Archive color photographs-particularly of sites where sediment is no longer available-are an underused resource that potentially contains detailed paleoenvironmental information. To investigate this potential, two sets of digital images were taken, at different times, of loess, glaciolacustrine, and deep-sea sediments. The first image set was taken using standard digital cameras. Lighting conditions and sediment surface preparation varied, in a similar way to characteristics likely to be encountered in archive photographs. The second image set was taken by a high-resolution, line-scan camera with an integrated light source. CIELAB (Commission Internationale de l'Eclairage) color reflectance data were obtained from both image sets and analyzed by cross correlation. Of the three reflectance parameters (L*, a*, and b*), L* reflectance is the most compromised by differences in ambient lighting or moisture content. Textural distinctions appear to be an important factor influencing the cross correlations and produce multiple, relatively weak solutions for the glaciolacustrine sediments, whereas the texturally uniform loess and deep-sea sediments produce a single, best-fit solution. Comparison of a* reflectance records from an undated marine sequence with a chronologically constrained sequence from the same site indicates the potential to apply color reflectance to produce preliminary age models.

Dating evidence for MIS 3 sites cited in the text

While there is extensive evidence for the Late Devensian, less is known about Early and Middle Devensian (approx. 110–30 ka) climates and environments in the UK. The Greenland ice-core record suggests the UK should have endured multiple changes, but the terrestrial palaeo-record lacks sufficient detail for confirmation from sites in the British Isles. Data from deposits at Finningley, South Yorkshire, can help redress this. A channel with organic silts, dated 40 314–39 552 cal a BP, contained plant macrofossil and insect remains showing tundra with dwarf-shrub heath and bare ground. Soil moisture conditions varied from free draining to riparian, with ponds and wetter vegetated areas. The climate was probably low arctic with snow cover during the winter. Mutual climatic range (MCR), based on Coleoptera, shows the mean monthly winter temperatures of −22 to −2°C and summer ones of 8–14°C. Periglacial structures within the basal gravel deposits and beyond the glacial limits indicate cold-climate conditions, including permafrost. A compilation of MCR reconstructions for other Middle Devensian English sites shows that marine isotope stage 3—between 59 and 28 ka—experienced substantial variation in climate consistent with the Greenland ice-core record. The exact correlation is hampered by temporal resolution, but the Finningley site stadial at approximately 40 ka may correlate with the one of the Greenland stadials 7–11.

Our understanding of how global climatic changes are translated into ice-sheet fluctuations and sea-level change is currently limited by a lack of knowledge of the configuration of ice sheets prior to the Last Glacial Maximum (LGM). Here, we compile a synthesis of empirical data and numerical modelling results related to pre-LGM ice sheets to produce new hypotheses regarding their extent in the Northern Hemisphere (NH) at 17 time-slices that span the Quaternary. Our reconstructions illustrate pronounced ice-sheet asymmetry within the last glacial cycle and significant variations in ice-marginal positions between older glacial cycles. We find support for a significant reduction in the extent of the Laurentide Ice Sheet (LIS) during MIS 3, implying that global sea levels may have been 30–40 m higher than most previous estimates. Our ice-sheet reconstructions illustrate the current state-of-the-art knowledge of pre-LGM ice sheets and provide a conceptual framework to interpret NH landscape evolution.

Mutual Climatic Range (MCR) calculation results for MIS 3 beetle based climate reconstrutions shown in figure 16. Data include comparison with modern day termperatures for Finningley.

Proglacial Lake Humber formed in the Vale of York and Ancholme Valley during the Late Devensian (Weichselian) glaciation, but its lake levels and their precise ages are uncertain. Three-dimensional geological modelling, based on 193 borehole sediment logs from the eastern part of the Vale of York, indicates that glaciolacustrine sediments extend no higher than 10 m O.D. By contrast, recent palaeoenvironmental reconstructions for the region that suggest Lake Humber had eight recessional shorelines, extending from 42 m to 5 m O.D. Above 10 m O.D., the sediments become more discontinuous, and comprise clay with occasional chalk and flint gravel, and matrix-rich diamicton interdigitated with sands and gravels. Sedimentary and geochemical analyses of sands and gravels from one of the putative shorelines, at 25 m O.D., indicate an easterly provenance for these sediments. They are interpreted here as colluvial deposits, sourced largely by periglacial weathering of Jurassic and chalk bedrock. Collectively, the geological evidence suggests that the highest level of Lake Humber during the Late Devensian did not exceed 10 m O.D., and therefore reconstructions invoking higher lake levels are thought to be unlikely. (C) 2018 The Geologists' Association. Published by Elsevier Ltd. All rights reserved.

