The Cochabamba Basin (Bolivia) is on the ancient road network connecting Andean and lowland areas. Little is known about the longevity of this trade route or how people responded to past environmental changes. The eastern end of the Cochabamba valley system constricts at the Vacas Lake District, constraining the road network and providing an ideal location in which to examine past human–environmental interactions. Multiproxy analysis of sediment from Lake Challacaba has allowed a c. 4000 year environmental history to be reconstructed. Fluctuations in drought tolerant pollen taxa and calcium carbonate indicate two periods of reduced moisture availability (c. 4000–3370 and c. 2190–1020 cal yr BP) compared to adjacent wetter episodes (c. 3370–2190 and c. 1020 cal yr BP–present). The moisture fluctuations broadly correlate to El Niño/Southern Oscillation variations reported elsewhere. High charcoal abundance from c. 4000 to 2000 yr ago indicates continuous use of the ancient road network. A decline in charcoal and an increase in dung fungus (Sporormiella) c. 1340–1210 cal yr BP, suggests that cultural changes were a major factor in shaping the modern landscape. Despite undisputable impacts of human populations on the Polylepis woodlands today, we see no evidence of woodland clearance in the Challacaba record.

During the mid-Holocene the eastern Taihu area, on China's Yangtze delta plain, was populated by advanced late Neolithic cultures supported by intensive domesticated rice cultivation. This agricultural system collapsed around 4200 cal yr BP, with severe population decline, the end of the Liangzhu culture, and about half a millennium of very low-scale human activity in the area before the re-establishment of agricultural production. Microfossil analyses from six sedimentary sequences, supported by AMS 14C dating, has allowed reconstruction of mid-Holocene hydrological conditions and salinity changes which would have had a major influence on agricultural viability and cultural history in the coastal wetlands. These data, allied to existing stratigraphic and sea-level records, show that chenier ridges that developed after ca. 7000 cal yr BP in the east of the area sheltered it from marine inundation and although still connected to the sea through tidal creeks, low-salinity conditions persisted throughout the Neolithic period. There is no evidence that marine flooding caused the collapse of Liangzhu culture. Marine influence was stable and evolved slowly. Social and cultural causes may also have been important, but if environmental change triggered the collapse of Neolithic agricultural society here, other natural forces must be sought to explain this event.

Multiproxy analysis of three littoral cores from western New Caledonia supports the hypothesis that the main controlling factors of environmental changes are sea-level change, ENSO variability and extra-tropical phenomena, such as the Medieval Warm Period (MWP) marked by a tendency for La Niña-like conditions in the tropical Pacific. The record starts during the late Holocene sea-level rise when the terrestrial vegetation indicated wet and cool conditions. The site was a coastal bay definitely transformed into a freshwater swamp at around 3400 cal yr BP, after the rapid drawdown of sea level to its current level. Sediments and foraminiferal assemblages indicated subsequent episodes of freshwater infillings, emersion or very high-energy conditions, likely related to climatic changes and mostly controlled by ENSO variability. Between 2750 and 2000 cal yr BP, relatively dry and cool climate prevailed, while wetter conditions predominated between ca. 1800 and 900 cal yr BP. The Rhizophoraceae peak between ca. 1080 and 750 cal yr BP, coeval with the MWP, may indicate a global phenomenon. Microcharcoal particles present throughout the record increased after 1500 cal yr BP, suggesting an anthropogenic source. From ca. 750 cal yr BP the appearance of current type of vegetation marks the human impact.

Geomorphic mapping in the upper Conejos River Valley of the San Juan Mountains has shown that three distinct periods of aggradation have occurred since the end of the last glacial maximum (LGM). The first occurred during the Pleistocene–Holocene transition (~ 12.5–9.5 ka) and is interpreted as paraglacial landscape response to deglaciation after the LGM. Evidence of the second period of aggradation is limited but indicates a small pulse of sedimentation at ~ 5.5 ka. A third, more broadly identifiable period of sedimentation occurred in the late Holocene (~ 2.2–1 ka). The latest two periods of aggradation are concurrent with increases in the frequency of climate change in the region suggesting that Holocene alpine and sub-alpine landscapes respond more to rapid changes in climate than to large singular climatic swings. Soil development and radiocarbon dating indicate that hillslopes were stable during the Holocene even while aggradation was occurring in valley bottoms. Thus, we can conclude that erosion does not occur equally throughout the landscape but is focused upslope of headwater streams, along tributary channels, or on ridge tops. This is in contrast to some models which assume equal erosion in headwater basins.

