Marine Geology 232 (2006) 1 – 15 www.elsevier.com/locate/margeo Sea-level change and inner shelf stratigraphy off Northern Ireland Joseph T. Kelley a,⁎, J. Andrew G. Cooper b , Derek W.T. Jackson b , Daniel F. Belknap a , Rory J. Quinn b a Department of Earth Sciences, University of Maine, Orono, ME 04469-5790, USA b School of Environmental Science, University of Ulster, Coleraine, BT52 1SA, Northern Ireland, UK Received 15 January 2006; received in revised form 22 March 2006; accepted 2 April 2006Abstract New seismic stratigraphic, vibracore and AMS 14C dates from two sites off the Northern Ireland coast yield information on thedeglacial to present sea-level history and shelf evolution of the region. A lowstand of sea level at about 30 m below present sealevel recorded by fossils in a lowstand shoreline deposit occurred around 13.4 cal ka B.P. following a period of rapid isostatic upliftassociated with a RSL fall of 6–7 cm/yr. Following the lowstand, contrasting styles of sedimentation characterized the two study sites. In the sheltered environment ofBelfast Lough, the lowstand shoreline was overtopped and buried by transgressive facies of intertidal and shallow sub-tidal mudand sandy mud. On the high-energy Portrush coast, the inner shelf sedimentary sequence is characterized by a basal conglomerateoverlain by well-sorted sands with occasional interbedded gravel.© 2006 Elsevier B.V. All rights reserved.Keywords: sea-level change; late Quaternary; Northern Ireland; seismic reflection; vibracores; isostasy1. Introduction resolution. With the relatively small model grid cells employed, extensive areas of formerly glaciated terrain Observations on changes in the sea level (Fairbanks, are represented by a single sea-level curve. This1989; Chappell and Polach, 1991; Barnhardt et al., problem becomes especially acute along continental1997) have led to sophisticated numerical models which margins where models are often unable to cope withestimate changes in sea level due to isostasy and eustasy local (short wavelength) variations in lithosphericon a global scale (Lambeck and Purcell, 2001; Peltier, thickness as well. Reliance on uncertain ice thickness2002; Clark and Mix, 2002). These models serve as a estimates and the timing of deglaciation can furtherproxy for observational data in most locations because weaken model accuracy. This proves frustrating to thoseof the scarcity of direct observations. This is especially gathering the field data that are used in models; whiletrue with respect to the position of local relative sea field observations were the fundamental data whichlevels at lower-than-present locations. A shortcoming of spawned models, the models often do not feed backthe global models, however, lies in their low spatial adequately to assist in solving regional geological and stratigraphic problems (McCabe, 1997). ⁎ Corresponding author. Tel.: +1 207 581 2162; fax: +1 207 581 In Ireland, many archeological and natural history2202. questions revolve around whether changing sea levels E-mail address: jtkelley@maine.edu (J.T. Kelley). permitted a land bridge to Great Britain (Mitchell and0025-3227/$ - see front matter © 2006 Elsevier B.V. All rights reserved.doi:10.1016/j.margeo.2006.04.0022 J.T. Kelley et al. / Marine Geology 232 (2006) 1–15Ryan, 2001; Devoy, 1995). Similarly, research on advancing from Scotland, and later ice confined to localglacial stratigraphy (Clark et al., 2004) and coastal ice caps in mountains (McCabe, 1987; McCabe et al.,evolution (Shaw and Carter, 1994) relies on an 2005). Glacial sediments crop out locally along theunderstanding of late Quaternary, relative sea-level coast, and a large moraine is believed to underliechanges. Models presently provide a crude framework Causeway Bank (Fig. 1).for understanding the sea-level history of this area Glacial-marine muddy sediment accumulated during(Lambeck and Purcell, 2001; Lambeck and Chappell, deglaciation (McCabe et al., 1994). It is widespread2001; Shennan et al., 2002a,b), but the timing and extent offshore (Cooper et al., 2002), and crops out along theof glaciation/deglaciation, the continental margin loca- coast and inland to about 20 m elevation (McCabe et al.,tion and the extensive perimeter of the island setting 1994; McCabe et al., 2005). It has not been dated where itpresent a challenge to accurate models (McCabe, 1997; is best exposed at Portballintrae (Fig. 1), but is a time-McCabe et al., 2005). transgressive deposit which, through its contained In 1998 we gathered some of the first seismic foraminiferal fauna, has provided high resolution tempo-reflection data from offshore of Northern Ireland, and ral constraints on the deglacial history of the regionspeculated on a sea-level lowstand position there at 30 m (McCabe and Clark, 2003; McCabe et al., 2005). Atwater depth (Cooper et al., 2002). Here we present the Portballintrae (Fig. 1) sand and gravel rest unconformablyinterpretation of additional seismic reflection data and on glacial-marine mud (McCabe et al., 1994). This offlapunderwater vibracores from east of this previous work, deposit is up to 3 m thick and dips offshore along the coastand data from new underwater vibracores. These (photographs in Cooper et al., 2002, their Fig. 3).vibracores provide ground truth to our remotely sensed The seafloor near Portrush is generally sandyobservations of Northern Ireland's inner shelf, and also (Lawlor, 2000). The sand is well sorted and relativelyrepresent the first published descriptions of the late fine, with a mean size between 1–3 phi. Bedrock cropsQuaternary stratigraphy of that region's inner shelf. In out locally near the cliffed coast and near the Skerries.addition, a series of new radiocarbon dates of vibracore Adjacent to rock outcrops and in constrictions formedsubsamples provides a chronological context for sea- by it, gravel dominates the