Thermoluminescence Age Estimates Research Articles

Heated flint TL-dating for Gruta da Oliveira (Portugal): dosimetric challenges and comparison of chronometric data

The Mousterian stratigraphy of Gruta da Oliveira (Torres Novas, Portugal) contains one of the latest occupations by Neanderthals on the Iberian Peninsula. Thermoluminescence (TL) dates on heated flints were obtained for layers 13 and 14, and compared with the radiocarbon (charcoal and burnt bone) and U/Th (secondary carbonates) dating results available for those layers and the succession as a whole. As the internal dose rates are very low, the TL age estimates are very dependent on the external dose rates, which show significant variation, including indications of spatial patterning. Through close inspection of the radiation geometries of samples and dosemeters, the appropriate gamma dosimetry is identified and the weighted mean TL ages (1-σ) calculated accordingly are 55 ± 7 ka for layer 13 and 77 ± 8 ka for layer 14. These ages are stratigraphically consistent with the radiocarbon chronology for layers 9–11 and with the termini ante and post quem provided by U-series dating for the beginning and end of the accumulation of the deposits. The TL results corroborate that all radiocarbon dates below layer 11 are minimum ages only and, given the constraints derived from the proxy data available, place layer 13 and 14 post MIS 5. Sample as well as dosemeter numbers are low at Gruta da Oliveira, which limits the precision of TL-dating of a prehistoric human activity. Further refinement of the chronostratigraphy will be based on U-series dating of non-human-related, interstratified secondary carbonates.

Luminescence dating of mid- to Late Wisconsinan aeolian sand as a constraint on the last advance of the Laurentide Ice Sheet across the Tuktoyaktuk Coastlands, western Arctic Canada

Luminescence dating of pre-glacial sand in the Tuktoyaktuk Coastlands, Northwest Territories, discounts an Early Wisconsinan age for the last Laurentide glaciation to cross the Arctic Coastal Plain and supports a Late Wisconsinan age. Aeolian dune sand from the Kittigazuit Formation near Cliff Point, on the southern shore of Liverpool Bay, predates till deposited during the Toker Point Stade. Potassium-rich feldspar from three stratigraphic sections ~35 km up-ice from the Toker Point glacial limit provides thermoluminescence age estimates that range from 18.7 ± 2.0 to 9.1 ± 1.0 ka and infrared stimulated luminescence age estimates of 23.8 ± 5.1 to 11.0 ± 2.1 ka. Quartz from four of the same samples provides optically stimulated luminescence age estimates of 16.5 ± 1.0 to 13.7 ± 0.9 ka. Collectively, these estimates reject the Early Wisconsinan age for the Toker Point Stade glaciation inferred hitherto from radiocarbon dating. A review of pre-glacial and post-glacial age estimates from the region indicates that during the Toker Point Stade ice advanced across the Tuktoyaktuk Coastlands no earlier than ~30 ka, and probably not before ~22 ka. Deglaciation had certainly commenced by 14.3 ka, and probably by ~16 ka. The Toker Point glaciation, therefore, dates approximately to the last glacial maximum, reinforcing the interpretation of the late rebuild up of the Laurentide Ice Sheet that characterized many parts of its margin.

Reassessment of TL age estimates of burnt flints from the Paleolithic site of Tabun Cave, Israel

The stratigraphy of Tabun Cave (Mt. Carmel), which comprises one of the longest sequences of Lower and Middle Paleolithic of the Near East, is widely used as a reference in debates on the evolution of Paleolithic industries and on the origin of modern humans and their relationship to the Neandertals. Considering the methodological improvements during the last ten years, the thermoluminescence (TL) dates of heated flints frequently quoted in the literature require an update. New TL results are discussed and compared with radiometric data recently obtained for this site, in particular by the ESR method, and with those obtained for other Levantine sites. The chronological framework previously proposed for the Paleolithic industries of the area is then re-examined.

