A team of researchers at BOKU University in Vienna, led by Gustav Firla, extracted a 30-meter-long drill core from the bank of the Salzach River near the Urstein hydroelectric power plant, just south of Salzburg. The core, recovered in 2013 with a percussion drilling rig, pierced through the full thickness of the lake sediments and into the coarser materials above. Using single-grain luminescence dating on six samples spread across the entire depth of the core, the team calculated that the sediments were deposited around 18,600 years ago, with an uncertainty of roughly 900 years. That places the lake phase squarely in the late glacial period, after the peak of the Last Glacial Maximum.

The finding matters because the Salzburger Seeton has long been a chronological puzzle. Earlier work relied on radiocarbon dating of peat layers intercalated within the lake sediments, which suggested a much older origin, potentially during the early part of the last glaciation more than 80,000 years ago. Other researchers, studying the geometry of the valley’s sedimentary fill, argued for deposition before the Last Glacial Maximum. The new luminescence ages contradict all of these older estimates and instead support a interpretation first proposed in 1979 by the Austrian geologist Dirk van Husen, who suggested the lake clays were of late Würmian age, deposited as the last ice age was winding down.

Luminescence dating works on a simple but powerful principle: when mineral grains such as quartz or feldspar are exposed to sunlight, the natural radiation dose stored in their crystal lattices is reset to zero. Once the grains are buried and shielded from light, ionizing radiation from surrounding sediments and cosmic rays gradually rebuilds this stored signal. In the laboratory, scientists can measure the light emitted when the grains are stimulated, calculate the total radiation dose absorbed since burial, and divide it by the rate at which that dose accumulated in the environment. The result is the time elapsed since the sediment last saw daylight, which in most settings corresponds closely to the moment of deposition.

For glacial sediments, however, the method faces a serious complication. Ice-transported and water-transported grains are often dumped into lakes without adequate exposure to sunlight, a phenomenon known as poor bleaching. When this happens, some grains retain residual signals from earlier burial episodes, and the measured age distribution becomes skewed toward values that are too old. The researchers observed exactly this pattern: all six samples produced right-skewed equivalent dose distributions, the classic fingerprint of incompletely bleached glacial material. To handle this, they applied a statistical tool called the three-parameter Minimum Age Model, which isolates the population of grains that were best reset at deposition and therefore yields the most accurate burial age.

The team measured potassium-rich feldspar grains individually using a Risø luminescence reader equipped with a single-grain attachment at the Vienna Laboratory for Luminescence Dating. They employed a two-step measurement protocol, detecting one infrared-stimulated signal at 50 degrees Celsius and a second, more thermally stable signal at an elevated temperature of 225 degrees Celsius. This post-infrared infrared stimulation approach is prized because the higher-temperature signal is far less susceptible to anomalous fading, a slow, athermal leakage of the stored signal that would otherwise make feldspar ages appear artificially young. Fading experiments confirmed that the 225-degree signal was essentially stable, while the 50-degree signal required a correction based on measured fading rates.

The comparison between the two signals proved highly informative. Ages derived from the faster-bleaching 50-degree signal were consistently younger than those from the 225-degree signal, exactly as expected if the sediments were incompletely bleached at deposition. Because the 50-degree signal resets more quickly in daylight, the team considers its fading-corrected ages the most faithful record of the last depositional event. Strikingly, the five samples from the lower and middle parts of the core all agreed within their uncertainties, showing no age trend from bottom to top. This pattern indicates that a substantial thickness of lake sediment, spanning roughly 27 meters of the core, accumulated in a geologically brief interval, with a combined average age of 18.6 plus or minus 0.9 thousand years.

The sedimentology of the core tells a coherent story alongside the dates. The lowest unit, extending from 30 meters down to about 14.6 meters, consists of silt-dominated fines with a coarsening-upward trend and the characteristic blue-greyish color described in classic accounts of the Salzburger Seeton. Above it lies a sandy unit, also coarsening upward, reaching to about 3 meters depth. The topmost unit is dominated by gravel and shows signs of human disturbance, including fragments of roof tiles and bricks. Radiocarbon dating of plant debris from this upper zone returned a modern age of 1966 to 1967, suggesting the surface sediments were reworked during construction of the nearby hydroelectric plant. A basal radiocarbon sample from the lake sediments, meanwhile, was beyond the reliable limit of the method, yielding only a minimum age of greater than 43,500 years.

That discrepancy between the radiocarbon and luminescence results carries its own scientific lesson. Radiocarbon dating measures when organic material last exchanged carbon with the atmosphere, but in glacial environments old plant matter can be eroded from older deposits and swept into younger lakes, where it is buried without any record of its true age. The researchers point to known interglacial deposits near Adnet, southeast of the drill site, as a plausible source of such reworked organic material. Luminescence dating, by contrast, dates the sediment grains themselves and their last exposure to sunlight, making it far more robust against this kind of recycling. The combination of right-skewed dose distributions and conflicting radiocarbon ages thus points convincingly to the presence of ancient, redeposited organic matter within the lake clays.

The study also resonates with parallel work on other overdeepened basins of the northern Alpine Foreland. In a companion investigation of a core from Neusillersdorf west of Salzburg, the same group found that a basin more than 100 meters deep filled rapidly, with no discernible age trend from its base to the top of its fine-grained fill. Together, these results paint a picture of Alpine valleys whose glacially carved depressions were quickly inundated and infilled as the ice retreated after the Last Glacial Maximum, rather than preserving long, slow sedimentary archives spanning multiple ice age cycles. For the lower Salzach Valley, the conclusion is clear: at least the shallow parts of this overdeepened trough were filled by post-glacial lake sediments around 18,600 years ago, closing a long-standing debate with the quiet glow of individual feldspar grains.

Subject of Research: Luminescence dating of glacio-lacustrine sediments in an overdeepened Alpine valley

Article Title: The depositional age of glacio-lacustrine sediments (Salzburger Seeton) from the lower Salzach Valley, Austria

Article References: Firla, G., Fiebig, M., Rauter, T., & Lüthgens, C. (2025). The depositional age of glacio-lacustrine sediments (Salzburger Seeton) from the lower Salzach Valley, Austria. E&G Quaternary Science Journal, 74(2), 213-218. https://doi.org/10.5194/egqsj-74-213-2025

Image Credits: AI Generated

DOI: 10.5194/egqsj-74-213-2025

Keywords: luminescence dating, Salzburger Seeton, Salzach Valley, glacio-lacustrine sediments, Last Glacial Maximum, overdeepened basins, feldspar, Quaternary geology, radiocarbon dating, Alpine foreland, late glacial, Austria

Floodplains are among the best record-keepers in the landscape. Every flood that overtops a riverbank deposits a thin veil of sediment, and the chemistry of that veil reflects what the water carried: eroded soil, organic matter and, crucially, any contaminants picked up along the way. By drilling transects of cores perpendicular to the channel at three sites along the Kinzig and its Wolf tributary, the researchers recovered sequences that tell a coherent story from bottom to top. At the base lie poorly sorted gravels deposited by a cold-climate, braided river of the late Pleistocene, when sparse vegetation, limited soil formation and physical weathering of the crystalline basement of the Black fed enormous loads of coarse bedload into the stream network.