seafloor. Mud occurslevel change. throughout much of Belfast Lough, although rock crops out in several places (Parker, 1982).2. Location and geological setting The seafloor dips gently seaward from the sandy beaches until about the 30 m isobath. Off high cliffs it The study sites are located in Northern Ireland, drops off more rapidly and the seafloor is highlyUnited Kingdom, between the 20 m and 40 m isobaths irregular (Cooper et al., 1998; Lawlor, 2000). Thein outer Belfast Lough, and between 15 m and 30 m Causeway Bank, off Giants Causeway, forms a linearisobaths offshore of Portrush, County Antrim, some bathymetric shoal parallel to the coast with 5–10 m of90 km northwest of Belfast (Fig. 1). The bedrock in this bathymetric relief (Fig. 1b). The muddy bottom ofregion is dominated by extensive flows of Tertiary basalt Belfast Lough very gently deepens from the inner to thethat form 80 m high cliffs between the two study sites outer Lough (Fig. 1c).(Wilson, 1972). The basalt overlies Cretaceous chalk The coast is wave dominated, with a significant wavethat also crops out on the western side of Belfast Lough height of around 2 m on the north coast, falling to 1.3 mand along the north coast. A dolerite sill forms a chain of at Belfast (Jackson et al., 2005). Wind is generally fromsmall islands (The Skerries) that partly shelter the the southwest.Portrush study area. A large beach (Curran Strand)exists in the lee of The Skerries, and smaller, pocket 3. Previous workbeaches occur in several other locations (Fig. 1b).Belfast Lough forms a 20 km-long, 10 km-wide Recent analysis of late glacial microfossils (McCabeindentation in the Northern Ireland coast that is the et al., 2005) has demonstrated major millennial-scalegeomorphic expression of a bedrock fault subsequently events during the deglacial period in which isostatic andmodified by ice erosion (Manning et al., 1970). Bedrock eustatic effects have combined to produce a complex,along the Belfast Lough shoreline includes material late Quaternary regional sea-level history. This field dataranging from Paleozoic to Tertiary age (Geological is at odds with numerical models of sea-level variabilitySurvey of Northern Ireland, 1997). (Lambeck and Purcell, 2001; McCabe et al., 2005) and Glacial ice covered this area in the late Quaternary. points to more complex cryo–lithosphere interactionsThe late glacial history is complex, with early ice than currently assumed. J.T. Kelley et al. / Marine Geology 232 (2006) 1–15 3Fig. 1. Map of the study areas. a) Map showing the study areas in Northern Ireland. Boxes are enlarged in b and c. b) Bathymetric map of NorthernIrish coast from Portrush to Giants Causeway; c) bathymetric map of Belfast Lough. The numbered lines are seismic tracks shown in later figures. Onb and c the numbered lines are seismic tracks shown in later figures, cores are labeled and located with an x. Bathymetric contours are labeled; thevertically rules area is intertidal. Carter (1982) summarized early work on Holocene lower, intermediate and upper units. The lower unitsea levels in the area. He recognized a higher-than- contains an intertidal fauna and remains of the intertidalpresent sea-level position in post-glacial times at +5 m seagrass, Zostera marina. It varies in thickness fromapproximately 5 ka BP, and speculated on a Holocene 15 cm to 4.3 m. The intermediate unit is characterizedlowstand around 15 m below present sea level at by abundant Pholas sp. and indicates increasing waterapproximately 10 ka BP. He noted, however, that there depth. The upper unit contains abundant molluscs ofwas no field data to support either the depth or time of diverse types, and is interpreted as deepwater estuarinethe lowstand. (Fig. 2). conditions with a water depth of 9 to 10 m (Manning et The transgressive stratigraphy in upper Belfast al., 1970). The basal peat was dated at 10.3 ka cal BPLough was recorded in many foundation investigations (9130 ± 120 BP, radiocarbon years) at − 12 m while(Manning et al., 1970) and comprises a freshwater peat wood from − 9 m dated at 9.8 ka cal BP (8715 ± 200 BP,up to 1 m thick at depths down to 15 m below present radiocarbon years) (Fig. 2) (Manning et al., 1970). Thesea level. This is overlain by a sequence of clays peat dates, while not sea-level indicators, constrain seasubdivided on the basis of their molluscan fauna into level below these depths at specific times.4 J.T. Kelley et al. / Marine Geology 232 (2006) 1–15 Lough (Manning et al., 1970). More recently, McDo- well et al. (2005) interpreted seismic reflectors 10 km west of Portrush as possible peat layers. A broadly similar deglaciation and sea-level history was developed for Scotland (Shennan et al., 2002a,b, 2005; Lloyd et al., 1999). The land-based record from locations near Northern Ireland appear well constrained, but suggest no excursion of sea level much below present in the late Pleistocene/early Holocene (Shennan and Horton, 2002). Many of the lowstand curves from here have no data from the time of the sea-level lowstand, however (Shennan and Horton, 2002). 4. Methods In June 1997 we gathered seismic reflection observations with a 2–16 kHz swept frequency Edge- tech X-Star chirp towfish. We gathered 150 km of tracklines from the Portrush study area and 70 km ofFig. 