Luminescence chronology of the Upper Pleistocene loess record at Kurtak in M iddle Siberia

The loess/palaeosol sequence of Kurtak, situated at the western bank of the upper Yenisei in Middle Siberia, represents one of the best developed Middle and Upper Pleistocene sediment records in Yenisei Siberia. More than 40 m thick loess and loess-like sediments intercalated by at least four pedocomplexes are exposed at a steep erosional slope at the bank of the Krasnoyarsk Water Reservoir. Infrared optically stimulated luminescence (IRSL) and thermoluminescence (TL) dating techniques have been applied on 38 fine grain samples from the upper 23 m of the profile, representing the penultimate and last interglacial–glacial cycle. The IRSL and TL age estimates are in good agreement with the geological estimates up to the last interglacial soil horizon (oxygen isotope substage (OIS) 5e). The luminescence ages show that the lowermost truncated palaeosol of kastanosjem-type is likely to have formed during the penultimate interglacial upon subaerial deposits. Three weak reddish brown palaeosols intercalated by reworked loess-like sediments correlate with early Upper Pleistocene interstadials (OIS5dion of -a), and a succession of humic horizons alternated by cryoturbation processes (Kurtak pedocomplex) is linked with OIS 3. Thick loess deposits between the pedocomplexes show now significant age increase with depth, indicating high accumulation rates at around 23 ka (OIS 2) and 60 ka (OIS 4).

Luminescence dates for the palaeolithic site of Piekary IIa (Poland): comparison between TL of burnt flints and OSL of a loess-like deposit

We report the thermoluminescence age-estimates for burnt flints excavated at the palaeolithic site of Piekary IIa and discuss attempts to date quartz grains extracted from the embedding deposit using the OSL technique with the SAR protocol. The results indicate that this loess-like deposit contains a mixture of grains, well bleached at different periods in the past, and that the study of small aliquots seems to be the only way to discriminate between the different fractions.

Luminescence dating and glacial stratigraphy in Westland, New Zealand

Westland is a key region in New Zealand for Quaternary stratigraphy; however, the regional glacial chronology is largely unsupported by numerical ages because of the unreliability of radiocarbon dating in this high‐rainfall region. Therefore, we tested the ability of thermoluminescence (TL) and infrared‐stimulated luminescence (IRSL) to produce stratigraphically consistent age estimates in this area. At Blue Spur Road, we obtained a stratigraphically reasonable TL age of 47 ± 7 ka in the second buried E horizon and a perhaps underestimated 53 ± 16 ka in the lower part of the fourth buried Bs horizon. A marine sand underlying Waimea outwash gave an incorrect (underestimated) IRSL age of c. 44 ka. In Saltwater Forest we collected five samples from two loess profiles overlying glacial moraines. Only one TL age (36 ± 3 ka in the lower part of the first Bs horizon) is acceptable. The next lower sample (c. 1.2 m), stratigraphically probably younger than 125 ka, gave a reasonable IRSL age of 89 ± 15 ka. In trench M1/1, on an older moraine, two TL age estimates were stratigraphically reversed, with a third result (87 ± 30 ka in unit L4) being reasonable. In contrast, IRSL produced ages of 65 ± 16 ka (1.1 m) and 145 ± 36 ka (2.0 m) for the upper two samples, removing the TL‐age reversal, although the age of the lower sample may be an overestimate. The third IRSL result (36 ± 5 ka, 2.7 m) is much too young, like the corresponding TL result. Thus, although TL is generally unreliable in the highly weathered loess‐paleosol sequences of Westland, the IRSL technique offers more promise, perhaps because of IRSL's feldspar specificity and its far greater signal‐zeroing efficiency.

Luminescence chronology of aeolianites from the section at Givat Olga — Coastal Plain of Israel

The Eastern Mediterranean Coastal Plain of Israel is composed mainly of loam, sand and gravel of Pleistocene to Holocene age, indicating an alternation of marine, coastal and continental environments. Owing to the complexity of the numerous exposures of aeolianites (“kurkar”) and soils (“hamra”) in the elongated ridges along the coastal plain, it is difficult to set up a reliable stratigraphy. A systematic luminescence dating study was carried out on loose sand, kurkar and hamra deposits in the coastal plain between Netanya and Haifa. Sixteen samples were taken and investigated from the section at Givat Olga. The chronological results show excellent agreement between the infrared optically stimulated luminescence (IRSL) and thermoluminescence (TL) techniques, indicating that the former sand dunes were well-bleached prior to deposition. The IRSL and TL age estimates range from about 75–3 ka for the aeolianites indicating that this kurkar ridge formed during an older period of the last glaciation, likely before 50 ka.

