The time series of volcanically produced sulfate from the GISP2 ice core is used to develop a continuous record of explosive volcanism over the past 110,000 yr. We identified ∼850 volcanic signals (700 of these from 110,000 to 9000 yr ago) with sulfate concentrations greater than that associated with historical eruptions from either equatorial or mid-latitude regions that are known to have perturbed global or Northern Hemisphere climate, respectively. This number is a minimum because decreasing sampling resolution with depth, source volcano location, variable circulation patterns at the time of the eruption, and post-depositional modification of the signal can result in an incomplete record. The largest and most abundant volcanic signals over the past 110,000 yr, even after accounting for lower sampling resolution in the earlier part of the record, occur between 17,000 and 6000 yr ago, during and following the last deglaciation. A second period of enhanced volcanism occurs 35,000–22,000 yr ago, leading up to and during the last glacial maximum. These findings further support a possible climate-forcing component in volcanism. Increased volcanism often occurs during stadial/interstadial transitions within the last glaciation, but this is not consistent over the entire cycle. Ages for some of the largest known eruptions 100,000–9000 yr ago closely correspond to individual sulfate peaks or groups of peaks in our record.

The Vedde Ash Bed (mid-Younger Dryas) and the Saksunarvatn Ash (early Holocene) are important regional stratigraphic event markers in the North Atlantic, the Norwegian Sea, and the adjacent land area. It is thus essential to date them as precisely as possible. The occurrence of the Saksunarvatn Ash is reported for the first time from western Norway, and both tephras are dated precisely by AMS analyses of terrestrial plant material and lake sediment at Kråkenes. The Vedde Ash has been previously dated at sites in western Norway to about 10,600 yr B.P. It is obvious in the Younger Dryas sediments at Kråkenes, and its identity is confirmed geochemically. The mean of four AMS dates of samples of Salix herbacea leaves adjacent to the tephra is 10,310 ± 50 yr B.P. The Saksunarvatn Ash is not visible in the early Holocene lake sediment at Kråkenes. After removal of organic material and diatoms, the identity of the tephra particles was confirmed geochemically, and their stratigraphic concentration was estimated. From curve matching of a series of seven AMS dates of terrestrial plant macrofossils and whole sediment, the radiocarbon age of the ash is 8930–9060 yr B.P., corresponding to an age of 9930–10,010 cal yr B.P. (7980–8060 cal yr B.C.).

Although post-glacial marine sediments of late Wisconsinan and early Holocene age are common in eastern Canada and the northeastern United States, remnants of older Pleistocene marine sediments are scarce. A fossiliferous marine clay that predates the classical Wisconsinan was recently discovered in the St. Lawrence Valley. A dominantly estuarine environment is inferred from the geochemistry of the shells (δ18O = −7.1) and from benthic foraminifer and ostracode assemblages. The clay indicates a marine invasion (Cartier Sea) shallower and probably shorter than that during the upper late Wisconsinan Champlain Sea episode (12,000–9,500 yr B.P.). The pollen content shows that regional vegetation during the marine episode began as open tundra, then became a Betula and Alnus crispa forest, reached a climatic optimum with Quercus, Corylus, and Abies, and concluded as a Pinus/Picea boreal forest. A corrected infrared stimulated luminescence age of 98,000 ± 9000 yr is compatible with the epimerization ratio of shells. The Cartier Sea resulted from a post-glacial glacio-isostatic marine invasion in the St. Lawrence lowlands. It probably occurred during late stage 5 and is tentatively assigned to the transition of oxygen isotope substages 5b/5a. This marine episode dates to stage 5 of the preceding continental glacier which extended to middle latitudes in NE America.

Lake records from northern Eurasia show regionally coherent patterns of changes during the late Quaternary. Lakes peripheral to the Scandinavian ice sheet were lower than those today but lakes in the Mediterranean zone were high at the glacial maximum, reflecting the dominance of glacial anticyclonic conditions in northern Europe and a southward shift of the Westerlies. The influence of the glacial anticyclonic circulation attenuated through the late glacial period, and the Westerlies gradually shifted northward, such that drier conditions south of the ice sheet were confined to a progressively narrower zone and the Mediterranean became drier. The early Holocene shows a gradual shift to conditions wetter than present in central Asia, associated with the expanded Asian monsoon, and in the Mediterranean, in response to local, monsoon-type circulation. There is no evidence of mid-continental aridity in northern Eurasia during the mid-Holocene. In contrast, the circum-Baltic region was drier, reflecting the increased incidence of blocking anticyclones centered on Scandinavia in summer. There is a gradual transition to modern conditions after ca. 5000 yr B.P. Although these broad-scale patterns are interrupted by shorter term fluctuations, the long-term trends in lake behavior show a clear response to changes in insolation and glaciation.