2. Sea-level reconstruction for Northern Ireland. Conventional and tracklines from the Belfast Lough site. Navigation wascalibrated (Calib, ver. 5.0) radiocarbon dates are shown for earlier logged from a DGPS receiver and was accurate tofreshwater peat dates in Belfast Lough (Manning et al., 1970) and theRiver Bann (Wilson and McKenna, 1996) discussed in text, for ± 10 m. Rossfelder-type vibracores were collected fromhighstand date on Rough Island (McCabe and Clark, 2003), and from the R/V Lough Foyle in June 2004 on the basis of thethis work. Dashed curve represents modeled sea-level record in seismic reflection profiles. The corer was a Rossfelderconventional radiocarbon years for area east of Belfast Lough by P-3 model with a steel barrel and 7.5 cm diameter plasticLambeck and Purcell (2001). liner. It was outfitted to collect up to 6 m long cores. After return of the corer to the deck, cores were cut Cooper et al. (1998, 2002) gathered the first into 1.5 m lengths if long enough. They were cappedextensive seismic reflection and side scan sonar records and stored upright on the deck and returned to thefrom off Northern Ireland. They recognized Quaternary sedimentology laboratory at the University of Ulsterstratigraphic units in the offshore directly seaward of within a week. In the lab, cores were opened,Portballintrae, where McCabe et al. (1994) had photographed and described. Samples were removeddescribed them on land. The section they interpreted for grain size analyses and all shells in one half of theincluded bedrock as basement, with patches of scattered core were removed for possible radiocarbon dating.till resting on it. Glacial-marine mud was common and Shells from the cores were identified at the Ulstercropped out commonly in water depths greater than Museum (Dr Julia Nunn personal communication).30 m. Surficial sand rested unconformably on the Samples selected for radiocarbon dating were stored inglacial-marine mud. Out to 32 m water depth, the sand aluminum foil until shipment. Radiocarbon dates areranged up to 10 m in thickness and extended as a sheet- reported as calibrated ages with a reservoir correction oflike deposit. Seaward of 30 m isobath, sand was worked 400 years (Stuiver and Reimer, 1993, Calib. versioninto sandwaves with several m of relief. Glacial-marine 5.0.2).mud appeared to crop out in swales between the sandwaves. 5. Results Cooper et al. (2002) suggested that the lowstand ofsea level in the Portrush study area was near 32 m below 5.1. Seaward of the Causeway Coastpresent sea level based on: 1) erosional notches cut in tilland bedrock (their Fig. 7), and 2) the termination of Seismic reflection observations landward of theofflap sand units observed on land (their Fig. 7). They Skerries were all similar. We interpret glacial-marinespeculated that the reflectors within the surficial sand sediment (Gm) to represent acoustic basement in mostunit might be peat deposits similar to peat that was areas (Figs. 3 and 4). Its surface is marked by a strong,recorded <10 km to the west in the River Bann estuary continuous reflector to about the 30 m isobath. Seawardby Wilson and McKenna (1996), as well as in Belfast of this depth, till and bedrock crop out near The Skerries J.T. Kelley et al. / Marine Geology 232 (2006) 1–15 5Fig. 3. Interpreted seismic reflection profile showing core targets off Runkerry. Locations of profiles and cores are shown on Fig. 1b. Gm is glacialmarine, Hs is Holocene sand.and Causeway Bank, but glacial-marine material is observations and surficial grab samples, we know thisobserved further offshore beneath large sandy bedforms acoustic unit is principally sand (Hs) (Cooper et al.,(Cooper et al., 2002). Unconformably overlying the 1998; Lawlor, 2000). This surficial sand unit is less thanglacial-marine mud is an acoustically transparent unit 5 m thick, and dips seaward. It abruptly thins andwith prominent internal reflectors. From side scan sonar terminates at the 30 m isobath in most locations (Cooper Fig. 4. Core logs for Skerries area. Lines show location of core photos shown in Fig. 5; core locations shown in Fig. 1b.6 J.T. Kelley et al. / Marine Geology 232 (2006) 1–15et al., 2002) (Fig. 3). The strong reflectors within the with the core lengths indicates that the cores should havesurficial sand unit generally dip seaward and terminate reached or penetrated the strong acoustic reflectors inon the glacial-marine contact at water depths between 15 the sand unit. Off Runkerry Strand, one core, SK9b,and 20 m. In some places these reflectors are closely penetrated 10 cm of fine sand and met refusal in coarsespaced and subparallel to the seafloor, as off the River basalt, chalk and flint gravel (Fig. 5b). At this locationBann to the west (McDowell et al., 2005); in other areas the acoustic reflector within the sand was believed to bethey take on a channel-like appearance (Cooper et al., nearly at the seabed.2002). They approach the surface in most locations andappear to crop out on or very near the seafloor off 5.2. Belfast LoughRunkerry Strand (Fig. 3). Three vibracores were gathered landward of the Seismic records from Belfast Lough contrast stronglySkerries, and three additional cores from off Runkerry with those from the Portrush area (Figs. 6 and 7).Strand (Figs. 1, 3, 4). Two of the latter cores, SK9a, Acoustic basement, interpreted as bedrock, is repre-SK9b, were from less than 100 m apart. Most cores were sented by a strong, continuous reflector with up to 15 mless than 1 m in length and all were dominated by fine, of sub-bottom relief over less than 0.5 km (Fig. 6). Awell-sorted sand. Rare shell fragments were preserved in non-stratified unit occasionally appears over bedrock.the cores, but no sedimentary structures were recorded, Its lower contact with bedrock is difficult to consistentlysave one burrow trace. The two longest cores, SK9a and detect, and it is interpreted as till (T). The thickestSK4, encountered some gravel and shell fragments at acoustic unit imaged, more than 10 m thick in places,depth (Figs. 4 and 5). Comparison of the seismic line typically rests on bedrock, though it was occasionallyFig. 