A new method is presented for using known ordering or other relationships between 14C samples to reduce 14C dating uncertainty. The order of sample formation is often known from, for example, stratigraphic superposition, dendrochronology, or crosscutting field relations. Constraints such as a minimum time between dates and limits from historical information are also readily included. Dendrochronologically calibrated calendric date histograms initially represent each date. The method uses Bayes theorem and the relational constraints to upweight date ranges in each date distribution consistent with the other date distributions and the constraints, and downweight unlikely portions. The reweighted date distributions retain all dating possibilities present in the initial calibrated date distributions, but each date in the result now reflects the extra information such as ordering supplied through the constraints. In addition, one may add information incrementally, and thus analyze systematically its effect on all the date distributions. Thus, the method can be used to assess the consistency of the quantitative data at hand. The Bayesian approach also uses the empirical calibrated date distributions directly, so information is not lost prematurely by summarized dates to a mean and variance or "confidence intervals." The approach is illustrated with data from two densely sampled paleoseismic sites on the San Andreas Fault in southern California. An average reduction in 14 C date distribution variance of 59% is achieved using ordering information alone, and 85% is achieved by also applying sedimentation rate constraints and historical information.

Twenty deep-sea sediment cores from the western Pacific between 30°N and 30°S provide evidence of sea-surface temperature (SST) changes throughout the tropics and subtropics. Glacial SSTs were estimated using the modern analog technique (MAT) applied to planktonic foraminifers and planktonic foraminiferal δ18O changes. We used δ18O to identify the last glacial maximum. The MAT method differs from the traditional transfer function approach in that it utilizes a global coretop database, and estimates paleotemperature by finding analogs from the modern coretop samples. In addition, the MAT approach appears to be less susceptible than the transfer function technique to biases introduced by carbonate dissolution. Our results indicate that tropical SSTs differed by less than 2°C from present; away from the tropics (30°N and 30°S) SSTs were at least 3°C cooler. Our results differ from those of previous studies in the western Pacific by using a set of well-preserved, high-sedimentation rate cores from shallow regions. The results of this study clearly indicate that a western Pacific warm pool existed during the last glacial maximum (LGM), providing a heat and moisture source for a Walker Circulation cell similar to that of today. We propose that a steeper lapse rate existed during the last glacial maximum and that this can explain at least part of the discrepancy between marine and terrestrial temperature estimates adjacent to New Guinea for the LGM.

Ice-walled melt ponds on the surfaces of active valley-floor rock glaciers and Matthes (Little Ice Age) moraines in the southern Sierra Nevada indicate that most of these landforms consist of glacier ice under thin (ca. 1 - 10 m) but continuous covers of rock-fall-generated debris. These debris blankets effectively insulate the underlying ice and greatly reduce rates of ablation relative to that of uncovered ice. Such insulation explains the observations that ice-cored rock glaciers in the Sierra, actually debris-covered glaciers, are apparently less sensitive to climatic warming and commonly advance to lower altitudes than do adjacent bare-ice glaciers. Accumulation-area ratios and toe-to-headwall-altitude ratios used to estimate equilibrium-line altitudes (ELAs) of former glaciers may therefore yield incorrect results for cirque glaciers subject to abundant rockfall. Inadvertent lumping of deposits from former debris-covered and bare-ice glaciers partially explains an apparently anomalous regional ELA gradient reported for the pre-Matthes Recess Peak Neoglacial advance. Distinguishing such deposits may be important to studies that rely on paleo-ELA estimates. Moreover, Matthes and Recess Peak ELA gradients along the crest evidently depend strongly on local orographic effects rather than latitudinal climatic trends, indicating that simple linear projections and regional climatic interpretations of ELA gradients of small glaciers may be unreliable.

The depression of snowlines, or equilibrium-line altitudes, of alpine glaciers is often used by glacial geologists to infer variations in mass balance. The climatic interpretation of snowline depression, however, is complicated by the number of factors that control glacier mass balance. The simple lapse-rate method of temperature interpretation ignores the effects of changes in radiation and snow accumulation. The statistical approach to temperature interpretation, which regresses precipitation and temperature against snowline altitude, neglects the effect of radiation. The most comprehensive approach for the climatic interpretation of snowline depression couples the heat and mass balances of a glacier surface. A sensitivity analysis that utilizes the coupled heat- and mass-balance approach indicates that the ∼1000-m variation in snowline of alpine glaciers on glacial-to-interglacial time scales could be a result of significant changes in temperature, and to a lesser extent changes in insolation. Snowline variations are sensitive only to relatively large changes in annual accumulation, which should also be evident in other proxy records of moisture change. The approaches outlined here provide glacial geologists with a summary of how various climatic forcings associated with glaciation may be quantified from snowline data.

