New cetacean r values for arctic north

Quaternary Science Reviews 91 (2014) 218e241 Contents lists available at ScienceDirect Quaternary Science Reviews journal homepage: www.elsevier.com/locate/quascirev New cetacean DR values for Arctic North America and their implications for marine-mammal-based palaeoenvironmental reconstructions  kowski b, 1, Roy D Coulthard c, Mark F.A Furze a, *, 1, Anna J Pien a b c Earth & Planetary Sciences Division, Department of Physical Sciences, MacEwan University, Edmonton, Alberta T5J 4S2, Canada School of Ocean Sciences, College of Natural Sciences, Bangor University, Menai Bridge, Anglesey, Wales LL59 5AB, UK Department of Earth & Atmospheric Sciences, University of Alberta, Edmonton, Alberta T6G 2E3, Canada a r t i c l e i n f o a b s t r a c t Article history: Received 24 March 2013 Received in revised form 25 July 2013 Accepted 24 August 2013 Available online 24 September 2013 Radiocarbon-dated marine mammal remains from emergent Arctic coastlines have frequently been used to reconstruct Holocene sea-ice histories The use of such reconstructions has hitherto been complicated by uncertain marine reservoir corrections precluding meaningful intercomparisons with data reported in calibrated or sidereal years Based on an exhaustive compilation of previously published marine mammal radiocarbon dates (both live-harvested materials and subfossils) from the Canadian Arctic Archipelago (CAA), new, statistically-derived d13C and DR values are provided Average d13C values are: À16.1 Ỉ 1.1& (bone collagen; n ¼ 193) for bowhead (Balaena mysticetus); 14.4 ặ 0.5& (n ẳ 44; dentine) for beluga (Delphinapterus leucas); 14.8 ặ 1.9& (teeth and tusks; n ẳ 18) and 18.0 ặ 4.7& (n ẳ 9; bone collagen) for walrus (Odobenus rosmarus) DR values are 170 Ỉ 95 14C years for bowhead (n ẳ 23) and 240 ặ 60 14C years for beluga (n ¼ 12) Scarce data preclude calculation of meaningful, statistically robust walrus DR Using the new DR values, an expanded and revised database of calibrated bowhead dates (651 dates; many used in previous CAA sea-ice reconstructions) shows pronounced late Quaternary spatio-temporal fluctuations in bone abundance Though broadly resembling earlier bowhead subfossil frequency data, analysis of the new expanded database suggests early- and mid-Holocene increases in whale abundance to be of longer duration and lower amplitude than previously considered A more even and persistent spread of infrequent low-abundance remains during “whale free” intervals is also seen The dominance of three eastern regions (Prince Regent Inlet & Gulf of Boothia; Admiralty Inlet; Berlinguet Inlet/Bernier Bay) in the CAA data, collectively contributing up to 88% of all subfossil remains in the mid-Holocene, is notable An analysis of calibrated regional sea-level index points suggests that severance of the Admiralty Inlet-Gulf of Boothia marine channel due to isostatically-driven regression may have played a significant role in enhanced whale mortality during this interval Comparisons between the newly calibrated bowhead data and other regional sea-ice proxy data further highlight spatial and temporal discrepancies, potentially due to regional asynchronicities and variable sensitivities in proxy response to climate and oceanographic forcing However, the limited number of deglacialepostglacial marine records continues to hamper extensive intercomparisons between marine mammal and other proxy datasets Nevertheless, an examination of assumptions inherent in linking bowhead subfossil frequencies, population densities, and sea-ice thickness and distribution, shows that such relationships are highly complex Factors such as broad sea-ice preferences, variable mortality rates and causes, long distance carcass transport, variable coastline and basin/channel geometries, and changing emergence rates all complicate the correlation of whale bone abundance to sea-ice histories Ó 2013 Elsevier Ltd All rights reserved Keywords: Radiocarbon calibration Reservoir correction DR Bowhead whale Balaena mysticetus Canadian Arctic Sea ice Introduction * Corresponding author Tel.: ỵ1 780 633 3918 E-mail address: furzem@macewan.ca (M.F.A Furze) All authors contributed equally to this publication 0277-3791/$ e see front matter Ó 2013 Elsevier Ltd All rights reserved http://dx.doi.org/10.1016/j.quascirev.2013.08.021 Radiocarbon-dated marine mammal remains have been widely used to reconstruct late Quaternary relative sea-levels, palaeoenvironments, and human occupation in Arctic North America (e.g., Dyke and Morris, 1990; Dyke et al., 1991, 1996a, 2011; Dyke M.F.A Furze et al / Quaternary Science Reviews 91 (2014) 218e241 and Savelle, 2001; Savelle et al., 2012) Such dates have historically been reported in uncorrected radiocarbon years, or have had assumed reservoir corrections (R) applied to them (Dyke et al., 1991, 1996a; Dyke and Savelle, 2001; Dyke and England, 2003), complicating meaningful intercomparisons with other proxy records reported in calibrated or sidereal years Coulthard et al (2010) discussed this problem regarding molluscan chronologies, providing statistically-based regional reservoir offset values (DRR) for calibration of Canadian Arctic Holocene radiocarbon dates Here, we provide the first, statistically-derived d13C and regional DR values for marine mammal-based radiocarbon chronologies from Arctic Canada, present an expanded, revised regional database of calibrated marine mammal dates, and conduct a spatio-temporal analysis of bowhead whale (Balaena mysticetus) remains as a palaeo-sea-ice proxy Historically, marine mammal dates have been considered of dubious quality due to the challenges associated with correcting for d13C fractionation for different species and different types of remains, besides the complexity of marine reservoir corrections for migrating species In the archaeological community, McGhee and Tuck (1976) and Tuck and McGhee (1984) argued that 14C dates on marine organisms should simply be ignored due to the difficulty of comparing them with terrestrial materials, despite Arundale’s (1981) call for a more moderate approach Dyke et al (1996b) recommended a reservoir correction of w200 years for bowhead dates whereas Dyke et al (1999) reported walrus dates without reservoir correction, as little data existed by which to derive a meaningful value Mangerud et al (2006) calculated reservoir corrections and DR from live-harvested 19th century whales in the European Atlantic, suggesting DR values of w7 years for both baleen and toothed whales, reflecting their food source in the ocean surface layer In the Canadian Arctic Archipelago (CAA), Dyke and Savelle (2009) proposed an increase in the bowhead reservoir correction (R) from 200 to 400 years, similar to changes proposed for molluscan R (Dyke et al., 2003) Notably, no statistically-based R or DR for CAA marine mammals has previously been attempted Nevertheless, radiocarbon-dated marine mammal remains from emergent Arctic coastlines have long been used to reconstruct deglacial to postglacial sea-level histories (Blake, 1961, 1970; Salvigsen, 1978; Dyke, 1979, 1980) Aside from marine molluscs and driftwood, bowhead remains form the basis of Holocene sealevel curves from the central and eastern CAA and Svalbard (Blake, 1975; Salvigsen, 1978; England, 1983; Dyke et al., 1991, 2011; Dyke, 1998) Given their comparative scarcity, walrus (Odobenus rosmarus), beluga (Delphinapterus leucas), and narwhal (Monodon monoceros) are seldom employed as Holocene sea-level index points While offshore sinking, landward crawling (walrus), and sea-ice push can complicate altitudinal relationships, careful construction of emergence curves derived from a range of different index points may identify and negate the distorting effects of “sinkers” and “crawlers” on sea-level reconstructions (Dyke, 1980, 1993; Dyke et al., 1991, 1999) The potential of marine mammals as Holocene