The Salinas del Bebedero occupies an isolated basin in the foreland of central Argentina at 33°S and was flooded repeatedly over past 25 ka. Isotopic evidence demonstrates that this flooding was due to overflow of the nearby Río Desaguadero with waters derived from the distant (≥300 km) central Andes between 28–34°S. Stratigraphic and shoreline evidence shows that floods occurred most frequently from 14.3 to 11.4 ka, followed by lesser events between 14.3 to 11.4 ka, and during the late Holocene from 2.6 to ca. 0.2 ka. Hydraulic modeling (2D HEC-RAS) shows that these floods could have originated from repeated subglacial drainage or sudden outbursts with a volume of >100 × 106 m3 and a peak discharge of >1,000 m3 s-1 each. The absence of flood deposits from 11 to 3 ka points to exceptionally dry and virtually ice-free conditions in the Andes between 28–34°S. The floods were probably caused by major rainfall or dammed-lake outbursts clustered largely during wet pluvial periods in the otherwise moisture-limited central Andes and Atacama Desert, such as when the Intertropical Convergence Zone was shifted southward. These include Central Andean pluvial events (CAPE) I (17–14.5 ka) and II (12.5–9 ka), and the Neoglacial/Formative archeological period 2500 ka to near-present.

A distinct feature of many of the earliest archaeological sites (13,000-11,200 cal yr BP) at the core of the Atacama Desert is that they lie at or just below the surface, often encased in desert pavements. In this study, we compare these sites and undisturbed desert pavements to understand archaeological site formation and pavement development and recovery. Our results indicate these pavements and their soils are poorly developed regardless of their age. We propose that this is because of sustained lack of rain and extreme physical breakdown of clasts by salt expansion. Thus, the core of the Atacama provides an example of the lower limits of rainfall (<50 mm/yr) needed to form desert pavements. At site Quebrada Maní 12 (QM12), humans destroyed the pavement. After abandonment, human-made depressions were filled with eolian sands, incorporating artifacts in shallow deposits. Small and medium-sized artifacts preferentially migrated upwards, perhaps due to earthquakes and the action of salts. These artifacts, which now form palimpsests at the surface, helped – along with older clasts - to restore surface clast cover. Larger archaeological features remained undisturbed on top of a deeper Byzm horizon. The vesicular A horizons (Av horizons) have not regenerated on the archaeological sites due to extreme scarcity of rainfall during the Holocene.

The iconic Plaza Tree of Pueblo Bonito is widely believed to have been a majestic pine standing in the west courtyard of the monumental great house during the peak of the Chaco Phenomenon (AD 850–1140). The ponderosa pine (Pinus ponderosa) log was discovered in 1924, and since then, it has been included in “birth” and “life” narratives of Pueblo Bonito, although these ideas have not been rigorously tested. We evaluate three potential growth origins of the tree (JPB-99): Pueblo Bonito, Chaco Canyon, or a distant mountain range. Based on converging lines of evidence—documentary records, strontium isotopes (87Sr/86Sr), and tree-ring provenance testing—we present a new origin for the Plaza Tree. It did not grow in Pueblo Bonito or even nearby in Chaco Canyon. Rather, JPB-99 originated from the Chuska Mountains, over 50 km west of Chaco Canyon. The tree was likely carried to Pueblo Bonito sometime between AD 1100 and 1130, although why it was left in the west courtyard, what it meant, and how it might have been used remain mysteries. The origin of the Plaza Tree of Pueblo Bonito underscores deep cultural and material ties between the Chaco Canyon great houses and the Chuska landscape.

