Ensemble simulations with the Northern Hemisphere ice sheets model are performed for the last glacial cycle (last 123 kyr). The time dependence of the atmospheric forcing is reconstructed based on two sets of borehole data — EPICA (The European Project for Ice Coring in Antarctica; the data exhibit weak millennium-scale variabilitу) and NGRIP (North Greenland Ice Core Project; the data show strong temperature variability at the indicated timescale). The time dependence is combined with the long-term means from theERA5 reanalysis(for the present-day state) and with the IPSL-CM4-V1-MR model output for the Last Glacial Maximum. We found that the simulated trajectory in MIS3 (Marine Isotope Stage 3; ~60–30 kyr before present) is very sensitive to the relatively small perturbation of thus constructed paleoscenarios. This sensitivity is exhibited both for EPICA(after perturbing the data by the stochastic autoregressive processes with decorrelation timescale up to 2 kyr) and NGRIP data (after removing the millenium-scale variability from the data by applying the Savitzky– Golay filter). In particular, depending on the existence of such variability in the forcing dataset and provided that this variability is strong enough, an artificial interglacial is developed during MIS3. The physical reason for the development of this interglacial is the location of the system trajectory near the glacial termination threshold owing to the specific Earth orbit parameter configuration in the time interval indicated. We conclude that both the EPICA and NGRIP data can be used to force the Northern Hemisphere ice sheet model in the last glacial cycle. However, it is sensible to exclude the millennium-scale variability from the NGRIP data before applying them to paleoscenario reconstruction. In addition, high sensitivity of the model to the applied forcing during MIS3 might indicate the trajectory splitting in the vicinity of the attractor basin boundary in this time interval.

Marine Isotope Stage (MIS) 3, the period from 57 to 30 thousand years ago (ka), is critical to understand how glaciers and ice sheets advanced into the Last Glacial Maximum (LGM; 30–17 ka). Previous estimates of ice-volume changes and their equivalent global mean sea level (GMSL) during MIS 3 remain uncertain, particularly regarding the extent of Northern Hemisphere ice-sheet reduction and exact GMSL values during this period. Here we present a new GMSL record from late MIS 3 to the LGM (ca. 39–18 ka) inferred from paleo-water depth estimates of photosymbiotic large benthic foraminiferal assemblages in upper slope sediment cores from the Great Barrier Reef. The relative sea-level-based GMSL record reconstructed using a Bayesian statistical model, revealed long-term mean trends characterized by incrementally lowering sea levels divided into five periods. Late MIS 3 GMSL decreased from –101 m (average with a 95% confidence interval (CI) of –89 to –113 m) at ca. 39 ka to –119 m (average with a 95% CI of –114 to –124 m) at ca. 30 ka, then finally reached down to –137 m (average with a 95% CI of –131 to –142 m) during the LGM. A relatively steep sea-level fall into the LGM—approximately 17 m in magnitude—began at ca. 35 ka and took ca. 7000 years, implying an acceleration of large-scale ice sheet advance into the LGM.

Physical Review . Tipping elements are large-scale components of the Earth system that may undergo abrupt and potentially irreversible transitions once critical thresholds are crossed. How these elements interact and coevolve under anthropogenic forcing remains poorly understood. Here, we introduce a Bayesian framework to quantify interdependence among tipping element regions based on lagged correlations, yielding a probabilistic measure of their connectivity that rigorously accounts for observational and sampling uncertainty. Using the preindustrial period as a baseline, we show that connectivity patterns among tipping elements are substantially weakened under high anthropogenic forcing, whereas the loss of coherence remains comparatively limited under low-forcing scenarios. This decorrelation is particularly pronounced in the Northern Hemisphere. We further identify the Atlantic meridional overturning circulation as a key contributor to the observed loss of connectivity, in line with recent evidence that it may be approaching a tipping point. In addition, we propose a simple toy model where tipping elements are seen as an adaptive dynamical system, losing coherence as a reaction to external perturbations. The results are robust across past decades and across different climate models, with deviations from the baseline strengthening into the future, especially under high-emission pathways.