Last Interglacial age of the Eva Forest Bed, Central Alaska, from thermoluminescence dating of bracketing loess

The Eva Forest Bed in the Yukon-Tanana Upland of east-central Alaska represents a frozen, buried, boreal forest. This bed consists of well-preserved peat lenses, sticks, roots, logs, as well as rooted and unrooted stumps of trees. Where preserved, the Eva Forest Bed is overlain by the Goldstream Formation and underlain by Gold Hill Loess. Almost all exposures have a prominent unconformity at the top of the Gold Hill Loess, even where the forest bed is not preserved. Attempts have been made over the last 50 yr to determine the age of this bed numerically. We report here details of thermoluminescence (TL) dating of 14 samples of loess from above and below this unconformity in the Fairbanks area. Together with knowledge about the climatic indicators from the Eva Forest Bed, these TL age estimates: (1) indicate that the most probable age of the Eva Forest Bed is ∼125 ka, the time of the warmest part of oxygen isotope stage 5; and (2) confirm stratigraphic inferences from tephra beds in the upper Gold Hill Loess that erosion has removed variable thicknesses of the original loess complex “just prior” to the development of the Eva Forest Bed.

Upper Pleistocene loess stratigraphy in Southern Germany

The detailed loess/paleosol sequences of Southern Germany provide a highly sensitive record of the regional climate and environmental changes of the Upper Pleistocene. Combined infrared stimulated luminescence (IRSL) and thermoluminescence (TL) dating was carried out on loess and loess derivatives at Upper Pleistocene-type sections from the Mainz Basin, the Middle Neckar Valley and the Alpine Foreland with the aim of setting up a reliable chronological framework for the past 130 000 years. The IRSL and TL age estimates are in agreement with the geological estimates up to 70 000 years. However, the chronostratigraphic position of the first truncated brown forest soil in the sections at Böckingen, Bönnigheim and Bobingen, considered to be the Eemian soil or equivalent to oxygen isotope (OI) substage 5e, remains uncertain. The methodological results indicate that IRSL dating is the more advantageous technique for the Upper and Middle Weichselian periglacial deposits, as demonstrated for the loess/paleosol sequence in Mainz-Weisenau.

Single-aliquot and single-grain optical dating confirm thermoluminescence age estimates at Malakunanja II rock shelter in northern Australia

Single-aliquot and single-grain optical dating confirm thermoluminescence age estimates at Malakunanja II rock shelter in northern Australia

Loess chronology of the Middle and Upper Pleistocene in Tadjikistan

The loess/paleosol sequences of Central Asia are continuous terrestrial records of the Quaternary period and enable detailed comparison with paleoclimatic archives such as marine and ice core records in order to reconstruct regional and global paleoclimatic and paleoecological development during the past 130 000 years. Thermoluminescence (TL) and infrared stimulated luminescence (IRSL) dating methods are applied to the extensively studied loess/paleosol sequence of the section at Darai Kalon/Chashmanigar, Tadjikistan, in order to determine a more accurate chronological framework and climatostratigraphic reconstruction for the last interglacial/glacial cycle. Luminescence dating suggests that the loess above the first pedocomplex from the top, PC1, accumulated during the last glacial period. A high accumulation rate of up to 1.20 m per 1000 years was determined for the last glacial loess, especially for the uppermost 5–8 m. PC1 formed during the last interglacial period (oxygen-isotope stage 5). The loess between PC1 and PC2 is designated to be of penultimate glacial deposition age. Infrared stimulated luminescence and TL age estimates are in agreement to 80 000 years before present (BP), indicating a long-distance transport of the aeolian dust prior to deposition. The upper numerical age-limit range is between 300 000 and 450 000 years. However, reliable dating of the loess older than 130 000 years is not possible due to age scatter between samples and an inadequate increase of paleodose with depth. This high-resolution dating study underlines the importance of the section at Darai Kalon and indicates that it is one of the most continuous loess/paleosol records of the Northern Hemisphere. The chronological results are particularly important for the reconstruction of the human evolution in Central Asia, suggesting much older age estimates than previously obtained for most of the archeological key sites associated with PC5 and PC4 in Tadjikistan.

