The progressive development of unusual rock-weathering features and soils and minimum-limiting radiocarbon dates provide a basis for subdividing four groups of late Pleistocene moraines on the west side of the Cordillera Blanca, northern Peru (9°20′1°000′S, 77°10′-77°30′W). Boulders on the youngest late Pleistocene moraines have 10 to 14-cm-tall weathering posts; soils on these moraines yield mean profile development index (PDI) values of 0.05 ± 0.04 (±1σ). These moraines date between ca. 13,500 and 9700 ± 500 yr B.P., older than previously postulated. The next older moraines have boulders with weathering-post heights between 20 and 25 cm and soils with PDI values of 0.08 ± 0.07, and were deposited prior to 13,280 ± 190 yr B.P., probably during the last glacial maximum (marine isotope stage 2). Moraines from an older glaciation have boulders with weathering posts between 39 and 50 cm high, soils that yield PDI values of 0.21 ± 0.07, and are older than 19,700 ± 340 yr B.P. Boulders on moraines from a still older glaciation have lost ca. 50% of their above-ground volume, and have weathering posts between 62 and 70 cm high. PDI values for soils on these moraines are 0.32 ± 0.06. Linear and logarithmic models of weathering-post and soil development with time are used to estimate minimum and maximum ages for the two oldest moraine groups. Linear models suggest that the second oldest moraines are between ca. 20,500 and 46,500 yr B.P., and that the oldest moraines are between ca. 29,000 and 72,000 yr B.P. In contrast, logarithmic models suggest ages of greater than ca. 75,500 yr B.P. and greater than ca. 500,000 yr B.P., respectively.

The Little Ice Age was the most extensive Neoglacial glacier advance in the Canadian Rockies. Evidence of earlier, less-extensive Neoglacial glacier advances is based on wood recovered from several glacier forefields. Wood flushed out of Athabasca Glacier (7550-8230 yr B.P., three dates) and Dome Glacier (6120-6380 yr B.P., two dates) indicates that forests occurred upvalley of present glacier termini during the Hypsithermal. Detrital logs from Peyto (14 dates), Saskatchewan (3 dates), Robson (3 dates), and Yoho (1 date) Glaciers, plus in situ slumps at Peyto and Robson Glaciers, have yielded 14 C dates between 2490 and 3300 yr B.P. (12 dates between 2800 and 2990). This wood is derived from sources at or upvalley from present glacier termini and represents forests overridden by glaciers between ca. 3100 and 2500 yr B.P. (Treeline was higher than present immediately prior to this advance.) This advance, which did not extend beyond the Little Ice Age maximum position, is designated the Peyto Advance and correlated with the Tiedemann Advance in western British Columbia. Earliest Little Ice Age advances at Peyto and Robson Glaciers are dated ca. 800-600 yr B.P. at positions ca. 500 m upvalley from Little Ice Age limits.

A review of published stratigraphic records of pollen, sediment grain size, diatoms, and organic matter composition from Lake Biwa, Japan, identifies four pre-Holocene episodes of milder climate, increased surface runoff, and enhanced aquatic productivity, indicating intervals of warmer and wetter conditions which are interpreted as being interglacial. Correlation of these episodes to times of marine interglacial periods revises the age scale of the Lake Biwa sediment sequence which has been based on fission-track dating. The revised chronostratigraphic scale proposes an age of ca. 430,000 yr B.P. for the base of the 250-m-thick T Bed instead of the former age of ca. 700,000 yr B.P.

