We use a core of salt-marsh sediment from Boston Harbor (Massachusetts, USA) to evaluate the sensitivity of late Holocene relative sea-level (RSL) reconstructions to the pre-treatment, graphitization, and upcore distribution of radiocarbon dates. The 87 radiocarbon dates from 58 unique depths in the 4.2-m long core show that ages from plant macrofossils (principally rhizomes and stems of Spartina patens and Distichlis spicata) are insensitive to pretreatment and graphitization regimes. Statistical resampling from the pool of radiocarbon dates generated many plausible chronologies and demonstrates that the precision of age-depth models increases as dates are added, but with diminishing returns. Estimated sample ages are not systematically biased by the density of dates. Identification of periods with slower sedimentation requires a higher density of dates than identifying periods with faster sedimentation. These results suggest that RSL variability reconstructed among sites and regions is unlikely to be the result of choices in the preparation and selection of samples from salt-marsh sediment for radiocarbon dating. We reconstructed paleomarsh elevation using a Bayesian transfer function trained on the observed relationship between salt-marsh foraminifera and tidal elevation in 212 surface-sediment samples. This model utilized informative prior information from bulk-sediment delta 13C values. During the past -4200 years RSL rose by -4.2 m in Boston Harbor at a rate of -0.9 mm/yr until the early 20th century when the rate increased to -3 mm/yr, which is consistent with estimates of glacio-isostatic adjustment and historic tide-gauge measurements in Boston Harbor.

We review geochronological data relating to the timing and rate of Laurentide Ice Sheet recession in the northeastern United States and model ice margin movements in a Bayesian framework using compilations of previously published organic 14 C ( n = 133) and in situ cosmogenic 10 Be ( n = 95) ages. We compare the resulting method‐specific chronologies with glacial varve records that serve as independent constraints on the pace of ice recession to: (1) construct a synthesis of deglacial chronology throughout the region; and (2) assess the accuracy of each chronometer for constraining the timing of deglaciation. Near the Last Glacial Maximum terminal moraine zone, 10 Be and organic 14 C ages disagree by thousands of years and limit determination of the initial recession to a date range of 24–20 ka. We infer that 10 Be inherited from pre‐glacial exposure adds 2–6 kyr to many exposure ages near the terminal moraines, whereas macrofossil 14 C ages are typically 4–8 kyr too young due to a substantial lag between ice recession and sufficient organic material accumulation for dating in some basins. Age discrepancies between these chronometers decrease with distance from the terminal moraine, due to less 10 Be inherited from prior exposure and a reduced lag between ice recession and organic material deposition. 14 C and 10 Be ages generally agree at locations more than 200 km distal from the terminal moraines and suggest a mostly continuous history of ice recession throughout the region from 18 to 13 ka with a variable pace best documented by varves.

Abstract Deposits of at least three glaciations are present in New Jersey and the New York City area. The oldest deposits are magnetically reversed. Pollen and stratigraphic relations suggest that they are from the earliest Laurentide advance at ~2.4 Ma. Deposits of a second advance are overlain by peat dated to 41 ka and so are pre-Marine Isotope Stage (pre-MIS) 2. Their relation to marine deposits indicates that they predate MIS 5 but postdate MIS 11 and may postdate MIS 7 or 9, suggesting an MIS 6 age. The most recent deposits are of MIS 2 (last glacial maximum [LGM]) age. Radiocarbon dates and varve counts tied to glacial-lake events indicate that LGM ice arrived at its terminus at 25 ka, stood at the terminus until ~24 ka, retreated at a rate of 80 m/yr until 23.5 ka, and then retreated at a rate of 12 m/yr to 18 ka. At 18 ka the retreat record connects to the base of the North American Varve Chronology at Newburgh, New York. The 25–24 ka age for the LGM is slightly younger than, but within the uncertainty of, cosmogenic ages; it is significantly older than the oldest dated macrofossils in postglacial deposits in the region.