Stratigraphy and chronology of Mississippi Valley loess in western Tennessee

The stratigraphy and chronology of loess deposits in the Mississippi Valley provide an important proxy record of the number and timing of glacial advances into the upper Mississippi drainage basin. The stratigraphy of loess and paleosols from two exposures in northwestern Tennessee reveals three major loess units that are correlated with the Loveland Loess, Roxana Silt, and Peoria Loess—from oldest to youngest. Detailed measurements of clay and secondary (pedogenic) iron contents suggest that the Loveland Loess may consist of multiple loess units which have been pedogenically welded to one another to form a composite Sangamon Soil. The mineralogy of the coarse silt fraction from one exposure reveals both primary mineralogical differences between the Peoria Loess and the Roxana Silt and pedogenic overprinting by the Sangamon Soil. Grain size and iron data suggest that the degree of pedogenic alteration increases from west to east, away from the loess source area, which probably reflects the role of sedimentation rate on the degree of pedogenic alteration of loess. However, only the Peoria Loess shows the expected exponential eastward decrease in thickness; the thickness of older loess units appears to be controlled more by the occurrence of localized erosion than by distance from source area. Most samples are resistant to solar resetting of the thermoluminescence signal, which, if unrecognized, could yield thermoluminescence ages that significantly underestimate the true age of the deposits. The best match between thermoluminescence age using the total bleach method and radiocarbon age of the Peoria Loess and Roxana Silt is achieved when either long durations of optical bleaching by sunlight (≥32 hr) or artificial ultraviolet (UV) light are applied to define the residual thermoluminescence level. Partial bleach analysis indicates that long-duration optical bleaching by sunlight or UV lamp bleaching does not systematically overbleach samples of Peoria Loess and Roxana Silt. Three thermoluminescence age estimates between 38 ± 4 (±1σ) and 53 ± 5 ka from the basal half of the Roxana Silt, and six age estimates between 18 ± 2 and 25 ± 3 ka from the basal half of the Peoria Loess are consistent with numerous finite radiocarbon dates from around the central Mississippi Valley. Thermoluminescence age estimates suggest that the oldest of the intra-Loveland Loess units is 127 ± 12 ka (one age) and may have been deposited during marine oxygen isotope stage 6. The younger two of these loess units were deposited between 113 ± 10 and 70 ± 6 ka (three ages). This latter age range for loess that underlies the top of the presumed Sangamon Soil is problematic because it requires that pedogenesis may have begun as late as approximately 70 ka. Moreover, loess deposited between about 120 and 70 ka suggests that the southern margin of the Laurentide ice sheet was within some part of the upper Mississippi Valley during this interval. This conflicts with some glacial geologic evidence, but it is consistent with global sea level and δ18O records of ice volume, which indicate several intervals of moderate glaciation during this time.

Diring Yuriakh: A Lower Paleolithic Site in Central Siberia

Lower Paleolithic artifacts have been recovered from a single occupation surface within stratified deposits at Diring Yuriakh, an archaeological site in central Siberia. Thermoluminescence age estimates from eolian sediments indicate that the cultural horizon is greater than 260,000 years old. Diring Yuriakh is an order of magnitude older than documented Paleolithic sites in Siberia and is important for understanding the timing of human expansion into the far north, early adaptations to cold climates, and the peopling of the Americas.