As the climate warmed into the Holocene, the catchment transformed. Forests stabilised the hillslopes, chemical weathering and soil formation intensified, and the supply of coarse material dwindled. The Kinzig responded by abandoning its braided character, incising into the Pleistocene gravels and settling into a single-thread channel flanked by newly formed floodplains. Above the gravels, the cores record a gradual transition through sandy deposits into the fine-grained flood loams, silty and clayey sediments, that dominate the upper metres of the floodplain today. These flood loams are the signature of hillslope erosion delivering fine particles to the river, a process that accelerated as humans began clearing and cultivating the land.

To pin dates on this sequence, the team employed optically stimulated luminescence dating, which measures the time since mineral grains were last exposed to sunlight. Because river sediments are notoriously incompletely bleached, the researchers applied minimum age models to skewed dose distributions, and cross-checked their chronology with radiocarbon dating of peat from an abandoned channel. The ages divide floodplain accumulation into three distinct phases with strikingly different sedimentation rates. From the late Pleistocene into the early Holocene, up to about 9,000 years ago, sediment accumulated at a languid 0.1 millimetres per year. Through the middle and late Holocene, from 9,000 to about 820 years ago, the rate roughly tripled to 0.3 millimetres per year. In the modern phase, spanning the last eight centuries or so, deposition surged to 1.1 millimetres per year, an eleven-fold increase over the natural baseline.

That acceleration is only half the story. The geochemistry of the overbank fines reveals an equally dramatic chemical fingerprint. Concentrations of barium, lead and copper rise steadily up through the cores, and enrichment factors calculated against deep, pre-impact background samples, normalised against the reference element vanadium, show clear contamination of the younger sediments. Barium is particularly telling: although it was never a targeted commodity, it occurs abundantly in the barite veins associated with the silver-lead-zinc ores of the Black Forest, making it a reliable proxy for mining intensity. At the Wolfach site, barium concentrations peak at nearly 2,000 milligrams per kilogram in the upper half metre, several times the natural background of a few hundred milligrams per kilogram recorded at the base of the cores.

The source of these metals is written in the documentary record of the Kinzig catchment. Small-scale Celtic mining likely began in the Bronze Age, with the first reliable evidence of ore smelting dated between roughly 225 BCE and 165 CE. Mining rights were first granted in the 11th century, and activity intensified through the medieval period, peaking in the 16th and 18th centuries. The correlation between the heavy metal peaks in the floodplain sediments and these documented phases of mining and ore processing is compelling. Mining did more than release metals directly: the voracious demand for timber to construct mine shafts, fuel smelting furnaces and produce charcoal drove widespread deforestation, destabilising hillslopes and multiplying the sediment load reaching the river, contamination and erosion travelling together.

The timber economy itself left physical scars in the floodplain. At the Schenkenzell site, lidar data and historic maps from 1816 and 1896 reveal a diverted channel known as the Schenkenzeller Holzkanal, a wood canal with weirs that served as a binding site for timber rafts. Every spring, oak and spruce felled in the winter were floated downstream through a network of sluices and weirs, requiring the clearing and modification of channel beds and banks. The researchers found poorly sorted deposits, cobbles embedded in a silty matrix, in the riparian forest, which they interpret as dredged material or erosion protection associated with these rafting operations. These are sediments placed and rearranged by human hands, not by natural floods.

Comparing the Kinzig with its neighbours sharpens the picture. Sedimentation rates in the Elz and Möhlin catchments remain below 1 millimetre per year in the upper floodplain deposits, roughly 0.6 times those of the Kinzig, and the modern Kinzig rates sit within the lower range compiled for Rhine tributaries, where extremes exceed 5 millimetres per year. The flood loam thicknesses along the Kinzig, between roughly 0.6 and 2.5 metres, match those documented in other Black Forest valleys, but the variations between tributaries of similar size demonstrate how sensitively each catchment responds to its own history of land use, geology and hillslope-to-channel connectivity. The message is that human impact on rivers was asynchronous and locally contingent: there was no single, Europe-wide moment when people took over.

The broader context is sobering. Roughly 95 percent of European floodplain ecosystems have been lost to intensive land use and river training, and the Kinzig itself was radically re-engineered in the early 19th century under the rectification programme led by the engineer Johann Gottfried Tulla. Yet the new study shows that the deep transformation began far earlier, long before concrete embankments and straightened channels. The gradual shift from a naturally controlled system to an anthropogenically altered one, intensifying over the last millennium, means that what we perceive as the natural baseline of Central European rivers is in fact a landscape already profoundly edited by centuries of mining, forestry and agriculture. For anyone seeking to restore rivers or assess their resilience under modern climate stress, the Kinzig floodplains deliver a clear warning: the human signature in these systems runs far deeper than the eye can see.

Subject of Research: Anthropogenic transformation of floodplain sediment dynamics in the Kinzig River catchment, southwestern Germany

Article Title: Silts with a human touch: the shift from natural to anthropogenically controlled fluvial dynamics in the Kinzig River floodplains, southwestern Germany

Article References: Engelmann, C. E., Preusser, F., Fülling, A., Wilk, J., Eiche, E., Quandt, D., Hergarten, S., & Blöthe, J. H. (2025). Silts with a human touch: the shift from natural to anthropogenically controlled fluvial dynamics in the Kinzig River floodplains, southwestern Germany. E&G Quaternary Science Journal, 74(2), 235-262. https://doi.org/10.5194/egqsj-74-235-2025

Image Credits: AI Generated

DOI: 10.5194/egqsj-74-235-2025

Keywords: floodplain sediments, Kinzig River, Black Forest, luminescence dating, heavy metal contamination, historical mining, deforestation, fluvial geomorphology, Holocene, Anthropocene, sedimentation rates, timber rafting

The technique at the heart of the study is post-infrared infrared stimulated luminescence, measured at 200 degrees Celsius, known as post-IR IRSL200. When feldspar grains are buried in sediment, background radiation from their surroundings and from within the crystals themselves steadily accumulates trapped electrical charge. Exposing the grains to light releases that charge as a faint glow, effectively resetting the clock. For luminescence dating to yield accurate ages, this resetting — called bleaching — must be complete before burial. In rivers, however, grains can travel in turbid water, in landslides, or in shadowed pulses, and sunlight may only partially wipe the signal. The leftover dose at the moment of burial, called the remnant dose, can inflate calculated ages, sometimes dramatically for young deposits.