Arid and Alpine ecosystems are known for extreme environmental changes during the Late Quaternary. We hypothesize that the world's largest Alpine arid ecosystem however, the Alpine Steppes of the Tibetan highlands, remained ecologically stable during the LGM and the mid-Holocene. This hypothesis is tested by distributional range of plant species, plant life forms and rate of endemism. The set of character species has a precipitation gradient between 50 and 350 mm/a, testifying for resilience to precipitation changes. 83% of the species have a wider vertical range than 1000 m used as a proxy for resilience to temperature changes. 30% of the species are endemic with 10 endemic genera, including plate-shaped cushions as a unique plant life form. These findings are in line with palaeo-ecological proxies (δ18O, pollen) allowing the assumption that Alpine Steppes persisted during the LGM with 3 to 4 K lower summer temperatures.

During the mid-Holocene, forests could have replaced Alpine Steppes in the upper catchments of the Huang He, Yangtze, Mekong, Salween and Yarlung Zhangbo, but not in the interior basins of the north-western highlands, because the basins were then flooded, suppressing forests and supporting the environmental stability of this arid Alpine grassland biome.

The formation of short-lived backswamps along the Carmel coast of Israel coincides with the rapid global sea-level rise during the late Pleistocene–early Holocene transition. The current study shows that the wetland phenomena originated around 10,000 yr ago and dried up shortly before the local Pre-Pottery Neolithic humans settled on the wetland dark clay sediments 9430 cal yr BP. Palaeontological and stable-isotope data were used in this study to elucidate previously published sedimentological reconstruction obtained from a core drilled into the western trough of the Carmel coastal plain. The water body contained typical brackish calcareous fauna, with variable numerical abundance and low species richness of ostracods and foraminifera. The δ18O and δ13C of the ostracod Cyprideis torosa show close similarity to the present Pleistocene coastal aquifer isotopic values. This study therefore concludes that the wetlands were shallow-water bodies fed by groundwater, with no evidence of sea-water mixing. It seems that they developed as the result of high groundwater levels, transportation of sediments landward, and deposition of sand bars at the paleo-river mouths. It is still not fully understood why these wetlands deteriorated abruptly and disappeared within less than 1000 yr.

Wetlands are a key archive for paleoclimatic and archeological work, particularly in arid regions, as they provide a focus for human occupation and preserve environmental information. The sedimentary record from 'Ayn Qasiyya, a spring site on the edge of the Azraq Qa, provides a well-dated sequence through the last glacial–interglacial transition (LGIT) allowing environmental changes in the present-day Jordanian desert to be investigated robustly through this time period for the first time. Results show that the wettest period at the site preceded the last glacial maximum, which itself was characterised by marsh formation and a significant Early Epipaleolithic occupation. A sedimentary hiatus between 16 and 10.5 ka suggests a period of drought in the region although seasonal rains and surface waters still allowed seasonal occupation of the Azraq region. Archeological evidence suggests that conditions had improved by the Late Epipaleolithic, about the time of the North Atlantic Younger Dryas. The changes between wet and dry conditions at the site show similarities to patterns in the eastern Mediterranean and in Arabia suggesting the Jordan interior was influenced by changes in both these regions through the LGIT climatic transition.

Botanical analyses of fossil and modern arctic ground squirrel (Urocitellus parryii) caches and nests have been used to reconstruct the past vegetation from some parts of Beringia, but such archives are understudied in Alaska. Five modern and four fossil samples from arctic ground squirrel caches and nests provide information on late Pleistocene vegetation in Eastern Beringia. Modern arctic ground squirrel caches from Alaska's arctic tundra were dominated by willow and grass leaves and grass seeds and bearberries, which were widespread in the local vegetation as confirmed by vegetation surveys. Late Pleistocene caches from Interior Alaska were primarily composed of steppe and dry tundra graminoid and herb seeds. Graminoid cuticle analysis of fossil leaves identified Calamagrostis canadensis, Koeleria sp. and Carex albonigra as being common in the fossil samples. Stable carbon isotopes analysis of these graminoid specimens indicated that plants using the C3 photosynthetic pathways were present and functioning with medium to high water-use efficiency. Fossil plant taxa and environments from ground squirrel caches in Alaska are similar to other macrofossil assemblages from the Yukon Territory, which supports the existence of a widespread mammoth steppe ecosystem type in Eastern Beringia that persisted throughout much of the late Pleistocene.