Thermoluminescence and new 14C age estimates for late quaternary loesses in southwestern Nebraska

Loess of late Quaternary age mantles most of Nebraska south of the Platte River Valley. At least five late Quaternary loesses are recognized: from oldest to youngest, one or more undifferentiated pre-Illinoian loesses, the Loveland Loess, the Gilman Canyon Loess, which exhibits a well developed soil and rests unconformably on the Sangamon soil, the Peoria Loess capped by the Brady soil, and the Bignell Loess, which is distributed discontinuously. Previous research shows that the Loveland Loess is Illinoian, the Gilman Canyon Loess and Peoria Loess are Wisconsin, and the Bignell Loess is Holocene. We present here the first thermoluminescence (TL) age estimates and new 14C ages for these late Quaternary loesses at two key sections in southwestern Nebraska, the Eustis ash pit and the Bignell Hill road cut. TL age estimates from all samples collected from Eustis ash pit and Bignell Hill were internally consistent. TL and 14C age estimates from these two sections generally agree and support previous age determinations. The TL age estimate on Loveland Loess indicates deposition at 163 ka. TL and radiocarbon age estimates indicate that Gilman Canyon Loess, believed to be deposited during the Farmdale interstade, first began to accumulate at about 40 ka: the lower part of the Gilman Canyon Loess is 36 ka at Eustis and the middle of the unit is 30 ka at Bignell Hill. The lower and upper parts of the Peoria Loess give age estimates of 24 ka and 17 ka, respectively. TL age estimates for deposition of the Bignell Loess are 9 ka near the base, in agreement with radiocarbon age estimates, and 6 ka immediately below the modern soil, substantiating its Holocene age. Comparisons of TL age estimates with δ 18O and insolation curves which show loess deposition during interglacial and interstadial as well as glacial periods, indicate that loess deposition on the Great Plains can occur under a variety of climatic conditions.

Age of Sheep Creek Tephra (Pleistocene) in Central Alaska from Thermoluminescence Dating of Bracketing Loess

The age of the Sheep Creek tephra (SCt), a widespread marker ash bed in eastern Alaska and western Yukon Territory, has been ambiguous and controversial. We have obtained three reliable thermoluminescence age estimates from bracketing loess near Fairbanks that imply a deposition age of about 190,000 ± 20,000 yr for SCt. Three of six loess samples near and closely bracketing the SCt beds near Fairbanks yielded younger age estimates (∼117,000 and ∼135,000 yr), most likely (based on field aspects) because of reworking and contamination by translocated grains. The new, reliable age assignment of 190,000 yr confirms independent stratigraphic evidence of a pre-last interglaciation age, and stratigraphic evidence from one site (Upper Eva Creek) that SCt is older than the more-widespread 140,000-yr-old Old Crow tephra. The SCt age also has implications for regional correlations of glacial and nonglacial deposits. In particular, it supports the stratigraphic and geomorphic interpretation that the Delta Glaciation in the east-central Alaska Range and the Reid Glaciation in western Yukon Territory are older than the last interglaciation (isotope substage 5e).

New evidence for the age of the youngest eruptions in the Valles caldera, New Mexico

New geochronologic data provide evidence for an age of about 50 to 60 ka for the youngest volcanic eruptions within the Valles caldera, New Mexico - an age that is significantly younger than most previous age determinations. Thermoluminescence age estimates for buried soils beneath the El Cajete pumice, a key stratigraphic marker in the region, range from 48 to 61 ka and {sup 14}C analyses of burnt logs within volcanic surge beds near the El Cajete vent yield similar ages of 50 to >58 ka. These data conflict with fissontrack, {sup 40}Ar/{sup 39}Ar, and K-Ar ages of 130 to >200 ka, but are supported by recent analyses by electron spin resonance. The results of this study reinforce the need to apply a variety of dating methods when evaluating the age of young volcanic events and support the hypothesis that the El Cajete eruptions were part of a new cycle of volcanic activity in the Valles caldera after an exceptionally long period of quiescence of nearly 460 ka. The new age constraints also suggest a previously unrecognized link between cycles of volcanism and pulses of hydrothermal activity in the caldera, such that hydrothermal outflow appears to decrease significantly following completion ofmore » eruptive cycles. 24 refs., 3 figs., 2 tabs.« less

Paleoclimatic significance of Late Quaternary eolian deposition on the Piedmont and High Plains, Central United States