sea-ice proxies was noted early in their use as sea-level indicators (Salvigsen, 1978; Dyke, 1979, 1980; Evans, 1989; Bednarski, 1990) Fundamental to this approach is the close relationship between sea-ice and the distribution of these animals, given their dependence on sea-ice for breeding, feeding, and resting (Moore and Reeves, 1993; Stirling, 1997; Laidre et al., 2008) Consequently, long- and short-term changes in sea-ice extent and thickness should affect species distributions (Vibe, 1967; Reeves et al., 1983) Therefore, the subfossil record should reflect local abundance and occupation changes (Dyke and Morris, 1990; Dyke et al., 1996a) This approach has been used extensively in the CAA, where decades of intense fieldwork, primarily by the Geological Survey of Canada (GSC), have generated 219 hundreds of elevationally-constrained radiocarbon dates on individual marine mammal specimens An even greater number of observations, the age of which can be determined by plotting their elevation on well-constrained sea-level curves, has been recorded (Dyke et al., 1996a; Savelle et al., 2000) Based on an exhaustive compilation of previously published marine mammal radiocarbon dates from Arctic North America, we here provide new d13C and DR values for bowhead (B mysticetus) and beluga (D leucas), and d13C values for walrus (O rosmarus) These values are based on both live-harvested (known age) materials (walrus, beluga), and co-occurring driftwood and bowhead whale subfossils The assembled database (Table S1) includes 809 marine mammal dates from the CAA, expanding previous compilations and enabling direct comparison with other marine and terrestrial data reported in terrestrial 14C, calibrated, and sidereal years We use this new dataset to perform a spatio-temporal analysis of subfossil B mysticetus occurrence as a presumed function of Holocene sea-ice variability, with valuable implications for marine mammal palaeoecology This approach permits comparisons with previous cetacean-based reconstructions and emerging alternative marine proxies, and allows the critical re-evaluation of marine mammals as sea-ice indicators Biology, ecology, and taphonomy of whales and walrus Most arctic marine mammals are poorly represented in the late Quaternary CAA, though bowhead and walrus are notable exceptions Proportionally, bowheads represent the largest subfossil component (>600), with lesser walrus (20 cm thick ice (Carroll and Smithhisler, 1980; George et al., 1989; Philo et al., 1993; Würsig and Clark, 1993) The degree to which ice entrapment represents a significant mortality cause in modern or Holocene populations is unknown (Philo et al., 1993) However, subfossil length distributions from Admiralty Inlet, Baffin Island, resemble modern populations, suggesting random, not size-selective mortality (Savelle et al., 2000) consistent with frequent and climate-dependent sea-ice entrapment throughout this 8000 year record It is unclear whether documented Arctic strandings occurred live or post-mortem, as carcasses can float for days to weeks (cf Schäfer, 1972; Marquette, 1978; Marquette and Bockstoce, 1980; Bogoslovskaya et al., 1982; Finley, 2001), potentially drifting significant distances with surface currents or freezing into, and being carried by, sea-ice Though carcasses frequently ground intertidally, as many as 50% of dated CAA fossils occur well below their contemporaneous isostatically raised shorelines (“sinkers”; Dyke and Morris, 1990) Arctic climate and permafrost, coupled with the large bowhead size (cf Schäfer, 1972; cf Espinoza et al., 1998) favours post mortem skeletal preservation, despite scavenging (Chesemore, 1968; Bentzen et al., 2007) 2.2 Delphinapterus leucas (Pallas 1776) (beluga, belukha, or white whale) and Monodon monoceros Linnaeus 1758 (narwhal) Beluga and narwhal are medium-sized (3e5 m length, 1500e 1900 kg) toothed whales, which inhabit sea-ice covered Arctic to sub-Arctic waters (Laidre et al., 2008), occur in pods of tens to Fig The Canadian Arctic Archipelago (CAA) with the modern range of the bowhead whale (Balaena mysticetus) after Moore and Reeves (1993), Reeves and Heide-Jørgensen (1996), Bogoslovskaya (2003), Koski et al (2006), and Ferguson et al (2010) Abbreviations used in the map are: Adm In ¼ Admiralty Inlet; A.R Is ¼ Amund Ringnes Island; Cam Is ¼ Cameron Is.; Cor Is ¼ Cornwallis Island; Eg Is ¼ Eglinton Island; E.R Is ¼ Ellef Ringnes Island; Grin Pen ¼ Grinnell Peninsula; Norw B ¼ Norwegian Bay; Stef Is ¼ Stefansson Island; Well Ch ¼ Wellington Channel M.F.A Furze et al / Quaternary Science Reviews 91 (2014) 218e241 hundreds (COSEWIC, 2004a, 2004b), and may interbreed (sexual maturity at w6 and w5e8 years, respectively; Heide-Jørgensen and Reeves, 1993; COSEWIC, 2004a, 2004b; Heide-Jørgensen and Laidre, 2006) Whereas adult beluga have 34e38 teeth (Tomilin, 1957), narwhal possess merely two upper canines, one of which develops into a prominent (2e3 m) tusk in most males (Best, 1981) Beluga teeth grow in dentine growth layer groups (GLGs) deposited as one layer per year (Stewart et al., 2006; earlier studies suggested two annual GLGs, Goren et al., 1987; Heide-Jørgensen et al., 1994) The beluga diet, though largely based on Arctic (Boreogadus saida) and polar cod (Arctogadus glacialis), varies with region, ice conditions, and season (Bluhm and Gradinger, 2008) Narwhal feeding, which relies on fish (A glacialis; B saida; Greenland halibut Reinhardtius hippoglossoides) and squid (Gonatus fabricii), may be maximized during winter (Laidre et al., 2003; Laidre and HeideJørgensen, 2005; Bluhm and Gradinger, 2008) Both cetaceans undertake extensive seasonal migrations and exhibit high site fidelity (Heide-Jørgensen et al., 2003a, 2003c; Luque and Ferguson, 2010) Beluga pods migrate seasonally between summering (coastal estuaries, rivers, bays) and wintering grounds (offshore, pack-ice-covered areas; Loseto et al., 2006) The total beluga population (w150,000) is distributed in several stocks which may overlap seasonally and interbreed (cf Suydam et al., 2001) Canadian stocks comprise the St Lawrence River (w1000 animals); Beaufort Sea (Alaska-Canada, w39,000); Ungava Bay (size unknown); eastern Hudson Bay (w2000e3000), western Hudson Bay (23,000); Cumberland Sound (1500); and eastern High Arctic-Baffin Bay (w21,000; Laidre et al., 2000; Innes et al., 2002; COSEWIC, 2004a; Jefferson et al., 2012a) The latter population migrates between wintering grounds in West Greenland (Disko Island-Maniitsoq) and CAA summering grounds (Prince Regent Inlet/Peel Sound estuaries; Innes et al., 2002; Heide-Jørgensen et al., 2003c), though w7900 animals overwinter in the North Water Polynya (cf Innes et al., 2002; Heide-Jørgensen et al., 2003c) and some individuals likely overwinter in CAA shore leads (Lancaster Sound, Jones Sound; Heide-Jørgensen et al., 2002) rather than travelling to Greenland Narwhal (total population w80,000) inhabit eastern Arctic Canada, Greenland, Russia, Svalbard and Jan Mayen (Gjertz, 1991; COSEWIC, 2004b; Jefferson et al., 2012b) They prefer heavily ice-covered offshore waters during winter (Baffin Bay; Heide-Jørgensen et al., 2003a; Laidre et al., 2004), while the primary Canadian summering ground is the CAA (w50,000e 70,000 animals; Peel Sound, Prince Regent Inlet/Gulf of Boothia, Admiralty Inlet, Eclipse Sound; Innes et al., 2002) Both species are relatively long-lived (30e50 years), and individual beluga as old 77 years have been reported (Harwood et al., 2002; Luque and Ferguson, 2010) Aside from senescence and hunting, mortality factors include predation (orca, polar bear), pathogens (Kenyon and Kenyon, 1977; Fay, 1978; Nielsen et al., 2001), and ice entrapment (Siegstad and Heide-Jørgensen, 1994; Suydam et al., 2001; Heide-Jørgensen et al., 2002), although both species are capable of breaking thin ice (Porsild, 1922; Siegstad and Heide-Jørgensen, 1994) Ice entrapment and stranding