We attempt to synchronize the North American Varve Chronology (NAVC) with ice core and calendar year timescales by comparing records of atmospherically produced 10Be fallout in the NAVC and in ice cores. The North American Varve Chronology (NAVC) is a sequence of 5659 varves deposited in a series of proglacial lakes adjacent to the southeast margin of the retreating Laurentide Ice Sheet between approximately 18 200 and 12 500 years before present. Because properties of NAVC varves are related to climate, the NAVC is also a climate proxy record with annual resolution, and our overall goal is to place the NAVC and ice core records on the same timescale to facilitate high-resolution correlation of climate proxy variations in both. Total 10Be concentrations in NAVC sediments are within the range of those observed in other lacustrine records of 10Be fallout, but 9Be and 10Be concentrations considered together show that the majority of 10Be is present in glacial sediment when it enters the lake, and only a minority of total 10Be derives from atmospheric fallout at the time of sediment deposition. Because of this, an initial experiment to determine whether or not 10Be fallout variations were recorded in NAVC sediments by attempting to observe the characteristic 11-year solar cycle in short varve sections sampled at high resolution was inconclusive: short-period variations at the expected magnitude of this cycle were not distinguishable from measurement scatter. On the other hand, longer varve sequences sampled at decadal resolution display centennial-period variations in reconstructed 10Be fallout that have similar properties as coeval 10Be fallout variations recorded in ice core records. These are most prominent in glacial sections of the NAVC that were deposited in proglacial lakes and are suppressed in paraglacial sections of the NAVC that were deposited in lakes lacking direct glacial sediment input. We attribute this difference to the fact that buffering of 10Be fallout by soil adsorption can filter out short-period variations in an entirely deglaciated watershed, but such buffering cannot occur in the ablation zone of an ice sheet. This implies that proglacial lakes whose watershed is mostly glacial may effectively record 10Be fallout variations. We attempted to match centennial-period variations in reconstructed 10Be fallout flux from two segments of the NAVC with ice core fallout records. For both records, it is possible to obtain matches that result in acceptable correlation between NAVC and ice core 10Be fallout records, but the best-fitting matches for the two segments disagree, and only one of them is consistent with independent calendar year calibrations of the NAVC and therefore potentially valid. This leaves several remaining ambiguities in whether or not 10Be fallout variations can, in fact, be used for synchronizing NAVC and ice core timescales, but these could most likely be resolved by higher-resolution and replicate 10Be measurements on targeted sections of the NAVC.

Varves – annual sediment layers, common in glacial lakes – are an important source of paleoclimate information. Manually recording their occurrence, typically by visual inspection, can be both time-consuming and prone to error, leading to several attempts in recent years to at least partially computerize the process. However, existing computerized methods of varve detection still require moderate to large amounts of user interaction — they are semi-automated, rather than fully automated. In light of that, this paper is a step towards fully automatic detection of varves. The presented program, DeepVarveNet - a glacial varve detector built on a convolutional neural network, is designed to automatically delineate annual layers in such lacustrine sediment in digital images of photographed sediment cores. To the best of the authors’ knowledge, this is the first approach that applies a convolutional neural network to this task. The performance of DeepVarveNet was assessed on a data set comprising images from seven sediment coring sites, of varying sedimentological properties. They represent three northeast U.S. glacial paleolakes, and glacial paleolake Ojibway. Our testing set contained 1415 identified varves, on which DeepVarveNet demonstrated sensitivity at a level of 0.986 and precision equal to 0.834, exceeding that of BMPix and ANFIS, the existing semi-automated varve identifiers.

