During the latest Quaternary, freshwater pluvial lakes intermittently formed in the topographically closed Willcox basin, Arizona. A lacustrine sequence of six separate high stands of Lake Cochise is documented by stratigraphic studies, 19 radiocarbon ages, and supplementary evidence. Two stands of pluvial Lake Cochise, older than 14,000 yr B.P., reached elevations above 1290 m. The prominent 1274-m shoreline of Lake Cochise, which circumscribes the basin, was largely created during a high stand between 13,750 and 13,400 yr B.P. During the Holocene, water filled the Willcox basin three times to an elevation slightly below the crest of the 1274-m shoreline. This occurred once during the early Holocene around or before 8900 yr B.P. and twice during the later part of the middle Holocene. Since the middle Holocene, only shallow ephemeral lakes have occupied the deflated central portion of ancient Lake Cochise, a depression known as the Willcox Playa. The lacustrine sequence of Lake Cochise provides an independent evaluation of late Quaternary paleoclimatic reconstructions for southern Arizona and the American Southwest.

The timing of deglaciation in the Lake Agassiz basin is critical in establishing the routing of meltwater and precipitation runoff from a 2,000,000-km2 region of central North America and in evaluating the influence this water had on rivers and oceans into which it drained. Dates of 12,400 ± 420 and 12,100 ± 160 yr B.P. for moss at the Rossendale site in Manitoba have long been a key for those advocating an “early” deglacial chronology in this region. However, new dates for wood from this site and paleoecological interpretations of ostracods, molluscs, and the dated moss all support a “young” deglacial scenario. Of particular significance is the fact that the dated moss, Scorpidium scorpioides, is a subaquatic type subject to contamination by old carbon dissolved from bedrock. In fact, most subaquatic moss may be unreliable for radiocarbon dating. For these reasons, the 12,400 and 12,100 yr B.P. dates are rejected. New dates of 9600 ± 70 and 9510 ± 90 yr B.P. for wood from the same organic-rich unit containing the dated moss, ostracods, and molluscs fit well with the “young” deglacial chronology of the southwestern Laurentide ice margin advocated by many. In short, the ice margin appears to have retreated into the southern Lake Agassiz basin after 12,000 yr B.P. and north of the Rossendale site by 11,000 yr B.P. About 10,000 yr B.P., following the Moorhead low-water phase, Lake Agassiz rose to the Campbell level. The dated organic matter at Rossendale was deposited in a lagoon behind the Campbell beach.

When considering the formation of tunnel valleys, most researchers have appealed to basal melting as the primary source of water. Erosion of the tunnel valleys can then be accomplished by steady-state drainage over a long period or by storage and subsequent catastrophic release. However, field relationships in a large system of tunnel valleys formed by the Superior lobe in central Minnesota indicate that another source of water must be considered. The Minnesota tunnel valleys are composed of individual segments 10–20 km long, and the segments commonly terminate in subaerial outwash fans at recessional ice margin positions. Eskers, associated with the tunnel valleys, are also composed of short segments, frequently beginning at moulin kames and terminating at the head of outwash fans. The dominant source of the water responsible for tunnel-valley formation was seasonal meltwater from the glacier surface that reached the bed through moulins and crevasses. The apparent continuity of the valleys resulted from the headward development of the englacial drainage system during ice retreat.