Previous work on this problem relied mostly on multi-grain measurements, in which thousands of crystals are stimulated together and their signals averaged. That approach risks masking the true diversity of the grain population: well-bleached, partially bleached, and saturated grains all blur into one number. The new research, led by Arindam Biswas of the University of Cologne and colleagues at Toulouse, took the systematic single-grain route that had never before been applied to modern fluvial analogues at this scale. The team measured up to 2,900 individual grains per sample for equivalent dose determination and up to 2,200 grains for residual dose analysis, mounting crystals in discs holding a ten-by-ten grid of 300-micrometre holes and stimulating each with an 830-nanometre infrared laser on an automated Risø reader.

The samples came from eleven catchments draining perpendicularly off the Andean range, with outlets at the foot of the main Cordillera to minimise the confounding effects of downstream storage and reworking. This positioning lets floodplain sediments record the millennial-scale erosion processes shaping the landscape. The region is a natural laboratory: active tectonics, volcanism, glaciers, steep slopes, and a pronounced climatic gradient from desert to Mediterranean conditions all vary across the study area, offering a wide range of lithologies and environments in which to test what controls bleaching.

In the laboratory, the team ran controlled bleaching experiments on four samples chosen for their diverse catchment characteristics and higher yields of luminescence-sensitive grains. Grains were first reset under a Hönle Sol2 solar simulator, then given a fixed regenerated dose of 30 grays before being exposed to simulated sunlight for durations ranging from one minute to 30,000 minutes. The results showed a strikingly uniform pattern at the sample-average level. After just one minute of exposure, the signal dropped by about 24 percent. By 100 minutes, roughly 79 percent of the initial signal had been depleted, and after 1,000 minutes only 10 percent remained. At 2,880 minutes — two full days — the signal stabilised at approximately 4 percent of its starting value, equivalent to about 1.2 grays, marking a bleaching plateau.

Individual grains, however, told a more complicated story. When the researchers classified grains into fast-, medium-, and slow-bleaching groups based on their response after one minute of light, all three populations behaved distinctly at first. Yet, remarkably, all converged toward the same asymptotic level between two days and 30,000 minutes of exposure. This convergence suggests a fundamental limit to signal reduction rather than insufficient bleaching time. Even more intriguing, prolonged light exposure homogenised the grain population: the interquartile range of normalised signal values shrank four- to six-fold, from about 0.16 to 0.18 at one minute down to 0.03 to 0.05 at 30,000 minutes. Extended sunlight, in other words, drives diverse grains toward increasingly uniform signal levels — a dose-homogenising effect with practical implications for how long sediments must be exposed to reset their clocks reliably.

What explains the variability in residual doses, which ranged from near zero to about 23 grays among individual grains and from 2.44 to 9.49 grays on average across samples? The team tested three candidate culprits. First, grain chemistry: using an electron microprobe, they measured major oxide concentrations in 121 individual grains, finding potassium oxide levels spanning from 0.20 to 16.60 weight percent. Despite this wide compositional range, no statistically significant correlation emerged between any major oxide — potassium, sodium, calcium, iron, barium, silicon, or aluminium — and either residual dose magnitude or bleaching rate. Mineral composition, including the potassium content that affects internal radiation dose, simply does not control how efficiently a grain bleaches. Second, catchment lithology: samples from catchments with similar dominant rock types showed markedly different doses, while samples from geologically different basins sometimes matched, ruling out provenance as a direct control.

The third candidate proved decisive. When the team compared the modal equivalent dose — the peak of the single-grain dose distribution, representing the best-bleached grain population — with the modal residual dose measured after two days of laboratory bleaching, a strong positive linear correlation emerged, with a coefficient of determination of 0.89. Grains carrying larger accumulated natural doses are progressively harder to bleach. Crucially, the intercept of the linear fit was indistinguishable from zero within uncertainty, implying a negligible unbleachable component in these samples — a contrast with earlier studies that reported small irreducible doses. The youngest samples, with low natural remnant doses, could plausibly reach zero residual dose, while older samples carrying more than 10 grays of remnant dose retained significant residuals even after two days of intense simulated sunlight.

These findings carry direct consequences for how scientists correct palaeodoses when dating sedimentary archives. The team found that remnant doses in their modern analogues, ranging from 0.88 to 25.50 grays, systematically exceeded both laboratory residual doses and the unbleachable component — meaning that subtracting a laboratory residual from a palaeodose would undercorrect the true inherited dose, potentially by 0.5 to 4 grays in the best cases and more than 20 grays in the worst. Subtracting the remnant dose of a single randomly chosen modern analogue is equally unreliable given the scatter. Instead, the authors evaluate three context-sensitive strategies: making no correction, using a distribution-based remnant dose estimate from multiple modern analogues, or applying the unbleachable component. Their recommendations depend on depositional setting — unbleachable-component corrections may suit well-bleached aeolian and coastal deposits, while fluvial and glacio-fluvial settings demand modern-analogue approaches — and on measurement protocol, aliquot size, and the statistical age model applied.

Beyond dating, the results open a path for luminescence-based sediment tracing, a technique that uses incomplete bleaching as a fingerprint of how sediment moved through a landscape. Identifying well-bleached grains is essential for tracing sediment pathways, and previous studies have used fixed thresholds, such as a 10-gray cutoff, to classify them. The new work suggests a more refined, sample-specific approach: classify grains as well-bleached if their natural dose falls within two standard deviations of the mean of that sample’s single-grain residual dose distribution, provided a clear bleaching plateau has been established. Though time-intensive and still requiring validation in field applications, this method promises more realistic thresholds tailored to each geomorphic setting. From the arid slopes of the high Andes to the rain-soaked basins of the south, the humble feldspar grain is proving to be not just a clock, but a recorder of its own journey through sunlight and stone.

Subject of Research: Bleaching behaviour of the post-IR IRSL200 luminescence signal in single-grain K-feldspars from modern fluvial deposits in the Southern Central Andes, Chile

Article Title: Novel insights into the post-IR IRSL200 signal bleachability of single-grain K-feldspars in fluvial modern analogues from the Southern Central Andes, Chile

Article References: Biswas, A., Riedesel, S., Karman-Besson, L., Hellers, M., Guyez, A., Bonnet, S., & Reimann, T. (2026). Novel insights into the post-IR IRSL 200 signal bleachability of single-grain K-feldspars in fluvial modern analogues from the Southern Central Andes, Chile. Geochronology, 8(2), 297-312. https://doi.org/10.5194/gchron-8-297-2026

Image Credits: AI Generated

DOI: 10.5194/gchron-8-297-2026

Keywords: luminescence dating, K-feldspar, post-IR IRSL, single-grain analysis, bleaching, residual dose, fluvial sediment, Southern Central Andes, Chile, sediment tracing, geochronology, modern analogues

The principle behind LRSED is as elegant as it is surprising. Minerals such as quartz and feldspar, buried deep inside a rock, are bombarded for hundreds of thousands of years by ionising radiation from their surroundings. This radiation fills so-called electron traps in the crystal lattice, and the trapped electrons store energy that can be released as luminescence in the laboratory. The moment sunlight strikes the rock surface, however, photons begin to knock those electrons free — a process called bleaching. Because light penetrates only a short distance into rock, the bleaching is strongest at the surface and fades with depth, producing a characteristic sigmoidal profile of luminescence signal versus depth. The shape and position of that profile encode how long the surface has been exposed and how fast it has been wearing away.