The North American Ice Sheet Complex (NAISC; consisting of the Laurentide, Cordilleran and Innuitian ice sheets) was the largest ice mass to repeatedly grow and decay in the Northern Hemisphere during the Quaternary. Understanding its pattern of retreat following the Last Glacial Maximum is critical for studying many facets of the Late Quaternary, including ice sheet behaviour, the evolution of Holocene landscapes, sea level, atmospheric circulation, and the peopling of the Americas. Currently, the most up-to-date and authoritative margin chronology for the entire ice sheet complex is featured in two publications (Geological Survey of Canada Open File 1574 [Dyke et al., 2003]; 'Quaternary Glaciations - Extent and Chronology, Part II' [Dyke, 2004]). These often-cited datasets track ice margin recession in 36 time slices spanning 18 ka to 1 ka (all ages in uncalibrated radiocarbon years) using a combination of geomorphology, stratigraphy and radiocarbon dating. However, by virtue of being over 15 years old, the ice margin chronology requires updating to reflect new work and important revisions. This paper updates the aforementioned 36 ice margin maps to reflect new data from regional studies. We also update the original radiocarbon dataset from the 2003/2004 papers with 1541 new ages to reflect work up to and including 2018. A major revision is made to the 18 ka ice margin, where Banks and Eglinton islands (once considered to be glacial refugia) are now shown to be fully glaciated. Our updated 18 ka ice sheet increased in areal extent from 17.81 to 18.37 million km(2), which is an increase of 3.1% in spatial coverage of the NAISC at that time. Elsewhere, we also summarize, region-by-region, significant changes to the deglaciation sequence. This paper integrates new information provided by regional experts and radiocarbon data into the deglaciation sequence while maintaining consistency with the original ice margin positions of Dyke et al. (2003) and Dyke (2004) where new information is lacking; this is a pragmatic solution to satisfy the needs of a Quaternary research community that requires up-to-date knowledge of the pattern of ice margin recession of what was once the world's largest ice mass. The 36 updated isochrones are available in PDF and shapefile format, together with a spreadsheet of the expanded radiocarbon dataset (n = 5195 ages) and estimates of uncertainty for each interval. (C) 2020 Elsevier Ltd. All rights reserved.

Abstract Recession of the Laurentide Ice Sheet from northern New Hampshire was interrupted by the Littleton-Bethlehem (L-B) readvance and deposition of the extensive White Mountain Moraine System (WMMS). Our mapping of this moraine belt and related glacial lake sequence has refined the deglaciation history of the region. The age of the western part of the WMMS is constrained to ~14.0–13.8 cal ka BP by glacial Lake Hitchcock varves that occur beneath and above L-B readvance till and were matched to a revised calibration of the North American Varve Chronology presented here. Using this age for when boulders were deposited on the moraines has enabled calibration of regional cosmogenic-nuclide production rates to improve the precision of exposure dating in New England. The L-B readvance coincided with the Older Dryas (OD) cooling documented by workers in Europe and the equivalent GI-1d cooling event in the Greenland Ice Core Chronology 2005 (GICC05) time scale. The readvance and associated moraines provide the first well-documented and dated evidence of the OD event in the northeastern United States. Our lake sediment cores show that the Younger Dryas cooling was likewise prominent in the White Mountains, thus extending the record of this event westward from Maine and Maritime Canada.

Deltas formed in Lake Hitchcock, a glacial lake that developed in the Connecticut River Valley, New England, between ∼18.3 and 12.5 ka. The heights of topset/foreset contacts of these deltas presently increase northward, linearly, at rate of ∼0.9 m/km. Others have interpreted this as indicating that isostatic rebound did not begin until after the lake drained, several kiloyears after glacial retreat began. However, (non-elastic) adjustment of Earth's lithosphere to changing loads is known to occur on time scales of years. Late-glacial shoreline features elsewhere in New England also increase in elevation with distance from the LGM margin at ∼0.9 m/km, suggesting that this is a result of fundamental properties of the crust and mantle, and independent of the history of glacier retreat. On the basis of a numerical model of flexure of the lithosphere beneath a circular load, we suggest that deflection of the lithosphere is remarkably linear in a zone 50–200 km wide between the retreating ice margin and a forebulge, and that initial rebound of this zone is spatially quite uniform for some kiloyears before differential rebound starts. Thus, lake shorelines, formed over a period of some centuries during deglaciation would, today, rise linearly northward.