Few glacial chronological studies have been conducted in the extremely cold and arid interior of the Tibetan Plateau, limiting a plateau-wide understanding of Quaternary glacial spatiotemporal patterns. This study presents an initial late Quaternary glacial chronology for six moraine ridges on the southwestern slope of Zangser Kangri using cosmogenic 10 Be and single-grain K-feldspar optically stimulated luminescence (OSL) dating. Ten ¹⁰Be ages yield ages of 146.35 ± 9.11 ka and 21.40 ± 1.37 ka for the outer two moraines, while four OSL ages indicate that the inner four moraines were deposited at 17.57 ± 2.27 ka, 9.30 ± 1.29 ka, 3.67 ± 0.46 ka, and 0.62 ± 0.05 ka, broadly corresponding to Marine Isotope Stage 6 (MIS 6), the global Last Glacial Maximum (LGM), Heinrich Stadial 1, and Bond events 6, 3, and 0, respectively. Evidence of MIS 3 glacial advances was not identified, suggesting that palaeoglaciers at Zangser Kangri may have been absent or underwent very limited expansion during this relatively warm interval, with any resulting moraines subsequently obliterated by later, more extensive glacier advances. With reference to independent palaeoclimatic archives, we argue that the MIS 6 and global LGM advances were primarily driven by glacial-scale global cooling, whereas the HS1 and Holocene advances presumably responded to millennial-scale North Atlantic cooling. Moisture delivered by the upper-level westerlies, combined with locally recycled moisture, acted as an important co-control on glacier expansion at Zangser Kangri in the extremely cold and arid interior of the Tibetan Plateau.

Understanding the regional characteristics of Paleo-Indian summer monsoon rainfall across short and long timescales (from a few decades to several millennia) helps explain and predict the monsoon system's sensitivity and response to various driving mechanisms. Multiple proxies (palynology, loss on ignition (LOI), and clay mineralogy) from three geoarchives (FaHien-Lena, Horton Plains, and Kabella-Lena) located at elevations ranging from 130 to 2330 m asl (meters above sea level) in southwestern Sri Lanka provide information on changing patterns of rainfall during the late Quaternary, spanning the past 45.6 ka (or part of this period; kilo-years). During the period 45.6-36.0 ka, which corresponds to the middle of MIS 3 (Marine Isotope Stage 3), the climate was characterized by periods of rainfall strengthening on a multi-millennial scale and weakening on a century-to-multidecadal scale, as indicated by the high abundance of characteristic rainforest pollen taxa, Shorea spp, Hopea spp, Vateria spp, Dipterocarpaceae, Durio spp, Anacardiaceae, and Meliaceae, from FaHien-Lena. These Intertropical Convergence Zone (ITCZ)-driven rainfall changes were strongly influenced by obliquity- and precession-driven changes in solar insolation, glacial boundary conditions in the Himalayan region and the Northern Hemisphere, and sea surface temperatures (SSTs) in the equatorial Pacific Ocean. As indicated by the absence of multiple proxies, the weakened monsoon at > 24.3 ka corresponds to a short-period, Heinrich stadial 2 (H2). The remarkably low accumulation of Upper Montane Rainforest (UMRF) pollen taxa (eg. Elaeocarpus spp, Syzygium spp, Symplocos spp and Calophyllum cf. walkeri ) with few exceptions, combined with palynological hiatuses, extremely low LOI, and events with increased clay mineralogy, all derived from Horton Plains, indicates that the prevalence of millennial- and multi-century-scale rainfall failures during the period 24.3-18.4 ka, which is identified as the Last Glacial Maximum (LGM), a sub-interval of MIS 2, was punctuated by multi-century- to century-scale wetting events. At the end of the LGM, rainfall increased stepwise at 18.4-18.0, 18.00-16.2, and 16.2-15.9 ka, as shown by increased pollen accumulation and LOI. Rainfall was weak at 15.90-13.70 ka and 13.0-11.9 ka and strengthened at 13.7-13.0 ka. The second weakened rainfall event here broadly matches the globally significant cool/dry Younger Dryas (YD) event. These rainfall changes, observed in all three archives, appear to be associated with precessional and millennial- and multi-century-scale postglacial temperature changes in high-latitude and Tibetan Plateau ice core records, as well as in North Atlantic and tropical ocean core records. The increasing accumulation of UMRF proxies, combined with clay mineralogy and LOI, indicates that early Holocene rainfall peaked between 11.5 and 8.8 ka in response to solar insolation driven by precession and obliquity. Several decadal-to-century-scale wet/dry events during this period appear linked to El Niño–Southern Oscillation (ENSO) conditions (e.g., La Niña and El Niño). A gradual decrease in middle Holocene rainfall persisted until 5.2 ka; after that, it weakened rapidly, with a significant drop in rainfall between 4.8 and 3.4 ka. The globally significant 4.2 ka megadrought events align with this period. The reactivated late Holocene rainfall was highly variable, with repeated periods of strengthening and weakening over century-to multi-century scales. These rainfall patterns were linked to ENSO, with drying conditions associated with El Niño-like conditions advancing in the latter part of the late Holocene.