We present a shoreline-based, millennial-scale record of lake-level changes spanning 12.8–2.3 ka for a large closed-basin lake system on the southwestern Tibetan Plateau. Fifty-three radiocarbon and eight U–Th series ages of tufa and beach cement provide age control on paleoshorelines ringing the basin, supplemented by nineteen ages from shell and aquatic plant material from natural exposures generally recording lake regressions. Our results show that paleo-Ngangla Ring Tso exceeded modern lake level (4727 m asl) continuously between ~ 12.8 and 2.3 ka. The lake was at its highstand 135 m (4862 m asl) above the modern lake from 10.3 ka to 8.6 ka. This is similar to other closed-basin lakes in western Tibet, and coincides with peak Northern Hemisphere summer insolation and peak Indian Summer Monsoon intensity. The lake experienced a series of millennial-scale oscillations centered on 11.5, 10.8, 8.3, 5.9 and 3.6 ka, consistent with weak monsoon events in proxy records of the Indian Summer Monsoon. It is unclear whether these events were forced by North Atlantic or Indian Ocean conditions, but based on the abrupt lake-level regressions recorded for Ngangla Ring Tso, they resulted in significant periodic reductions in rainfall over the western Tibetan Plateau throughout the Holocene.

Black mats are prominent features of the late Pleistocene and Holocene stratigraphic record in the southern Great Basin. Faunal, geochemical, and sedimentological evidence shows that the black mats formed in several microenvironments related to spring discharge, ranging from wet meadows to shallow ponds. Small land snails such as Gastrocopta tappaniana and Vertigo berryi are the most common mollusk taxa present. Semiaquatic and aquatic taxa are less abundant and include Catinellids, Fossaria parva, Gyraulus parvus, and others living today in and around perennial seeps and ponds. The ostracodes Cypridopsis okeechobi and Scottia tumida, typical of seeps and low-discharge springs today, as well as other taxa typical of springs and wetlands, are common in the black mats. Several new species that lived in the saturated subsurface also are present, but lacustrine ostracodes are absent. The δ13C values of organic matter in the black mats range from −12 to −26‰, reflecting contributions of tissue from both C3 (sedges, most shrubs and trees) and C4 (saltbush, saltgrass) plants. Carbon-14 dates on the humate fraction of 55 black mats fall between 11,800 to 6300 and 2300 14C yr B.P. to modern. The total absence of mats in our sample between 6300 and 2300 14C yr B.P. likely reflects increased aridity associated with the mid-Holocene Altithermal. The oldest black mats date to 11,800–11,600 14C yr B.P., and the peak in the 14C black mat distribution falls at ∼10,000 14C yr B.P. As the formation of black mats is spring related, their abundance reflects refilling of valley aquifers starting no later than 11,800 and peaking after 11,000 14C yr B.P. Reactivation of spring-fed channels shortly before 11,200 14C yr B.P. is also apparent in the stratigraphic records from the Las Vegas and Pahrump Valleys. This age distribution suggests that black mats and related spring-fed channels in part may have formed in response to Younger Dryas (YD)-age recharge in the region. However, the inception of black mat formation precedes that of the YD by at least 40014C yr, and hydrological change is gradual, not rapid.

Patterns of climate and C4plant abundance in the southwestern United States during the last glaciation were evaluated from isotopic study of herbivore tooth enamel. Enamel δ13C values revealed a substantial eastward increase in C4plant consumption for Mammuthusspp., Bisonspp., Equusspp., and Camelopsspp. The δ13C values were greatest in Bisonspp. (−6.9 to +1.7‰) and Mammuthusspp. (−9.0 to +0.3‰), and in some locales indicated C4-dominated grazing. The δ13C values of Antilocaprids were lowest among taxa (−12.5 to −7.9‰) and indicated C3feeding at all sites. On the basis of modern correlations between climate and C4grass abundance, the enamel data imply significant summer rain in parts of southern Arizona and New Mexico throughout the last glaciation. Enamel δ18O values range from +19.0 to +31.0‰ and generally increase to the east. This pattern could point to a tropical or subtropical source of summer rainfall. At a synoptic scale, the isotope data indicate that interactions of seasonal moisture, temperature, and lowered atmospheric pCO2determined glacial-age C4abundance patterns.