1. Center for Earth and Environmental Science, SUNY Plattsburgh, 101 Broad Street, Plattsburgh, NY 12901, 518.564.4033, franzida@plattsburgh.edu 2. Department of Earth and Ocean Sciences, Tufts University, Lane Hall, Medford, MA 02155, 617.627.3494, Jack.Ridge@tufts.edu 3. Hanley Sustainability Institute and College of Arts and Sciences, University of Dayton, 300 College Park, Dayton, OH 45469, 937.229.3295, dpair1@udayton.edu 4. DeSimone Geoscience Investigations, Petersburg, NY 12138, 518.686.9809, djdesimone@gmail.com 5. Department of Geological Science, SUNY New Paltz, 1 Hawk Drive, New Paltz, NY 12561, 845.257.3767, rayburnj@newpaltz.edu 6. Department of Geology, SUNY Cortland, P.O. Box 2000, Cortland, NY 13045, 607.753.2921, david.barclay@cortland.edu POST-VALLEY HEADS DEGLACATION OF THE ADIRONDACK MOUNTAINS AND ADJACENT LOWLANDS

Fluvial, glacial, and estuarine deposits in the Delaware Valley record the response of the Delaware River to glaciation, sea-level change, and glacioisostasy during the Quaternary. Incision following an early Pleistocene glaciation created the present valley, which is inset into a Pliocene strath and fluvial plain. Middle and upper Pleistocene and Holocene deposits were laid down in this inset valley. Estuarine terraces in the lower valley and bayshore at +20m (probably Marine Isotope Stage [MIS] 11), +8m (MIS 5e), and +3m (MIS 5a or c), and a fluvial deposit that correlates to offshore MIS 3 marine deposits at −20m are at elevations consistent with glacioisostatic models. Successive incisions during lowstands in the middle and late Pleistocene lengthened, deepened, and narrowed the channel in the lower valley and shifted the channel westward in Delaware Bay. During MIS 2 glaciation, from 25 to 18ka, the Delaware was diverted to the Hudson Shelf Valley by glacioisostatic tilting. Most glacial sediment was trapped in fluvial-lacustrine valley fills north of the terminal moraine. Incision of the valley fill was accomplished during the early stage of rebound, between 17 and 12ka. Drainage to the Delaware shelf was restored between 15 and 13ka as the forebulge collapsed. During incision, multiple postglacial terraces formed where the valley was perpendicular to rebound contours and so was steepened and elevated northward; and a single terrace formed where the valley paralleled the contours, and there was no differential elevation or steepening. About 65% of the original volume of MIS 2 glacial sediment remains in the main valley, and most of the eroded volume is in the channel in the lower valley beneath Holocene estuarine fill. Little glacial sediment reached the Delaware or Hudson shelf. Overbank deposition on the lower postglacial terrace and modern floodplain spans the Holocene. The volume of Holocene sediment in the estuary and bay yields a basinwide denudation rate of about 20m/my.

Click to increase image sizeClick to decrease image size AcknowledgementsThe Varve Working Group has been supported by IGBP-PAGES. The VWG Workshop in Manderscheid received additional support from the Deutsche Forschungsgemeinschaft (DFG) and was organized by Bernd Zolitschka. We thank Arndt Schimmelmann for proofreading the manuscript.

New glacial varve records from long cores combined with records from key surface exposures and new radiocarbon ages have allowed the correction, consolidation, expansion, and calibration of Ernst Antevs' original New England Varve Chronology (NEVC) in the Connecticut Valley of New England, U.S.A. The varve records have been reformulated, with corrections and a new numbering system, as the new North American Varve Chronology (NAVC), which is a continuous 5659-yr varve sequence that spans most of the last deglaciation (18,200-12,500 yr BP) in the northeastern United States. Rates of ice recession for separate intervals terminated by abrupt glacial stillstands and readvances have been determined for western New England. Ice recession history is coupled to varve thickness changes that depict changes in meltwater production in the Connecticut Valley and show the relationship of changes in ablation rate (summer climate variation) to glacial readvances and periods of halted and rapid ice recession (up to 300 m/yr). Comparison of varve thickness records to Greenland ice-core climate records show that after 15,000 yr BP, climate changes of sub-century and longer scales recorded in both records appear identical and synchronous. After 15,000 yr BP, therefore, there was a link between North Atlantic climate and marginal processes of the southeastern sector of the Laurentide Ice Sheet (LIS). Prior to 15,000 yr BP, when the LIS was closer to an equilibrium condition, retreat rates were generally lower and changes in varve thickness and ablation were more subtle, but can still be linked to ice sheet activity. Only weak relationships between varve thickness changes and Greenland climate are evident suggesting that changes in the southeastern LIS during this time may have been significantly influenced by climate patterns unique to the North American continent or ice dynamics.