Despite its promise, LRSED has struggled to spread beyond a handful of specialist laboratories. The mathematical analysis of luminescence depth profiles has been hampered by subjective choices: researchers traditionally picked the deep, saturated portion of a profile by eye to normalise their data, and different studies adopted different kinetic models and fitting strategies. The result was a patchwork of results that were difficult to compare and sometimes difficult to reproduce. CoRSEER, a MATLAB application released openly on Zenodo and GitHub, was built specifically to close these gaps by automating the entire workflow from raw luminescence measurements to exposure ages and erosion rates, complete with formal uncertainty estimates.

The software operates through three linked modules. The first, CoRSEER-I, tackles the normalisation problem head-on. Instead of relying on a human analyst to judge where the saturation plateau begins, it fits a weighted three-parameter logistic sigmoidal curve to the measured profile. The upper asymptote of that curve defines the saturation level objectively, and the inflection point — the depth at which half the saturation signal is reached — becomes a key diagnostic of exposure history. When the team reanalysed a published calibration profile from a road-cut sample with a known 57-year exposure history, the automated renormalisation revealed that the original manual processing had under-normalised the data by roughly 8.5 percent. Correcting for this raised the quality of the model fit, measured by the coefficient of determination, from 0.868 to 0.928, and dramatically tightened the uncertainty bounds on the fitted bleaching parameters.

Those bleaching parameters — the effective detrapping rate at the rock surface and the attenuation coefficient that governs how quickly light dies away with depth — are the linchpin of the whole method. They depend on rock type, mineralogy, surface orientation and even the accumulation of rock varnish, so they must be calibrated for each site and lithology using samples of known age. CoRSEER-I performs this calibration through a Monte Carlo inversion: it simulates thousands of candidate profiles across a broad parameter space, scores each against the measured data using a chi-square misfit, and retains the statistically plausible solutions. The second module, CoRSEER-II, then holds these calibration parameters fixed and inverts unknown profiles for exposure age, while the third, CoRSEER-III, searches for steady-state erosion histories that can explain profiles whose apparent ages fall far short of independently known ages.

The physics embedded in the forward model is what separates CoRSEER from earlier attempts. The governing equation tracks three simultaneous processes: the growth of the luminescence signal driven by ambient environmental radiation, the depth-dependent bleaching caused by attenuating sunlight, and the advection of the entire profile toward the surface as erosion removes material. Crucially, the model supports general-order kinetics rather than only first-order decay. This matters because feldspar infrared stimulated luminescence signals, widely used in the field, bleach more slowly and non-exponentially than first-order models assume. Previous work by the same research community has shown that assuming first-order kinetics can underestimate exposure ages and overestimate erosion rates, in some cases by up to an order of magnitude. CoRSEER solves the full equation numerically with a finite-difference scheme, using a second-order upwind method for the advection term.

To demonstrate that the platform works, the authors reanalysed published datasets from multiple regions, including 23 calibration samples. The results were encouraging on the most important metric. Apparent exposure ages computed by CoRSEER-II for 41 unknown samples clustered tightly around the one-to-one line relative to previously published values, spanning roughly one year to 8,500 years, with a correlation against the 1:1 line of 0.889. In a benchmark case — a rock-avalanche boulder independently dated to about 4,500 years old — CoRSEER-II returned an apparent age of only about nine years, correctly signalling that the boulder surface had been eroding. CoRSEER-III then estimated a steady-state erosion rate of about 0.107 millimetres per year, broadly consistent with the published value of 0.066 millimetres per year, and placed the onset of that erosion within decades of exposure.

The compilation also exposed honest limitations. Calibration parameters derived by CoRSEER-I sometimes shifted systematically from literature values, which the authors attribute mainly to differences in normalisation and fitting strategy rather than to any flaw in the underlying physics. Erosion-onset times showed larger departures than erosion rates themselves, reflecting the inherent difficulty of pinning down when erosion began from a single profile. The authors are candid that the tool’s performance depends heavily on data quality: profiles need dense millimetre-scale sampling, more than half of the measurement points should lie on the sigmoidal transition rather than the saturated plateau, and measurement scatter must be controlled. They also recommend using young, uneroded calibration surfaces wherever possible, because erosion on older calibration samples can bias the fitted parameters by large factors.

Why does this matter beyond the technical community? Exposure and erosion histories are the raw material of geomorphology, linking climate change, tectonics and the surface processes that sculpt mountains, coastlines and river valleys. The established alternative, terrestrial cosmogenic nuclide dating such as beryllium-10 measurement, works best on millennial timescales and suffers from a fundamental ambiguity: a single nuclide concentration can be explained by infinitely many combinations of exposure duration and erosion rate. LRSED offers an independent chronometer sensitive to decadal-to-millennial events — glacier retreat, coastal boulder transport, fault-scarp rupture, quarrying — and can be coupled with cosmogenic methods to break that ambiguity. By replacing subjective analyst choices with a transparent, standardised and openly available inversion framework, CoRSEER aims to make results from different laboratories genuinely comparable. If it succeeds, the quiet testimony written in light and depth on every sunlit rock surface will become a far more reliable archive of how our landscapes have changed — and how fast they are changing still.

Subject of Research: An open-source calculator for luminescence rock surface exposure dating that estimates rock exposure ages and erosion rates from luminescence depth profiles.

Article Title: CoRSEER: the calculator of rock surface exposure age and erosion rates for luminescence rock surface exposure dating

Article References: Pathan, A. N., Biswas, R. H., & Kumar, D. (2026). CoRSEER: the calculator of rock surface exposure age and erosion rates for luminescence rock surface exposure dating. Geochronology, 8(4), 607-625. https://doi.org/10.5194/gchron-8-607-2026

Image Credits: AI Generated

DOI: 10.5194/gchron-8-607-2026

Keywords: luminescence dating, rock surface exposure dating, erosion rates, geomorphology, geochronology, CoRSEER, open-source software, Monte Carlo inversion, feldspar IRSL, cosmogenic nuclide dating, landscape evolution, sigmoidal fitting

Glacially overdeepened basins are among the most powerful sediment traps in the geological record. Carved by a combination of glacial abrasion and pressurized, sediment-laden meltwater flowing beneath the ice, these structures erode far below the normal fluvial base level of a drainage system. Once the glacier retreats, the basin begins to fill, preserving nested sequences of glacial, glaciolacustrine and glaciofluvial deposits that can span multiple glacial cycles. Because such basins act as archives of Pleistocene glaciation history, the international ICDP-DOVE project, short for Drilling Overdeepened Alpine Valleys, has been systematically drilling and surveying these structures along the northern edge of the Alps to quantify the age, extent and environmental impact of past Alpine glaciations.