may promote predation by polar bears (Smith and Sjare, 1990; HeideJørgensen et al., 2002) Like bowheads, beluga and narwhal carcasses float and drift with currents (Martineau et al., 2002) and can be frozen into sea-ice (Porsild, 1922) 221 1982; Gjertz et al., 2001; Schreer et al., 2001) During winter to early spring, they occupy pack-ice interspersed with leads and polynyas In summer, females and juveniles stay within pack-ice while males occupy coastal sites (Fay, 1985; Ray et al., 2006) Walrus primarily feed on bivalves (Mya truncata, Hiatella arctica, Serripes groenlandicus; Sheffield et al., 2001; Born et al., 2003) though occasionally they feed on seabirds (Mallory et al., 2004), seals (Lowry and Fay, 1984), fish, and whales (Fay, 1985) Whether this represents aberrant behaviour or opportunism remains debated The two recognized subspecies, O rosmarus rosmarus (eastern Canadian Arctic, Greenland, Svalbard, western Russian Arctic) and O rosmarus divergens (Bering-Chukchi seas, coastal Alaska, Wrangel Island, Beaufort Sea; Fay, 1985; Knutsen and Born, 1994; Wiig and Gjertz, 1996) are presumed separated by central CAA multi-year sea-ice (Harington, 1966), despite some evidence of Holocene genetic exchange (Andersen et al., 1998) The modern population, much decimated by harvesting (pre-1931; Born et al., 1995), is highest in the Pacific (w200,000 animals; Ray et al., 2006) Extant Canadian Atlantic stocks are found in south and east Hudson Bay (500? walrus), northern Hudson Bay-Davis Strait (6000), Foxe Basin (5500), and Baffin Bay (1500; Born et al., 1995; COSEWIC, 2006) The walrus life span is w30 years (COSEWIC, 2006); mortality factors include senescence, predation (polar bear, orca; Calvert and Stirling, 1990), pathogens (Fay, 1978; Nielsen et al., 2001; Serhir et al., 2001), and violent death by other walrus at haul-out sites (Loughrey, 1959; Fay and Kelly, 1980) Mass mortality has been attributed to exhaustion from sustained open sea exposure due to sea-ice loss (Fischbach et al., 2009) Walrus can break ice ( 20 cm thickness) with their tusks or skull (underwater), and can also travel ( km) over ice or land to find open water though this puts them at risk of predation, starvation, freezing, and disorientation (Richard and Campbell, 1988; Calvert and Stirling, 1990) Dead walrus landward of their contemporaneous shorelines, resulting from either anomalous behaviour or disorientation, are documented from the CAA (Thorsteinsson, 1958; Dyke, 1979; Dyke et al., 1999), Svalbard (Lauritsen et al., 1980), Russia (Perfil’ev, 1970), and the Champlain Sea (Grant, 1989; Harington et al., 1993), similar to pinnipeds from modern Antarctic settings (Stirling and Kooyman, 1971; Banks et al., 2010) If dying offshore, walrus typically undergo “bloat and float” (floating, sinking, floating with decomposition gas build-up; Schäfer, 1972; Espinoza et al., 1998); carcasses float for weeks to months with surface currents before beaching (on ice-free coasts), potentially travelling 100’s of kilometres (Fay, 1978) Upon beaching of cadavers, or where death occurs at shoreline haul-outs, physical and chemical processes (sea-ice, waves) disarticulate and scatter bones (Espinoza et al., 1998; Dyke et al., 1999) Most common Holocene walrus remains are isolated tusks, crania, and mandibles Individuals that die inland experience lesser post mortem disarticulation and bone scattering compared to littoral carcasses and are overrepresented in the Holocene record (Dyke et al., 1999) Aside from shoreline mortality and carcass stranding (at sea-level), and landward crawling (above sea-level), significant numbers of CAA Holocene walrus are offshore sinkers, complicating their use as sealevel indicators (Dyke et al., 1999) DR derivation and calibration of radiocarbon dates 2.3 Odobenus rosmarus (Linnaeus 1758) (walrus) 3.1 Marine mammal samples and database construction Walrus are tusked, gregarious pinnipeds (Atlantic regions: max w3.2 m length, 1100 kg) that inhabit Arctic continental shelves year-round (COSEWIC, 2006) They are restricted by their shallow diving abilities ( 80 m depth), using mobile sea-ice for resting and breeding, and swimming between land and sea-ice haul outs (Fay, We conducted a comprehensive literature search of marine mammal radiocarbon dates from Arctic Canada and northwest Greenland (Table S1) The primary sources of bowhead data are GSC publications (e.g., Dyke and Morris, 1990; Dyke et al., 1996b) 222 M.F.A Furze et al / Quaternary Science Reviews 91 (2014) 218e241 Walrus dates were primarily compiled from Dyke et al (1999) and archaeological references therein Data from live-collected beluga are from Stewart et al (2006) and Campana (pers comm 2012) Narwhal dates are predominantly from GSC surficial geology maps (e.g., Dyke and Hooper, 2000) Remaining dates (seal) are chiefly from the archaeological literature (e.g., Morrison, 1989, and references therein) All data were cross-referenced with Harington (2003) and the Canadian Archaeological Radiocarbon Database (Morlan, 2005), which provided additional information (e.g., coordinates, material) not reported in the original publications Original sources are given for each date in Table S1 Our compiled database encompasses 651 bowhead whale (on 609 individuals), 103 walrus (98 individuals), 21 narwhal (21 individuals), 33 seal (31 individuals), and one beluga dates 3.2 Marine mammal d13C For calculation of average d13C values, considered representative of a species, only dated specimens with measured d13C were selected and samples considered non-finite were excluded Our selection encompasses bone collagen measurements from 193 bowhead and walrus samples (Tables and S1) Additionally, 18 d13C measurements on walrus tusks and teeth (primarily collagen) are available For beluga, 44 measured d13C values on dentine are available (Stewart et al., 2006) Arithmetic averages of d13C were calculated based on available data for each group (Table 1) For bowhead bone collagen, the standard deviation of the average was calculated assuming data were representative for the entire population using Equation (1) The small sample size of all other groups, requires the assumption that only a subset of the population is represented, as per Equation (2) s Pn i ẳ x xị sẳ n (1) after Ward and Wilson (1978) sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Pn i ẳ x xị sẳ n ẳ (2) after Ward and Wilson (1978) For both equations, s is the pooled standard deviation, x is the measured d13C, x is the average d13C value, and n is the number of d13C measurements The resulting average d13C values are reported in Table Of the total 628 bowhead collagen dates, 193 dates have measured and recorded d13C values and 378 dates were formerly non-normalized due to lack of measured d13C and could not previously be directly compared with conventional radiocarbon dates Using our new d13C values, we report all our dates (Table S1) as conventional radiocarbon dates (normalized to d13C ¼ À25&; Donahue et al., 1990; Reimer et al., 2004; spreadsheet available at http://intcal.qub.ac uk/calib/fractionation.html, Stuiver et al., 2013) Table Average d13C values for bowhead whale, walrus and beluga in the Canadian Arctic Archipelago Original d13C values are included in Table S1 Species Common name Material d13C s n Balaena mysticetus Balaena mysticetus Delphinapterus leucas Odobenus rosmarus Odobenus rosmarus Bowhead whale Bowhead whale Beluga whale Atlantic walrus Atlantic walrus Bone collagen Bone apatite Dentine Teeth & tusks Bone collagen À16.1 À12.7 À14.4 À14.8 À18 1.1 0.5 1.9 4.7 193 44 18 3.3 DR calculation 3.3.1 Bowheads Dyke et al (1996b) compared driftwood and whale bone dates from the same raised shorelines to estimate bowhead R for the CAA, assuming contemporaneity of both materials In addition to these data, we use driftwood dates by Blake (1975), Dyke (1998, 2000), Dyke and Savelle (2009), and Dyke et al (2011) We calculate and report only DR; more appropriate for calibration than R and subject to lesser temporal variability (Stuiver et al., 1986; Coulthard et al., 2010) Apparent DR of accordant bone and driftwood samples is calculated using the following procedure: Calibration of driftwood date into calibrated years (Calib 6.0, Stuiver et al., 2013) Identification of the global ocean modelled radiocarbon age contemporaneous with the calibrated driftwood date (equivalent Marine09 age in 14C years; Table 2) from the Marine09 (Reimer et al., 2009) spreadsheet (www.radiocarbon.org/ IntCal09%20files/marine09.14c) Subtraction of conventional whale bone age (in 14C years) from the equivalent Marine09 age (in 14C years) to determine apparent DR Aside from accordant boneedriftwood pairs, many sites have these materials elevationally separated by a small amount (typically m) In such cases, apparent DR may still be determined by correcting the bone date for the elevation-equivalent age difference using published relative sea-level curves (e.g., Blake, 1975; Dyke et al., 1991, 2011; Dyke, 1998) prior to calculating DR (Step 3.) sensu Dyke et al (1996b) In calculating apparent DR, the following assumptions are made: i To a first approximation, whales and trees (e.g., Picea, Pinus, Larix) have similar life spans (Viereck and Johnston, 1990; George et al., 1999) Therefore, whale boneedriftwood pairs are assumed to have fixed their carbon at approximately the same time ii Whale bone and driftwood record the age of the shoreline from which they are collected iii All whale bone in our comparisons is assumed to belong to a single population that can be described by a single DR value Whereas significant differences may exist between individual DR values for whales from the western (Beaufort Sea/Pacific) and eastern (Baffin Bay/Atlantic) populations, all driftwoode whale pairs are most likely from eastern whale stocks given their geographic distribution (Harington, 1966; Dyke et al., 1996a) iv Once apparent DR values are calculated, sample pairs that contradict Assumption ii are identified As the Eastern CAA bowhead population lives in Arctic Canada and Baffin Bay waters throughout the year (Heide-Jørgensen et al., 2003b; Boertmann et al., 2009), bounding whale DR values can be assumed by adopting the regional maximum and minimum DR values for marine organisms in equilibrium with seawater bicarbonate (molluscs; Coulthard et al., 2010) These values are 335 Ỉ 85 14C years for the NW CAA (Coulthard et al., 2010) and À10 Ỉ 80 14C years for eastern Baffin Bay (West Greenland, Table S2; McNeely et al., 2006) Therefore only comparisons yielding an apparent DR between À10 and 335 14 C years are considered valid As whales migrate within these regions, we apply this assumption only for eastern Canadian Arctic subfossils, but not for western Arctic/Beaufort Sea bowheads DR values outside the accepted range have Table Bowheadedriftwood pairs used to derive individual values of apparent DR Wood lab codeb 14 C age wood 14 C yrs BP Calibrated Equivalent Whale wood age marine09c age lab codeb,d cal yrs BP 14C yrs BP 14 Russell Island GSC-4002 3820 Ỉ 35 4214 4140 Ỉ 26 S-2662 3822 Ỉ 111 14.7 14.6 0.1 Russell Island GSC-2240 3630 Ỉ 30 3942 3950 Ỉ 27 S-2662 3822 Ỉ 111 14.3 14.6 À0.3 Cape Richard Collinson GSC-4387 4070 Ỉ 30 4558 4410 Æ 26 S-2919 4697 Æ 92 16.5 0.75 Cape Richard Collinson GSC-4343 8680 Ỉ 45 9627 8979 Ỉ 28 S-2964 8702 Ỉ 166 59 57 29 Hollist Point GSC-3962 3660 Ỉ 30 3984 3995 Ỉ 26 S-2589 4302 Ỉ 101 12 12 N/A Hollist Point GSC-3936 8230 Ỉ 55 9200 8544 Ỉ 30 S-2588 9022 Ỉ 136 58.5 58 0.5 29.41 Prescott Island GSC-4503 3470 Ỉ 35 3751 3803 Ỉ 26 S-2921 3462 Ỉ 77 11 Add 500 yrs for m of emergence Guillemard Bay GSC-3989 4400 Ỉ 70 4965 4737 Ỉ 26 S-2861 4652 Ỉ 87 16 18 À2 Guillemard Bay GSC-3989 4400 Æ 70 4965 4737 Æ 26 S-2600 5017 Æ 97 16 17 Foss Fiord GSC-5077 4680 Ỉ 40 5404 5031 Æ 25 S-3345 5257 Æ 92 36.5 Easter Cape GSC-239 940 Ỉ 65 849 1294 Ỉ 25 S-3099 1257 Ỉ 62 Lavoie Point GSC-5428 4170 Ỉ 30 4715 4504 Ỉ 26 S-3427 Lavoie Point GSC-5428 4170 Ỉ 30 4715 4504 Æ 26 Crown Prince Frederick Is GSC-5294 4210 Æ 35 4740 4510 Ỉ 25 C age whale 14 C yrs BP Elevation Emergence Whale age Whale Wood correction rate elevation elevation diff m/ka m m m 17.25 Add 25 yrs for 0.1m of emergence Subtract 75 yrs for 0.3m of emergence Add 150 yrs for 0.75m of emergence Add 69 yrs for m of emergence Corrected whale age 14 C yrs BP Apparent DR Reference 14 C yrs (dates and emergence rates) 3847 Ỉ 111 À293 Ỉ 114 Dyke et al 1991; 1996 3747 Ỉ 111 À203 Ỉ 114 Dyke et al 1991; 1996 4847 Ỉ 92 Dyke et al 1991; 1996 437 Ỉ 95 8771 Ỉ 166 À208 Ỉ 168 Dyke et al 1991; 1996 4302 Ỉ 101 307 Ỉ 105 Dyke et al 1991; 1996 Add 17 yrs for 9039 Ỉ 136 0.5 m of emergence 495 Ỉ 139 Dyke et al 1991; 1996 3975 Ỉ 77 172 Ỉ 81 Dyke et al 1996 Subtract 328 4324 Ỉ 87 yrs for m of emergence À413 Ỉ 90 À1 Subtract 164 4853 Ỉ 96 yrs for m of emergence 116 Ỉ 100 Dyke et al 1991; 1996 37.5 À1 10 Subtract 100 5160 Æ 92 yrs for m of emergence 129 Æ 95 Dyke et al 1996 À1 4057 Ỉ 141 8.5 0.5 3.07 S-3414 4077 Ỉ 131 9.25 0.25 3.07 UCIAMS53044 4815 Ỉ 20 34.5 34.5 3.51 N/A 972 Ỉ 62 À322 Ỉ 67 Subtract 285 yrs for m of emergence Subtract 165 3892 Ỉ 141 À612 Ỉ 143 yrs for 0.5 m of emergence 3996 Ỉ 131 À508 Ỉ 134 Subtract 81 yrs for 0.25m of emergence 4815 Ỉ 20 305 Ỉ 32 Valid? Dyke et al 1991; 1996 M.F.A Furze et al / Quaternary Science Reviews 91 (2014) 218e241 Localitya Dyke et al 1996 Dyke et al 1996 Dyke et al 1996; Dyke 2000 Dyke et al 2011 223 (continued on next page) 224 Table (continued ) Wood lab codeb 14 C age wood 14 C yrs BP Calibrated Equivalent Whale wood age marine09c age lab codeb,d cal yrs BP 14C yrs BP 14 C age whale 14 C yrs BP Elevation Emergence Whale age Whale Wood correction rate elevation elevation diff m/ka m m m Crown Prince Frederick Is GSC-5294 4210 Æ 35 4740 4510 Æ 25 UCIAMS64616 4585 Æ 15 34.5 Crown Prince Frederick Is GSC-5295 5870 Ỉ 35 6694 6248 Ỉ 28 S-3433 Truelove Inlet S-431 5280 Ỉ 100 6070 5660 Ỉ 26 Providence Mountain GSC-5768 3050 Ỉ 25 3283 Owen Point GSC-5771 3750 Ỉ 40 Owen Point GSC-5810 Owen Point Corrected whale age 14 C yrs BP 4430 Æ 15 Apparent DR Reference 14 C yrs (dates and emergence rates) À80 Ỉ 29 37.5 À3 19.35 Subtract 155 yrs for 3m of emergence 6497 Ỉ 131 68 72 À4 25.81 Subtract 155 6342 Ỉ 131 yrs for m of emergence 94 Ỉ 134 Dyke et al 1996 S-432 6247 Ỉ 126 11 11 6247 Ỉ 126 587 Ỉ 129 Dyke et al 1996 3397 Ỉ 27 S-3522 2740 Ỉ 160 12.25 170 Ỉ 162 Dyke et al 1996 4111 4074 Ỉ 25 S-3528 3510 Ỉ 150 16 3350 Ỉ 80 3590 3681 Ỉ 26 S-3528 3510 Æ 150 15 GSC-5815 1400 Æ 25 1309 1778 Æ 26 S-3529 2175 Ỉ 140 Porden Point GSC-5782 3800 Ỉ 80 4192 4130 Ỉ 26 S-3532 3955 Ỉ 150 20 18 Porden Point GSC-5786 3740 Ỉ 30 4097 4071 Ỉ 26 S-3532 3955 Ỉ 150 17.5 Porden Point GSC-5847 2280 Æ 30 2313 2621 Æ 26 S-3533 Porden Point GSC-5811 2250 Ỉ 30 2232 2540 Ỉ 27 Porden Point GSC-5812 2060 Ỉ 30 2029 2398 Ỉ 26 N/A 3.25 3.93 Add 827 yrs for 3.25 m of emergence 3567 Ỉ 160 15 4.51 Add 222 yrs for m of emergence 3747 Ỉ 150 À327 Ỉ 152 15 3510 Ỉ 150 À171 Ỉ 152 N/A Dyke et al 2011 Dyke et al 1996; Dyke 1998 Dyke et al 1996; Dyke 1998 3.52 Subtract 71yrs for 0.25m of emergence 2104 Ỉ 140 326 Ỉ 142 Dyke 1998 4.73 Add 423 yrs for m of emergence 4378 Æ 150 248 Æ 152 Dyke et al 1996; Dyke 1998 18 À0.5 4.73 Subtract 106 3854 Ỉ 150 À217 Æ 152 yrs for 0.5 m of emergence Dyke et al 1996; Dyke 1998 2820 Ỉ 150 13.5 13 0.5 4.73 Add 106 yrs for 0.5 m of emergence 2926 Æ 150 305 Æ 152 Dyke et al 1996; Dyke 1998 S-3533 2820 Ỉ 150 11.75 13 À1.25 4.73 Subtract 264 2556 Ỉ 150 yrs for 1.25 m of emergence 16 Ỉ 152 Dyke et al 1996; Dyke 1998 S-3534 2110 Ỉ 170 3.95 Add 253 yrs for m of emergence 2363 Ỉ 170 À35 Ỉ 172 Dyke et al 1996; Dyke 1998 9.25 9.5 À0.25 Valid? M.F.A Furze et al / Quaternary Science Reviews 91 (2014) 218e241 Localitya GSC-5846 1850 Ỉ 30 1785 2178 Ỉ 27 S-3534 2110 Ỉ 170 8 Porden Point GSC-5816 690 Ỉ 25 661 1109 Ỉ 