We investigated the lithology and tephra of the 87-metre-long drill core (WD1) from the Walsdorfer maar (Eifel, Germany) in order to reconstruct the sequence of volcanic activity and its integration into palaeoclimatic developments between Marine Isotope Stage (MIS) 7 and MIS 6. Based on tephra counts, mineralogical and geochemical analyses, we have identified 11 primary tephra layers in the WD1 core that have not previously been described in the Eifel, with an age range of approximately 196–158 ka, alongside 13 reworked tephra layers, particularly in the older part of the core. Leucite was detected in the important marker tephra Hüttenberg Tephra (HBT) at 215 ka and Glees Tephra (GLT) at 151 ka. The basal Walsdorfer Maar Tephra (WDT) dates the outbreak to 217,843 years ago. Our results thus extend the existing ELSA–23–Tephra–Stack for the period 220–130 ka. The results for organic components show that, during the period 215–185 ka (MIS 7c–7a), a predominantly boreal forest consisting mainly of thermophilic deciduous species was established, whilst in the range of 170–152 ka, during the earlier half of MIS 6, hardly any pollen and less organic material have been preserved. The data point to changes in productivity and vegetation dynamics, which are reflected in the composition of the organic matter. The liptinite-rich intervals with higher fluorescence primarily characterise warmer subphases within MIS 7, whereas cool-humid conditions during the glacial MIS 6 favoured the formation of huminite-rich peats. The presence of calcareous layers in the 200–190 ka range indicates formerly wet conditions, suggesting MIS 7a. Palaeotemperature calibrations were analysed by brGDGT and provide the temperature of MAF comparable to Holocene levels. The data from the WD1 core (Eifel) confirm the climatic conditions during MIS 7a as inferred from other European archives, add a new region to the picture, and, because of their high resolution, enable the identification of climatic fluctuations that could not be detected in previous studies. We reconstruct a tephrostratigraphic chronology as well as the palaeoclimatic evolution of the period 220–132 ka in the Walsdorfer maar.

Numerous sedimentary deposits are present along the coast of Brittany (W France). It is assumed that these sequences are related to glacioeustatic variations of the Late Pleistocene, but the scarcity of reliable ages prevent accurate chronological constraints to correlate their deposition with a clearly defined paleoclimate context. In this study, the analysis of Optically Stimulated Luminescence (OSL) signals from multi-grain quartz aliquots has been conducted for the calculation of a burial age of 21 sedimentary sequences along this coast. Considering the limitations of fast-OSL signal saturation at high doses, the palaeodose estimation and dose recovery tests were analysed using Bayesian models implemented in the R package BayLum and the results compared with those obtained using frequentist procedures and models such as the central age model (CAM) or the average dose model (ADM). The calculated ages in this work, ranging from the end 350,000 to 160,000 years, define a timeline for the formation of Upper Pleistocene sediments, providing further evidence for morphological changes along the coastline during the last glacial episode. The fossil raised beaches included in this record also provide further insight into the highstands reached during the Eemian transgression (ca. 125,000 years ago), within the penultimate interglacial episode.