The northeastern sector of the Rub' al-Khali desert in the eastern United Arab Emirates (UAE) is dominated by large NE–SW trending dune ridges orientated perpendicular to the currently prevailing northwesterly wind regime. In this study, extensive use has been made of artificially exposed sections through these major dune ridges that reveal internal sedimentary structures and allow an intensive, high-resolution sampling programme to be carried out. Here, we present the optical dating results for samples from 7 sections. The results indicate that dune activity and preservation occurred within the periods 7–3 ka, 16–10 ka and 22–20 ka with evidence of earlier preservation during marine oxygen isotope stages MIS 3 and 5, with net accumulation rates in the range 2.2–25 m.ka− 1. In several instances, hiatuses in the preserved record of dune accumulation coincide with stratigraphic bounding surfaces visible in the exposed section profiles with associated truncation of internal sedimentary structures. Caution must be exercised when interpreting such gaps in the recorded accumulation chronologies of these dunes since these may simply constitute phases of low preservation potential rather than phases of low aeolian activity. Other factors such as sediment supply and availability in relation to sea-level dynamics may be significant and are also considered.

The Andean piedmont of Mendoza is a semiarid region covered by extensive and partially vegetated dune fields consisting of mostly inactive aeolian landforms of diverse size and morphology. This paper is focused on the San Rafael plain (SRP) environment, situated in the distal Andean piedmont of Mendoza (34° 30′S), and reports the sedimentology and OSL chronology of two representative exposures of late Quaternary deposits, including their paleoenvironmental and paleoclimatic significance. Eleven facies, including channel, floodplain, fluvio–aeolian interaction, and reworked pyroclastic and aeolian deposits, were described and grouped into two facies associations (FA1 and FA2). FA1 was formed by unconfined sheet flows, minor channelized streams and fluvial–aeolian interaction processes. FA2 was interpreted as aeolian dune and sand-sheet deposits. OSL chronology from the SRP sedimentary record indicates that between ca. 58–39 ka and ca. 36–24 ka (MIS 3), aggradation was governed by ephemeral fluvial processes (FA1) under generally semiarid conditions. During MIS 2, the last glacial maximum (ca. 24–12 ka), a major climatic shift to more arid conditions is documented by significant aeolian activity (FA2) that became the dominant sedimentation process north of the Diamante–Atuel fluvial system. The inferred paleoenvironmental conditions from the SRP sections are in broad agreement with regional evidence.

Although glacial landscapes have previously been used for the reconstruction of late Quaternary glaciations in the Central Andes, only few data exist for the Eastern Cordillera in Bolivia. Here, we present results from detailed morphostratigraphic mapping and new data of surface exposure dating (SED), optically stimulated luminescence (OSL), and radiocarbon dating (14C) from the Huara Loma Valley, Cordillera de Cochabamba (Bolivia). Discrepancies between individual dating methods could be addressed within the context of a solid geomorphic framework. We identified two major glaciations. The older is not well constrained by the available data, whereas the younger glaciation is subdivided into at least four major glacial stages. Regarding the latter, a first advance dated to ~ 29–25 ka occurred roughly contemporaneous with the onset of the global last glacial maximum (LGM) and was followed by a less extensive (re-)advance around 20–18 ka. The local last glacial maximum (LLGM) in the Huara Loma Valley took place during the humid lateglacial ~17–16 ka, followed by several smaller readvances until ~10–11 ka, and complete deglaciation at the end of the Early Holocene.