Uncertainty about the geological processes that deposited syngenetically frozen ice-rich silt (yedoma) across hundreds of thousands of square kilometres in central and northern Siberia fundamentally limits our understanding of the Pleistocene geology and palaeoecology of western Beringia, the sedimentary processes that led to sequestration of hundreds of Pg of carbon within permafrost and whether yedoma provides a globally significant record of ice-age atmospheric conditions or just regional floodplain activity. Here, we test the hypotheses of aeolian versus waterlain deposition of yedoma silt, elucidate the palaeoenvironmental conditions during deposition and develop a conceptual model of silt deposition to clarify understanding of yedoma formation in northern circumpolar regions during the Late Pleistocene. This is based on a field study in 2009 of the Russian stratotype of the 'Yedoma Suite', at Duvanny Yar, in the lower Kolyma River, northern Yakutia, supplemented by observations that we have collected there and at other sites in the Kolyma Lowland since the 1970s. We reconstruct a cold-climate loess region in northern Siberia that forms part of a vast Late Pleistocene permafrost zone extending from northwest Europe across northern Asia to northwest North America, and that was characterised by intense aeolian activity.Five litho-and cryostratigraphic units are identified in yedoma remnant 7E at Duvanny Yar, in ascending stratigraphic order: (1) massive silt, (2) peat, (3) stratified silt, (4) yedoma silt and (5) near-surface silt. The yedoma silt of unit 4 dominates the stratigraphy and is at least 34 m thick. It is characterised by horizontal to gently undulating subtle colour bands but typically lacks primary sedimentary stratification. Texturally, the yedoma silt has mean values of 65 +/- 7 per cent silt, 15 +/- 8 per cent sand and 21 +/- 4 per cent clay. Particle size distributions are bi- to polymodal, with a primary mode of about 41 mu m (coarse silt) and subsidiary modes are 0.3-0.7 mu m (very fine clay to fine clay), 3-5 mu m (coarse clay to very fine silt), 8-16 mu m (fine silt) and 150-350 mu m (fine sand to medium sand). Semidecomposed fine plant material is abundant and fine in-situ roots are pervasive. Syngenetic ice wedges, cryostructures and microcryostructures record syngenetic freezing of the silt. An age model for silt deposition is constructed from 47 pre-Holocene accelerator mass spectrometry (AMS) C-14 ages, mostly from in-situ roots and from three optically stimulated luminescence (OSL) ages of quartz sand grains. The C-14 ages indicate that silt deposition extends from 19 000 +/- 300 cal BP to 50 000 cal BP or beyond. The OSL ages range from 21.2 +/- 1.9 ka near the top of the yedoma to 48.6 +/- 2.9 ka near the bottom, broadly consistent with the C-14 age model.Most of the yedoma silt in unit 4 at Duvanny Yar constitutes cryopedolith (sediment that has experienced incipient pedogenesis along with syngenetic freezing). Mineralised and humified organic remains dispersed within cryopedolith indicate incipient soil formation, but distinct soil horizons are absent. Five buried palaeosols and palaeosol 'complexes' are identified within cryopedolith on the basis of sedimentary and geochemical properties. Magnetic susceptibility, organic content, elemental concentrations and ratios tend to deviate from average values of these parameters at five levels in unit 4. The cryopedolith-palaeosol sequence accreted incrementally upwards on a vegetated palaeo-land surface with a relief of at least several metres, preserving syngenetic ground ice in the aggrading permafrost. Pollen spectra dated to between about 17 000 and 25 000 C-14 BP characteristically have frequencies of 20-60 per cent tree/shrub pollen (mainly Betula and Pinus) and 20-60 per cent graminoids, predominantly Poaceae, plus forbs, whereas spectra dated to about 30 000-33 000 C-14 BP have lower values of woody taxa (about 10%) and are dominated by graminoids (mainly Poaceae), forbs (particularly Caryophyllaceae and Asteraceae) and Selaginella rupestris. The latter are more typical of Last Glacial Maximum (LGM) samples reported elsewhere in Siberia, and the unusually high arboreal pollen values in the LGM yedoma at Duvanny Yar are attributed to long-distance transport of pollen.Three hypotheses concerning the processes and environmental conditions of yedoma silt deposition at Duvanny Yar are tested. The alluvial-lacustrine hypothesis and the polygenetic hypothesis are both discounted on sedimentary, palaeoenvironmental, geocryological and palaeoecological grounds. The loessal hypothesis provides the only reasonable explanation to account for the bulk of the unit 4 yedoma silt at this site. Supporting the loessal interpretation are sedimentological and geocryological similarities between the Duvanny Yar loess-palaeosol sequence and cold-climate loesses in central and northern Alaska, the Klondike (Yukon), western and central Siberia and northwest Europe. Differences between loess at Duvanny Yar and that in western and central Siberia and northwest Europe include the persistence of permafrost and the abundance of ground ice and fine in-situ roots within the yedoma. Modern analogues of cold-climate loess deposition are envisaged at a local scale in cold, humid climates where local entrainment and deposition of loess are generally restricted to large alluvial valleys containing rivers that are glacially sourced or drain areas containing Late Pleistocene glacial deposits, and thus glacially ground silt. The Duvanny Yar yedoma shares sedimentological and geocryological features with yedoma interpreted as ice-rich loess or reworked loess facies at Itkillik (northern Alaska) and in the central Yakutian lowland, and with yedoma in the Laptev Sea region and the New Siberian Archipelago. It is therefore suggested that many lowland yedoma sections across Beringia are primarily of aeolian origin (or consist of reworked aeolian sediments), although other depositional processes (e.g. alluvial and colluvial) may account for some yedoma sequences in river valleys and mountains.A conceptual model of yedoma silt deposition at Duvanny Yar as cold-climate loess in Marine Isotope Stage (MIS) 3 and MIS 2 envisages summer or autumn as the main season of loess deposition. In summer, the land surface was snow-free, unfrozen and relatively dry, making it vulnerable to deflation. Graminoids, forbs and biological soil crust communities trapped and stabilised windblown sediments. Loess accretion resulted from semicontinuous deposition of fine background particles and episodic, discrete dust storms that deposited coarse silt. Winter was characterised by deep thermal contraction cracking beneath thin and dusty snow covers, and snow and frozen ground restricted deflation and sediment trapping by dead grasses. Sources of loess at Duvanny Yar potentially include: (1) sediments and weathered bedrock on uplands to the east, south and southwest of the Kolyma Lowland; (2) alluvium deposited by rivers draining these uplands; and (3) sediments exposed in the Khallerchin tundra to the north and on the emergent continental shelf of the East Siberian Sea. Glacially sourced tributaries of the palaeo-Kolyma River contributed glacially ground silt into channel and/or floodplain deposits, and some of these were probably reworked by wind and deposited as loess in the Kolyma Lowland.The palaeoenvironmental reconstruction of the sedimentary sequence at Duvanny Yar is traced from MIS 6 to the late Holocene. It includes thermokarst activity associated with alas lake development in the Kazantsevo interglacial (MIS 5e), loess accumulation, pedogenesis and syngenetic permafrost development, possibly commencing in the Zyryan glacial (70 000-55 000 cal BP) and extending through the Karginsky interstadial (55 000-25 000 cal BP) and Sartan glacial (25 000-15 000 cal BP), cessation of yedoma silt deposition during the Lateglacial, renewed thermokarst activity in the early Holocene, and permafrost aggradation in the mid to late Holocene.Beringian coastlands from northeast Yakutia through the north Alaskan Coastal Plain to the Tuktoyaktuk Coastlands (Canada) were characterised by extensive aeolian activity (deflation, loess, sand dunes, sand sheets, sand wedges) during MIS 2. Siberian and Canadian high-pressure cells coupled with a strengthened Aleutian low-pressure cell would have created enhanced pressure gradient-driven winds sufficient to entrain sediment on a regional scale. Summer winds are thought to have deflated sediment exposed on the East Siberian Sea shelf and deposited silt as a distal aeolian facies to the south. Additionally, stronger localised winds created by local downslope gravity flows (katabatic winds) may have entrained sediment. Local katabatic winds in summer may have transported silt generally northwards towards the Kolyma Lowland, particularly during times of extended upland glaciation in the North Anyuy Range to the east during the Zyryan (MIS 4) period, whereas winter winds carried limited amounts of silt generally southwards as a result of pressure gradient forces. The Duvanny Yar yedoma is part of a subcontinental-scale region of Late Pleistocene cold-climate loess. One end member, exemplified by the yedoma at Duvanny Yar, was loess rich in syngenetic ground ice (Beringian yedoma). The other, exemplified by loess in northwest Europe, was ice-poor and subject to complete permafrost degradation at the end of the last ice age. These end members reflect a distinction between enduring cold continuous permafrost conditions leading to stacked ice-rich transition zones and large syngenetic ice wedges in much of Beringia versus conditions oscillating between cold permafrost, warm permafrost and seasonal frost, leading to repeated permafrost thaw and small ice-wedge pseudomorphs in northwest Europe. Copyright (C) 2015 John Wiley & Sons, Ltd.