The Schäftlarn core, retrieved in 2017 by the Bavarian Environment Agency as part of this effort, sits within the former ablation area of the Isar-Loisach paleo-glacier, roughly one kilometer south of the ice-marginal position of the Last Glacial Maximum. The core revealed a striking three-part structure: about 83 meters of fine-grained silt and clay at the base, roughly 111 meters of coarse gravel-dominated deposits above, and a thin cap of diamictic sediments at the top. The carbonate content of the fine-grained lower unit remained remarkably constant throughout, indicating a stable catchment area dominated by Molasse bedrock and the Northern Calcareous Alps, with no evidence of a major glacial transgression from the Inn Valley into the foreland during deposition.

Dating this sequence was the central challenge, and the research team led by Gustav Firla of the University of Natural Resources and Life Sciences Vienna turned to single-grain feldspar luminescence dating, a technique that uses mineral grains as natural dosimeters to determine when sediment was last exposed to sunlight. Water-lain glacial sediments are notoriously difficult to date this way because grains are often transported in turbid meltwater and never fully reset their luminescence signal. By measuring individual grains of potassium-rich feldspar rather than bulk aliquots, the team could isolate the small, well-bleached population of grains that faithfully recorded the moment of deposition, applying a post-IR IRSL protocol and a bootstrapped Minimum Age Model to handle the complex, right-skewed dose distributions typical of poorly bleached glacial deposits.

The results were unambiguous and, in places, startling. Five samples from the base of the core clustered at 294,000 years, plus or minus 11,000 years, placing the start of basin infilling in early Marine Isotope Stage 8. A second cluster of samples from sand layers within the overlying gravel unit dated to late MIS 8, around 247,000 years ago. This means the entire lower half of the basin, from quiet glaciolacustrine silts to the first pulse of coarse glaciofluvial gravel, was deposited within a single glacial cycle that most researchers had assumed left almost no terrestrial trace in this region. Hughes and colleagues had previously termed MIS 8 a “missing glaciation” because it was part of a comparatively weak glacial-interglacial cycle, and known deposits from this interval are vanishingly rare in Europe, with only the Meikirch record in northern Switzerland previously documented in the Northern Alpine Foreland.

The team also investigated the Münchner Klettergarten outcrop near Baierbrunn, a celebrated exposure of glaciofluvial gravels on the southern margin of the Munich gravel plain that was first described by Penck and Brückner in 1909 as a classic example of their glacial series model. The lower sections of the outcrop had long been attributed to the Donau or Günz glaciations, which Bavarian stratigraphy places before MIS 12, making them potentially well over 450,000 years old. The new luminescence ages tell a very different story: the basal gravels date to late MIS 8, matching the coarse-grained deposits in the Schäftlarn core, while overlying sediments date to MIS 6, the penultimate glaciation. The long-assumed great antiquity of these deposits simply does not hold.

Corroborating evidence came from an unexpected direction. Gravel petrographic analysis of the Schäftlarn core showed a sharp shift in clast composition, with crystalline content rising from less than one percent at the base to nearly six percent higher up, and the limestone-to-dolomite ratio changing in parallel. The upper samples match the signature of last glacial deposits in the Wolfratshausen region almost exactly, suggesting the top of the gravel unit is far younger than the MIS 8 base. Even more remarkably, the sediments displayed unusually low potassium-40 and thorium-232 activities, verified by independent gamma spectrometry at two laboratories and by downhole spectral gamma logging, and this distinctive geochemical fingerprint allowed the team to correlate the Schäftlarn gravels directly with the Baierbrunn outcrop.

Reflection seismic surveys conducted in autumn 2023 added a crucial third dimension. Three profiles, using both P-wave and SH-wave configurations, revealed not one but two cross-cutting glacial basins beneath the site. The drill core was recovered from the older basin, the one filled during MIS 8, while a younger basin of undetermined age cuts directly through it. Based on comparison with other dated overdeepenings across the Alpine foreland, which consistently show fine-grained MIS 6 infills, the researchers tentatively attribute the younger basin to the penultimate glaciation. The seismic data also show that the younger glacial advance excavated parts of the older MIS 8 sediments and the underlying Molasse bedrock, then filled its own basin with a similar sedimentary succession.

Not every dating method cooperated. Cosmogenic nuclide burial dating of quartz cobbles from the outcrop produced beryllium-10 and aluminum-26 concentrations so low, and with aluminum-26 to beryllium-10 ratios between 0.7 and 4.5, that they could not be converted into reliable depositional ages. Instead, the team interpreted these signals as evidence that the cobbles were last exposed at the surface in the Pliocene or earlier, buried deeply for a prolonged period, and then rapidly redeposited by high-discharge proglacial processes, a reminder of how violently glacial systems can recycle sediment. A palynological screening of the fine-grained basin sediments likewise yielded mostly reworked Neogene and Early Pleistocene palynomorphs, preventing any reconstruction of the vegetation cover during MIS 8.

The broader implications reach to the foundations of Alpine Quaternary stratigraphy. The fourfold division of the Baierbrunn outcrop can no longer be mapped onto the four classical glaciations of Penck and Brückner; instead, it most likely records different stages of just the last two glacial cycles, with MIS 8 deposits at the base, MIS 6 deposits above a weathering horizon, and last glacial sediments on top. This reinforces the idea that the Munich gravel plain, an apparent exception to the terrace stratigraphy of the Bavarian foreland, follows a normal geological sequence but is significantly younger than previously expected. Together, the Schäftlarn core, the Baierbrunn outcrop and the seismic surveys provide the first steps in a re-evaluation of the established depositional chronology of the Bavarian Alpine foreland, and they demonstrate that even a glaciation once considered missing has left its signature, waiting patiently in the dark beneath the gravel plains of Munich.

Subject of Research: Luminescence dating of a glacially overdeepened basin fill from Marine Isotope Stage 8 in the northern Alpine foreland

Article Title: An MIS 8 terrestrial record retrieved from a glacially overdeepened basin in the northern foreland of the European Alps

Article References: An MIS 8 terrestrial record retrieved from a glacially overdeepened basin in the northern foreland of the European Alps. (n.d.). https://doi.org/10.5194/egqsj-75-85-2026

Image Credits: AI Generated

DOI: 10.5194/egqsj-75-85-2026

Keywords: MIS 8, glacial overdeepening, luminescence dating, Alpine foreland, Quaternary stratigraphy, ICDP-DOVE, Isar-Loisach glacier, Bavaria, Pleistocene glaciations, Schäftlarn core, Baierbrunn outcrop, reflection seismics

The stakes are far from academic. Cities such as Auckland, Mexico City, Naples and Clermont-Ferrand sit inside or beside volcanic fields where future eruptions are expected but their timing is unknown. Long-term forecasting also shapes decisions as consequential as where to site nuclear waste repositories. Yet the chronological record is startlingly thin. A survey of the southwestern United States found that of more than 2,000 identified Quaternary volcanoes in 37 fields, only about 25 percent have been dated at all, and that figure is comparatively good by global standards. Many published ages come without analytical details or error estimates, or were produced by methods since shown to be flawed.