26 S-3536 1080 Ỉ 150 2.5 2.8 À0.3 3.95 1004 Ỉ 150 À105 Ỉ 152 Subtract 76 yrs for 0.3m of emergence Dyke et al 1996; Dyke 1998 Dyke et al 1996; Dyke 1998 Porden Point GSC-5816 690 Ỉ 25 661 1109 Ỉ 26 S-3559 1425 Ỉ 140 2.5 À0.5 3.95 Subtract 127 1298 Ỉ 140 yrs for 0.5 m of emergence 189 Æ 142 Dyke et al 1996; Dyke 1998 Porden Point GSC-5847 2280 Ỉ 30 2313 2622 Ỉ 26 S-3556 3250 Æ 150 13.25 14 À0.75 4.31 Bere Bay GSC-5805 2090 Æ 30 2062 2424 Æ 26 S-3537 2725 Æ 200 10.5 À2.5 3.83 Dyke et al 1996, Dyke 1998 Dyke et al 1996 Point Refuge GSC-1952 3070 Ỉ 35 3295 3401 Ỉ 27 S-3564 3320 Ỉ 150 16.5 16.5 454 Ỉ 152 Subtract 174 3076 Ỉ 150 yrs for 0.75m of emergence Subtract 653 2072 Ỉ 200 À352 Ỉ 202 yrs for 2.5 m of emergence 3320 Ỉ 150 À81 Ỉ 152 Triton Bay GSC-5952 1760 Ỉ 35 1666 2076 Ỉ 27 S-3568 2260 Ỉ 200 Lovell Point GSC-5861 7780 Ỉ 40 8559 8099 Ỉ 27 S-3598 7850 Ỉ 180 41 40 12.82 Cape Storm GSC-839 4390 Ỉ 30 4945 4715 Ỉ 26 GSC-1021-2 4580 Ỉ 30 17.5 16.5 6.59 Cape Storm GSC-1545 6540 Ỉ 130 7442 6946 Ỉ 27 GSC-1498-1 7260 Ỉ 40 38 38 Murray Bay UCIAMS-42168 6195 Ỉ 20 7081 6550 Ỉ 27 UCIAMS43982 7115 Ỉ 20 33.5 34.5 À1 11.96 Murray Bay UCIAMS-42171 6515 Ỉ 25 7433 6939 Ỉ 27 UCIAMS43982 7115 Æ 20 32 34.5 À2.5 11.96 Murray Bay UCIAMS-42168 6195 Æ 20 7081 6550 Æ 27 UCIAMS44005 6805 Æ 30 33.5 31 2.5 11.96 Murray Bay UCIAMS-42171 6515 Ỉ 25 7433 6938 Ỉ 27 UCIAMS44005 6805 Ỉ 30 32 31 11.96 Murray Bay UCIAMS-42170 5130 Ỉ 20 5906 5528 Æ 27 UCIAMS49808 5195 Æ 20 24.5 22.5 6.67 7.25 0.25 2110 Ỉ 170 N/A N/A 2.14 Dyke et al 1996; Dyke 1998 301 Ỉ 202 Dyke et al 1996; Dyke 1998 Add 78 yrs for 7928 Ỉ 180 À171 Ỉ 182 m of emergence Dyke et al 1996; Dyke 1998 4734 Ỉ 60 19 Æ 65 Blake 1975 7260 Æ 80 314 Æ 84 Blake 1975 7031 Ỉ 20 481 Ỉ 34 Dyke et al 2011 6906 Ỉ 20 À33 Ỉ 34 Dyke et al 2011 7014 Ỉ 30 464 Ỉ 40 Dyke et al 2011 6889 Ỉ 30 À49 Ỉ 40 Dyke et al 2011 5495 Ỉ 20 À33 Ỉ 34 Dyke et al 2011 Add 117 yrs for 0.25m of emergence Add 154 yrs for m of emergence N/A Subtract 84 yrs for 1m of emergence Subtract 209 yrs for 2.5 of emergence Add 209 yrs for 2.5m of emergence Add 84 yrs for m of emergence Add 300 yrs for 2m emergence 2377 Ỉ 200 À68 Ỉ 172 M.F.A Furze et al / Quaternary Science Reviews 91 (2014) 218e241 Porden Point (continued on next page) 225 Table (continued ) Wood lab codeb 14 C age wood 14 C yrs BP Calibrated Equivalent Whale wood age marine09c age lab codeb,d 14 cal yrs BP C yrs BP 14 C age whale C yrs BP Elevation Emergence Whale age Whale Wood correction rate elevation elevation diff m/ka m m m 14 Corrected whale age C yrs BP 14 Apparent DR Reference 14 C yrs (dates and emergence rates) UCIAMS-42169 4620 Æ 20 5422 5038 Æ 25 UCIAMS49808 5195 Æ 20 23 22.5 0.5 6.67 Add 75 yrs for 5270 Ỉ 20 0.5m of emergence 232 Ỉ 32 Dyke et al 2011 Murray Bay UCIAMS-42169 4620 Ỉ 20 5422 5038 Ỉ 25 UCIAMS44004 5205 Ỉ 25 23 21.5 1.5 6.67 Add 225 yrs for 1.5m of emergence 5430 Ỉ 25 392 Æ 35 Dyke et al 2011 Murray Bay UCIAMS-42166 3955 Æ 20 4425 4315 Æ 26 UCIAMS49826 4560 Æ 20 19 19.5 À0.5 6.17 4479 Ỉ 20 Subtract 81 yrs for 0.5 m of emergence 164 Ỉ 33 Dyke et al 2011 Murray Bay UCIAMS-42165 4125 Ỉ 20 4657 4475 Æ 25 UCIAMS49826 4560 Æ 20 18 19.5 À1.5 6.17 À158 Ỉ 32 Dyke et al 2011 Murray Bay UCIAMS-42166 3955 Ỉ 20 4425 4315 Ỉ 26 UCIAMS49810 4960 Ỉ 20 19 20.5 À1.5 6.17 Subtract 243 4317 Ỉ 20 yrs for 1.5m of emergence Subtract 243 4717 Ỉ 20 yrs for 1.5m of emergence 402 Ỉ 33 Dyke et al 2011 Murray Bay UCIAMS-42165 4125 Ỉ 20 4657 4475 Æ 25 UCIAMS49810 4960 Æ 20 18 20.5 À2.5 6.17 Subtract 405 4555 Ỉ 20 yrs for 2.5m of emergence 80 Ỉ 32 Dyke et al 2011 Murray Bay UCIAMS-42173 4065 Ỉ 20 4546 4404 Ỉ 26 UCIAMS43983 4720 Ỉ 15 18 18 N/A 4720 Ỉ 15 316 Ỉ 30 Dyke et al 2011 Murray Bay UCIAMS-42165 4125 Ỉ 20 4657 4475 Ỉ 25 UCIAMS43983 4720 Ỉ 15 18 18 N/A 4720 Ỉ 15 710 Ỉ 30 Dyke et al 2011 Murray Bay UCIAMS-42164 3670 Ỉ 20 4014 4010 Ỉ 26 UCIAMS43983 4720 Ỉ 15 17 18 À1 6.17 Subtract 162 yrs for 1m of emergence 4558 Ỉ 15 83 Ỉ 29 Dyke et al 2011 Murray Bay UCIAMS-42166 3955 Ỉ 20 4425 4315 Ỉ 26 UCIAMS43983 4720 Æ 15 19 18 6.17 872 Æ 30 Dyke et al 2011 Murray Bay UCIAMS-42173 4065 Ỉ 20 4546 4404 Ỉ 26 UCIAMS49806 4200 Ỉ 15 18 17.5 0.5 6.17 4882 Ỉ 15 Add 162 yrs for 1m of emergence Add 81 yrs for 4281 Ỉ 15 0.5 m of emergence À123 Ỉ 30 Dyke et al 2011 Murray Bay UCIAMS-42164 3670 Ỉ 20 4014 4010 Ỉ 26 UCIAMS49806 4200 Ỉ 15 17 17.5 À0.5 6.17 Subtract 81 4119 Æ 15 yrs for 0.5m of emergence 109 Æ 30 Dyke et al 2011 Murray Bay UCIAMS-42164 3670 Ỉ 20 4014 4010 Ỉ 26 UCIAMS43984 4190 Ỉ 15 17 17 4190 Ỉ 15 180 Ỉ 30 Dyke et al 2011 N/A M.F.A Furze et al / Quaternary Science Reviews 91 (2014) 218e241 Murray Bay Valid? 226 Localitya Latitude and longitude for whale bone samples is included in Table S1 Laboratory codes: GSC ¼ Geological Survey of Canada, Ottawa; S ¼ Saskatchewan Research Council, Saskatoon; UCIAMS ¼ WM Keck Carbon Cycle Accelerator Mass Spectrometry Laboratory, Irvine S- and UCIAMS-dates were reported by the laboratory at the 1s confidence interval GSC-dates were reported at the 2s confidence interval, but are here reported as conventional radiocarbon dates (1s) c Marine09 values are from, and interpolated from, Reimer et al (2009) d All S-dates on whale bone were originally reported as non-normalized radiocarbon dates They are reported here as conventional radiocarbon dates, normalized to d13C ¼ À25& according to Stuiver et al (2013) and Coulthard et al (2010) They are also reported in Table S1, this paper a b Dyke & Savelle 2009 97 Ỉ 29 1104 Ỉ 15 Subtract 296 yrs for 1m of emergence 3.38 1007 Ỉ 25 607 King William Island GSC-5820 660 Ỉ 25 UCIAMS-29234 1400 Ỉ 15 À308 Ỉ 30 1968 Ỉ 15 3.52 1400 Ỉ 15 2276 Ỉ 26 1879 King William Island GSC-5819 1930 Ỉ 40 17 UCIAMS49807 UCIAMS29234 4010 Æ 26 4014 Murray Bay UCIAMS-42164 3670 Æ 20 4355 Æ 20 17 N/A Add 568 yrs for 2m of emergence 4355 Ỉ 20 345 Ỉ 33 Dyke et al 2011 Dyke & Savelle 2009 M.F.A Furze et al / Quaternary Science Reviews 91 (2014) 218e241 227 likely been affected by complicating factors including: postdepositional up/down slope remobilization (wood, whale bone); shallow water stranding (bone); deep-water sinking (bone; Dyke et al., 1996b) Once apparent individual whale DR values (DRI) are calculated and screened, an error-weighted mean (DRR) can be calculated using equations (3) to (5), following Coulthard et al (2010) Pn DRIi i ¼ s2 I DRR ¼ Pn (3) i i ¼ s2 I i after Bevington (1969) vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi u u spooled ¼ tPn (4) i ¼ s2 I i after Bevington (1969) vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi u Pn DRI ÀDRR 2 u u n i¼1 sWi u sR ¼ t Pn nÀ1 i¼1 (5) sI i after Bevington (1969) For Equations (3)e(5), sI is the standard deviation for DRI, spooled is the standard deviation of the mean, and sR is the square root of the weighted average variance The larger of spooled or sR is considered the standard deviation of for DRR The internal variability of the data is described by c2 values 3.3.2 Beluga Beluga DR values are based on 12 radiocarbon dates on dentine from four live-harvested whales (Stewart et al., 2006; Campana, pers com 2012; Table 3) Sclerochronology permits radiocarbon dating of individual, annual GLGs of known age, back-calculated from the date of harvesting All used GLG radiocarbon dates predate 1955 and are thus unaffected by bomb radiocarbon Following Coulthard et al (2010), Equation (6) is used to calculate individual dentine DRI, where individual 14C ages are assigned to the midpoint of the dated GLGs and the global average marine reservoir age (Marine09age) for that year is determined from Reimer et al (2009) Equations (3)e(5) are then used to calculate beluga DR DRI ¼ 14 Cage À Marine09age (6) after Stuiver et al (1986) 3.3.3 Walrus The possibility of walrus and driftwood stranding on the same raised shoreline is less likely than for bowheads Walrus have been known to die well inland above contemporaneous sea-level (“crawlers”; Richard and Campbell, 1988; Calvert and