The continental slope in Orphan Basin, northeast of Newfoundland, is underlain by several seaward-thinning debris-flow wedges alternating with acoustically stratified, regionally extensive, mainly hemipelagic sediments. δ18O stratigraphy and volcanic ash layers in a 11.67-m core indicate that the uppermost debris-flow wedge formed during the last of several sea-level lowstands in isotopic stages 2–4. Similarly, seismic reflection correlation of dated levels at DSDP Site 111 with the Orphan Basin succession suggests that two deeper debris-flow wedges were deposited during oxygen isotopic stages 6 and 8. The oldest of the debris-flow deposits in at least three of the wedges formed well into the corresponding glacial cycle, after ice sheets had reached the edge of the continental shelf. Slower deposition by hemipelagic processes and ice rafting formed the acoustically stratified units, including Heinrich layers. The youngest three debris-flow wedges each have volumes of 1300–1650 km3. Approximately two-thirds of this material is attributed to glacial erosion of Mesozoic and Tertiary strata beneath the Northeast Newfoundland Shelf. The remainder is believed to have been derived by glacial erosion of older bedrock that now forms the island of Newfoundland. The observed sediment volumes and the inferred basal and upper ages of the debris-flow wedges imply an average glacial denudation rate of about 0.13 mm/yr for this older bedrock, and an average of about 60 m of glacial bedrock erosion since oxygen isotope stage 22. This denudation rate is similar to estimates from the Barents Sea region off Norway.

Aspartic acid (Asp) racemization occurs at a significantly higher rate than isoleucine epimerization and consequently provides better temporal resolution of Arctic marine deposits (from Alaska, Spitsbergen, and Baffin Island). Heating experiments (at 100°C) on the bivalves Mya and Hiatella show the Asp racemization rate decreases with increasing D/L values, as is typical for biogenic carbonates. Based on these experimental racemization rates and rates determined from racemization of samples radiocarbon dated to ca. 10,000–12,000 yr B.P., activation energies for Mya and Hiatella are estimated to be 30.6 and 30.0 kcal/mol, respectively, for Asp racemization, and 29.0 and 29.5 for isoleucine epimerization. Analysis of a time series of Plio–Pleistocene Hiatella from the north coast of Alaska shows that last-interglacial mollusks can be readily distinguished from modern samples by Asp but not by isoleucine. D/L Asp values indicate a younger age for the Fishcreekian transgression than does isoleucine epimerization. For Spitsbergen, D/L Asp shows a slight age difference (ca. 12,000 yr) between two units of the “episode B” interstadial and suggests that the age of these units may be closer to 65,000 than to 80,000 yr B.P., two possible ages suggested by other evidence. The age of the Loks Land Interstadial on Baffin Island is likely to be greater than that indicated by radiocarbon ages. Within deposits from each region, D/L Asp values are less variable among individual shells than isoleucine epimerization values. This may indicate better reliability of Asp for geochronology.

Analyses of pollen, plant macrofossils, macroscopic charcoal, mollusks, magnetic susceptibility, and geochemical content of a sediment core from Farewell Lake yield a 11,000-yr record of terrestrial and aquatic ecosystem changes in the northwestern foothills of the Alaska Range. Between 11,000 and 8500 yr B.P., the regional landscape was dominated by a Betula shrub tundra, in which Populus-Salix communities were common. Abundant charcoal in sediments indicates that fires were common in the lake catchment during this period, and high mineral accumulation rates, allogenic elemental content, and magnetic susceptibility suggest intense soil erosion. In addition, mollusks, pollen and macrofossils of aquatic macrophytes, and biogenic silica provide evidence that the lake was substantially shallower and more productive 11,000–8500 yr B.P. than later. Low lake level and high aquatic productivity might have been caused by warm and dry summers associated with early postglacial insolation maximum in northern high latitudes. About 8000 yr B.P., Picea glauca arrived within the lake catchment, forming a forest tundra association until ca. 6000 yr B.P. Alnus shrub thickets established in the region ca. 6500 yr B.P., and Betula papyrifera arrived ca. 6000 yr B.P. Closed P. glauca forests developed ca. 6000 yr B.P. Picea mariana became important subsequently and replaced P. glauca as the dominant tree species in the region ca. 4000 yr B.P. An increase in authigenic Fe/Mn ratios suggests that the development of waterlogged soils accompanied this vegetation change. Fires increased in importance at this time and might have accelerated soil erosion. The establishment of P. mariana forests probably reflected complex responses of forest ecosystems to the onset of cooler and wetter climate conditions during the late Holocene.