Abstract The fluvial sequences of the Milton and the Letchworth formations in the south Midlands of England and neighbouring regions represent at least two pre-existing rivers, the Milton and Brigstock streams, underlying Middle Pleistocene glacial sediments. The Milton Formation includes sand sourced from the Midlands bedrock. This implies that both streams were aligned in a northwest to southeast direction. This direction parallels the contemporaneous courses of the rivers Thames and Trent, the former turning towards the east and northeast to enter the North Sea. Their alignments indicate that the Milton and Letchworth streams formed left-bank tributaries of the Thames, joining the river in Hertfordshire and Essex, as illustrated in the article. This reconstruction has important implications for the interpretation of the proto-Soar river of the south Midlands, represented by the Baginton Formation. Although originally thought to represent a late Middle Pleistocene line, this southwest to northeast aligned system was reinterpreted as the headwaters of a pre-Anglian ‘Bytham river’, a1ligned towards East Anglia. However, recent work has shown that this river could not have existed in the pre-Anglian since there is no link between the Midlands and East Anglian spreads. Recent re-recognition that the Baginton Formation deposits do represent a later, post-Anglian drainage line is reinforced by the identification of the Milton and Letchworth streams, whose catchments occupied the area later drained by the proto-Soar. Overall, the main drainage alignment in southern England during the pre-Anglian period was dominated by northwest–southeast-draining consequent rivers adjusted to the regional geological dip. After widespread drainage disruption caused by the Anglian glaciation, northeast–southwest-orientated subsequent streams eroded frost-susceptible clay bedrock under periglacial and permafrost conditions, and beheaded the courses of some of the older consequent streams.