Geologic records of fluctuations of the Laurentide ice sheet margin following the most recent glacial maximum (ca. 20,000 14C yr B.P.) identify fundamental differences in ice-sheet behavior depending on subglacial bed conditions. Rapid and irregular icemargin fluctuations occurred only over areas of deforming sediment, indicating nonclimatic forcing controlled by the inherent instability of coupled ice sheet-deforming sediment dynamics. In contrast, largely uninterrupted ice-margin retreat with no evidence of significant readvance occurred over rigid-bed areas, indicating stable behavior. Unstable ice-sheet behavior was most pronounced from 15,000 until 10,000 14C yr B.P., by which time most of the ice margin had retreated onto a rigid bed. Unstable ice-sheet behavior would have been an integral component in controlling variable fluxes of icebergs and meltwater, as well as meltwater routing, to the North Atlantic, thus affecting thermohaline circulation. The abrupt climate oscillations in the North Atlantic region that ended at 10,000 14C yr B.P. may thus have their origin in the inherently unstable behavior of the Laurentide ice sheet overriding deforming sediment.

Uranium-series age determinations by mass spectrometric methods were done for travertines and associated carbonate veins related to clastic deposits of the last glaciation (Pinedale) in the northern Yellowstone area. Dramatic variations in the hydrologic head are inferred from variations in the elevation of travertine deposition with time and are consistent with the expected hydrologic effects of glaciation. We determine the following chronology of the Pinedale Glaciation, with the key assumption that travertine deposits (and associated carbonate veins) perched high above present thermal springs were deposited when glaciers filled the valley below these perched deposits: (1) the early Pinedale outlet glacier advanced well downvalley between 47,000 and 34,000 yr B.P.; (2) the outlet glacier receded to an interstadial position between 34,000 and 30,000 yr B.P.; (3) an extensive Pinedale ice advance occurred between 30,000 and 22,500 yr B.P.; (4) a major recession occurred between 22,500 and 19,500 yr B.P.; (5) a minor readvance (Deckard Flats) culminated after 19,500 yr B.P.; and (6) recession from the Deckard Flats position was completed before 15,500 yr B.P. This chronology is consistent with the general trend of climatic changes in the northern hemisphere as revealed by recent high-resolution ice-core records from the Greenland ice sheet.

Two major meltwater events are documented in cores from the NW Labrador Sea. One occurred ca. 20,000 14C yr B.P. in association with deposition of a major detrital carbonate unit. Both prior to and after this event, δ18O values of near-surface planktonic foraminifera were 4.5%, indicating fully enriched glacial values. A younger event (ca. 14,000 14 C yr B.P.) is characterized by a dramatic change in δ18O from 4.5 to 2.0% and coincided with the retreat of ice from the outer SE Baffin Shelf, possibly into Hudson Strait. These meltwater events coincide with Heinrich (H) layers 1 and 2 from North Atlantic sediments. The 14,000 14C yr B.P. meltwater event indicates that the eastern margin of the Laurentide Ice Sheet also underwent rapid retreat at approximately the same time as other ice sheet margins around the NE North Atlantic. A third major detrital carbonate event at the base of HU87-033-009, possibly correlative with Heinrich layer 3, occurred ca. 33,960 ± 675 14 C yr B.P.; however, this is older than the accepted date for H-3 of 27,000 14C yr B.P. and may be H-4.