5. a) Photograph of Core SK4. Uniform fine sand dominated all cores except for this area near the core bottom where several shell fragments anda 2.5 cm diameter basalt fragment were noted. b) Photograph of Core SK9b. The core was collected where a strong acoustic reflector within thesurficial sand deposit appeared to crop out at the seafloor (Fig. 3). The longest dimension of one cobble exceeded the diameter of the core barrel.Locations of photos within cores are shown in Fig. 4. J.T. Kelley et al. / Marine Geology 232 (2006) 1–15 7Fig. 6. Interpreted seismic reflection profile showing core locations in the shallow water of Belfast Lough. Inset box is enlargement of line near basin.Locations of profiles and cores are shown on Fig. 1c. Gm is glacial marine, T is till, Hm is Holocene mud.recorded on till (Figs. 6 and 7). It is acoustically are draped over the underlying units and mimic theirtransparent and contains strong, coherent reflectors that relief. We interpret this acoustic unit as glacial-marineare continuous over several kilometers. These reflectors sediment (Gm) based on its similarity to other publishedFig. 7. Interpreted seismic reflection profile showing core locations in the deep water of Belfast Lough. Cores were located on a break-in-slope inglacial-marine sediment that we interpret as the lowstand shoreline. Inset shows detail of the disappearance of the reflector that represents anunconformity on the surface of the glacial-marine sediment. Locations of profiles and cores are shown in Fig. 1c. Gm is glacial marine, T is till, Hm isHolocene mud.8 J.T. Kelley et al. / Marine Geology 232 (2006) 1–15 Fig. 8. Core descriptions of BL1, BL2, from inner Belfast Lough. Lines show locations of core photos in Fig. 9. Cores are located in Fig. 7.descriptions (Barnhardt et al., 1997; McDowell et al., sediment dips very gently seaward, and bears no2005) and occurrence in outcrop on nearby land relationship to underlying reflectors. Internal reflectors(McCabe et al., 1994). The surface of the glacial-marine within the glacial-marine sediment are truncated by itsFig. 9. a) Photograph of typical contact between the sand and mud beds. Location of photo shown in Fig. 8. b) Photograph of abundance and variety ofintact shells in typical core interval. Location of photo shown in Fig. 8. J.T. Kelley et al. / Marine Geology 232 (2006) 1–15 9Table 1 Table 1 (continued)List of shells identified in cores Core BLVC04-03Core BLVC04-01 110–130 cm 0–15 cm Circumphalus casina (Venus) Capulus ungerius Parvicardium sp. Turitella communis 170–180 cm 15–35 cm (numerous fragments) Pectinidae family Balanus sp. ( 1 intact shell radiocarbon dated) Turitella communis Nucula sp. (3 intact shells radiocarbon dated) Dosinia exoleta Astarte sp. Anomiidae family Mactridae family (1 large fragment radiocarbon dated) Pecten maximus Core BLVC04-04Core BLVC04-02 45–85 cm 0–10 cm Queen sp. Turitella communis Dosinia exoleta Corbula sp. Timoclea ovata Hinea incresata? Corbula gibba 10–20 cm Mactridae family Turitella communis Anomiidae family Corbula sp. Venerupis senegalensis Nucula sp. Parvicardium sp. 52–70 cm Abra sp. Aporrhais pes-pelican (truly sublittoral) Antalis sp. (sublittoral) Aequipecten sp. Identifications were made by Dr. Julia Nunn of the Ulster Museum, Chlamys varia Belfast. All shells live in a range of depths from shallow sub-tidal to Corbula gibba continental shelf settings. Parvicardium sp.Core part B surface reflector, suggesting that the surface reflector 60–80 cm Turitella communis represents the basal unconformity (reflector Ub, Fig. 6). Corbula gibba (muddy sediment only) In rare locations, the glacial-marine unit is cut by aCore part C channel or basin whose surface is also reflector Ub. 45–85 cm Acoustic reflections within this unit (H3, Fig. 6) are Modiolus modiolus coherent and sub-horizontal and appear enhanced by Chlamys distorta Turitella communis natural gas (Judd and Hovland, 1992). Above the Can galena glacial-marine sediment in most locations are two Nucula sp. additional acoustic units. The lower unit, (H2, Fig. 6), Anomiidae family is 2–4 m in thickness and lacks coherent internalCore BLVC04-03 reflections. It possesses numerous small areas of 0–20 cm Garifervensis sp. enhanced reflectivity, giving the unit a diffuse or Abra prismatica? “cloudy” impression (Fig. 6). The uppermost acoustic Abra nitida? unit, H1, extends to the seafloor. It is acoustically Dosinia exoleta (muddy gravel) transparent and corresponds well with muddy sediment Mactridae family previously described from Belfast Lough (Manning et 100–110 cm Arctica islandica al., 1970). It is present on all profiles, though it rarely Nucula exceeds 5 m in thickness. 110–130 cm In deeper water (Fig. 7), signal attenuation appears to Aequipecten sp. obscure deeper reflections. Bedrock is occasionally Chlamys varia recognized, but till is apparently thicker and the contact Timoclea ovata Ridged scallop (septumradiatum) between the two is unclear. Reflections within the Paleolim tylmostratum glacial-marine mud are also less clear. The unconfor- Hyatella arctica mity on the surface of the glacial-marine sediment Venerupis senegalensis remains relatively strong until about 30 m water depth, Nucula sp. when it becomes weaker and discontinuous (Fig. 8). It Corbula gibba Mactridae family disappears by 40 m below present sea level. Anomiidae family Two cores that were gathered within 100 m of one another in 15 m depth in Belfast Lough do not appear to10 J.T. Kelley et al. / Marine Geology 232 (2006) 1–15have penetrated the unconformity on the surface of contact in core BL3 at 32.5 m depth (Fig. 11a). The sandglacial-marine mud (Figs. 7 and 9). The cores are similar deposit is well sorted, with a mean size 2.1 phi and ato one another in stratigraphy with alternating beds of high proportion of shells and shell fragments. The sandmud and sandy mud (Fig. 9a). Upper and lower contacts was 11 cm thick in BL3, but the equivalent sand layer inbetween beds are mostly abrupt transitions. Carbonate BL4 was only 3 cm thick. Beneath the sandy material incontent in the cores is high, and a wide variety of intact BL3, 20–30 cm of laminated red mud and orange sandshells were