Abstract Contemporary landslide susceptibility modeling commonly utilizes machine learning (ML) approaches and algorithms. These modeling approaches typically use landslide inventories as a primary component, along with a conditioning feature selection method. Although model results may be sensitive to variability in input data, resulting effects are seldom investigated in detail. We leveraged a susceptibility approach based on logistic regression (LR) to assess the sensitivity of model results to landslide inventory size. Using an inventory of 1054 landslides, we evaluated sample size at 50-landslide increments (to the maximum possible value of 1050), generating 1000 random inventory subsets for each sample size (i.e., Monte Carlo simulation) to quantify variation in model performance due to randomized train and holdout selection. Initial visual inspection of susceptibility map outputs produced using these model results revealed extreme overprediction. The spurious models indicate a false correlation between model accuracy and realism. To address this, our sensitivity analysis workflow utilizes a performance metric correction factor and stricter feature selection to promote consistent, realistic, and reproducible models. The workflow presented here quantifies model variability and assesses the sensitivity of susceptibility maps to changing model inputs.
This study reconstructs Late Holocene vegetation and Indian summer monsoon (ISM) variability over the past similar to 2700 years from the Bandhavgarh Tiger Reserve, located in central India's Core Monsoon Zone (CMZ). High-resolution pollen data reveal five distinct vegetational phases, with the period between similar to 2700 and 2300 cal yr BP characterized by a dense moist deciduous forest dominated by Shorea robusta, indicating strong monsoonal activity. The interval between similar to 2300 and 1600 cal yr BP shows four distinct sub-phases: a warm and humid episode (similar to 2300-2200 cal yr BP), a pronounced dry spell (similar to 2200-2030 cal yr BP), an ameliorating phase (similar to 2030-1680 cal yr BP), and renewed monsoon intensification (similar to 1680-1600 cal yr BP). A moderately strong and stable ISM persisted until similar to 820 cal yr BP, followed by a dry, open vegetation regime (similar to 820-400 cal yr BP) coinciding with increased anthropogenic indicators. Since similar to 400 cal yr BP, the resurgence of moist deciduous taxa reflects climatic recovery and forest regeneration. Pollen-based climate reconstructions provide valuable insights into long-term monsoon-vegetation interactions in tropical India.
Abstract Sediment cores from Dyken and Shaver ponds on the Rensselaer Plateau, eastern New York state, USA, were analyzed for sediment chemistry, pollen, plant macrofossils, and diatoms to reconstruct the ecological and climate history since Laurentide ice sheet retreat of this previously unstudied region. Forests were established by 13,040 cal yr BP, and pollen records follow the well-documented northeastern U.S. sequence: Allerød interstade (mixed boreal and thermophilous taxa, including Tsuga ), Younger Dryas stadial cooling (rise of Alnus, Betula , and Picea ), Early Holocene warmth and dryness ( Pinus dominance), increased moisture and Tsuga rise, Mid-Holocene Tsuga decline followed by Late Holocene recovery, neoglacial cooling in the past two millennia ( Picea rise), and European settlement ( Ambrosia rise). Superimposed on these longer-term trends are centennial- to sub-millennial-scale climate variations marked by shifts in vegetation and changes in diatom species abundance. The combined datasets indicate that Mid-Holocene Tsuga decline coincided with several severe droughts as well as climatic cooling events between ca. 6100 and ca. 3700 cal kyr BP.
Luminescence dating has developed over the last ∼60 years as a powerful technique for placing environmental and anthropogenic change into a secure temporal framework. However, over time, many have forgotten, or were never introduced to, the history of how of the method developed, particularly the role of unique instruments built in-house that enabled key methodological advances. In this paper we provide a concise history of the technique’s evolution, drawing on our own experiences.