Presently stabilized dune systems on the piedmont of eastern Colorado and adjacent High Plains have been repeatedly re-activated during the past 20,000 years. Radiocarbon and thermoluminescence age estimates indicate eolian activity late in the last glacial cycle ca. 20,000–12,000 yr B.P. and subsequent episodes of dune reactivation at ca. 6000, 4500 and 1000 yr B.P. Pollen analysis from aggraded buried soil A horizons show a shift from grasses and shrubs to goosefoot, a disturbance indicator. The association of maximum goosefoot levels with the coarsest part of the buried A horizon immediately prior to burial by eolian sand indicates a substantial reduction in grass and dominance of shrubs with onset of eolian activity. The vegetation change and eolian depositional sequence indicates a reduction in plant coverage with regional drought, possibly augmented by bison grazing and surface heating effects. We infer an increase in summer monsoonal precipitation between 13,000 and 9000 yr B.P. reflecting a heightened land-to-sea temperature gradient associated with rising summer solar-insolation values and a meltwater cooled Gulf of Mexico. Dune reactivation in the middle and late Holocene appears to be independent of summer insolation values, but rather reflects a small (< 10°) easterly shift of the Bermuda High and western ridge aloft, difficult parameters to link to a cause and to resolve with climate models.

The kinetics of TL peaks and its effect on the dose versus temperature plot

The central procedure in thermoluminescence (TL) dating is the determination of the radiation dose, D E, that a sample has received in its geological past. Any random or systematic errors in the assessment of D E are directly reflected in the corresponding TL age estimate. In many TL dating studies, the D E value is derived from plateaux in the dose vs temperature plot. This study presents simulations of TL glow curves containing ideal first- and second-order kinetics TL peaks. The TL intensities and corresponding doses were selected to lie exactly on a single saturating exponential function. The results imply that there is a wide range of possibilities to estimate incorrect D E values from plateaux in the dose vs temperature plot, even when only first-order kinetics TL peaks are involved. However, it can also be shown that the regeneration method or its refinement, the slide method, can in principle produce correct D E estimates for first- and second-order kinetics TL peaks.

Comparison of TL and IRSL age estimates of feldspar coarse grains from waterlain sediments

The infrared stimulated luminescence (IRSL) and thermoluminescence (TL) dating methods are applied to alkali feldspar coarse grains from shallow marine, estuarine, shallow lacustrine and sublittoral sediments in the range 10–200 ka. No significant difference was observed between the equivalent doses obtained using IRSL and TL. Most luminescence dates are consistent with the expected geological age estimates obtained from radiocarbon, amino acid and ESR dates. This study provides evidence of the reliability of IRSL age estimates on alkali feldspars from well-bleached waterlain sediments older than 100 ka, provided that a correction for long-term thermal stability is applied in each case.

Thermoluminescence dating tests for lacustrine, glaciomarine, and floodplain sediments from western Washington and British Columbia

To help further develop reliable procedures for accurate thermoluminescence (TL) dating of Quaternary waterlaid sediments, we tested TL dating procedures on sediment types rarely examined: six glaciolacustrine samples, three samples of glaciomarine drift, and eight samples of floodplain deposits. We used the partial-bleach (R-beta/gamma) technique applied to fine-silt polymineral grains. Results from our younger known-age glaciolacustrine sediments confirm earlier observations that only the clayey laminae are generally suitable for TL dating. A clayey lamina older than ca. 140 – 150 ka produced no age underestimation, and this result (ca. 300 ka) suggests that such older lake sediments are probably suitable for TL dating. Two proximal samples of glaciomarine drift produced large TL age overestimates, whereas the single distal glaciomarine-drift sample yielded an expected age (177 ± 38 ka), suggesting that follow-up studies are warranted. Our results for eight flood-plain samples confirm that zeroing of light-sensitive TL is more likely to be effective for sediments deposited in quiet, ponded water on the floodplain than for proximal sediments deposited from turbid floodwater. TL age estimates for floodplain sediments of the regional Whidbey Formation are consistent with its expected last-interglacial age, and those for two samples from stratigraphically older beds are consistent with deposition near or beyond 200 ka. Two of our younger samples (one lacustrine and one floodplain) gave TL age underestimates, perhaps because of use of ultraviolet TL emissions for these samples.