New palynological and sedimentological data from St. Lawrence Island present a rare view into late-glacial and Holocene environments of the central Bering Land Bridge. The late glaciation was a time of dynamic landscape changes in south-central Beringia, with active thermokarst processes, including the formation and drainage of thaw lakes. The presence of such a wet, unstable substrate, if widespread, probably would have had an adverse impact on food sources and mobility for many of the large mammal populations. The establishment of Betula shrub tundra on the island suggests late-glacial summers that were warmer than present, consistent with regional paleoclimatic interpretations. However, the increasing proximity to the Bering Sea, as postglacial sea levels rose, modified the intensity of warming and prevented the establishment of deciduous forest as found in other areas of Beringia at this time. The mid- to late Holocene is marked by more stable land surfaces and development of Sphagnum and Cyperaceae peat deposits. The accumulation of organic deposits, decline of shrub Betula, and decrease in thermokarst disturbance suggest that conditions were cooler than the previous. A recent decline in peat accumulation at the study sites may relate to local geomorphology, but similar decreases have been noted for other arctic regions.

The type section silts of the late Pleistocene Wilson Creek Formation at Mono Lake contain outsized clasts, dominantly well-rounded pebbles and cobbles of Sierran lithologies. Lithic grains > 425 μm show a similar pattern of variability as the > 10 mm clasts visible in the type section, with decreasing absolute abundance in southern and eastern outcrops. The largest concentrations of ice-rafted debris (IRD) occur at 67–57 ka and 46–32 ka, with strong millennial-scale variability, while little IRD is found during the last glacial maximum and deglaciation.

Stratigraphic evidence for high lake level during high IRD intervals, and a lack of geomorphic evidence for coincidence of lake and glaciers, strongly suggests that rafting was by shore ice rather than icebergs. Correspondence of carbonate flux and IRD implies that both were mainly controlled by freshwater input, rather than disparate non-climatic controls. Conversely, the lack of IRD during the last glacial maximum and deglacial highstands may relate to secondary controls such as perennial ice cover or sediment supply. High IRD at Mono Lake corresponds to low glacial flour flux in Owens Lake, both correlative to high warm-season insolation. High-resolution, extra-basinal correlation of the millennial peaks awaits greatly improved age models for both records.

Loess accumulated on a Bull Lake outwash terrace of Marine Oxygen Isotope Stage 6 (MIS 6) age in southern Jackson Hole, Wyoming. The 9 m section displays eight intervals of loess deposition (Loess 1 to Loess 8, oldest), each followed by soil development. Our age-depth model is constrained by thermoluminescence, meteoric 10Be accumulation in soils, and cosmogenic 10Be surface exposure ages. We use particle size, geochemical, mineral-magnetic, and clay mineralogical data to interpret loess sources and pedogenesis. Deposition of MIS 6 loess was followed by a tripartite soil/thin loess complex (Soils 8, 7, and 6) apparently reflecting the large climatic oscillations of MIS 5. Soil 8 (MIS 5e) shows the strongest development. Loess 5 accumulated during a glacial interval (~ 76–69 ka; MIS 4) followed by soil development under conditions wetter and probably colder than present. Deposition of thick Loess 3 (~ 43–51 ka, MIS 3) was followed by soil development comparable with that observed in Soil 1. Loess 1 (MIS 2) accumulated during the Pinedale glaciation and was followed by development of Soil 1 under a semiarid climate. This record of alternating loess deposition and soil development is compatible with the history of Yellowstone vegetation and the glacial flour record from the Sierra Nevada.

A prominent lake formed when glaciers descending from the Kodar Range blocked the River Vitim in central Transbaikalia, Siberia. Glacial Lake Vitim, evidenced by palaeoshorelines and deltas, covered 23,500 km2 and held a volume of ~ 3000 km3. We infer that a large canyon in the area of the postulated ice dam served as a spillway during an outburst flood that drained through the rivers Vitim and Lena into the Arctic Ocean. The inferred outburst flood, of a magnitude comparable to the largest known floods on Earth, possibly explains a freshwater spike at ~ 13 cal ka BP inferred from Arctic Ocean sediments.

Understanding the timing of mountain glacier and paleolake expansion and retraction in the Great Basin region of the western United States has important implications for regional-scale climate change during the last Pleistocene glaciation. The relative timing of mountain glacier maxima and the well-studied Lake Bonneville highstand has been unclear, however, owing to poor chronological limits on glacial deposits. Here, this problem is addressed by applying terrestrial cosmogenic 10Be exposure dating to a classic set of terminal moraines in Little Cottonwood and American Fork Canyons in the western Wasatch Mountains. The exposure ages indicate that the main phase of deglaciation began at 15.7 ± 1.3 ka in both canyons. This update to the glacial chronology of the western Wasatch Mountains can be reconciled with previous stratigraphic observations of glacial and paleolake deposits in this area, and indicates that the start of deglaciation occurred during or at the end of the Lake Bonneville hydrologic maximum. The glacial chronology reported here is consistent with the growing body of data suggesting that mountain glaciers in the western U.S. began retreating as many as 4 ka after the start of northern hemisphere deglaciation (at ca. 19 ka).