Conventional thought holds that in formerly glaciated areas straying of anadromous fish from nearby unglaciated areas established contemporary salmon populations. An additional explanation for patterns of salmon life-history diversity and population structure derives from isolation of populations in proglacial lakes. We evaluate evidence for these potentially complementary hypotheses in chum salmon from two previously glaciated North American regions: the southern Alaska Peninsula/upper Cook Inlet and the Salish Sea of northwestern Washington and southern British Columbia. Some chum salmon populations in the southern Alaska Peninsula are genetic outliers compared with other nearby populations, while Salish Sea chum salmon populations have greater region-wide genetic divergence and lower gene diversity. Within-population genetic diversity and among-population divergence in both study areas support a hypothesis of salmon persistence relying on cryptic isolation and freshwater-resident (trout-like) life histories in proglacial lakes. We find that ice age adaptation of salmon to a trout life history helps explain aspects of contemporary population structure and life-history diversity.
Four new Pleistocene track-bearing aeolianite surfaces have been identified on South Africa's Cape south coast, each portraying evidence of tortoise tracks. Together, they add to and buttress previous reports of tortoise tracks and trackways from the region. Globally, this remains the only area from which fossilized tortoise tracks have been recorded, and for the first time we illustrate the preservation of typical tortoise trackway morphology (involving a 'tramline' pattern with a wide straddle and closely spaced tracks), as observed in the trackways of extant tortoises. One site provides further evidence for the inferred presence of a very large tortoise trackmaker from the region during the Pleistocene. This tortoise was substantially larger than the largest extant tortoises in southern Africa, which bolsters the inference of either an extinct very large tortoise or a large chrono-subspecies of the extant leopard tortoise (Stigmochelys pardalis). The mismatch between the body fossil record and trace fossil record with respect to the presence of large tortoises in the southern Cape persists. One trackway was probably registered by a smaller leopard tortoise, and the other trackways may have been registered by an angulate tortoise (Chersina angulata).
We present isotopic data from mammalian megafauna from the Jirau Paleontological Site (Itapipoca, Cear & aacute; State, Brazil) and Rio Miranda (Mato Grosso do Sul State), both located in the Brazilian Intertropical Region, dating to the Mid-Late Holocene. The isotopic composition (delta 13C) of eight tooth fragments was determined for the following taxa: E. laurillardi, N. platensis, T. platensis, S. populator, X. bahiense, and P. major. Results indicate that the herbivorous taxa had mixed diets, consistent with deciduous to semi deciduous forest and wooded savannah environments. S. populator likely preyed upon herbivores with mixed diets and inhabited wooded savannahs. E. laurillardi, N. platensis, and T. platensis exhibited generalist feeding behavior with a high proportion of C-3 plants in their diet, associated with the fragmentation and reduction of open environments (savannah and wooded savannah) and the concurrent expansion of forested areas during the Holocene Climatic Optimum. X. bahiense and P. major exhibited browser-type diets in Itapipoca, suggesting adaptation to the expansion of deciduous and semi deciduous forests during the Holocene Climatic Optimum. By comparing isotopic data with paleoecological, palynological, and paleobiogeographical evidence, we infer that the Intertropical Region represented one of the last environmental refuges for extinct meso- and megamammal faunas during the Holocene.
Indian banyan trees, with their complex network of secondary trunks, support entire ecosystems and are often favored for large rural gatherings due to their vast canopy. However, dating these heritage trees in South Asia has traditionally relied on written records and local lore, resulting in imprecise age estimates that limit our understanding of their historical significance. This study establishes a precise minimum age for a monumental banyan tree located on the Indian Tobacco Company campus, founded in 1901 CE in Munger, Bihar, India. According to local lore, the tree has long been a gathering place for both rulers and commoners. Radiocarbon dating of the pith from an exposed secondary trunk and a primary branch of the Munger Banyan were 276 +/- 36 and 652 +/- 37 years BP, respectively, indicating when each sample of wood stopped exchanging carbon with the atmosphere. These findings suggest a minimum radiocarbon age of ca. 700 years, making it the oldest accurately dated Ficus benghalensis in the world. The study refines radiocarbon dating approaches for tropical hardwoods by emphasizing precise pith targeting, a method rarely applied due to indistinct growth-ring boundaries. By integrating advanced calibration techniques, the study enhances chronological accuracy and improves understanding of the longevity and significance of heritage trees and tropical forest ecosystems.