The review concentrates on dispersed, monogenetic volcanism, in which each small batch of magma, typically less than a cubic kilometre, ascends along a fresh conduit and erupts in a new location over less than a century. Individual volcanoes are short-lived, but a field can remain active for millions of years, and because each new vent opens in a new place, hazard assessment demands spatiotemporal models of the entire field. All the dating methods covered exploit one of the few natural processes that run at constant rates over geological time: radioactive decay. Four of them date volcanic products directly, potassium-argon and argon-argon dating, uranium-series methods, cosmogenic nuclide exposure dating and luminescence dating, while radiocarbon dates organic material killed or bracketed by eruptions.

Argon dating is the workhorse for lava. Potassium-40 decays to the noble gas argon-40 with a half-life of 1,250 million years, and once a potassium-bearing mineral cools below its closure temperature, radiogenic argon becomes trapped in the crystal lattice. The classic K/Ar method measures potassium and argon on separate aliquots, which biases results in heterogeneous samples. The argon-argon variant sidesteps this by converting potassium-39 to argon-39 through neutron irradiation in a reactor, then measuring all argon isotopes at once. Step heating, in which gas is released in increments, reveals whether a sample has remained closed since eruption: an ideal sample yields a plateau age across at least three contiguous steps representing over half the released argon-39. In Argentina’s Payenia Basaltic Province, one basalt flow produced a textbook plateau and an atmospheric argon ratio consistent with the present-day value of 295.5, while a second, apparently very young sample showed a poorly constrained isochron, illustrating how close young basalts push the method’s roughly 1,500-year lower limit.

The method’s power is best shown at the other end of the precision spectrum. Renne and colleagues dated sanidine phenocrysts from the pumice of Vesuvius’s famous 79 CE eruption, documented by Pliny the Younger, and obtained 1.925 plus or minus 0.094 thousand years, agreeing with the historical record despite hints of excess argon. In Australia’s Newer Volcanic Province, argon-argon ages reproduced older K/Ar results but with far better precision, and some ages came out significantly younger, exposing the excess-argon problems the older method could not detect. Even so, the authors caution that ignimbrites, altered lavas and glasses often carry anomalously high atmospheric argon, and that argon-39 recoil during irradiation can distort fine-grained samples.

Radiocarbon dating, by contrast, never dates the eruption itself. It dates organisms, trees burnt by lava, charcoal in pyroclastic flows, or peat layers bracketing a deposit, whose death coincided with or bounds the event. Carbon-14 decays with a half-life of 5,730 years, setting an upper limit near 55,000 years, and accelerator mass spectrometry now delivers sub-percent precision on milligram samples. But the atmosphere’s carbon-14 production has never been constant, so laboratory ages must be calibrated against curves built from tree rings, varved sediments, corals and ice cores. Reservoir effects complicate matters further: magmatic carbon dioxide contains no carbon-14 at all, and modern leaves around the Solfatara crater in Italy’s Campi Flegrei show radiocarbon ages spanning 5,000 years. At Taupo in New Zealand, such magmatic bias can overestimate eruption ages by more than 200 years.

The most famous radiocarbon controversy involves Santorini. Archaeological chronologies place the great Late Bronze Age eruption around 1520 BCE, but more than 100 radiocarbon measurements on short-lived material, plus a dated olive branch from the island itself, point to roughly 1627 to 1600 BCE, a century or more older. Decades of debate have implicated magmatic degassing, calibration-curve anomalies and archaeological assumptions, and the review presents it as a cautionary tale: even a mature method with exquisite precision can yield contested ages when calibration curves are imperfect. Meanwhile, applied work keeps proving its worth. At Vulcano in Italy, radiocarbon combined with tephrostratigraphy revealed at least 19 eruptions between 900 and 1550 CE, one every 35 years, far more frequent than the previously assumed 130-year recurrence. High-density dating of sediments from Lake Motosu similarly refined Mount Fuji’s Holocene history and uncovered previously unknown eruptions.

Uranium-series methods exploit the decay chains of uranium-238, uranium-235 and thorium-232, in which intermediate isotopes fall out of secular equilibrium during melting and crystallisation and then rebalance at known rates. Plotting mineral phases on an isochron yields crystallisation ages, as demonstrated for the Puy de Dôme in France, where a 12.1 plus or minus 1.0 thousand-year age agrees with earlier luminescence and radiocarbon results. Because crystals can linger in a magma chamber long before eruption, these ages are best treated as maxima. The zircon double-dating technique closes that gap by applying both uranium-thorium disequilibrium and helium accumulation to the same crystals, delivering eruption ages from 2 to 1,000 thousand years. At Las Tres Vírgenes in Baja California, this approach produced a corrected eruption age of 36 plus or minus 6 thousand years, overturning a lone radiocarbon date and an ambiguous eighteenth-century historical reference.

Cosmogenic nuclide exposure dating reads a different clock: the buildup of rare isotopes such as helium-3, neon-21, chlorine-36 and beryllium-10 as cosmic-ray secondaries bombard freshly exposed rock surfaces. New lava flows are ideal because they start with no inherited signal, though identifying the original surface and scaling production rates across latitude and altitude remain tricky, especially with few calibration sites in the Southern Hemisphere. Applications range from helium-3 and neon-21 dating of olivine in Argentine basalts, confirming activity as recent as 5,000 years ago, to chlorine-36 dating of dacitic flows on Pico de Orizaba in Mexico, where high chlorine content pushed uncertainties to nearly 20 percent. Luminescence dating, finally, measures trapped charge in quartz and feldspar, reset by heat or light and rebuilt by environmental radiation. It can date xenoliths baked by magma, sediments beneath lava flows, and even fragments shattered by phreatic explosions, as shown for the Eifel maars in Germany, though volcanic minerals are notoriously prone to anomalous fading that protocols like post-infrared infrared stimulated luminescence are designed to avoid.

The review’s overarching advice is blunt: never take a published volcanic age at face value. Each method dates a slightly different moment, cooling, exposure, crystallisation or death, and the most trustworthy chronologies are those reproduced by two or more independent techniques. Detailed reporting of sampling, procedures and stratigraphy, the authors argue, is what allows ages to be re-evaluated decades later. As analytical instruments improve and more eruptions worldwide are dated, the foundation for forecasting the next eruption of an Auckland or a Naples grows firmer, one carefully vetted number at a time.