Two processes contribute to inaccurate 14C age estimates of carbonates precipitated within the Lahontan basin, NevadaCalifornia: low initial 14C/C ratios in lake water (reservoir effect) and addition of modern carbon to calcium carbonate after its precipitation. The mast reliable set of 14C ages on carbonates from elevations > 1310 m in the Pyramid and Walker Lake subbasins indicate that lakes in all seven Lahontan subbasins coalesced ∼14,200 14C yr B.P. forming Lake Lahontan. Lake Lahontan achieved its 1330-m highstand elevation by ∼13,800 14 C yr B.P. and receded to 1310 m by ∼13,700 14C yr B.P. Calculations, based on measured carbonate-accumulation rates, of the amount of time Lake Lahontan exceeded 1310 and 1330 m (500 and 50 yr) are consistent with this chronology. The timing of the Lake Lahontan highstand is of interest because of the linkage of highstand climates with proximity to the polar jet stream. The brevity of the Lahontan highstand is interpreted to indicate that the core of the southern branch of the polar jet stream remained only briefly over the Lahontan basin.

Taking China as a whole, lake expansion occurred 500 B.C. (?)-0 A.D., 650-950 A.D., and 1250-1650 A.D. over the past 3000 yr. The expansions were mainly due to climatic fluctuations and correspond to frequent flood reports. Three regional characteristics of lake development have been recognized. (1) Fluctuations of near-coastal lakes, and those in other regions as well, were out of phase during 500 B.C.-500 A.D., which possibly resulted from out-of-phase fluctuations of climate and sea level on 10- to 100-yr time scale. (2) As the lower Yellow River channel migrated southward, the lakes in the south of the North China Plain formed and reached their largest sizes much later than those in the north. (3) For the lakes in the middle and lower Yangtze River valley, the higher a lake basin lies above a neighboring river and the further the lake is from the river's mouth, in general, the younger the lake is. Through land reclamation, the ancient Chinese have changed most of the lakes greatly. However, human activities were also constrained in certain ways by the natural changes of the lakes. Reclamation usually was banned and farmland was abandoned to repair reservoirs while lakes were expanding.

Detailed mapping and provisional numerical age determinations of glacial deposits in the South Chiatovich Creek Basin of the White Mountains provide an opportunity to evaluate the ability of conventional soil parameters to discriminate first- and second-order glacial events. Sampling and analytical procedures were designed to minimize variation in climate and lithology. When lithology and climate are similar among sites, age trends are more pronounced in both field and chemical soil properties. Profile development indices (PDIs), adjusted by removing melanization and pH, systematically increase with greater soil age, and discriminate first-order, but not second-order, glacial events. The best-fit curve for adjusted PDI data assumes an exponential form and suggests that the rate of soil formation in this region decreases over time, similar to the rate of weathering-rind development. Variation in eolian influx and surface erosion, which are dominant processes affecting soils of the basin, cause major uncertainties in establishing soil age and, hence, soil-development rates. Even on the youngest glacial deposits, soil age is probably significantly less than deposit age due to these geomorphic processes. Soil and weathering parameters imply that these field techniques can be inexpensively employed to define relative chronologies and to assess surface degradation and its impact on surface exposure ages. Results from this study indicate that site-selection strategy for establishing glacial chronologies should be reevaluated. Working with stable residual bedrock surfaces and associated low-relief outwash fans and terraces may prove more productive than focusing on relatively unstable moraine surfaces in tectonically active mountain systems.

A chronological sequence of soils formed on a series of alluvial depositional surfaces ranging in age from late-middle Miocene to late Pleistocene was characterized to clarify soil-geomorphic relations and provide a basis for allostratigraphic subdivision of the inner Coastal Plain. On Quaternary river terraces, Ultic Hapludalfs containing abundant weatherable mineral species and clast types are estimated to have formed in 60,000-120,000 yr, whereas Typic Hapludults greatly depleted in weatherable minerals and showing strong weathering of clast types are estimated to be 700,000-1,600,000 yr old. Typic Paleudults with incipient plinthite, duripan, and ferricrete development characterize interfluves that have been little eroded since early Pliocene time (3.4-5.3 myr ago). Typic-Plinthic Paleudults with intense weathering of siliceous clasts and moderate to strong duripan and ferricrete development are found on surfaces that formed near the beginning of late Miocene time (10.8-13.0 myr ago). Chemical weathering in the chronosequence may be classified into three progressive stages: (1) decomposition of unstable sand- and silt-sized minerals into a mixed (stable + unstable) clay-mineral suite (stable Fe + Al/Si bulk chemical composition, < 106 yr); (2) transformation of mixed clay-mineral suite into a stable suite (increasing Fe + Al/Si bulk chemical composition, 106 - 107 yr); and (3) transformation of stable suite into ultrastable clay-mineral suite (increasing Fe/Si bulk composition, > 107 yr). Not all soil properties show unidirectional development, nor is a steady state of pedon development observed even after approximately 107 yr of chemical weathering. Soil development in the chronosequence is episodic. The transition from one phase to the next is marked by a change in rate, and sometimes a reversal in the direction, of development of one or more soil properties.