Parts of the East Antarctic Ice Sheet rest on a bed below sea level, making them vulnerable to ice dynamic changes and instability, but few glaciers have been studied in detail. Here we report on glaciers draining into Porpoise Bay, Wilkes Land, overlying the Aurora Subglacial Basin. We find ice surface thinning, grounding line retreat and ice surface velocity increase over recent decades, consistent with warm ocean waters intrusion. This highlights this region's vulnerability to future warming.

Nature Geoscience, Published online: 30 September 2026; doi:10.1038/s41561-026-02102-1 The Antarctic Ice Sheet is committed to mass loss by 2100, with higher greenhouse gas emissions increasing its contribution to sea-level rise according to observation-constrained projections. Emissions cuts can limit, but not prevent, this loss.

Nature Reviews Earth & Environment, Published online: 30 September 2026; doi:10.1038/s43017-026-00827-6 The Wilkes Subglacial Basin is a marine-based sector of the East Antarctic Ice Sheet that is vulnerable to ocean-forced melting and climate warming. This Review assesses changes in ice-sheet extent in the Wilkes Subglacial Basin under past and projected warming and notes large uncertainties in temperature-related and climate-related thresholds for its major retreat.

Abstract Studies on the effects of long-term climate change on marine communities are crucial to our understanding of ecosystem resilience. The Late Quaternary fossil record, dominated by extant taxa, provides a valuable archive of past biotic responses to major climate fluctuations. While most studies on faunal turnover so far have focused on taxonomic identity, trait-based approaches are becoming increasingly common as they provide a stronger link to ecosystem functioning. We applied biological traits analysis based on a set of ecologically relevant traits to study functional turnover in nearshore molluscan assemblages from the Late Quaternary of the Po-Adriatic system (Italy). We found that the functional composition of fossil assemblages from the last two interglacials was statistically indistinguishable. In contrast, the assemblage from the last glacial was characterized by higher functional diversity, as indicated by dispersion, richness, and evenness metrics. Lower diversity in interglacial assemblages is related to the high dominance of the bivalve Lentidium mediterraneum , which strongly impacts the distribution of functional profiles. Functionally distinct species are more evenly distributed in the glacial assemblage, probably because of greater habitat variation coupled with individualistic responses of species to temperature changes. While functional similarity between the assemblages from the last interglacial and the Holocene indicates resilient response to natural climatic shifts, more extreme scenarios of future global warming will likely induce elevated rates of species turnover. Furthermore, additional anthropogenic pressures, including bottom trawling, eutrophication, and invasive species can also affect functional diversity to a higher degree. Thus, biotic responses to Late Quaternary climate change documented herein provide a valuable baseline of natural functional variability against which modern and future shifts in nearshore ecosystems of the Adriatic Sea driven by higher rates of warming and other human pressures can be assessed.

Holocene relative sea-level (RSL) reconstructions from far-field regions constrain past ice volume and global mean sea-level change, yet records from the South China Sea remain debated, particularly regarding the existence of a mid-Holocene highstand. Here, we present a revised RSL history for the Pearl River Delta based on six new radiocarbon-dated index points from an ancient Saccostrea cuccullata oyster reef, integrated into an updated, quality-controlled database of 292 data points. Ecological, petrographic, and geochronologic analyses confirm that the reef formed in situ within its habitat limits during a sustained RSL highstand. The combined dataset reveals rapid early Holocene RSL rise, reaching + 2.3 ± 0.5 m by 5.4 ka and declining to + 1.6 ± 0.5 m at 3.3 ka before reaching present level. Comparison with glacial isostatic adjustment models suggests delayed early Holocene meltwater input dominated by Northern Hemisphere sources, consistent with far-field records across Asia, Oceania, Africa, and the Americas. Our findings resolve longstanding debates over mid-Holocene RSL evolution in the South China Sea and provide robust constraints for Holocene sea-level budgets.