Focusing on lowland Britain and the southern North Sea Basin, this article reviews the sedimentary and geomorphic evidence for the main glacial lakes inferred during the Middle and Late Pleistocene and evaluates their impacts on drainage-basin development. Glacial lakes are best known from glaciations during Marine Isotope Stages (MIS) 12, 6 and 2, although glacial lakes have also been inferred during MIS 10 and 4. Some lakes – for example, Bosworth, low-level Humber and the lakes of the eastern Fenland margin – are reconstructed from unequivocal sedimentary evidence, including rhythmites and subaqueous outwash, whereas others lakes – for example, Lapworth and Fenland – are inferred mostly from erosional features and remain to be substantiated. The largest known glacial lake developed in the southern North Sea Basin between an ice sheet to the north and a chalk bedrock ridge in the Strait of Dover area, first during the Anglian/Elsterian glaciation of MIS 12 and again during the late Wolstonian/late Saalian Drenthe glaciation of MIS 6. The palaeohydological impacts of lake drainage are thought to include cutting of the Strait of Dover as a result of catastrophic drainage from the North Sea Lake during MIS 12, incision of a number of gorges and river valleys in England, and diversion or even reversal of major rivers such as the Thames and the proto-Soar/Avon system. Recently, varve chronologies have been correlated with the Greenland ice-core record, although caution is needed to discriminate between varves and non-annual rhythmites. Future work on Pleistocene glacial lakes needs to test chronologies of lake development by luminescence dating of glaciolacustrine sediments deposited in non-ice-proximal locations – (1) fine-grained rainout deposits and (2) wave-rippled sands deposited in shallow water – and to model the impacts of glacial isostasy in order to reconstruct lake extents. All of this work should be based on the rigorous application of sedimentology to interpret sedimentary facies and depositional environments.

The sedimentary sequence through the Hemingbrough Formation exposed at two sites in the central part of the Vale of York, south of the Escrick moraine ridge, is described and used to reconstruct the palaeoenvironmental history of Glacial Lake Humber. Interbedded wave ripples and laminated silts and clays at both sites indicate that Lake Humber was characterised by fluctuating water levels, often no deeper than wave base. Optically stimulated luminescence ages of 21.0±1.9, 21.9±2.0, and 24.1±2.2kyr returned from two wave-rippled sandy beds within the glaciolacustrine sequence at Hemingbrough, c. 10km south of the Escrick moraine ridge, provide the first direct chronological determination for the low-level phase of Lake Humber. As these beds are principally attributed to glacial meltwater emanating from the Vale of York ice lobe of the British Ice Sheet, when its margin was at or near the Escrick moraine ridge, this corroborates the interpretation that this ridge marks the LGM ice limit.