Large-magnitude flooding of the Mississippi River from proglacial lakes Agassiz and Superior most likely occurred between 11,300 and 10,900 and 9900 and 9500 yr B.P. The Charleston alluvial fan, a depositional remnant of one of these floods, is located at the head of a wide alluvial plain near Charleston, Missouri. The fan is an elongate, convex-up sand body (16 × 24 km) composed of medium- and fine-grained sand at least 8 m thick. This sand contrasts with the older coarse-grained sand of the braided stream surface to the west and south and younger silty clay of the meandering stream level to the north and east. A weakly developed soil separates the underlying braided steam deposits from the alluvial fan. A bulk-soil radiocarbon date of 10,590 ± 200 yr B.P. from the contact between the fan and clays of the meandering stream system indicates that the Charleston fan was deposited near the end of the early interval of flooding from Lake Agassiz about 10,900 yr B.P. If the Charleston fan is the last remnant of deglacial flooding in the lower Mississippi Valley, then deposition of significant quantities of sediment from largemagnitude floods between 10,000 and 9500 yr B.P. did not extend into the lower Mississippi Valley through Thebes Gap.

Late-Pleistocene glacial sequences at McGee, Pine, and Bishop creeks, in the east-central Sierra Nevada, are resolvable into distinct relative-age groups on the basis of soil development, the weathering of surface and subsurface clasts, and geomorphic criteria. The data differentiate moraines from Tioga and Tahoe glaciations at McGee Creek, and Tioga, Tahoe, and pre-Tahoe/post-Sherwin glaciations at Pine and Bishop creeks; no moraines from a Tenaya glaciation are differentiated by the data, perhaps because the soil-geomorphic methods are not sensitive enough to resolve small age differences between moraines. The lack of fine age resolution is probably due to a combination of factors, including (1) slow rates of soil development, (2) low amounts of atmospheric dust added to the soils, (3) the susceptibility of crest soils to erosion, and (4) the susceptibility of footslope soils to burial by colluvium. Age resolution is improved by evaluating soils at both the moraine crests and the relatively wetter footslope sites, and by basing age assignments on a combination of macro- and micromorphologic soil properties, the disintegration of subsurface clasts, and parameters of surface-clast weathering.

Several Holocene tephra deposits of Hayes volcano constitute a marker horizon in southern and east-central Alaska. Their identification is aided by high amphibole/pyroxene ratio and biotite in trace amounts, unique among Holocene tephra deposits of the region. However, correlations are obscured by chemical heterogeneity of the glass which occurs at a scale less than the size of a lapilli. Single-shard analyses confirm that the heterogeneity is due neither to fractionation nor to plagioclase microlites. The heterogeneity may be due to mixing of magmas prior to eruption. It is proposed that the deposits be informally called Hayes tephra set H.

A peaty marsh soil is sharply overlain by a sand sheet and intertidal mud at tidal marshes near Tofino and Ucluelet, Vancouver Island, British Columbia. Foraminifera and vascular plant fossils show that the buried soil was submerged suddenly and was covered quickly by sand. Radiocarbon ages place this event between 100 and 400 yr ago. The coastal subsidence suggested by the submergence occurred in an area of net late Holocene emergence, perhaps during the most recent great earthquake on the northern part of the Cascadia subduction zone. The sand sheet overlying the peaty soil records the tsunami triggered by this earthquake. Similar stratigraphic sequences of about the same age have been reported from estuaries along the outer coasts of Washington and northern Oregon, suggesting that hundreds of kilometers of the Cascadia subduction zone may have ruptured during one, or a series of plate-boundary earthquakes less than 400 yr ago.

The concentrations of 15 elements and mineral grain-size distributions were determined for Late Quaternary loess sediments from the Chinese Loess Plateau. Particle-size distributions of the loess were approximately log-normal for each of three modes (i.e., 0.04 to 1 μm diameter, 2 to 20 μm, and 20 to 200 μm) that integrate the transport and deposition of eolian dust. The mass particle-size distributions of typical crustal elements (Si, Ca, Fe, Ti, K, Mn, and V) in paleosol and loess samples show patterns similar to those of mineral aerosols, but higher concentrations occur in the submicrometer sediment particles. Absolute principal component analysis of the elemental data suggests that the loess can be described in terms of synsedimentary and postdepositional concentration and subvertical migration of CaCO3 and various elements through the deposit. A variety of geochemical indicators in the loess are apparently sensitive to Asian monsoon variations; these include the mass ratio of course to fine particles, the concentration ratios of certain trace elements, the elemental fluxes associated with eolian dust, and the bulk concentration of Ca in the sediments.