recorded (Fig. 9b; Table 1). The muddy sand occurred (Fig. 11). The laminations range from mm-sizedeposits were notably enriched with Turitella communis, partings of sand to centimeter-thick layers. Beneath thewhich was abundant in all muddy sands. The lowermost laminated material, massive red clay continued to the0.7 m in BL2 was very fine-grained and noticeably bottom of the core.denser than other layers in the cores. It was bluish gray In core BL4, the abrupt contact between the overlyingwith black laminations of reduced sediment with several sandy mud and the sand deposit was at about 220 cm inmudballs up to 5 cm in diameter associated with rare the core, or 31 m water depth (Fig. 11). The sand wasshell fragments at a possible erosional contact near the well sorted and the same size, as in the adjacent core.bottom of the core. No intact shells were observed. Though abundant shell fragments existed in this sandy Two additional cores were gathered within 100 m of deposit, none were whole and most were sand-sizeone another between 29–31 m water depth (Fig. 10). fragments. In sharp contact with the sand, massive redComparison of the cores and the seismic record suggests clay continued beneath the sand. One shell of Spisula sp.that they each penetrated through the Holocene mud to (Table 1) was observed at 255 cm within the core.the unit interpreted as glacial-marine mud. Each of these Four radiocarbon dates were obtained from the twocores collected more than 1 m of fossiliferous, sandy deep cores (Table 2). All of the shells that were datedmud, but no finer grained mud layers were observed. were fresh (Fig. 12), though none were articulated or inAlthough many fossils were present, Turitella was not life position. Three of the shells from the 11 cm thickrecorded in the deeper water cores. sand deposit in BL3 produced similar AMS 14C ages Just below 1.5 m sub-bottom depth in each of the around 13.4 ka cal B.P. (Table 2). The three Nucula anddeeper water cores, pebbles appeared above a sandy one Balanus shells were essentially indistinguishableshelly layer (Fig. 10). The transition between the with overlapping calibrated age ranges of 13.3–13.4 caloverlying sandy mud and the sand was abrupt, and an B.P.; the Mactridae (mussel) fragment was approxi-angular, 2 cm diameter clast of red clay occurred at the mately 200 years younger than the other shells (13.13 ka Fig. 10. Core descriptions of BL3, BL4, from outer Belfast Lough. Lines show locations of core photos in Fig. 11. Cores are located in Fig. 7. J.T. Kelley et al. / Marine Geology 232 (2006) 1–15 11Fig. 11. a) Photo of contact between sandy bed interpreted as a “beach” in core BL3, with overlying sandy mud. The red mud clast is identical to thelower, glacial-marine mud, and is interpreted as a rip-up clast b) laminated sand and mud beneath “beach” deposit in core BL4. Photographs arelocated in Fig. 10.cal B.P.) (Table 2). The Spisula shell was coincident in (Fig. 4). Reflectors within the upper sand unit wereage with the Nucula and Balanus. earlier postulated to be peat (Cooper et al., 2002; McDowell et al., 2005), based on submarine terrestrial6. Discussion peat deposits cored from a nearby estuary (Wilson and McKenna, 1996). Our new cores correlate gravel with6.1. Causeway Coast those reflectors where they were reached (Fig. 5). The refusal our corer met in < 2 m of sand in most locations Cores from seaward of the beaches of the Causeway is also suggestive of a hard bottom. On one coringCoast contain fine sand texturally similar to the beaches attempt, the hardened, stainless steel core cutter came upTable 2Radiocarbon dated samplesSample # a Material calibrated age b 13 C/12C Conventional 14 C age Comment Range 1 sigma (2 sigma)202998 Nucula sp. +1.2 11,990 ± 40 BP Lowstand beach, shell was fresh, intact 13298–13425 (13246–13542)202999 Balanus sp. +0.7 12,000 ± 40 BP Lowstand beach, shell was fresh, intact 13303–13434 (13254–13557)203000 Mactridae family +0.7 11,670 ± 40 Lowstand beach, shell was fresh 13059–13195 (12961–13229)203001 Spisula sp. +0.9 12,030 ± 40 Glacial-marine mud 13314–13468 (13277–13594) a Beta Analytic Lab. b Calib 5.0.2.12 J.T. Kelley et al. / Marine Geology 232 (2006) 1–15 was the deposit that contained them indicative of a shore line, but the mud containing the foraminifer correlates with a shoreline feature nearby at +20 m. We believe that following this highstand, sea level dropped rapidly to approximately 30 m below present sea level by 13.4 ka cal B.P., based on the continuity of the later-formed transgressive unconformity to that depth in Belfast Lough and the shoreline deposits we interpret from cores BL3, and BL4 (Fig. 13). The unconformity clearly truncates acoustic reflectors in the glacial-marine material and marks the post-lowstand rise in sea level. Similar interpretations were recognized elsewhere (Barnhardt et al., 1997; Kelley et al., 1992;Fig. 12. Photograph of the delicate shells from the beach deposit. Belknap et al., 2004; Stea et al., 1994). We interpret theArrows point to those that were radiocarbon dated. Such fine shells sand deposits near the 30 m isobath in the two deeperwould not survive a long transport distance. Shells are from sandy cores as low-energy beach deposits slightly seaward of“beach” deposit seen in Fig. 11a. N, Nucula sp., B, Balanus sp., M, an eroding bluff of glacial-marine mud. The well sortedMactridae family. fine sand and abundance of shells of varied species is what is commonly observed in such a modern settingwith many dents in its cutting edge, again indicating a (Doerjes et al., 1986). The angular rip-up clast of red,rocky bottom. glacial-marine mud (Fig. 11a), as well as the pristine Gravel deposits are not a surprise along the inner shells (Fig. 12), are suggestive of rapid deposition of theshelf, as gravel commonly forms layers on the fine sand material, possibly in a