Glacial environments are subject to drastic oscillations in energy regime that rapidly modify the local environment. The impact of glacial phenomena on the distribution, abundance, and evolution of biota based on trace-fossil evidence is the focus of this chapter. Arthropod trackways, shallow horizontal burrows, and fish trails dominate the glacial and periglacial trace-fossil assemblages preserved in terrestrial and glaciolacustrine sedimentary sequences, while nonspecialized feeding burrows that are diminutive when compared to normal-marine settings comprise glaciomarine ichnofaunas. As a consequence of rapid meltwater discharge, freshwater conditions prevail in some fjord settings during deglaciation, allowing for the establishment of suites more typical of freshwater or brackish-water conditions. Despite changes in the composition of the trace making community through time, ichnofacies relationships and ecological niche occupation are similar between the Paleozoic and Cenozoic, an indication of the constancy of the interplay between the biotic community and glacial processes.

Trace fossils as paleoecological and paleobiogeographical tools in Pleistocene glaciolacustrine sediments have been largely ignored. Combining high resolution varve stratigraphy with trace fossil data can lead to refined paleoenvironmental interpretations during times of rapid climate change. Based on trace fossils from glaciolacustrine varves in the Connecticut and Merrimack River valleys, a new timeline for the reinhabitation of New England by fish and associated invertebrates has been constructed. In addition, it appears that colonization of the recently deglaciated environments occurred in at least four successional stages. If these stages can be recognized elsewhere, it may be possible to accurately model other postglacial migrations using trace fossils and varve chronostratigraphy.

New trace fossil material from Late Pleistocene glaciolacustrine varves of the Connecticut River Valley, New England, USA represent the first evidence of freshwater sculpin in glacial Lake Hitchcock. Paleobiogeographic data constrain the timing of the last reinhabitation of freshwater cottids from a Wisconsinan glacial refugium. Freshwater sculpin were present in the area of study by similar to 13.7 ky BP, moving approximately 400 km in 5000 yr, and following the ice margin at distances as close as 35 km. The trace fossils warrant erection of a new ichnospecies, Broomichnium flirii isp. nov. Comparison of this new ichnospecies to Broomichnium permianum reveals distinct similarities, and it is possible that the Permian examples of B. permianum also were made by fish, which would reconcile a long-running controversy. Many groups of fish are demersal and make ventral body contact with the substrate and could potentially leave similar traces. Identifying new forms of trace fossils made by fish that use alternative modes of locomotion will prove useful in paleoenvironmental interpretations. (C) 2008 Elsevier B.V. All rights reserved.

New trace fossils found in the Late Pleistocene glaciolacustrine varves of the Connecticut River Valley, Vermont, USA represent the first known notostracan presence in glacial Lake Hitchcock. These unique trace fossils warrant a new ichnogenus and ichnospecies Surculichnus bifurcauda. The New England Varve Chronology (NEVC) constrains the initial presence of S. bifurcauda at similar to 13.3-13.2 kyr. The morphology of S. bifurcauda correlates well with notostracan characteristics and behavior. Sieving of bedding planes that contained S. bifurcauda produced one chitinous fossil that is suggestive of a notostracan telson. Neoichnological experimentation was conducted with the species Triops longicaudatus. During the subadult stage, T longicaudatus produced traces representative of locomotion and feeding behaviors, and at the adult stage reproduced S. bifurcauda, as well as Rusophycus. Possible explanations for the productions of these traces are egg laying or predation behaviors that are both related to maturity. Further research in the paleo- and modern ecology of glacial and nearctic lakes may shed more light on the maker of S. bifurcauda. (C) 2008 Elsevier B.V. All rights reserved.