The Asian Monsoon, which brings ~80% of annual precipitation to much of the Tibetan Plateau, provides runoff to major rivers across the Asian continent. Paleoclimate records indicate summer insolation and North Atlantic paleotemperature changes forced variations in monsoon rainfall through the Holocene, resulting in hydrologic and ecologic changes in plateau watersheds. We present a record of Holocene hydrologic variability in the Yarlung Tsangpo (YT) valley of the southern Tibetan Plateau, based on sedimentology and 14C dating of organic-rich black mats’ in paleowetlands deposits, that shows changes in wetlands extent in response to changing monsoon intensity. Four sedimentary units indicate decreasing monsoon intensity since 10.4 ka BP. Wet conditions occurred at ~10.4 ka BP, ~9.6 ka BP and ~7.9–4.8 ka BP, with similar-to-modern conditions from ~4.6–2.0 ka BP, and drier-than-modern conditions from ~2.0 ka BP to present. Wetland changes correlate with monsoon intensity changes identified in nearby records, with weak monsoon intervals corresponding to desiccation and erosion of wetlands. Dating of in situ ceramic and microlithic artifacts within the wetlands indicates Epipaleolithic human occupation of the YT valley after 6.6 ka BP, supporting evidence for widespread colonization of the Tibetan Plateau in the early and mid-Holocene during warm, wet post-glacial conditions.

Five stratigraphic units and five soils of late Pleistocene to Holocene age crop out in dissected badlands on Corn Creek Flat, 30 km northwest of Las Vegas, Nevada, and at Tule Springs, nearer to Las Vegas. The record is dominantly fluvial but contains evidence of several moister, marsh-forming periods: the oldest (Unit B) dates perhaps to the middle Wisconsin, and the more widespread Unit D falls between 30,000 and 15,000 yr B.P. Unit D therefore correlates with pluvial maximum lacustrine deposits elsewhere in the Great Basin. Standing water was not of sufficient depth or extent during either period to form lake strandlines. Between 14,000 and 7200 yr B.P. (Unit E), standing surface water gradually decreased, a trend also apparent in Great Basin pluvial lake chronologies during the same period. Groundwater carbonate cementation and burrowing by cicadas (Cicadae) accompany the moist-phase units. After 7200 yr B.P., increased wind action, decreased biotic activity, and at least 25 m of water-table lowering accompanied widespread erosion of older fine-grained deposits. Based on pack-rat midden and pollen evidence, this coincides with major vegetation changes in the valley, from sagebrush-dominated steppe to lower Mohave desertscrub.

Badland exposures in the Indian Springs Valley, southern Nevada, contain evidence of formerly widespread spring and seep discharge. The stratigraphic position and appearance of most of these deposits suggests correlation with late Wisconsin (30,000 to ca. 10,000 yr B.P.) marsh sediments in nearby Las Vegas Valley. Previously, all these deposits have been loosely described as lacustrine because of the presence of extensive green mudstones associated with aquatic mollusks. However, this association also typifies modern groundwater discharge environments in many basins of northeast Nevada such as the Steptoe Valley, basins often without hydrographic closure. Such analogs best explain the origin of late Wisconsin fine-grained deposits in the unclosed southwestern arm of the Indian Springs Valley. Key features of these depositional systems are the lack of shoreline deposits, the presence of a broad belt of subaerially deposited plae-brown silts surrounding spring, “wet meadow,” and marsh deposits, and the intermixture of terrestrial and aquatic mollusks in most horizons where mollusks occur.

Evidence from shoreline and deep-lake sediments show Laguna Cari-Laufquén, located at 41°S in central Argentina, rose and fell repeatedly during the late Quaternary. Our results show that a deep (> 38 m above modern lake level) lake persisted from no later than 28 ka to 19 ka, with the deepest lake phase from 27 to 22 ka. No evidence of highstands is found after 19 ka until the lake rose briefly in the last millennia to 12 m above the modern lake, before regressing to present levels. Laguna Cari-Laufquén broadly matches other regional records in showing last glacial maximum (LGM) highstands, but contrasts with sub-tropical lake records in South America where the hydrologic maximum occurred during deglaciation (17–10 ka). Our lake record from Cari-Laufquén mimics that of high-latitude records from the Northern Hemisphere. This points to a common cause for lake expansions, likely involving some combination of temperature depression and intensification of storminess in the westerlies belt of both hemispheres during the LGM.