storm, with no opportunity forbeaches of Ireland (Orford et al., 1999, 2003). Gravel degradation of the fragile shells or rounding of the clast.occurs from the edge of the frontal dune out onto the Thus, although the shells were not animals that lived onshallow nearshore zone when beaches experience a beach, we believe they were deposited on a beach at,erosional periods (Orford et al., 2003). Thus, the or shortly after the time of their death. We believe thereflectors we now interpret as gravel deposits aresuggested to mark the presence of a littoral depositsout to approximately the 30 m water depth. Thisobservation, coupled with erosional notches out to thesame depth (Cooper et al., 2002), suggests a lowstandposition around 30 m water depth along the CausewayCoast. We found no sub-tidal sedimentary record of themid-Holocene fall of sea level to its present location.6.2. Belfast Lough Fossils and sedimentary facies recorded within theBelfast Lough cores correspond well with earlier work(Manning et al., 1970). The cores and seismic observa-tions from Belfast Lough contain a clear record ofchanging sea level, as observed in other paraglaciallocations (Barnhardt et al., 1997; Stea et al., 1998).Glacial-marine mud was deposited initially in relativelydeep-water in late glacial times. In Northern Ireland, thelast late glacial higher-than-present stand of localrelative sea level was at about + 20 m at 14.2 ka cal B.P. about 25 km south of the study area (Fig. 2) (McCabeand Clark, 1998, 2003). McCabe and Clark (1998) dated Fig. 13. Shoreline change in Belfast Lough. a) Highstand in late glacialElphidium clavatum in marine mud draped over a time; b) lowstand in the latest Pleistocene; c) Holocene highstand, d)drumlin. These animals are not specific to any depth, nor present sea level. J.T. Kelley et al. / Marine Geology 232 (2006) 1–15 13shells did not travel far (Flessa, 1998), and that their clear. It is a muddy material, but denser than the surfaceradiocarbon dates mark the time sea level reached mud. Rip-up clasts within it may cause the diffuse− 30 m in Belfast Lough. reflections within the unit. The apparently erosional The dated Spisula shell from core BL4, in what is surface of this unit could represent a response to theinterpreted as glacial-marine sediment (Fig. 10), pre- lowering of sea level and an increase in wave energy assents a potential problem because it is only slightly older sea level from the mid-Holocene highstand, but morethan the beach deposit. Although we say “glacial- cores are needed to better define this unit.marine” mud, in its uppermost layer, the red mud maysimply be a marine deposit formed from reworking of 6.3. Summary and implications for further workthe red glacial-marine muddy sediment abundant in thearea. Thus, the Spisula sp. shell may have lived in This study underscores the importance of collectingshallow water, contemporaneously and just seaward of core samples to verify geological interpretations basedthe beach. The alternating laminations of red mud and on seismic reflection profiles. Earlier workers off Thefine sand below the sandy unit in BL3 are suggestive of Skerries (Fig. 1) (Cooper et al., 2002) and River Bann,a nearshore deposit (Reineck and Singh, 1980). 15 km west of the Skerries (McDowell et al., 2005), Although we have selected 30 m below present sea interpreted acoustic reflectors within the surficial sandlevel as the lowstand depth, it is possible that sea level of the inner shelf as buried peat deposits. This was afell somewhat lower than this depth. In Belfast Lough plausible interpretation because freshwater peat depositsthe unconformity does not end abruptly at 30 m water below present sea level had been cored in this coastaldepth, but it appears diminished in acoustic contrast at zone and used to establish a lower-than-present sea-levelgreater depths. It is likely that a ravinement unconfor- position (Wilson and McKenna, 1996). In Maine, USA,mity formed seaward of the beach deposit to a paleo- we had also cored peat deposits beneath the shorefacedepth of 10 m or more. Only more cores along a transit and used them in sea-level reconstructions (Kelley et al.,into deeper water will provide further insight to that 2005). On the exposed coast of Northern Ireland, gravelquestion. deposits are also interbedded with beach sand (Orford et As we interpret the record, the 30 m below present al., 2003), so it is not surprising that gravel layers formsea-level lowstand requires a relatively rapid fall in sea prominent acoustic reflectors on the shoreface and out tolevel from the + 20 m high stand of 14.2 ka cal B.P. A the lowstand position. Although these materials cannotminimum of 50 m of uplift in approximately 800 years be dated and used for a sea-level curve, their presencerequires local relative sea-level fall of about 6.3 cm/yr. establishes confidence in the sea-level reconstruction byThe timing of the uplift coincides with ice wasting and is demonstrating that shoreline deposits extend to the 30 mlikely to have been a period of rapid isostatic isobath. Although we did not gather sufficient seismicreadjustment (McCabe et al., 2005). Compared to data to quantify the geometry of the gravel layers, such amodern and ancient rates of post-glacial isostatic land study might permit distinction of onlap and offlapadjustment in Scandinavia of 8 cm/yr (Berglund, 2004) deposits that resulted from the Holocene highstand andunder similarly recent deglaciation, the suggested rate subsequent fall to present sea level.for Northern Ireland is not too extreme. Such rapid uplift Because of its sheltered setting, Belfast Loughmust have been associated with major crustal instability allows accumulation of Holocene mud deposits. Thesein the region. are more readily cored through than the sand and gravel After sea level topped the bluff of glacial-marine off The Skerries and