Thousands of Paleolithic artifacts have been recovered from Paleolithic sites in the Luonan Basin, in the upper South Luohe River of central China. Their discovery suggests that the basin was an important area for hominin settlement during the Pleistocene. However, the initial timing of this occupation and the environmental conditions for this period are still largely unknown. In addition, the sediments are not well dated and most of the artifacts lie on the surface. In an attempt to resolve these issues, a new systemic paleomagnetic analysis was carried out on the loess deposits that contain in situ stone tools. Our detailed loess–paleosol analyses of the stratigraphy of different sites in the basin and Chinese Loess Plateau shows the accumulation of the loess since at least 1.1 million years (Ma) ago. Moreover, recently discovered in situ cores, flakes and retouched stone tools in these deposits show that hominins used this region repeatedly from 0.8–0.7 Ma to 0.4–0.3 and 0.2–0.1 Ma. Pedostratigraphic analyses, magnetic susceptibility and carbon isotope analyses also indicate that these hominins lived in a subtropical to warm-temperate climate with broad-needle-leaf forest vegetation mixed with grasses.

Previous studies have suggested that Marine Isotope Stage (MIS) 13, recognized as atypical in many paleoclimate records, is marked by the development of anomalously strong summer monsoons in the northern tropical areas. To test this hypothesis, we performed a multi-proxy study on three marine records from the tropical Indian Ocean in order to reconstruct and analyse changes in the summer Indian monsoon winds and precipitations during MIS 13. Our data confirm the existence of a low-salinity event during MIS 13 in the equatorial Indian Ocean but we argue that this event should not be considered as “atypical”. Taking only into account a smaller precession does not make it possible to explain such precipitation episode. However, when considering also the larger obliquity in a more complete orbitally driven monsoon “model,” one can successfully explain this event. In addition, our data suggest that intense summer monsoon winds, although not atypical in strength, prevailed during MIS 13 in the western Arabian Sea. These strong monsoon winds, transporting important moisture, together with the effect of insolation and Eurasian ice sheet, are likely one of the factors responsible for the intense monsoon precipitation signal recorded in China loess, as suggested by model simulations.

Tight Entrance Cave (TEC) in southwestern Australia provides a Pleistocene sequence documenting the extinction of 14 large mammal species. This record has been interpreted as indicating that extinctions did not occur during or before the penultimate glacial maximum (PGM) and that humans played a primary role in the extinctions. However, it remains possible that the majority of extinct megafauna persisted no later than the PGM. The TEC extinctions correspond with vegetation change, a cooling/drying trend, increased biomass burning, and increasingly unstable small mammal communities. The initiation of these trends predates human arrival on the continent and implies environmentally mediated extinctions.

Many paleoclimate and landscape change studies in the American Midwest have focused on the Late Glacial and early Holocene time periods (~ 16–11 ka), but little work has addressed landscape change in this area between the Last Glacial Maximum and the Late Glacial (~ 22–16 ka). Sediment cores were collected from 29 new lake and bog sites in Ohio and Indiana to address this gap. The basal radiocarbon dates from these cores show that initial ice retreat from the maximal last-glacial ice extent occurred by 22 ka, and numerous sites that are ~ 100 km inside this limit were exposed by 18.9 ka. Post-glacial environmental changes were identified as stratigraphic or biologic changes in select cores. The strongest signal occurs between 18.5 and 14.6 ka. These Midwestern events correspond with evidence to the northeast, suggesting that initial deglaciation of the ice sheet, and ensuing environmental changes, were episodic and rapid. Significantly, these changes predate the onset of the Bølling postglacial warming (14.8 ka) as recorded by the Greenland ice cores. Thus, deglaciation and landscape change around the southern margins of the Laurentide Ice Sheet happened ~ 7 ka before postglacial changes were felt in central Greenland.