Experiments carried out under temperatures and relative humidities that approximate ambient conditions show that the rate of hydration of obsidian is a function of the relative humidity, as well as of previously established variables of temperature and obsidian chemical composition. Measurements of the relative humidity of soil at 25 sites and at depths of between 0.01 and 2 m below ground show that in most soil environments, at depths below about 0.25 m, the relative humidity is constant at 100%. We have found that the thickness of the hydrated layer developed on obsidian outcrops exposed to the sun and to relative humidities of 30-90% is similar to that formed on other portions of the outcrop that were shielded from the sun and exposed to a relative humidity of approximately 100%. Surface samples of obsidian exposed to solar heating should hydrate more rapidly than samples buried in the ground. However, the effect of the lower mean relative humidity experiences by surface samples tends to compensate for the elevated temperature, which may explain why obsidian hydration ages of surface samples usually approximate those derived from buried samples.

A study of Wisconsinan loess in part of southeastern Minnesota confirms earlier suggestions that much of the loess in this region was not derived flora the floodplain of the Mississippi River. Two Wisconsinan loess units, the Peoria Loess and Roxana Silt, occur in the study area. Peoria Loess, 1-8 m thick, fines systematically eastward from an abrupt western border toward the Mississippi. There are no apparent grain-size trends away from other adjacent rivers. Peoria Loess thickness generally decreases eastward, but is highly variable, probably because of differential erosion. Potential sources for this unit are pre-Wisconsinan sediments on the Iowan Erosion Surface immediately west of the border of thick loess and Wisconsinan glacial sources tens or hundreds of kilometers to the west. The underlying Roxana Silt, up to 1 m thick, occurs only near the Mississippi and fines away from that river. The Roxana Silt deposit could reflect aggradation of the Mississippi floodplain because of glacial activity upstream and does not have clear implications for upland vegetation near the study area. By contrast, transport of the Peoria Loess from distant glacial sources or wind erosion of upland surfaces closer to the loess deposit both imply extensive, very sparsely vegetated surfaces west of the study area when the Peoria Loess accumulated.

Amino acid epimeric (aIle/Ile) values from terrestrial molluscs are used to define and correlate three aminozones in loess sequences exposed across the central Mississippi Valley, in Arkansas and Tennessee. Three superposed aminozones are defined at Wittsburg quarry, Arkansas, primarily using aIle/Ile values from total hydrolysates of the gastropod genus Hendersonia: Peoria Loess (aIle/Ile = 0.07 ± 0.01), Roxana Silt (0.14 ± 0.02), and a third loess (0.28 ± 0.06). Loess units at Wittsburg quarry can be correlated on lithologic characteristics eastward across the Mississippi Valley to the Old River section, near Memphis, Tennessee; however, only one loess unit is fossil-bearing (Peoria Loess, aIle/Ile = 0.05) at that section. Radiocarbon analyses of charcoal from the upper Roxana Silt (ca. 26,000 to 29,000 yr old) and mollusc shell carbonate from the basal Roxana Silt (ca. 39,000 yr old) are used to calibrate amino acid epimeric data for the central Mississippi Valley. These data, applied to the apparent parabolic kinetic model of R. M. Mitterer and N. Kriausakul (1989, Quaternary Science Reviews 8, 353-357), suggest an Illinoian (>120,000 yr) age for the third loess in the central Mississippi Valley that is correlative with part of the Loveland Loess in Illinois and Iowa.

The extent of isoleucine epimerization in a calcite speleothem was determined to evaluate the amino acid racemization method in abiotic calcite. A 5.5-cm-thick flowstone slab from Hamarnesgrotta, northern Norway, was analyzed for amino acid concentration, composition, and isoleucine epimerization at 26 levels through the sequence. U-series dates provide an independent chronologic control. Epimerization increases monotonically with stratigraphic depth and linearily with U-series age, independent of amino acid concentrations. The rate of epimerization is calibrated against the U-series dates, and extrapolation into lower strata beyond the U-series limit yields absolute age estimates that are consistent with paleomagnetic data from the same speleothem. The results suggest that, if adequately calibrated, amino acid dating is applicable to speleothem material reaching time spans beyond the range of conventional dating methods. Amino acids in the speleothem were probably derived from surface soils and are associated with brown humic stains in the calcite.