14C dating of shoreline deposits of closed-basin lake Baqan Tso in the western Tibetan Plateau shows that lake level regressed from the undated highstand (46 m above modern, 4.3 × modern surface area) of likely earliest Holocene age by 11.5 ka, and remained larger than modern until at least ≈ 5.0 ka. The shoreline record broadly matches other regional climate records, with lake level closely following Northern Hemisphere summer insolation overprinted by sub-millennial lake-level oscillations. A model coupling modern land runoff and lake surface heat closely reproduces estimated modern precipitation of ≈ 240 mm/yr. We estimate that the Baqan Tso basin required ≈ 380 mm/yr precipitation to sustain the maximum early Holocene lake area, a 55% increase over modern. Precipitation increases, not glacial meltwater, drove lake-level changes, as Baqan Tso basin was not glaciated during the Holocene. Our estimate assumes early Holocene insolation (≈ 1.3% overall increase), and mean annual increases of 2°C in temperature, and 37% in relative humidity. We additionally developed a Holocene precipitation history for Baqan Tso using dated paleolake areas. Using the modern and early Holocene model results as end-members, we estimate precipitation in the western Tibetan Plateau which was 300–380 mm/yr between 5.0 and 11.5 ka, with error of ± 29–57 mm/yr (± 12–15%).

We have developed an 87Sr/86Sr, 234U/238U, and δ18O data set from carbonates associated with late Quaternary paleolake cycles on the southern Bolivian Altiplano as a tool for tracking and understanding the causes of lake-level fluctuations. Distinctive groupings of 87Sr/86Sr ratios are observed. Ratios are highest for the Ouki lake cycle (120–95 ka) at 0.70932, lowest for Coipasa lake cycle (12.8–11.4 ka) at 0.70853, and intermediate at 0.70881 to 0.70884 for the Salinas (95–80 ka), Inca Huasi (~ 45 ka), Sajsi (24–20.5 ka), and Tauca (18.1–14.1 ka) lake cycles. These Sr ratios reflect variable contributions from the eastern and western Cordilleras. The Laca hydrologic divide exerts a primary influence on modern and paleolake 87Sr/86Sr ratios; waters show higher 87Sr/86Sr ratios north of this divide. Most lake cycles were sustained by slightly more rainfall north of this divide but with minimal input from Lake Titicaca. The Coipasa lake cycle appears to have been sustained mainly by rainfall south of this divide. In contrast, the Ouki lake cycle was an expansive lake, deepest in the northern (Poópo) basin, and spilling southward. These results indicate that regional variability in central Andean wet events can be reconstructed using geochemical patterns from this lake system.

Widespread, organic-rich diatomaceous deposits are evidence for formerly wetter times along the margins of the central Atacama Desert, one of the driest places on Earth today. We mapped and dated these paleowetland deposits at three presently waterless locations near Salar de Punta Negra (24.5°S) on the western slope of the Andes. Elevated groundwater levels supported phreatic discharge into wetlands during two periods: 15,900 to ~ 13,800 and 12,700 to ~ 9700 cal yr BP. Dense concentrations of lithic artifacts testify to the presence of paleoindians around the wetlands late in the second wet phase (11,000?–9700 cal yr BP). Water tables dropped below the surface before 15,900 and since 8100 cal yr BP, and briefly between ~ 13,800 and 12,700 cal yr BP. This temporal pattern is repeated, with some slight differences, in rodent middens from the study area, in both paleowetland and rodent midden deposits north and south of the study area, and in lake level fluctuations on the adjacent Bolivian Altiplano. The regional synchroneity of these changes points to a strengthening of the South American Monsoon — which we term the "Central Andean Pluvial Event" — in two distinct intervals (15,900–13,800 and 12,700–9700 cal yr BP), probably induced by steepened SST gradients across the tropical Pacific (i.e., La Niña-like conditions).