Analysis of phytoliths and macroscopic charcoal from natural soils near Manaus, Brazil shows that the central Amazonian terra firme forest has not been stable during the Holocene. Phytolith taxonomic composition and radiocarbon ages indicate that a closed forest has continuously existed in the area since at least 4600 yr B.P., but that forest composition and species abundance changed significantly sometime during the Holocene Epoch. There is no evidence that the vegetation experienced swidden cultivation. The distribution and dates of the charcoal in the soils indicate that forest fires between 1795 and 550 yr B.P. burned an area of about 200 hectares. Vegetational changes and fires appear to be a result of climatic drying that may have affected large areas of the Amazon Basin over the past 5000 to 7000 years. Soil phytolith analysis is a promising tool for the study of environmental history in the humid tropics and may reduce the previous dependence on lake and swamp sediments, at least for the Holocene Epoch.

Millennium-long tree-ring chronologies have recently been derived fromThuja occidentalis L. growing on the exposed cliffs of the Niagara Escarpment, Canada. Lengthening of these chronologies may be possible by incorporating tree-ring series from subaerially exposed and submerged subfossil wood, provided that its anatomy is not significantly influenced by time or treatment. In order to determine if Holocene-age samples would be suitable for incorporation into very long tree-ring chronologies, a study was undertaken comparing wood structure of living trees, dead trees exposed to the air during the late Holocene, and dead trees submerged since the early Holocene. The results show that the impact of time and environment on wood anatomy is considerably less than expected. The anatomy of annual rings in Holocene trees is similar to that found in modern specimens. High magnification scanning electron micrographs show that tracheid walls lose their rectangular appearance after 8000 yr, but few other signs of deterioration are present. The results mean that Holocene wood can be used to add to the long dendrochronological record that has already been produced from the Niagara Escarpment.

A brief cold–warm–cold climate change during the middle Wisconsinan is described for the first time in North America, based on fossil beetle assemblages at Titusville, Pennsylvania. AMS dating of insect chitin and wood suggests the change occurred between 39,000 and 43,500 yr B.P. Basal peats in river terrace deposits contain arctic–subarctic beetle species representing a climate similar to that found at treeline in Canada, where mean July temperatures range from 10° to 13°C. These cold-adapted beetles were replaced by species representing a climate similar to the mixed coniferous–deciduous forests of southern Canada, where mean July temperatures range from 18° to 20°C. In turn, these warmer-adapted beetles were replaced by arctic–subarctic species in the upper part of the section, representing a climate with mean July temperatures in the range 10°–13°C. The brief warm interval is informally designated the Titusville interstade. The difference in temperature between the colder and warmer intervals is about 5°–7°C. The Titusville interstade is correlated with the Upton Warren Interstade of the British Isles which is of similar age and intensity. It is also correlated with the high sea-surface temperature interval between Heinrich events 4 and 5 represented in North Atlantic sediment cores.

Deformation structures interpreted as vertebrate hoof- and foot-tracks occur within upper-Pleistocene eolian sand-sheet deposits in the stabilized Kantishna sand sea of central Alaska and in the Nushagak lowland of southwestern Alaska. Exposures of tracks are generally limited to cross sections, which reveal concave-up deformation structures in which displacement of preexisting strata diminishes downward. Deposits in both areas contain tracks that are 6 to 16 cm in diameter and are divided by a central ridge, reflecting formation by artiodactyl (even-toed) ungulates. Larger (21–34 cm) tracks without a central ridge, observed in the Nushagak lowland, were formed by proboscideans, probably woolly mammoth. Large vertebrate tracks occur within irregularly stratified sand and silty sand that accumulated upon partially vegetated sand sheets, and within thin, even wind-ripple laminae of unvegetated sand sheets. The presence of tracks at multiple stratigraphic levels and preservation of roots and rhizocretions within the eolian deposits suggest that vegetated sand sheets may have formed a locally important grazing habitat for large herbivores during at least part of the last glaciation. Recognition that vertebrate tracks are preserved in eolian sand-sheet deposits, and in deposits of other environments as well, opens a new source of stratigraphic and paleoecological data to aid reconstruction of the vanished ecosystems of Beringia.

Fossil shells of the boring mussel Lithophaga lithophaga provide a means for dating changes in relative sea level. These bivalves, being among the first colonizers of bare calcareous substrates, can mark the earliest stages of marine submergence of caves. Here we report data concerning the deepest submerged speleothem presently sampled in a temperate area, at 48 m below present sea level off the west coast of Italy (Mediterranean Sea). A fossil mussel shell beneath encrusting layers from later marine colonists gave an AMS age of 9580 ± 35 14C yr B.P. (10,253 ± 72 cal yr B.P.).