Stirling, 1990; Dyke et al., 1999) The only two available direct walrusdriftwood comparisons (Dyke et al., 1996b) indicate an apparent negative DR, i.e their ocean reservoir age is less than that of the global ocean - highly unlikely in the CAA given the restricted ventilation and the regional molluscan DR (335 Ỉ 85 years; Coulthard et al., 2010) A single live-collected (AD 1915) radiocarbon dated walrus bone (K-347; 590 Ỉ 50 14C years; Tauber, 1979; Table S1) from Thule, Greenland, produced an apparent DR of 228 M.F.A Furze et al / Quaternary Science Reviews 91 (2014) 218e241 Table Radiocarbon-dated dentine from beluga teeth used to derive DR All dates were originally reported by Stewart et al (2006), with further data provided by Campana (pers com 2012) Sample codea B92-105 MNR5 B92-105 MNR5 old B92-108 MNR2 B92-108 MNR2 B92-108 MNR2 old B97-037 MNR2 B97-037 MNR2 B97-037 MNR2 old LH91-27 MNR2 LH91-27 MNR2 LH91-27 MNR2 LH91-27 MNR2 old a b NOSAMS lab code (OS-) d13C & 14 14 48673 48657 48752 48788 48794 48820 48833 48671 48790 48795 48821 48796 À14.81 À15.09 À13.97 À14.06 À14.3 À14.35 À14.31 À14.54 À14 À14.09 À13.96 À13.93 755 760 650 690 735 725 645 830 660 675 645 770 Ỉ C age C yrs BP 35 50 35 35 40 30 35 35 30 30 40 30 Year of carbon fixing Years of dated GLGs 1953e1954 1950e1952 1951e1953 1954e1956 1948e1950 1953e1955 1949e1952 1943e1948 1945e1946 1951e1955 1947e1950 1940e1944 1954 1951 1952 1955 1949 1954 1951 1946 1945 1953 1949 1942 Marine09b age C yrs BP Ỉ 473 470 471 474 468 473 470 465 464 472 468 462 24 24 24 24 24 24 24 23 23 24 24 23 14 Ỉ DR 14 C yrs 282 290 179 216 267 252 175 365 196 203 177 308 42 55 42 42 47 38 42 42 38 38 47 38 Sample codes with “old” suffix are the oldest growth layer groups from an individual tooth Marine09 values are from, and interpolated from Reimer et al (2009) 125 Ỉ 55 14C years (assuming 30 year life span and bone collagen carbon fixing on sexual maturity at age w5) The original date was re-calculated by Olsson (1980) as 645 Ỉ 50 14C years, yielding an apparent DR of 180 Ỉ 55 14C years Both values seem excessive given the DRR of west Greenland molluscs (À10 Ỉ 80 14C years; Table S2; McNeely et al., 2006) As Thule is located at the transition between West Greenland waters and Arctic Ocean waters, this single sample may not be diagnostic of either region or may be otherwise incorrect As no other data are available (locally or regionally), we strongly discourage the use of this or any other single measurement as the basis for regional DR (DRR) as this may lead to inaccurate chronologies (cf Coulthard et al., 2010) should be considered a minimum because Beaufort Sea whales overwinter in the Bering and Chukchi Seas where molluscan DR may exceed 450 14C years (McNeely et al., 2006) We additionally analysed six whale boneedriftwood pairs from southwestern Victoria Island (Table 5; Dyke and Savelle, 2000b, 2000c, 2003, 2004) which indicate a DR of w300 14C years may be more appropriate for this region We consider this value to be unacceptably uncertain due to an insufficient number of sample sites and slow emergence rates If this value is correct, whale dates from the Beaufort Sea population calibrated using our (eastern) CAA value would decrease by w140 years Spatio-temporal patterns in calibrated bowhead dates 3.4 d13C and DR results 4.1 Previous work Results of d13C and DR calculations are tabulated in Tables and 4, respectively Bowhead bone collagen produces an average d13C value of 16.1 ặ 1.1& (n ẳ 193) Walrus teeth and tusks yield an average d13C value of À14.8 Æ 1.9& (n ¼ 18), whereas walrus bone collagen shows an average d13C value of 18.0 ặ 4.7& (n ẳ 9) For live-harvested beluga from SE Baffin Island, 44 d13C dentine measurements yield an average of À14.4 Ỉ 0.5& Insufficient data and uncertain data quality preclude the calculation of reliable d13C and DR values for other groups (seal, narwhal; Table S1) DR values (Table 4) are 170 Ỉ 95 14C years for bowheads (n ẳ 23) and 240 ặ 60 14C years for beluga (n ¼ 12) Notably, bowhead DRR would be 175 ặ 360 14C years (n ẳ 57) if the data had not been screened As this average is nearly identical, our screening procedure appears justified Importantly, whereas beluga were sampled from a single restricted area (SE Baffin Island), bowheadedriftwood pairs were collected over a much larger geographic region, encompassing much of the central and eastern CAA For western CAA bowheads (Amundsen Gulf, west/southwest Victoria Island, Prince Patrick and Eglinton islands), our DR value The derivation of a CAA bowhead DR value (170 Ỉ 95 14C years) permits the calibration of B mysticetus dates in the expanded database (Table S1) and thus permits a new spatio-temporal analysis of bowhead whale subfossil distribution as a presumed function of Holocene sea-ice variability Similar previous analyses (in 14 C years; Dyke and Morris, 1990; Dyke et al., 1996a) indicate a bimodal distribution in bowhead remains, interpreted as a fourfold palaeoclimatic division characterized by fluctuating bone abundance, and thus sea-ice conditions centred on M’Clintock Channel and Viscount Melville Sound (Fig 1) Highly abundant remains from the immediate deglacial (!8.5 14C ka BP) are attributed to limited sea-ice due to high-volume meltwater export Critically, the M’Clintock Channel/Viscount Melville Sound plug of multi-year sea-ice, thought to isolate Pacific and Atlantic stocks (Harington, 1966), is considered absent Central CAA bowheads at this time are assumed to be of Pacific stock, Atlantic whales being excluded by a still-glaciated eastern Parry Channel (Dyke et al., 1996a) The rapid decline in remains at w8.5e5 14C ka BP is Table Recommended DRR values for bowhead whales from the Canadian Arctic Archipelago, and for beluga from SE Baffin Island Species Balaena mysticetus Delphinapterus leucas a b c d Number of samples n DRR weighted 23 12 171 242 mean spooled 14 sR c2(n À 1):0.05 c2/n À smeasa C yrs BP 10 12 94 61 86.87 < 33.92 23.19 < 19.68 3.95 2.11 46 42 Population mean spop 186 243 101 61 sextb stotalc smaxd DRR recommended values 89 44 90 46 94 61 Estimated standard error is the pooled mean error multiplied by the square root of n, where n is the number of samples; smeas ¼ spooled*On External variance is found by subtracting measurement variance from total population variance; sext ¼ O(s2pop À s2meas) Uncertainty includes external variance; stotal ẳ O(s2pooled ỵ s2ext) smax ẳ highest of spooled, stotal or sR 14 C yrs BP 170 Æ 95 240 Æ 60 1029 GSC-6542 1140 Æ 50 Cape Back Latitude and longitude for whale bone samples is included in Table S1 Laboratory codes: GSC ¼ Geological Survey of Canada, Ottawa; S ¼ Saskatchewan Research Council, Saskatoon; TO ¼ IsoTrace Laboratory, University of Toronto S- and TO-dates were reported by the laboratory at the 1s confidence interval GSC-dates were reported at the 2s confidence interval, but are here reported as conventional radiocarbon dates (1s) c Marine09 values are from, and interpolated from Reimer et al (2009) d S-dates on whale bone were originally reported as non-normalized radiocarbon dates They are reported here as conventional radiocarbon dates, normalized to d13C ¼ À25& according to Stuiver et al (2013) and Coulthard et al (2010) They are also reported in Table S1, this paper 2878 GSC-6517 2780 Ỉ 60 Cape Back a AA-40409 1790 Ỉ 44 1475 Ỉ 26 680 740 Æ 50 Williams Point GSC-4504 1524 GSC-6373 1630 Æ 80 Cape Baring b AA-40848 4471 Ỉ 43 3115 Ỉ 27 2.5 0.5 1.96 Subtract 1103 yrs for 2.5 m 3368 Ỉ 43 of emergence Add 255 yrs for 0.5 m of 2045 Ỉ 44 emergence 2.27 À2.5 11.5 1.96 À1 S-2995 1145 Ỉ 26 1537 Ỉ 72 1.96 2.5 " 1719 GSC-6370 1790 Ỉ 80 Cape Baring " " 1972 Ỉ 27 5.5 1.96 2.5 " 5.5 1134 GSC-6322 1210 Ỉ 60 Cape Baring " " 2140 Æ 27 1027 Æ 72 " 2614 Æ 80 284 Ỉ 84 Dyke and Savelle, 2000b; 2004 474 Ỉ 84 Dyke and Savelle, 2000b; 2004 642 Ỉ 84 Dyke and Savelle, 2000b; 2004 À118 Ỉ 77 Sharpe 1992; Dyke and Savelle, 2000c, 