Thirty packrat (Neotoma spp.) middens collected from boulder fields near Catavi & ntilde;a, Baja California, Mexico, at 640-680 m elevation provide the first long chronology of macrofossils and pollen spanning the late Quaternary in the Central Desert of Baja California. Midden plant macrofossil and pollen assemblages document a rich chaparral/woodland assemblage during the last glacial and early Holocene dominated by Parry pinyon (Pinus quadrifolia) and California juniper (Juniperus californica) until 11,630 cal yr BP. This indicates chaparral/woodland had a much more extensive distribution in what are now desert elevations in northern and central Baja California. In contrast to late glacial and early Holocene midden records from northeastern Baja, the Catavi & ntilde;a middens of the same age lack plants adapted to warm season precipitation, suggesting that decreased temperatures and evapotranspiration during the growing season and enhanced winter precipitation, with little contribution from summer rains, supported the lowering of chaparral/woodland species distributions in central Baja California. Catavi & ntilde;a middens also record endemic desert plant taxa mixed in with chaparral/woodland species during the Pleistocene, persisting throughout the Holocene, followed by the quick arrival of other desert species after similar to 11,000 cal yr BP. Baja California remains a high-potential yet poorly sampled area for packrat midden research in North America.
Slack-water deposits are archives of paleoflood frequency, magnitude, and provenance. In the eastern Himalaya, deposits along the Siang River document Quaternary outburst megafloods originating from southeastern Tibet. Here we present new observations of slack-water deposits within aggradational terraces of the Siyom River, the Siang's largest tributary. Terrace stratigraphy reveals distinct, regionally extensive sedimentary facies including laminated sands, clays, and detrital organic-rich deposits consistent with slack-water deposition from temporarily impounded waters. Radiocarbon dates from clay and organic horizons range from 34,020 to 10,630 cal yr BP, overlapping with age constraints for Tibetan paleolake deposits. Detrital zircon (U-Th)/Pb geochronology confirms a local source for the underlying fluvial facies, whereas event-deposit silts contain young zircons derived from Tibet, supporting their interpretation as megaflood deposits. This evidence, combined with the deposits' temporal overlap with Tibetan paleolakes and distinctive slack-water sedimentology, demonstrates that event facies formed through megaflood backflooding sourced from southeastern Tibet. The results point to the likelihood of similar deposits in other tributaries, providing a framework for regional investigation. Our findings further show that megafloods in steep terrain can produce substantial deposition and terrace formation tens of kilometers upstream in tributaries-far beyond the main stem floodway-revealing an overlooked geomorphic imprint of extreme floods.
The Sauce Grande River Basin (Buenos Aires Province, Argentina) presents a late Pliocene-Holocene sedimentary succession that preserves key evidence of Quaternary paleoenvironmental change in the southern Pampas. This study integrates stratigraphy, sedimentology, paleopedology, geochemistry, mineralogy, and stable isotope data to reconstruct paleoenvironmental evolution from the late Pliocene to the Late Pleistocene. Three paleosol types (Calcisols, Calcic Protosols, and Protosols) were identified and characterized through field descriptions, micromorphology, and molecular indices. Their development reflects shifts in landscape stability, sediment supply, and soil moisture regimes, consistent with glacioeustatic fluctuations and climatic oscillations during the late Plio-Pleistocene transition and the Quaternary. Stable isotope analyses of pedogenic carbonates reveal a trend from C-3-dominated vegetation under more humid conditions in the late Pliocene-Early Pleistocene, to more arid Late Pleistocene-Holocene environments, with increased delta 13C values indicating reduced vegetation density and a higher potential contribution of C-4 plants. These findings align with palynological and faunal evidence, highlighting the value of paleosols as sensitive indicators of environmental change. The multiproxy approach adopted here provides new insights into soil formation dynamics and Quaternary palaeoecological transitions in non-glaciated midlatitude settings of South America.