The major uncertainty in relating cosmogenic-nuclide exposure ages to ages measured by other dating methods comes from extrapolating nuclide production rates measured at globally scattered calibration sites to the sites of unknown age that are to be dated. This uncertainty can be reduced by locating production rate calibration sites that are similar in location and age to the sites to be dated. We use this strategy to reconcile exposure age and radiocarbon deglaciation chronologies for northeastern North America by compiling 10Be production rate calibration measurements from independently dated late-glacial and early Holocene ice-marginal landforms in this region. 10Be production rates measured at these sites are 6–12% lower than predicted by the commonly accepted global 10Be calibration data set used with any published production rate scaling scheme. In addition, the regional calibration data set shows significantly less internal scatter than the global calibration data set. Thus, this calibration data set can be used to improve both the precision and accuracy of exposure dating of regional late-glacial events. For example, if the global calibration data set is used to calculate exposure ages, the exposure-age deglaciation chronology for central New England is inconsistent with the deglaciation chronology inferred from radiocarbon dating and varve stratigraphy. We show that using the regional data set instead makes the exposure age and radiocarbon chronologies consistent. This increases confidence in correlating exposure ages of ice-marginal landforms in northeastern North America with glacial and climate events dated by other means.

Over the last 20 years there have been major improvements to our understanding of the Quaternary glaciation of New England. Numerous Accelerator Mass Spectrometer (AMS) 14C ages of terrestrial plant fossils have eliminated some of the errors associated with 14C ages of lake-bottom bulk sediment samples and marine fossils. The rebirth of Antevs' New England varve chronology and its 14C calibration have added precision to the Late Wisconsinan deglaciation chronology and a means of precisely testing correlations of deglacial events. Palaeomagnetic studies of both declination and polarity have allowed precise correlation between New England and adjacent areas. The exact chronology of pre-Late Wisconsinan glaciation is still poorly constrained because of difficulties in determining numerical ages. The terrestrial glacial record of New Jersey dictates that New England was glaciated at least twice in pre-Wisconsinan time, at least once when the geomagnetic field had a reversed polarity (pre-Illinoian, Marine Isotope Stage (MIS) 22 or older, 850 ka) and again during the Illinoian (160–180 ka BP, MIS 6). Controversy exists concerning pre-Late Wisconsinan till in central and northern New England with its weathering supporting an Illinoian age and amino acid ages from reworked marine fossils in Boston Harbor supporting an Early Wisconsinan age. In the Connecticut Valley of southern New Hampshire and Vermont the advance of Late Wisconsinan ice buried advance outwash and lake beds were deposited in tributaries impounded by advancing ice. Late Wisconsinan ice reached its limit on islands along the southern coast at 24.0-20.0 14C ka BP (28.0-23.7 cal ka BP). The overall pattern of deglaciation was one of spurts of ice recession punctuated by readvances and end moraine building, and acceleration of ice recession over time. The early chronology of deglaciation (20.0-15.0 14C ka BP, 23.7-18.0 cal ka BP) has been inferred from correlations to Greenland ice core records and varve sequences have been used to crudely constrain the chronology. The later part of deglaciation (15.0-11.5 14C ka BP, 18.0-13.4 cal ka BP) has been determined precisely where basal ice-proximal varves are matched to the 14C-calibrated New England varve chronology in the Connecticut, Merrimack, Passumpsic, and Winooski Valleys. The resulting precise chronology of deglaciation in central and northern New England is about 1500 years younger than in previous models largely based on bulk sediment lakebottom 14C ages. A comparison of terrestrial 14C ages for ice margins in the Merrimack Valley with marine 14C ages for contemporaneous marine ice margins in southern Maine suggest that a 600-1300-yr marine reservoir correction should be applied to the marine chronology. Correlations between New England and New York have been formulated with the matching of independent palaeomagnetic declination records from both areas and the connection of icefront positions that appear to align geographically. There appears to be a contemporaneity of readvances in both regions that also match cold intervals on Greenland ice core records. These observations support a rapid response of the last ice sheet's ablation system to cold events. Following deglaciation a delayed isostatic tilting occurred across central to northern New England with water planes dipping 0.85–0.94 m/km towards the south-south-east. The nonglacial remnants of glacial lakes in the upper Connecticut Valley persisted until at least 10.5 14C ka BP 12.5 cal ka BP) and may have been seen by the first humans in the area. While isostatic uplift was well underway in southern and central New England recession of ice in Canada allowed the invasion of marine water into the Champlain Basin at about 11.0 14C ka BP (13.0 cal ka BP). This age estimate is consistent with previous studies that attempted to remove a reservoir error from marine 14C ages of Champlain Sea fossils.