Packrat (Neotoma spp.) middens from the White Mountains indicate climatic and plant community conditions for the last 19,000 yr. During full-glacial times (ca. 19,000 yr B.P.) and at the Pleistocene/Holocene boundary juniper woodlands were 600 m lower in elevation than at present. Midden assemblages and stable-isotope data suggest lower temperatures and increased precipitation relative to the present into the early Holocene. Two early Holocene middens (ca. 8000 yr B.P.) from lower elevations of the White Mountains contain fragments of pinyon pine and indicate that both pinyon and juniper grew at a site that today supports only a pinyon woodland. Two middle Holocene middens (ca. 5000 yr B.P.) indicate that there was an upward migration of pinyon-juniper woodlands along a high-elevation ecotone, but little change in the middle of the pinyon-juniper woodland. Middens from the late Holocene indicate that present-day plant communities were in place by ca. 2000 yr B.P. or before.

Amino acid ratios of gastropods provide a useful basis for correlation and approximate dating of middle to late Quaternary fluvial deposits from the northern Española basin, New Mexico. Sparsely fossiliferous slackwater deposits in the Rio Chama-Rio Grande floodplain were buried episodically by piedmont alluvium during periods of climatic change as the axial river system cut down 120 m during the past 620,000 yr. Alloisoleucine/isoleucine (alle/Ile) ratios in the total hydrolysate and free fraction of amino acids in Succinea and Vallonia are strongly correlated with elevation. Ratios in Succinea range from 0.01 to 0.79 in the total hydrolysate and from 0.00 to 1.15 in the free fraction for deposits that range in age from modern to ⩾620,000 yr old. Amino acid ratios are tightly clustered for deposits that contain the 620,000-yr-old Lava Creek B tephra layer, demonstrating their utility for correlation and providing a calibration point for a local dating curve. Two 14C ages from younger deposits and limiting ages for erosion surfaces that lie above gastropod-bearing units tend to support the validity of much of the curve. Combining the dating curve with the geologic constraints suggests that aggradation events, separated by episodes of net incision, occurred at about 620,000, 310,000 ± 70,000, 170,000 ± 40,000, 95,000 ± 15,000 yr ago, after 19,000 yr ago, and during two minor periods between 80,000 and 25,000 yr ago.

The paleoecologic record of the northern Rocky Mountains shows considerable spatial variability, which has been difficult to explain as a unified response to large-scale changes in effective moisture or temperature through time. Comparison of past and present climatic patterns suggests that the spatial heterogeneity in the fossil record arises from a local response to the effects of greater summer solar radiation during the early Holocene. Solar radiation directly determines evapotranspiration and indirectly controls the strength of the summer monsoon and the eastern Pacific subtropical high-pressure system. The Yellowstone region provides an example of an area where the indirect effects of solar radiation are manifest by the juxtaposition of two contrasting climatic regimes today and where these regimes were amplified during the early Holocene. Fossil-pollen data suggest that central and southern Yellowstone (the summer-dry area) became drier in the early Holocene, while northern Yellowstone (the summer-wet area) became wetter. The boundary between regimes, however, is controlled by topography and has not changed significantly in the Holocene. The spatial variations of Holocene vegetation and climatic changes in the northern Rocky Mountains may therefore be explained as a trade-off between strengthened monsoonal circulation and a stronger subtropical high in the early Holocene and the relative importance of these precipitation regimes on a topographically complex landscape.