A dense network of 3.5 kHz and Uniboom seismic profiles revealed the geological evolution of the Ierissos-Alexandroupolis Shelf area of the northern Aegean during the latest sea-level rise. Near the end of the Pleistocene, ca. 14,000 yr B.P., the sea was at about − 120 m, and almost 5300 km2 of shelf was exposed to subaerial erosion. Two permanet lakes existed in the Ierissos and Kavalla Gulfs, as well as a number of ephemeral lakes scattered throughout the rest of the area. Kavalla Gulf was drained by the Nestos River which joined the Strymon River at the outer Strymonikos Plateau. On the flat Samothraki Plateau were dune fields, marshes, and a number of minor seasonal rivers, while the Evros River flowed east of the plateau. By ca. 13,000 yr B.P. the sea had risen to − 70 m and covered 30% of the previously exposed shelf. The lakes at Ierissos and Strymonikos Gulfs were shallower and the sea approached to within about 5 km of them. Kavalla Gulf and the adjacent eastern Strymonikos Plateau lay 10 m above sea level, while the sea intruded along river mouths at the Samothraki Plateau. During this transgression of the sea river-bed gravels and sands were covered by silts and clays. At the Pleistocene-Holocene transition, ca. 10,500 yr B.P., the sea lay at about − 50 m and in areas of steep slope the coast was close to its present position. At Kavalla Gulf, the sea advanced along the paleochannel of the Nestos River, and the island of Samothraki was separated from the mainland. By ca. 7500 yr B.P. the sea was only 15 m below its present level and the northeastern Aegean shelf assumed nearly its present morphology. The Nestos River changed course to the east and Thassos Island was separated from the mainland. Coarse sediment formed wedges nearshore, whereas fine-grained sediments were distributed widely by current action. In many areas, relict sediments are present.

Many proxy indicators of past climate integrate the effects of temperature and precipitation, and not all indicators integrate these effects in the same way. This poses a problem in the reconstruction of past climate. A study of the response of tree growth around Lake Superior and the water supplies to that lake to summer climate shows differences and similarities in response. Tree growth is positively correlated with precipitation—although the response is reduced when conditions are very wet—and negatively correlated with temperature. In contrast, the water supplied to the lake are positively correlated with both precipitation and temperature; the correlation with temperature is, however, not very strong. Multiple regression models developed to assess the relative magnitudes of the similarities and differences show that the similarity in response to precipitation accounts for about 30% of the variance in water supplies, the nonlinear response of trees—represented by a log transformation—accounts for an additional 15%, and the difference in response to temperature is equivalent to about 30% of the variance in water supplies.

Late-Quaternary paleoecological and glacial evidence from the higher latitudes of the Southern Hemisphere implies overall uniformity of large-scale glacial and interglacial climatic fluctuations for the past 40,000 yr. Climate of the last glacial maximum, variously dated between 30,000 and 11,000 yr B.P., was relatively cold and dry compared with the warmer, more humid climate of the Holocene and the interstade preceding the last glacial maximum. Conditions were apparently coldest during millennia centered around 20,000 yr B.P. and warmest in the early Holocene. Recorded small-scale fluctuations, frequently variable for any given time interval, are less consistent. A cold late-glacial episode, estimated as occurring between ca. 13,000 and 11,000 yr B.P. in Antarctica, possibly was coeval with the Younger Dryas Stade in northwestern Europe and may be correlative with a climatic episode in southern South America and perhaps in New Zealand and South Georgia; however, there is no evidence for the event in Tasmania. General atmospheric circulation models for the polar latitudes at the time of the last glacial maximum show an intensification of the southern westerlies, apparently a result of the expansion of ice cover in Antarctica and of sea ice in the Southern Ocean.

Three sites on the Point Reyes Peninsula, California, provide evidence of major environmental change during the Pleistocene-Holocene transition. A 12,300-yr-long pollen record from Coast Trail Pond reveals a change from closed canopy Pseudotsuga-Abies forest to coastal scrub and grassland about 10,000 yr B.P. The same change is also evident in pollen records from two sea cliff exposures that contain a series of buried soils interstratified with fluvial and debris-flow deposits. Eight radiocarbon dates from these exposures indicate that aggradation at both sites began ca. 12,000 yr B.P. and that the change from forest to scrub and grassland occurred between 10,300 and 9400 yr B.P. High percentages of alder pollen and bracken fern spores reflect repeated geomorphic disturbance at ca. 10,400 yr B.P. The occurrence of major aggradation accompanied by recurrent debris flows between ca. 12,000 and 10,000 yr B.P. suggests that this was a period of more frequent high-intensity storms.