Subject of Research: Numerical dating methods for establishing eruption chronologies of Quaternary volcanism

Article Title: Review of numerical methods for dating Quaternary volcanism

Article References: McCreary, W., May, V., Mueller, D., Serra, E., & Preusser, F. (2026). Review of numerical methods for dating Quaternary volcanism. E&G Quaternary Science Journal, 75(1), 131-162. https://doi.org/10.5194/egqsj-75-131-2026

Image Credits: AI Generated

DOI: 10.5194/egqsj-75-131-2026

Keywords: volcanism, geochronology, argon-argon dating, radiocarbon dating, uranium-series dating, cosmogenic nuclide exposure dating, luminescence dating, volcanic hazard, Quaternary, tephrochronology, zircon double dating, eruption recurrence

The authors argue that much of the confusion surrounding anomalous fading can be traced back to insecure mathematical footing in the field’s seminal papers, which has since been papered over by a wealth of recirculated measurement protocols, spreadsheets, and data-reduction packages, all patching the theoretical ambiguities in numerically divergent ways. One striking example concerns the work of Ann Wintle, who in 1973 first documented anomalous signal loss in mineral samples and stated that the decay curves did not conform to any simple time dependence. Guralnik and King show that her data in fact exhibit a clean logarithmic decay, consistent with more than a century of phosphorescence research stretching back to Becquerel in 1867, and with tunneling-driven kinetic models developed in the late 1950s and 1960s. An erroneous statement, they suggest, stalled progress for decades.

The problems do not stop there. Rudiger Visocekas, who coined the term logarithmic decay in the 1970s and 1980s, defined his decay constant in a way that was at odds with his own subsequent derivation, producing a model that inevitably predicts negative concentrations of trapped charge, an unphysical outcome. The authors show that a self-consistent integration of the underlying power-law decay instead yields an exponential-of-a-logarithm expression, in which concentration remains positive at all times and which foreshadows a more physically grounded model published by David Huntley in 2006. They also demonstrate that a supposedly new equation proposed by Lamothe and colleagues in 2003, which extended the popular Huntley and Lamothe correction beyond the linear part of the dose response curve, is in reality merely a reparametrized, single-iteration restatement of the original 2001 formula, a tautology rather than a new model. The older ages it returned owe nothing to fading itself, but arise from an implicit inversion of a nonlinear dose response, a step formalized only later by Wallinga and colleagues in 2007.

At the heart of the reappraisal is a matter of units and dimensional hygiene. The familiar fading rate, the g-value, is conventionally reported in the pseudo-units of percent per decade of time, a normalization the authors describe as a violation of core metrological principles established since Maxwell’s era. They recast the fading constant as a dimensionless sensitivity, defined as the fractional loss of trapped charge per fractional progression of time, and show that all the familiar results of the field can be derived cleanly from this single definition. They further supply, for the first time in a verifiable publication, the formula for rescaling a g-value from one reference time to another, a relationship that had previously existed only in published code citing a private communication, which the authors call the hallmark of grey literature underpinning an ambiguous primary source.

The practical payoff comes in the form of new closed-form analytical expressions for the two most widely used age correction schemes, those of Huntley and Lamothe from 2001 and of Kars and colleagues from 2008. The Huntley and Lamothe correction, which assumes linear signal growth, turns out to be solvable exactly using the Lambert W function, a special function familiar from problems where an unknown appears both raw and inside a logarithm. The authors derive an explicit age equation together with a fully analytical propagation of uncertainties, replacing the iterative and Monte Carlo approaches that practitioners previously relied upon. When tested against the published feldspar ages of a Japanese marine sediment archive, their explicit equations reproduce the implicit iterative results to sixteen-digit numerical precision, and their analytical uncertainties match Monte Carlo simulations to within a few percent, at a fraction of the computational cost.

The Kars model, which couples quantum tunneling to the gradual refilling of traps by environmental radiation, is more demanding because it requires integration over a distribution of electron-hole separation distances. Guralnik and King show that this integration can be approximated with sub-percent accuracy by extending the concept of an effective tunneling radius, yielding a compact explicit age equation and, for the first time, an analytical formula for its uncertainty. Until now, there existed no alternative to Monte Carlo simulation for estimating the age uncertainties of the Kars correction, making the new expression a significant practical advance for laboratories worldwide.

Perhaps the most intriguing part of the paper is its exploration of unorthodox model combinations that had somehow never been assembled before. The authors couple fading to signal growth obeying general order kinetics and to the one-trap one-recombination center model, and they propose a new correction based directly on the nearest-neighbor distribution model of tunneling, bypassing the logarithmic decay law altogether. When benchmarked against an independent age constraint, the Toya tephra in northern Japan, dated to roughly one hundred thousand years, the corrected luminescence ages of the underlying sediment spanned a considerable range, but the models built on the physically grounded nearest-neighbor distribution decay clustered closest to the independent chronology. Notably, the choice of retrapping model, whether first-order or higher-order, appeared to exert only a second-order effect on the corrected ages.

The authors are careful to warn against overinterpreting such benchmarks. They point out that even in a well-constrained sedimentary archive bristling with tephra markers, distinguishing between fading models by appealing to independent ages can quickly descend into circular arguments, particularly given systematic uncertainties of up to twenty percent in dose rate estimates. With a more realistic ten percent uncertainty on the tephra age itself, every fading-corrected luminescence age in their test proved compatible with the best available independent chronology. Their stated aim, they emphasize, is not to crown a winning model but to state all the existing models in a mathematically literate manner and present correct, transparent calculations, leaving it to practitioners to use them responsibly.

The reappraisal extends beyond sediment dating into luminescence thermochronology. Using samples from the KTB deep borehole in Germany, where rocks have been stored isothermally for times far exceeding the response time of the luminescence system, the authors combine their new athermal fading corrections with a thermal equilibrium factor in a single one-line expression. This compact model quantitatively predicts observed infrared stimulated luminescence ages across twelve depths in the borehole, spanning roughly one and a half orders of magnitude in apparent age, and replaces extensive Monte Carlo simulations with a straightforward analytical formula. For a field in which the arithmetic of its founding papers has gone largely unchallenged for fifty years, the message of this work is clear: the mathematics of anomalous fading needed a housecleaning, and now it has one.

Subject of Research: Mathematical correction of anomalous fading in feldspar luminescence dating

Article Title: Anomalous fading correction in luminescence dating – a mathematical reappraisal

Article References: Guralnik, B., & King, G. E. (2026). Anomalous fading correction in luminescence dating – a mathematical reappraisal. Geochronology, 8(3), 589-606. https://doi.org/10.5194/gchron-8-589-2026

Image Credits: AI Generated

DOI: 10.5194/gchron-8-589-2026

Keywords: luminescence dating, anomalous fading, feldspar, geochronology, quantum tunneling, Lambert W function, dose rate, thermochronology, uncertainty propagation, Kars model, Huntley and Lamothe, trapped charge

The team, led by Eren Şahiner of Ankara University’s Earth Sciences Application and Research Center, together with Tamer Koralay, Bilal Söğüt, Tunç Sezgin, and Yusuf Kağan Kadıoğlu, dated five specimens drawn from the Frigidarium’s fabric. Optically stimulated luminescence, or OSL, works by measuring the trapped charge accumulated in mineral grains such as quartz since they were last exposed to sunlight or heat. When a brick is fired or mortar quartz is exposed during preparation, its luminescence clock resets; from that moment, background ionizing radiation from the surrounding sediments, cosmic rays, and the sample’s own radioisotopes steadily winds the clock forward. In the laboratory, controlled optical stimulation releases the stored signal, and the intensity reveals the equivalent dose, the total radiation absorbed since the resetting event. Dividing that dose by the environmental dose rate yields an age.

