

(Figure 45)." data-name="images/P945243.jpg">(Figure 44) The Aptian to Turonian succession in well 49/16-A5, at the Kinsale Head Gasfield. Adapted after Colley et al. (198 1 ). For location see (Figure 45).

(Figure 45) Distribution and thickness of the Upper Cretaceous in the North Celtic Sea Basin. From Tucker and Arter (1987).

(Figure 46) Upper Cretaceous successions in wells in the South Celtic Sea Basin and southern England. For offshore locations see (Figure 45).

(Figure 47) Depth below sea level to the base of the Chalk.

(Figure 48) Distribution of Tertiary strata, and the depth to their base.

(Figure 49) The distribution of Tertiary basins along, or adjacent to, the Sticklepath–Lustleigh fault zone in Devon and the Bristol Channel.

(Figure 50) Log of Tertiary sediments in the Mochras borehole. For location see (Figure 48).

(Figure 51) Seismic-reflection profile, with interpretation, across part of the Cardigan Bay Basin. The interpreted profile is extended to the Mochras Fault. For location see (Figure 48).

(Figure 52) Sedimentary cycles in the Mochras borehole. Modified after Dobson and Whittington (1987).

(Figure 53) Depositional model for the Tertiary in the Cardigan Bay Basin. The alluvial fans were active only during the early part of sediment accumulation. The meander belt is restricted to the area between the dashed lines, and includes an active levee zone breached by crevasse splays which carried sediment on to the floodplain. There may also have been small-scale overbank invasion. The overall depositional environment is interpreted as a high-sinuosity river with tributaries and a swamp-dominated floodplain. From Dobson and Whittington (1987).

(Figure 54) Seismic-reflection line SWAT-2 across the St George's Channel Basin, showing speculative seismostratigraphical relationships of the Tertiary across the St George's Fault. For location see (Figure 48).

(Figure 55) Tertiary successions in wells from the St George's Channel Basin. For locations see (Figure 48).

(Figure 56) BGS borehole BH89/10 from the South Celtic Sea Basin. For location see (Figure 48).

(Figure 57) Distribution of Tertiary rocks in and around the Stanley Bank Basin, with logs of BGS horeholes drilled into the basin.

(Figure 58) The thickness of Quaternary sediments in the report area and adjacent areas.

(Figure 59) Simplified bathymetry of the report area and surrounding areas.

(Figure 60) The distribution of major incisions in the report area. Only those incisions whose shoulder-to-base depths are in excess of approximately 100 m are shown.

(Figure 61) Quaternary stratigraphy, and a schematic cross-section through part of the report area. Correlation of stages after West (1977); absolute ages after Bowen et al. (1986).

(Figure 62) Cross-sections through the Quaternary deposits of the report area. For explanation of abbreviations, see (Figure 61).

(Figure 63) Summary logs of the Quaternary successions in selected BGS boreholes. For explanation of abbreviations, see (Figure 61).

(Figure 64) Thickness of the St George's Channel Formation.

(Figure 65) Cartoons illustrating Weichselian and Holocene sea-level changes.

(Figure 66) The distribution of sea-bed sediments in the report area and adjacent areas. Based on BGS maps.

(Figure 67) The distribution of both active and relict bedforms.

(Figure 68) Water movements and sediment transport. a.Maximum surface tidal streams, after Sager and Sammler (1968). b. 50-year maximum wave heights, after Draper (1973). c. Bottom stress, after Pingree and Griffiths (1979). d. Net sand transport, after Stride (1982).

(Figure 69) Sand and gravel dredging locations in the Bristol Channel, together with proposed locations for the Severn Barrage.