Stable carbon isotopes in Holocene soils provide key insights into past climate and ecosystems. This study presents high-resolution isotope analyses of pedogenic carbonates and organic carbon from modern loess-derived soils in northern Iran across a strong precipitation gradient (150-850 mm mean annual precipitation [MAP]). Eight soil profiles span five soil orders: Alfisols, Mollisols, Inceptisols, Aridisols, and Entisols. The mean delta(13)Cpc values show strong linear relationships with MAP (delta(13)Cpc = -0.0093 & times; MAP + 1.8878; R-2 = 0.98) and with the ratio of precipitation to potential evapotranspiration, P/PET (delta(13)Cpc = -8.6842 & times; P/PET + 1.608; R-2 = 0.99), indicating that delta(13)Cpc reliably reflects moisture availability during carbonate formation. Values range from -6.2 parts per thousand in wetter sites to -0.1 parts per thousand in dry areas, reflecting changes in soil respiration and CO2 flux. delta(13)Coc values (-25.6 parts per thousand to -23.3 parts per thousand) indicate dominant C-3 vegetation and exhibit a bimodal response to precipitation, increasing from arid to semiarid zones and decreasing in wetter forests. Oxygen isotopes in carbonate (delta(1)(8)Opc = -7.9 parts per thousand to -6.6 parts per thousand) show limited climate sensitivity, reflecting precipitation mainly from the Caspian Sea with minimal evaporative enrichment. Overall, delta(13)Cpc, delta(13)Coc, and delta(1)(8)Opc provide robust proxies for soil moisture, vegetation structure, and water sources, supporting paleoenvironmental reconstruction in loess systems.
The 7.7-m loess-paleosol sequence in Trzebnica records a short (ca. 3 ka) but intense phase of loess accumulation that preceded the onset of the Weichselian Late Glacial (MIS 2). The stratigraphy contains a well-defined erosional unconformity, marking a hiatus prior to the deposition of the youngest loess unit (L1LL1). The Trzebnica loess contains numerous signals of slope redeposition, as evidenced by lithological indicators. In this paper, we present optically stimulated luminescence (OSL) and radiocarbon data for the Trzebnica site, as well as granulometric analyses and end-member modeling of grain-size data. Our findings refine the regional loess stratigraphy and challenge an earlier hypothesis regarding the age of the Trzebnica 2 archeological site, which pointed to its evolution beginning around MIS 11. Limitations of 14C dating and its comparison to OSL dates are discussed.
Low optically stimulated luminescence (OSL) sensitivity is commonly observed in quartz from tectonically active catchments, suggesting limited or short-lived conditions favorable for sensitization. We characterize the OSL sensitivity of quartz sand within a small, tectonically active catchment in Sicily using modern fluvial samples and a hillslope soil sample. We investigate how OSL sensitivity varies with bedrock lithology, weathering proxies, and topographic metrics. OSL sensitivity spans three orders of magnitude (60-2800 counts/Gy/mm3) with no clear linkage to bedrock source. The soil sample exhibits the highest OSL sensitivity, and positive relationships between OSL sensitivity, magnetic susceptibility, and weathering intensity suggest that pedogenic hillslope processes enhance quartz OSL sensitivity. In contrast, fluvial sediments show low OSL sensitivity and a modest inverse relationship to channel steepness and hypsometry. OSL sensitivity decreases downstream, suggesting that highly sensitized grains from hillslope soils are progressively diluted by low OSL sensitivity sediment likely generated by rapid bedrock erosion in the catchment. These results highlight a hierarchy of controls: bedrock lithology sets the initial OSL sensitivity, hillslope processes enhance it, rapid erosion dilutes it, and fluvial transport modulates it through mixing, explaining why tectonically active catchments rarely preserve quartz with high OSL sensitivity.