because of this we reached thesediment, shallow, and then progressively deeper water transgressive unconformity and associated shorelineled to the observed succession of fauna (Fig. 13). The deposits with cores. The striking transition of thevariation of muddy and sandy mud beds in the shallow transgressive unconformity to conformity at depthscores might mark episodic storm events. River and below the lowstand position (Fig. 7) lends strongestuarine channels, and possibly lake basins, that existed credence in our interpretation of the thin, sandy depositin Belfast Lough at the time of the lowstand, were in about 30 m water depth as the lowstand shorelinedrowned by the rising water (Fig. 13). Preservation of position. The dated shells are good, but less than idealthe organic-rich estuarine or lake sediment has locally sea-level indicators because they appear to have washedled to methane generation (Fig. 6) as in other glaciated onto the beach and are not in life position. The complexsettings (Rogers et al., 2006). The origin of the radiocarbon calibration curve at the time of thewidespread acoustic unit, Hm2 (Fig. 6), directly above lowstand also lends some uncertainty to the age of thethe transgressive unconformity in Belfast Lough is less lowstand because there are multiple intersections of the14 J.T. Kelley et al. / Marine Geology 232 (2006) 1–15conventional radiocarbon date with the calibration Penobscot Bay, Maine. In: FitzGerald, D.M., Knight, J. (Eds.),curve. Despite this, Belfast Lough has potential to High Resolution Morphodynamics and Sedimentary Evolution of Estuaries. Springer, The Netherlands, pp. 335–360.yield more shells from lowstand positions, as well as Berglund, M., 2004. Holocene shore displacement and chronology inshells in life position and wood fragments along the Angermanland, eastern Sweden, the Scandanavian glacio-isostaticsurface of the transgressive unconformity, as has been uplift center. Boreas 33, 48–60.shown elsewhere (Kelley et al., 1992). Carter, R.W.G., 1982. Sea level changes in Northern Ireland. Proc. Geol. Assoc. 93, 7–23. Two implications of the new sea-level lowstand off Chappell, J.M., Polach, H., 1991. Post glacial sea level rise from aNorthern Ireland are related to the question of a coral record at Huon Peninsula, Papua, New Guinea. Nature 349,landbridge to Scotland (Devoy, 1995) and the problem 147–149.with numerical models of sea-level change in this region Clark, U., Mix, A.C., 2002. Ice sheets and sea level of the last glacial(McCabe, 1997). A landbridge to Scotland is ruled out maximum. Quat. Sci. Rev. 21, 1–7.for this area of Northern Ireland because water depths Clark, P.U., McCabe, A.M., Mix, A.C., 2004. Rapid sea-level rise at 19,000 years ago and its global implications. Science 304,> 50 m exist between the lowstand shoreline and 1141–1144.Scotland. Thus, the oldest archeological site in Ireland, Cooper, J.A.G., Kelley, J.T., Belknap, D.F., 1998. New seismic<50 km from the Skerries area (Fig. 1) along the River stratigraphic and side scan sonar evidence for a sea-level lowstandBann, was not a site that was reached directly by people off the north coast of Ireland: a preliminary appraisal. J. Coast. Res.on foot. Spec. Issue 26, 129–133. Cooper, J.A.G., Kelley, J.T., Belknap, D.F., Quinn, R., McKenna, J., Recent numerical models of sea-level change along 2002. Inner shelf seismic stratigraphy off the north coast ofthis coast clearly need adjustment to accommodate the Northern Ireland: new data on the depth of the lowstand. Mar.data reported in this paper and elsewhere (McCabe, Geol. 186, 369–387.1997). The rapidity of the late Pleistocene fall in local Devoy, R.J.N., 1995. Deglaciation, earth crustal behavior and sea-levelrelative sea level, presumably owing to isostatic uplift of changes in the determination of insularity: a perspective from Ireland. In: Preece, R.C. (Ed.), Island Britain: A Quaternarythe land, and the depth of the lowstand are not captured Perspective. Geol. Soc. London Spec. Paubl., vol. 96, pp. 181–208.by existing models (Lambeck and Purcell, 2001). Doerjes, J., Frey, R.W., Howard, J.D., 1986. Origin of, andSimilar rapid changes in relative sea level in the Gulf mechanisms for, mollusk shell accumulations on Georgia beaches.of Maine and Scotian Shelf, Canada (Barnhardt et al., Senckenbergiana Marit. 18, 1–43.1997; Stea et al., 1998; Shaw et al., 2002) are observed, Fairbanks, R.G., 1989. A 17,000 year glacio-eustatic sea level record- influence of glacial melting rates on the Younger Dryas event andbut not well accounted for by the coarse nature of deep ocean circulation. Nature 342, 637–642.existing numerical models (Peltier, 1998). Hopefully, Flessa, K.W., 1998. Well-travelled cockles: shell transport during thethe creation of more refined and spatially diverse sea- Holocene transgression of the southern North Sea. Geology 26,level curves such as the one presented herein will 187–190. Geological Survey of Northern Ireland, 1997. Geological Map ofpromote more realistic numerical models of late Northern Ireland: Solid Geology. 