Rodent (Abrocoma, Lagidium, Phyllotis) middens collected from 2350 to 2750 m elevation near Arequipa, Peru (16°S), provide an ∼9600-yr vegetation history of the northern Atacama Desert, based on identification of >50 species of plant macrofossils. These midden floras show considerable stability throughout the Holocene, with slightly more mesophytic plant assemblages in the middle Holocene. Unlike the southwestern United States, rodent middens of mid-Holocene age are common. In the Arequipa area, the midden record does not reflect any effects of a mid-Holocene mega drought proposed from the extreme lowstand (100 m below modern levels, >6000 to 3500 yr B.P.) of Lake Titicaca, only 200 km east of Arequipa. This is perhaps not surprising, given other evidence for wetter summers on the Pacific slope of the Andes during the middle Holocene as well as the poor correlation of summer rainfall among modern weather stations in the central Andes-Atacama Desert. The apparent difference in paleoclimatic reconstructions suggests that it is premature to relate changes observed during the Holocene to changes in El Niño Southern Oscillation modes.

A late Quaternary ichthyofauna from Homestead Cave, Utah, provides a new source of information on lake history in the Bonneville basin. The fish, represented by 11 freshwater species, were accumulated between ∼11,200 and ∼1000 14C yr B.P. by scavenging owls. The 87Sr/86Sr ratio of Lake Bonneville varied with its elevation; 87Sr/86Sr values of fish from the lowest stratum of the cave suggest they grew in a lake near the terminal Pleistocene Gilbert shoreline. In the lowest deposits, a decrease in fish size and an increase in species tolerant of higher salinities or temperatures suggest multiple die-offs associated with declining lake levels. An initial, catastrophic, post-Provo die-off occurred at 11,300–11,200 14C yr B.P. and was followed by at least one rebound or recolonization of fish populations, but fish were gone from Lake Bonneville sometime before ∼10,400 14C yr B.P. This evidence is inconsistent with previous inferences of a near desiccation of Lake Bonneville between 13,000 and 12,000 14C yr B.P. Peaks in Gila atraria frequencies in the upper strata suggest the Great Salt Lake had highstands at ∼3400 and ∼1000 14C yr B.P.

Both aquatic and land snails are common in the geologic record, but their utility in dating is greatly restricted by their well-documented tendency to yield14C dates inconsistent with true14C ages. In this study, we examine the use of14C ages from (1) small, previously unstudied, terrestrial snails to date hosting spring deposits and from (2) cooccuring aquatic snails to constrain groundwater travel times during the last glacial period. Our study area in the southern Great Basin encompasses Yucca Mountain, site of the proposed high-level nuclear waste repository, where information on the age and extent of past high water tables and on groundwater flow times is crucial to several licensing issues. Our results show that shells of small terrestrial snails belonging toValloniasp. yield14C dates consistent with14C ages of associated carbonized wood. These results imply that these taxa can provide reliable14C age control on the broadly distributed deposits in which they have been described. In contrast, cooccurring aquatic snails from fossil spring deposits yield14C ages generally greater than the control age. This is because the aquatic shells often formed in spring waters that had an initial14C deficiency. However, the magnitude of the deficiency is much less than that observed in nearby modern springs, arguing for much higher average14C contents in late Pleistocene groundwaters in these basins. If representative, this implies shorter groundwater travel times through aquifers in southern Nevada during late-glacial time.

Lacustrine deposits exposed around Lake Aricota, Peru (17° 22′S), a 7.5-km2 lake dammed by debris flows, provide a middle to late Holocene record of lake-level fluctuations. Chronological context for shoreline deposits was obtained from radiocarbon dating of vascular plant remains and other datable material with minimal 14C reservoir effects (<350 yr). Diatomites associated with highstands several meters above the modern lake level indicate wet episodes. Maximum Holocene lake level was attained before 6100 14C yr B.P. and ended ∼2700 14C yr B.P. Moderately high lake levels occurred at 1700 and 1300 14C yr B.P. The highstand at Lake Aricota during the middle Holocene is coeval with a major lowstand at Lake Titicaca (16°S), which is only 130 km to the northeast and shares a similar climatology. Comparisons with other marine and terrestrial records highlight emerging contradictions over the nature of mid-Holocene climate in the central Andes.