New cosmogenic 10Be exposure ages and a proglacial lake sediment archive provide the first record of local ice cover following the deglaciation of the Cordilleran Ice Sheet (CIS) in southeast Alaska. Exposure ages from Necker Bay corroborate existing evidence for a CIS deglaciation age of similar to 15-14 ka from the outer coast of Baranof Island. We date retreat farther inland on the western and eastern flanks of the island to the Early Holocene, providing evidence for an ice cap persisting on Baranof Island similar to 3 ka after CIS retreat. Baranof Lake sediment cores document continued local ice cover until similar to 10.4 ka, after which glaciers receded to their Holocene minima until similar to 8 ka. Glaciers grew through the remainder of the Holocene, reaching their maxima during the last millennium before retreating rapidly during the last century. Remote sensing analysis of glacial change around Baranof Lake from 1948 to 2023 CE shows that the rate of glacier area loss increased by an order of magnitude after 1986 CE, from -0.03 km2/yr to -0.29 km2/yr. This trend in glacier area loss is reflected across Alaska and western Canada, highlighting the sensitivity of Beringian glaciers to climate changes and the significant contribution they will make to sea-level rise this century.
The Kovrizhka sites are some of the most studied and highly informative for the entire northern Cis-Baikal region of Siberia and illustrate the history and development of ancient cultures in the Vitim area during the Late Upper Paleolithic to Early Neolithic. To better understand human settlement practices during this time, we constructed a model of Late Quaternary landscape formation and human occupation in the Vitim River valley based on a geomorphological study and radiocarbon dating of archaeological sites Kovrizhka I-VI in the Baikal-Pathom Highlands. The model reconstructs human habitation of the valley from 19.9 to 6.7 ka and connects settlement patterns to general landscape features, stone (mineral) and food resources, the flood regime of the Vitim River, and dynamics of landscape formation. A secondary focus of this study is to assess the timing and geomorphological remnants of megafloods originating from breakthroughs of the Muya (Vitim) glacial paleolake in Marine Isotope Stages (MIS) 3 and 2, and their impact on human settlement. The last megaflood could not have been later than the earliest settlement on Kovrizhka II (19.9 ka). However, erosive flood activity is observed at all stages, especially a shift in floods at the Pleistocene-Holocene boundary.
The efficacy of using luminescence dating on glacial deposits is tested for a portion of the Marine Isotope Stage (MIS) 6 Laurentide Ice Sheet margin in southwestern Indiana. We assess small-aliquot quartz optically stimulated luminescence (OSL) and feldspar infrared-stimulated luminescence (IRSL) dating of glaciofluvial, glaciodeltaic, and aeolian sediments against a well-established soil stratigraphy and a cosmogenic Be-10 depth profile. Results indicate that standard blue-light OSL regenerative protocols used on MIS 2 glacial sediments in the region warrant caution when duplicated for MIS 6 sediments. Quartz OSL ages underestimate age by up to 50% compared with cosmogenic and feldspar post-IR IRSL200 ages. Presence of unstable or hard-to-bleach OSL signal components that cannot be removed with modified preheat protocols yields unreliable data. While dates obtained using post-IR IRSL200 protocols on feldspar are affected by partial bleaching and anomalous fading, these factors can be accounted for. Discrimination of negligible-fading small-aliquot data allowed us to obtain post-IR IRSL200 ages between 103 +/- 12 and 241 +/- 28 ka. Post-IR IRSL200 ages are mostly consistent with Be-10 depth-profile dating and stratigraphic constraints and represent a viable option to study glaciofluvial sedimentation during MIS 6 and older glaciations in the region.