2004 253 Ỉ 51 Dyke and Savelle, 2003, 2004 570 Ỉ 51 Dyke and Savelle, 2003, 2004 1850 Ỉ 80 Add 510 year for m of emergence Add 1274 years for 2.5 m of emergence Add 1274 years for 2.5 m of emergence Subtract 510 yrs for m of emergence 1.96 1340 Æ 80 TO-8363 1566 Æ 27 Corrected whale Apparent Reference (dates; age 14C yrs BP DR 14C yrs emergence rates) Elevation Emergence Whale age correction Whale C age whale Wood rate m/ka C yrs BP elevation elevation diff m m m 14 14 Whale Lab C age wood Calibrated wood Equivalent Codeb,d C yrs BP age cal yrs BP Marine09c age 14C yrs BP 14 14 Wood lab codeb Localitya Table Analysed whale boneedriftwood pairs from southwestern Victoria Island based on data from Dyke and Savelle (2000b; 2000c; 2003; 2004) Analysis indicates that DR of w300 14C years may be more appropriate for this region though this value is unacceptably uncertain due to an insufficient number of sample sites and slow emergence rates M.F.A Furze et al / Quaternary Science Reviews 91 (2014) 218e241 229 interpreted as an exclusion of whales due to the establishment of Holocene oceanography, resulting in multi-year sea-ice accumulation in southern M’Clintock Channel and Gulf of Boothia A mid-Holocene resurgence in bowhead subfossils is inferred to signal climatic amelioration and a diminished M’Clintock ice plug Admiralty Inlet shows an unprecedented subfossil peak at 4e 3.5 14C ka BP, interpreted as high-density summer bowhead occupation Notably, this fjord system remains connected to southern Prince Regent Inlet via Bernier Bay and Berlinguet Inlet until w2 14C ka BP (Hooper, 1996) due to isostatic depression Subsequent (0, less than the global ocean R), but did not provide an applicable value due to insufficient live-collected samples on which to base calculations Most recently, Dyke et al (2011) suggested a bowhead correction of w400 14C years, or an approximate “delta R” of Ỉ 50 14C years They note that >400 14C years may be excessive, potentially bringing archaeological sites and driftwood dates into conflict with cetacean dates, which may appear “too old” relative to their elevation when projected on regional isostatic sea-level curves Our new bowhead DR of 170 Ỉ 95 14C years seemingly contradicts Dyke et al.’s (2011) conclusions However, it is important to note that the modelled global average marine reservoir effect (R) is time-variant (Stuiver et al., 1986; Reimer et al., 2009) Throughout the mid to late Holocene (w2000e5000 years ago) when most whale bone and archaeological sites occur (Dyke et al., 2011), global R averaged w330 14C years globally, or w500 Ỉ 100 14C years for CAA bowheads This is statistically indistinguishable from Dyke et al.’s (2011) proposed R (w400 Ỉ 50 14C years), and removes this apparent contradiction While it is tempting to always place whale bone on or below a sea-level curve, our data suggest this may not always be warranted Our whaleedriftwood pairs (Table 2) indicate that the likelihood of whale bone recording the age of a shoreline (n ¼ 23) is about equal to the likelihood of upslope movement of bowhead remains (e.g., sea-ice push) or downslope transport of driftwood (e.g., gelifluction), indicated by apparent DR < À10 14C yrs (n ¼ 23) Each of these is approximately twice as common as downslope movement of whales (e.g., gelifluction) or “sinkers”, or upslope transport 232 M.F.A Furze et al / Quaternary Science Reviews 91 (2014) 218e241 of driftwood (e.g., sea-ice push) indicated by apparent DR > 335 14C yrs (n ¼ 11; Table 2) In the Norwegian North Atlantic, Mangerud et al (2006) found that baleen and toothed whales with known age of death showed nearly identical DR values Our CAA dataset demonstrates that baleen whales (bowhead) differ somewhat in DR from toothed whales (beluga), but are statistically indistinguishable at 1s (170 Ỉ 95 vs 240 Ỉ 60 14C years) However, carbon sources and pathways influence the isotopic composition, and hence DR, of a given material Therefore, DR differs between bowhead bone collagen and beluga dentine (e.g., Table 1) due to their different diets, trophic niches, and metabolic pathways (Schell et al., 1989; Hobson et al., 1996; Hobson and Schell, 1998) The known age beluga data provide a useful check on the bowhead results and demonstrate that bowhead data fall within a reasonable age range For beluga, regional and population dynamics need to be further addressed to determine the applicability of our calculated beluga DR to populations elsewhere Dyke et al (1999) discussed the reservoir effect for walrus in Arctic Canada and concluded that walrus likely have a positive reservoir age that is less than that of molluscs In practice, Dyke et al (2011) have applied their bowhead whale reservoir correction to walrus remains This may be problematic in that unlike bowhead whales, which are migratory, walrus tend to spend most of their lives within a single region, thus requiring region-specific walrus DRR values The lack of abundant live-collected walrus remains from which to derive relevant DRR values prevents any truly quantitative analysis Unless more pre-bomb, live-collected walrus dates become available, we recommend against using walrus dates in chronologies We were unable to derive DR for narwhal or seal due to lack of known age collections or of sufficient co-occurring driftwoodebone pairs It may be possible to derive DR based on narwhal tusk sclerochronology if suitably long-lived animals can be obtained Furthermore, given the similarity of narwhal to beluga (diet, potential interbreeding; see x2.2), DR values of these two species may be similar Nevertheless, without additional data, we recommend against using seal or narwhal from the CAA in essential chronological work We note that Dyke et al (2011) assign their mollusc DR value to polar bears (Ursus maritimus) We disagree with this approach, as polar bears derive their carbon from different sources than molluscs, namely seals (Bentzen et al., 2007; COSEWIC, 2008) The lack of definitive data suggests that like narwhal and seal data, radiocarbon dated polar bear should be avoided in calibrated chronologies Furthermore, in the absence of evidence to the contrary, we emphasize that applying DR from one species or material to another may lead to unnecessary uncertainty and error, and may complicate or invalidate otherwise robust chronologies Following Coulthard et al (2010), we specifically stress the crucial importance of using only regional DRR values to calibrate marine radiocarbon dates This is particularly important where individual calibration sites are based on paired driftwood-bone dates such as are applied here The use of individual site-specific DR values for marine mammals, molluscs (e.g., Vickers et al., 2010; Ross et al., 2012) or other marine materials is not a statistically rigorous approach, unnecessarily complicating regional comparisons and potentially leading to erroneous chronologies and correlations We further consider it preferable that chronologies from oceanographically non-analogous intervals (e.g., pre midHolocene CAA) be calibrated using statistically established lateHolocene DRR values (e.g., Coulthard et al., 2010), or simply reported in conventional radiocarbon years, as opposed to poorlyconstrained site-specific corrections If a site-specific correction must be used, it should be reported alongside, and not as a substitute for, the late Holocene value 5.2 Marine mammals as sea-ice indicators Our newly-calibrated dataset expands previous compilations (Dyke and Morris, 1990; Dyke et al., 1996a) and permits direct comparisons with alternatively derived chronologies Since the majority of our datapoints are likely in common with earlier studies,2 our spatio-temporal patterns broadly agree, as should be expected, both at local and Archipelago scales Specifically, the early-Holocene peak and subsequent decline (Fig 4) are in keeping with earlier patterns (Dyke and