A model of pollen deposition on the surface of an entire basin is developed to estimate pollen source area, and results are compared with those for a point at the center of a basin (I. C. Prentice, 1985, Quaternary Research 23, 76-86; 1988, "Vegetation History," (pp. 17-42, Kluwer Academic). This model is more appropriate for approximating the source area of pollen in lake sediment, since mixing in lake water and focusing of sediment redistribute pollen originally deposited over the entire surface. In general, the pollen source radius for the entire basin surface is 10-30% smaller than the source radius for a point at the center; the difference in the source radius is more profound for heavier pollen types such as spruce and sugar maple than for lighter types such as oak and ragweed. The average pollen input to the entire surface is more strongly influenced by nearby pollen sources than pollen deposition at the center. The pollen record from a lake may therefore provide different spatial resolution than the record from a bog of similar radius.

The world's peatlands contain about 450 GT of readily decomposeable organic carbon. Peat-forming systems have two main layers, of which the lowest is the thickest and includes the peat proper. The upper layer fixes carbon by photosynthesis, loses it by selective decay, and passes on about 15% to the lower zone; here decay continues, although very slowly. One consequence is that as for Point Escuminac, New Brunswick, the relation of age to depth may be concave. Although the surface of the peatland is as productive as ever, the true rate of carbon accumulation is decreasing; after 10,000 yr it is only 33% as efficient at sequestering carbon as it was when the peatland began to grow. Peatlands are usually thought to represent sinks for carbon, but a warming climate could make some peatlands carbon sources rather than sinks, thus initiating positive feedback.

Tree-ring data from subalpine conifers in the southern Sierra Nevada were used to reconstruct temperature and precipitation back to A.D. 800. Tree growth of foxtail pine (Pinus balfouriana) and western juniper (Juniperus occidentalis ssp. australis) is influenced by nonlinear interactions between summer temperature and winter precipitation. Reconstruction of the separate histories of temperature and precipitation is feasible by explicitly modeling species and site differences in climatic response using response surfaces. The summer temperature reconstruction shows fluctuations on centennial and longer time scales including a period with temperatures exceeding late 20th-century values from ca. 1100 to 1375 A.D., corresponding to the Medieval Warm Period identified in other proxy data sources, and a period of cold temperatures from ca. 1450 to 1850, corresponding to the Little Ice Age. Precipitation variation is dominated by shorter period, decadal-scale oscillations. The long-term record presented here indicates that the 20th century is anomalous with respect to precipitation variation. A tabulation of 20- and 50-yr means indicates that precipitation equaling or exceeding 20th-century levels occurred infrequently in the 1000+-yr record.

During the past 8 centuries, light rings (LRs) have occasionally formed in black spruce (Picea mariana) at treeline near Bush Lake, northern Quebec (L. Filion, S. Payette, L. Gauthier, and Y. Boutin, 1986, Quaternary Research 26, 272-279; A. Delwaide, L. Filion, and S. Fayette, 1991, Canadian Journal of Forest Research 21, 1828-1832). New analyses of climate data compiled during the period of overlapping tree-ring and instrumental records show that years of LR formation at Bush Lake have unusually cool May, June, August, and September temperatures. The analyses also show that August-September temperatures strongly correlate with May-July temperatures. Thus, late spring and entire growing-season temperatures influence LR formation at subarctic treeline. LRs formed in at least 5% of the trees at Bush Lake when May-September mean temperatures at Inukjuak fell below 4.2°C and August-September mean temperatures fell below 6.7°C. These threshold temperature/LR relationships can be used to infer limiting summer temperatures during the period preceding instrumental records. For example, the LR record suggests that May-September temperatures at northern Quebec treeline dropped below 4.2°C in A.D. 1601 after a major volcanic eruption of unknown source. Visual assessments of LR occurrence provide a new approach for extracting quantitative paleoclimatic information from tree rings.