The prehistoric cultural material at the Lodoso Site, a Middle Archaic to Late Prehistoric campsite near Driscoll, Texas, is dominated by irregularly shaped heat-hardened fragments of earth called burned clay objects (BCOs). These artifacts are a common component of coastal plain archaeology in Texas and elsewhere in the state. Many such sites, especially those situated in sandy soils, do not preserve charcoal well, which renders dating the occupations challenging. The clayey matrix of the Lodoso Site does preserve charcoal, which presents an opportunity to assess the suitability of luminescence dating of burned clay objects by thermoluminescence (TL) and optically stimulated luminescence (OSL), and directly compare those results with charcoal collected from the same excavation provenience, as well as OSL dating of the sedimentary matrix. The TL and OSL results generally track together well but are consistently slightly older than charcoal collected from the same level. The charcoal dates suggest the midden formed between approximately 1500 and 2300 years ago, whereas the luminescence ages suggest formation occurred between 2000 and 3000 years ago. Where the luminescence ages occur out of stratigraphic order, so do the paired charcoal samples, indicating that this is a result of pedoturbation. The results of both radiocarbon and luminescence dating indicate that the midden is a diachronic feature resulting from use over a prolonged period rather than the product of a single burning event.
Mountain permafrost is a climatically sensitive but poorly constrained component of the terrestrial cryosphere. We use 108 U-Th dates from speleothems in two limestone caves in the Uinta Mountains of northern Utah, USA, to reconstruct a 600,000 year history of permafrost presence and absence in this alpine setting. Speleothem growth in both caves is confined almost entirely to discrete intervals that align with interglacial conditions of Marine Isotope Stages (MIS) 5e, 7e, 9e, 11c, and 13. In contrast, growth hiatuses correspond to glacial periods, when permafrost apparently inhibited infiltration of liquid water. Regional lapse rates indicate that cooling of similar to 3 degrees C to 5 degrees C relative to present would be sufficient to generate permafrost above the caves. The persistence of age clustering across multiple speleothems in two independent cave systems suggests that mountain permafrost repeatedly formed and degraded in concert with orbitally paced climate cycles. Comparison with nearby paleoclimate records confirms that these changes reflect regional-scale climate forcing. Our results provide rare empirical evidence for the extent of alpine permafrost over Quaternary timescales. As mountain permafrost degrades under modern warming, such records are essential for refining climate models, assessing geomorphic and hydrologic risks, and placing recent changes within the context of natural variability.
Climate oscillations may strongly modify continental precipitation and temperature patterns, therefore understanding their history is relevant for comprehending effects of past and ongoing climate changes. For this purpose, temperature and precipitation reconstructions beyond the instrumental record are extremely useful. As widespread terrestrial archives, loess-paleosol sequences are viable targets for such analyses. Consequently, cost-efficient geophysical proxies have gained increasing attention, but little is known about their capability to reflect even narrow climatic differences. Here we assess the sensitivity of rock-magnetic and photo-spectrometric properties of topsoil samples (n = 50) along uncorrelated, mean annual precipitation (MAP: 525 +/- 1 mm/yr to 584 +/- 1 mm/yr) and mean annual temperature (MAT: 10.8 +/- 0.1 degrees C to 11.2 +/- 0.1 degrees C) gradients across the Ba & ccaron;ka Loess Plateau (Serbia) and test a multivariate approach. Most proxies are sensitive to MAP <565 +/- 1 mm/yr, especially anhysteretic remanent magnetization (r(2) = 0.81). Applying a multivariate approach to hysteresis data reveals a robust relationship between precipitation (r(2) = 0.63), aridity (r(2) = 0.67) and physical properties over the entire MAP range. Although the approach needs to be further tested considering different climates, regression analyses, and timescales, our study indicates that multi-proxy approaches may increase the robustness with respect to single-proxy measurements for MAP and aridity reconstructions.