The lower carnassial lengths of spotted hyenas (Crocuta crocuta) in 12 late Pleistocene samples from Britain indicate that, on average, local hyenas of the last (Devensian) glaciation were significantly larger than their last-interglaciation (Ipswichian) counterparts. Together with the tendency for spotted hyena carnassial length to increase with latitude in present-day Africa, this suggests that spotted hyena body size is inversely related to temperature, as predicted by Bergmann's rule. The implication is that spotted hyena carnassial length can be used as an independent gauge of Pleistocene temperature variation, though the combined African and British data imply that the relationship between carnassial length and temperature is curvilinear, such that as temperature declines, equal amounts of further decline produce progressively smaller increases in average carnassial length.

Leisey Shell Pit, on the southeastern edge of Tampa Bay, has produced the largest early Pleistocene vertebrate fauna in North America, consisting of about 30,000 cataloged specimens. The predominantly terrestrial and freshwater vertebrate fauna was transported into an estuarine environment during a regressive phase in an otherwise marine, bay-bottom sequence of sandy shell beds. The biochronological age of the Leisey Shell Pit local fauna is late Sappan (or late early Irvingtonian) based on the presence of such mammalian species as Mammuthus meridionalis and Smilodon gracilis, the absence of any Blancan holdovers, and early records of Lutra, Nothrotheriops, Palaeolama, and a new small glyptodont. Magnetostratigraphy indicates that the entire Pleistocene section at Leisey is reversed and thus presumably accumulated in the post-Olduvai part of the Matuyama Magnetochron (i.e., 1.66 myr or younger). Strontium isotope ratios (87Sr/86Sr) sampled from aragonitic bivalve shells (Chione cancellata) trend higher with increasing height in the section, yielding a mean age estimate for the two major bone beds of 1.7 ± 0.35 myr. In combination, these three geochronologic methods provide a high-resolution approach toward age determination and delimit the age of Leisey Shell Pit local fauna as between 1.66 and 1.4 myr.

Three radiocarbon dates along with relative-dating criteria place limits on the deglaciation history of Manachaque Valley, Cordillera Central. Ice retreated from the late-glacial maximum by at least 12,100 yr B.P. During ice retreat numerous moraines were deposited throughout the valley. Glacier cover was reduced to about half that of the last glacial maximum by at least 9700 yr B.P. and to less than a tenth by at least 6450 yr B.P. Because all dates are minimum, the dates and field data are consistent with little or no ice remaining by early Holocene. No unambiguous Younger Dryas moraines are present.

In 1983 pingos were reported in the Vestfold Hills, Antarctica. The features form dome-shaped hills up to 4 m high and 12 m in diameter and occur on an ice-cored moraine. Their form, structure, and sedimentary processes observed on the surface of the moraine, together with a consideration of the environmental requirements for pingo growth preclude their origin as pingos. They appear to be residual landforms that have formed on the surface of the moraine as the ice core has been destroyed by ablation and thermal erosion. The previous interpretation of the features has relied on the shape of the hills and their occurrence in a similar geologic setting to pingos in the Northern Hemisphere. The interpretation appears to have misinterpreted tension cracks and coherent slumping of sediment as dilation cracks and as evidence for the extrusion of a central sediment plug. Pingos on supraglacial debris have not been described from any other locations and it seems that their formation is inconsistent with the ablation of the ice core of moraines.

Four Holocene alluvial fills are present in Nebraska's South Loup River valley. Fill IV, the oldest and thickest, was deposited between 10,200 and 4800 14C yr B.P.; Fill III has an age of about 3000 14C yr B.P.; Fill II is younger than 2100 and older than 900 14C yr B.P.; and Fill I is younger than 900 14C yr B.P. Regional contemporaneity of valley alluviation in the eastcentral Great Plains suggests that climate has controlled long-term sediment storage in the South Loup River valley.