What distinguishes this study is the rigor of its dose-rate protocol. The researchers documented every step, including grain-size-dependent attenuation of beta radiation, a carbonate water-equivalence correction specifically designed for lime-bound mortars, and an explicit assessment of gamma dose-rate heterogeneity. Lime mortars are notoriously difficult subjects for luminescence dating because their carbonate content alters the way water content and radiation attenuation interact, and the team applied the Carb-style modeling approach to handle this. Gamma dose-rate heterogeneity is a further subtlety: in a wall of compound masonry, where bricks and mortar alternate, the radiation field is not uniform, and a sample’s dose rate depends on the geometry of its immediate surroundings. Rather than treating this uncertainty as a nuisance to be ignored, the authors modeled the gamma-field geometry as a Bayesian nuisance parameter, propagating the dosimetric uncertainty of compound masonry directly into the final chronological posteriors.

The five OSL ages were then synthesized in a constrained Bayesian phase model, a framework adapted from the radiocarbon dating community, carrying only stratigraphic information and terminus post quem priors. In plain terms, the model was told only the relative order of events and the earliest possible dates, and then asked to reconcile the luminescence measurements into coherent phases of the building’s life. Three phases emerged. Primary construction of the Frigidarium falls in the late first to early second century CE, with a posterior median of 162 CE and a 95.4 percent highest-posterior-density interval spanning 60 to 323 CE. This is consistent with the flourishing of Roman bath architecture in Caria during the Antonine and Severan periods, when cities of western Anatolia competed to monumentalize their civic centers.

The second phase captures a repair episode, dated with a mode of 375 CE within an interval of 365 to 499 CE. The coincidence is striking: the great Eastern Mediterranean earthquake of 365 CE, generated near Crete and felt across the entire region, devastated coastal cities and is one of the most extensively documented seismic events of antiquity. The repair date aligns closely with that catastrophe, suggesting that the Frigidarium was damaged and subsequently restored in its aftermath. Archaeoseismology, the discipline that reads earthquake damage in ancient structures, has long catalogued tilted walls, displaced masonry blocks, and collapsed colonnades across the Mediterranean, but assigning absolute dates to such damage has remained difficult. Here, the luminescence clock provides an independent chronological anchor for a restoration that historical sources alone could never have pinned to this provincial city.

The third phase concerns the building’s terminal collapse. Under dosimetric priors alone, the model placed the collapse at a mode of 545 CE within a broad window of 405 to 802 CE, a range too wide to discriminate among the known seismic events of the period. The Eastern Mediterranean experienced what some scholars describe as seismic clustering between the fourth and sixth centuries, with major earthquakes in 494 CE and in 554 to 558 CE, as well as events around 610 to 630 CE. Any of these could, in principle, have delivered the fatal blow. The Bayesian model alone could not choose between them, and the authors were careful not to overstate what the luminescence data could resolve.

The decisive breakthrough came from archaeology rather than physics. Coins minted in 613 CE were found sealed beneath the collapse debris, providing a hard terminus post quem: the building could not have fallen before those coins were deposited. When this numismatic evidence was incorporated into the Bayesian model, the collapse window narrowed dramatically to 613 to 850 CE, with a mode of 633 CE. Within that refined window, the circa 610 to 630 CE seismic anchor dominates, achieving an overlap probability of 0.31 at a 25-year tolerance. The statistical conclusion is that a seismic trigger is the most probable cause of the terminal loss of the Frigidarium’s monumental function in the early seventh century CE, a period of chronic instability for the Eastern Roman Empire marked by Persian wars, plague, and persistent tectonic activity in western Anatolia.

Stratonikeia itself is an ideal laboratory for such work. Located in the Yatağan district of Muğla Province, the city has been the focus of long-running excavations directed from Pamukkale University, and previous studies have documented earthquake relics in the ancient city and at the nearby sanctuary of Hekate at Lagina. Earlier archaeometric investigations of Stratonikeia’s bricks, mortars, and building stones, including luminescence dating of bricks from the Erikli Basilica and characterization of the theater’s architectural stones, laid the groundwork for the present study. The Frigidarium, the cold room where bathers began or ended their circuit of the Roman baths, is a structurally massive space whose thick walls preserve a legible record of construction, repair, and destruction.

The methodological implications extend well beyond one building. Mortar dating by luminescence has matured over the past two decades, and reviews of the field have emphasized the need for transparent, reproducible dose-rate assessment, particularly for carbonate-rich materials. By publishing a fully documented protocol, treating gamma-field geometry as an explicit model parameter, and releasing the R and Python scripts implementing the dose-rate calculation, equivalent-dose modeling, and Bayesian phase chronology, together with full Monte Carlo posterior samples, the team has set a benchmark for how archaeoseismic chronologies should be constructed and reported. The approach converts scattered, imprecise age estimates into a coherent probabilistic narrative of a building’s biography.

For the wider public, the study offers something rare: a precise, quantified answer to the question of when a great Roman monument was born, when it was wounded, and when it died. The Frigidarium of Stratonikeia rose in the high empire, was patched after the great earthquake of 365 CE, and stood for roughly another two and a half centuries before an early seventh-century tremor ended its life as a functioning bath. In the layered ruins of Caria, the marriage of luminescence physics and Bayesian statistics is now capable of reading that biography directly from the walls themselves, giving voice to buildings that survived their own civilization by more than a millennium.

Subject of Research: Bayesian archaeoseismic chronology of the Stratonikeia Frigidarium using optically stimulated luminescence dating

Article Title: Bayesian archaeoseismic chronology of the Stratonikeia Frigidarium from construction to seismic collapse

Article References: Şahiner, E., Koralay, T., Söğüt, B., Sezgin, T., & Kadıoğlu, Y. K. (2026). Bayesian archaeoseismic chronology of the Stratonikeia Frigidarium from construction to seismic collapse. Archaeological and Anthropological Sciences, 18(10), Article 203. https://doi.org/10.1007/s12520-026-02564-9

Image Credits: AI Generated

DOI: 10.1007/s12520-026-02564-9

Keywords: luminescence dating, OSL, Stratonikeia, archaeoseismology, Bayesian modelling, Roman architecture, frigidarium, Roman baths, earthquake chronology, mortar dating, Anatolia, 365 CE earthquake