1/250,000.Quaternary land and sea-level changes. Jackson, D.W.T., Cooper, J.A.G., del Rio, L., 2005. Geological control of beach morphodynamic state. Mar. Geol. 216, 297–314.Acknowledgements Judd, A.G., Hovland, M., 1992. The evidence of shallow gas in marine sediments. Cont. Shelf Res. 12, 1081–1095. We are grateful to the Visiting Scholar Program at the Kelley, J.T., Dickson, S.M., Belknap, D.F., Stuckenrath, R., 1992. Sea- level change and the introduction of late Quaternary sediment toUniversity of Ulster-Coleraine for support for Kelley. the Southern Maine inner continental shelf. In: Wehmiller, J.,We acknowledge Matt Service for assistance in the field Fletcher, C. (Eds.), Quaternary Coasts of the United States. Soc.and in providing access to the R/V Lough Foyle. Bill Econ. Paleo. and Mineralogists, Spec. Pap., vol. 48, pp. 23–34.Clarke is acknowledged for his assistance in the field as Kelley, J.T., Barber, D.C., Belknap, D.F., FitzGerald, D.M., vanis Sam Kelley. Thanks also for constructive reviews of Heteren, S., Dickson, S.M., 2005. Sand budgets at geological, historical and contemporary time scales for a developed beachthis manuscript from Roland Gehrels and an anonymous system, Saco Bay, Maine, USA. Mar. Geol. 214, 117–142.reviewer. Lambeck, K., Chappell, J., 2001. Sea-level change through the last glacial cycle. Science 292, 679–686.References Lambeck, K., Purcell, A.P., 2001. Sea-level change in the Irish Sea since the last glacial maximum: constraints from isostaticBarnhardt, W.A., Belknap, D.F., Kelley, J.T., 1997. Sequence modeling. J. Quat. Sci. 16, 497–506. stratigraphy of submerged river-mouth deposits in the northwest- Lawlor, D., 2000. Inner shelf sedimentology off the north coast of ern gulf of maine: responses to relative sea-level changes. Geol. Northern Ireland. Unpublished Ph. D. Thesis, University of Ulster. Soc. Amer. Bull. 109, 612–630. Lloyd, J.M., Shennan, I., Kirby, J.R., Rutherford, M.M., 1999.Belknap, D.F., Gontz, A.M., Kelley, J.T., 2004. Paleodeltas and Holocene relative sea-level changes in the inner Solway Firth. preservation potential on a paraglacial coast: evolution of eastern Quat. Int. 60, 83–105. J.T. Kelley et al. / Marine Geology 232 (2006) 1–15 15Manning, P.I., Robbie, J.A., Wilson, H.E., 1970. Geology of Belfast Reineck, H.E., Singh, I.B., 1980. Depositional Sedimentary Environ- and the Lagan Valley. Memoirs of the Geological Survey. HMSO, ments. Springer-Verlag, NY. 549 pp. Belfast. Rogers, J., Kelley, J.T., Belknap, D.F., Gontz, A., Barnhardt, W.A.,McCabe, A.M., 1987. Quaternary deposits and glacial stratigraphy in 2006. Shallow water pockmark formation in temperate estuaries: a Ireland. Quat. Sci. Rev. 6, 259–299. consideration of origins in the western Gulf of Maine with specialMcCabe, A.M., 1997. Geological constraints on geophysical models emphasis on Belfast Bay. Mar. Geol. 225, 45–62. of relative sea-level change during deglaciation of the western Irish Shaw, J., Carter, R.W.G., 1994. Coastal peats from northwest Ireland; Sea basin. J. Geol. Soc. Lond. 154, 601–604. implications for late Holocene relative sea-level change andMcCabe, A.M., Clark, P.U., 1998. Ice sheet variability around the North shoreline evolution. Boreas 23, 74–91. Atlantic Ocean during the last deglaciation. Nature 392, 373–377. Shaw, J., Gareau, P., Courtney, R.C., 2002. Paleogrography of AtlanticMcCabe, A.M., Clark, P.U., 2003. Deglacial chronology from County Canada 13–0 kyr. Quat. Sci. Rev. 21, 1861–1878. Donegal, Ireland: implications for deglaciation of the British-Irish Shennan, I., Horton, B., 2002. Holocene land- and sea-level changes in sector ice sheet. J. Geol. Soc. Lond. 160, 847–855. Great Britain. J. Quat. Sci. 17, 511–526.McCabe, A.M., Carter, R.W.G., Haynes, J.R., 1994. A shallow marine Shennan, I., Peltier, W.R., Drummond, R., Horton, B., 2002a. Global emergent sequence from the northwestern sector of the lst British to local scale parameters determining relative sea-level changes ice sheet, Portballintrae, Northern Ireland. Mar. Geol. 117, 19–34. and the post-glacial isostatic adjustment of Great Britain. Quat. Sci.McCabe, A.M., Clark, P.U., Clark, J., 2005. AMS14C dating of Rev. 21, 397–408. deglacial events in the Irish Sea Basin and other sectors of the Shennan, I., Hamilton, S., Hillier, C., Woodroffe, S., 2002b. A 16,000- British-Irish Ice sheet. Quat. Sci. Rev. 24, 1673–1690. year record of near-field, relative sea-level changes, northwestMcDowell, J., Knight, L., Quinn, J., 2005. High resolution Scotland, United Kingdom. Quat. Int. 133–134, 95–106. geophysical investigations seaward of the Bann estuary, Northern Shennan, I., Hamilton, S., Hillier, C., Woodroffe, S., 2005. A 16,000 Ireland coast. In: FitzGerald, D.M., Knight, J. (Eds.), High year record of near-field relative sea level changes, northwest Resolution Morphodynamics and Sedimentary Evolution of Scotland, United Kingdom. Quat. Int. 133–134, 95–106. Estuaries. Springer, The Netherlands, pp. 11–31. Stea, R.R., Boyd, R., Fader, G.B.J., Courtney, R.C., Scott, D.B.,Mitchell, F., Ryan, M., 2001. Reading the Irish landscape. Town Pecore, S.S., 1994. Morphology and seismic stratigraphy of the House, Dublin, Republic of Ireland. 392 pp. inner continental shelf off Nova Scotia, Canada: evidence for aOrford, J.D., Cooper, J.A.G., McKenna, J.M., 1999. Mesoscale −65 m lowstand between 11,650 and 11,250 14C yr B.P. Mar. temporal changes to foredunes at Inch Spit, southwest Ireland. Geol. 117, 135–154. Zeitshr. Fur Geomorph. 43, 439–461. Stea, R., Piper, D.J.P., Fader, G.B.J., Boyd, R., 1998. WisconsinanOrford, J.D., Murdy, J.M., Wintle, A.G., 2003. Prograded Holocene glacial and sea level history of Maritime Canada and the adjacent beach ridges with superimposed dunes in north-east Ireland: continental shelf: a correlation of land and sea events. Geol. Soc. mechanisms and timescales of fine and coarse beach sediment Amer. Bull. 110, 821–845. decoupling and deposition. Mar. Geol. 194, 47–64. Stuiver, M., Reimer, P., 1993. Extended 14C database and revisedParker, J.G., 1982. Grain-size characteristics of recent sediments in CALIB radiocarbon calibration program. Radiocarbon 32, Belfast Lough. Mar. Geol. 50, 143–154. 215–230.Peltier, W.R., 1998. Postglacial variations in the level of the sea: Wilson, H.E., 1972. Regional Geology of Northern Ireland. HMSO, implications for climate dynamics and solid earth geophysics. Rev. Belfast. 115 pp. Geophys. 36, 603–690. Wilson, P., McKenna, J., 1996. Holocene evolution of the River BannPeltier, W.R., 2002. On eustatic sea level history: last glacial maximum estuary and adjacent coast. Proc. Geol. Assoc. 107, 241–252. to Holocene. Quat. Sci. Rev. 21, 377–396.