Four late-Quaternary alluvial fills and terraces are recognized in Wolf Creek basin, a small (163 km2) drainage in the Kansas River system of the central Great Plains. Two terraces were created during the late Pleistocene: the T-4 is a fill-top terrace underlain by sand and gravel fill (Fill I), and the T-3 is a strath terrace cut on the Cretaceous Dakota Sandstone. Both Fill II (early Holocene) and Fill III (late Holocene) are exposed beneath the T-2, a Holocene fill-top terrace. The T-1 complex, consisting of one cut and three fill-top terraces, is underlain by Fills III and IV. A poorly developed floodplain (T-0) has formed within the past 1000 yr. As valleys in Wolf Creek basin filled during the early Holocene, an interval of soil formation occurred about 6800 yr B.P. Early Holocene fill has been found only in the basin's upper reaches, indicating that extensive erosion during the middle Holocene removed most early-Holocene fill from the middle and lower reaches of the basin. Valley filling between 5000 and 1000 yr B.P. was interrupted by soil formation about 1800, 1500, and 1200 yr B.P. As much as 6 m of entrenchment has occurred in the past 1000 yr. Holocene events in Wolf Creek basin correlate well with those in other localities in the central Great Plains, indicating that widespread changes in climate, along with adjustments driven by complex response, influenced fluvial activity.

In a dissected, mostly sandy terrain along the Rio Grande valley border of southern New Mexico, a reddish-brown coarse-loamy Haplargid in an eolian deposit has a stage II carbonate horizon that is typical of soils of the Leasburg surface (8000-15,000 yr B.P.). The eolian deposit buries sediments and soils of late Pleistocene age or older. Thus, it does not fit in the chronological framework of the stepped sequence of geomorphic surfaces along the valley border, in which age of the surfaces and their soils increases with increasing elevation of the steps. Because the eolian deposit occurs in isolated areas where its age relative to the stepped sequence cannot be demonstrated geomorphically, soil characteristics are major tools for placing the deposit and its soil in the chronological scheme for the area. On the basis of the stage II carbonate horizon, the eolian deposit is considered to be an eolian analog of Leasburg alluvium. One of the Leasburg Haplargids in eolian sediments on a ridge crest has a total of 23 kg/m2 of pedogenic carbonate. Downslope in an area that receives runoff from higher areas, a fine-loamy Haplargid in Leasburg alluvium has 186 kg/m2 of carbonate; a nearby coarse-loamy Haplargid that does not receive runoff has only 46 kg/m2 of carbonate. Thus, landscape position and texture have a major influence on the total amount of pedogenic carbonate, which can range to more than eightfold in soils of Leasburg age.

A paleoclimatic record for a southern Great Plains locality (the Aubrey Clovis site in north-central Texas) has been established using stable carbon and oxygen isotopes. Detailed composite stratigraphic sections, constrained by 14 C ages, place the age of these deposits between 14,200 and 1600 yr B.P. Calcium carbonate samples of lacustrine and pedogenic origin were analyzed. Oxygen isotopic compositions of most of these in situ carbonates reflect equilibrium precipitation from local meteoric waters. Oxygen isotope values reflect changes in the composition of meteoric waters tied to changes in the isotopic composition of moisture derived from the Gulf of Mexico. Oxygen isotopic variability at the Aubrey site is coincident with marine isotopic records from the gulf that vary due to changes in Laurentide ice sheet volume and meltwater influx. The stable carbon isotopic record, reflecting changing biomass through time, corroborates humid versus arid interpretations based on sedimentology and rates of alluviation. A middle Holocene arid period was in contrast to moist early and late Holocene climate, affirming interpretations of other workers studying southern Great Plains Holocene climate history.

The petrographic and morphologic differences between calcite precipitated below, at, or above the present water table and uranium-series dating were used to reconstruct a chronology of water-table fluctuation for the past 120,000 yr in Browns Room, a subterranean air-filled chamber of Devils Hole fissure adjacent to the discharge area of the large Ash Meadows groundwater flow system in southern Nevada. The water table was more than 5 m above present level between about 116,000 and 53,000 yr ago, fluctuated between about +5 and +9 m during the period between about 44,000 and 20,000 yr ago, and declined rapidly from +9 to its present level during the past 20,000 yr. Because the Ash Meadows groundwater basin is greater than 12,000 km2 in extent, these documented water-table fluctuations are likely to be of regional significance. Although different in detail, water-level fluctuation recorded by Browns Room calcites generally correlate with other Great Basin proxy palcoclimatic data.