Morris, 1990; Dyke et al., 1996a) Nevertheless, the overall trends are driven by three key areas (regions 13e15; Figs and 5) Indeed, the pronounced mid-Holocene (w3750 cal ka BP) peak is wholly attributable to Admiralty Inlet (Region 14) However, additional datapoints in our compilation result in an apparently prolonged interval of whale abundance and reduce the palaeoenvironmental significance of the 3e5 14C ka BP peak of Dyke et al (1996a) Using calibrated chronologies coupled with the rigorous selection of appropriate dates (x4.2), results in a more even chronological distribution, attenuating peaks of earlier reconstructions and producing a more regular and persistent spread of infrequent low-abundance remains during “whale free” periods (Fig 4; regions Barrow Strait, 12 Peel Sound) Further minor discrepancies between this study and earlier work are likely due to differences in defining geographic regions 5.3 Cetacean reconstructions versus other proxy data Comparison between whale and other proxy data are limited due to the few detailed, long (deglacialemodern) records from areas with abundant subfossil bowheads Nevertheless, recent palaeoceanographic reconstructions (IP25, dinocyst modernanalogue-technique (MAT) transfer functions, diatoms, multiproxy studies) permit a late-Quaternary perspective of the CAA centred on east-central Parry Channel and the mainland coast (Short et al., 1994; Vare et al., 2009; Belt et al., 2010; Ledu et al.,  kowski et al., 2012, 2013) However, although a pre2010; Pien liminary comparison with whale-derived Holocene palaeoenvironments is possible, attempts to fully reconcile cetacean and non-cetacean interpretations must, at this stage, remain the “fool’s errand” of Dyke and Morris (1990, p.12) Following deglaciation (w10.8 cal ka BP), proxy data suggest significant biological productivity and reduced annual sea-ice in  kowski et al., Barrow Strait (Vare et al., 2009; Ledu et al., 2010; Pien 2012), in keeping with the early Holocene bowhead occupation of deglaciating channels and meltwater-export-driven circulation in the central CAA (Dyke et al., 1996a) Conversely, cetacean-based and other proxy-based interpretations diverge during the subsequent interval (!4.5 cal ka BP) Where microfossil and biomarker data imply ameliorated (but highly fluctuating) conditions associated with climatic warming and the establishment of southeastward oceanic circulation (Short et al., 1994; Vare et al., 2009; Belt et al., 2010; Ledu et al., 2010), reduced bowhead remains are interpreted as an exclusion of whales due to increased sea-ice (Dyke et al., 1996a) Dyke et al (1996a) attribute this interval to the establishment of “modern” CAA circulation and multi-year seaice accumulation in large basins, though earlier oceanic through kowski et al (2012, 2013, 2014) Noneflow is suggested by Pien theless, the trend towards modern sea-ice conditions by the midHolocene indicated by microfossils and biomarkers agrees with the proposed establishment of the M’Clintock ice plug (M’Clintock Channel, Viscount Melville Sound) as a barrier to marine mammals Dyke et al (1996a) not provide a list of utilized dates M.F.A Furze et al / Quaternary Science Reviews 91 (2014) 218e241 Fig Spatio-temporal distribution of Canadian Arctic Archipelago (CAA) bowhead whale (Balaena mysticetus) subfossil remains Whale frequencies are plotted in 250 calibrated year age bins according to region (Fig 2) 233 234 M.F.A Furze et al / Quaternary Science Reviews 91 (2014) 218e241 Fig Detailed spatio-temporal distribution of bowhead whale (Balaena mysticetus) subfossil remains from regions 13 Prince Regent Inlet & Gulf of Boothia, 14 Admiralty Inlet, and 15 Berlinguet Inlet/Bernier Bay These areas and their constituent subregions (Fig 3) contribute the majority of whale remains throughout the Holocene M.F.A Furze et al / Quaternary Science Reviews 91 (2014) 218e241 Increased subfossil abundances at 4.5e2.9 cal ka BP (z5.0e 3.0 14C ka BP; Dyke et al., 1996a), considered indicative of favourable conditions, are at odds with microfossil- and biomarker-based studies which suggest increasing sea-ice (Vare et al., 2009;  kowski et al., 2012, 2013) MAT-based dinocyst data (Ledu Pien et al., 2010), meanwhile, point to decreased seasonal sea-ice concentration and extent favourable to increased marine mammal densities Similar contradictions are seen in the southwest, where IP25 and dinocyst MAT records (from the same core) disagree (Belt et al., 2010), at a time when bowheads are considered excluded from Amundsen Gulf (Dyke and Savelle, 2001) Such divergent proxy records, coupled with differing chronostratigraphic approaches (age depth-model construction, marine reservoir effect assumptions) hinder comparisons with marine mammal-based environmental histories None of the records from this period can confidently provide summer sea-ice reconstructions specific enough to accurately predict sea-mammal occupation (given that bowheads will occupy to >95% summer sea-ice; Moore et al., 2000; Ferguson et al., 2010) Nonetheless, the highest frequency of Holocene whale remains seen throughout the CAA (Region 14 Admiralty Inlet), interpreted as highly ameliorated conditions, is difficult to reconcile with nearby proxy records showing limited (though decreasing, stable, or marginally increasing) sea-ice The late-Holocene (2.9e0 cal ka BP) of Dyke et al (1996a) is characterized by spatial variations in whale bone abundances, interpreted as a function of regionally disparate sea-ice regimes Only within M’Clintock Channel and environs (Viscount Melville Sound, Franklin Strait, Victoria Strait) is a continuous late-Holocene sea-ice increase proposed, being considered an expansion of the M’Clintock ice plug to late Little Ice Age (LIA) dimensions Elsewhere, sea-ice increase is either interrupted by amelioration or absent Cetacean-based reconstructions from Wellington Channel (Dyke et al., 1996a) and Belcher Channel/Norwegian Bay indicative of conditions favourable for whale occupation broadly agree with Barrow Strait proxy records that suggest an amelioration  kowski et al., 120 cm; Ferguson et al., 2010) Dyke et al (1996a) consider the alternative e that bone-barren intervals indicate occupation with low or zero mortality e to be flawed due to a lack of mechanism to explain sudden mortality rate changes However, given the low subfossil frequencies throughout the CAA (Fig 4), chronological trends in remains fail to demonstrate abrupt or pronounced changes Simultaneously, regions known to support large modern and historic seasonal bowhead populations (West Greenland and Baffin Island coasts; Reeves et al., 1983; Rugh et al., 2003) show rare remains The assumption that absent remains indicate absent animals fails to accommodate the effects of mid- to late-Holocene isostatic transgression effectively removing older remains from the terrestrial record or the effects of precipitous coastlines preventing carcass stranding 236 M.F.A Furze et al / Quaternary Science Reviews 91 (2014) 218e241  Sea-ice entrapment is the primary mortality cause prior to commercial harvesting The assumed relationship between subfossil frequencies and sea-ice thickness and distribution implies enhanced mortality rates at the ice edge, likely due to sea-ice entrapment The high contribution of sea-ice entrapment to bowhead death assemblages has been demonstrated based on subfossil sizeeage profiles in regions of high bone concentration (e.g., Admiralty Inlet; Savelle et al., 2000), though such an approach is problematic in other areas low in bones Additionally, senescence, contrary to entrapment, is considered the primary mortality cause in the cetacean literature (Philo et al., 1993; George et al., 1999) Ice entrapment may only be common under particular physiographic, climatological, oceanographic, or ecological circumstances  Holocene mortality rates are constant (prior to harvesting by humans