As the Arctic warms, the Greenland Ice Sheet is losing mass, with implications for global sea levels. We analyzed biomarkers from sediments beneath 1387 m of ice at Camp Century, Greenland, deposited when the ice sheet was smaller. Reconstructed summer temperature and precipitation isotopes from ~ 400 000 years ago show summers ~ 4.7 °C warmer than today, with more local moisture from reduced ice. This result highlights that both how warm and how long warming lasts matter to reduce ice loss.

Primary productivity (PP) of the ocean is a crucial factor controlling of the carbon cycle of the Earth. The Last Glacial Maximum (LGM) provides a valuable end-member climate state for evaluating its control. Numerical models, however, disagree on how PP in the North Pacific responded to LGM climate, limiting mechanistic interpretation of proxy records. We compare LGM and pre-industrial simulations from MIROC-ES2L and MITgcm-REcoM2 with proxy reconstructions, and use a high-resolution ROMS-CoSiNE simulation to diagnose interactions among iron supply, mixing, circulation, and macronutrient limitation. MIROC-ES2L simulates widespread LGM PP enhancement and a pronounced increase in subarctic surface nitrate, whereas MITgcm-REcoM2 simulates slightly lower subarctic PP and higher subtropical PP, in closer agreement with the reconstructed spatial contrast. The models also differ strongly in mixed-layer depth, background nitrate, and dissolved iron, indicating that their productivity responses cannot be explained by dust deposition alone. Seasonal diagnostics suggest that local vertical mixing controls nutrient seasonality, while contrasting background nutrient inventories and large-scale redistribution shape the intermodel differences. ROMS-CoSiNE further shows that enhanced iron availability can shift the subarctic Pacific from iron limitation toward nitrogen or phosphorus limitation. Productivity increases where physical processes replenish macronutrients, including the western subarctic Pacific and Kuroshio Extension, but decreases in parts of the northeastern Pacific where intensified nutrient consumption produces macronutrient depletion. These results reconcile the apparent paradox of greater glacial iron supply but lower productivity at some subarctic sites: the response depends on the coupled balance among iron fertilization, macronutrient resupply, and upper-ocean physics. Accurate glacial productivity simulations therefore require realistic initial nutrient fields, circulation, mixing, and ecosystem limitation, not dust forcing alone.

The oxygen isotope (δ 18 O ice ) record from a ∼310-meter core drilled on the Guliya ice cap (GIC) in 2015 (GP2015) demonstrates reproducibility with a ∼309-meter core drilled nearby in 1992. Construction of the timescale for a composite of the GIC cores used 10 Be and 36 Cl to locate the 41–thousand year Laschamp geomagnetic excursion and a match to a δ 18 O atm -dated core from the Guliya Summit. Comparison of the δ 18 O record from an Asian Monsoon speleothem with the GIC δ 18 O ice supports a timescale back to 128 thousand years, the oldest known so far from Tibet. The δ 18 O ice record shows orbital variations in phase with June insolation and polar ice core stable isotope records through the last glacial cycle, although the relationship is reversed since the deglaciation. Future investigation is required to resolve this observation. Radiocarbon-dated ice from the glacier margin supports the presence of glacial-stage ice in the GIC. The 36 Cl/ 10 Be ratios in GP2015 confirm that GIC ice is older than 100 thousand years.

arXiv:2609.36077v1 Announce Type: cross Abstract: The cryosphere regulates global cycles of water, energy, and carbon, affecting societies and ecosystems. By assimilating observations into models, we can infer the past trajectory of state variables, generating a cryospheric reanalysis. Existing cryospheric reanalyses typically treat each water year independently, with no pooling of information across years. With sparse and noisy observations, such 'climate-assumed' reanalyses perform poorly, reverting to an assumed yet uncalibrated background climatology. Conversely, Bayesian calibration with static parameters completely pools information across water years, inferring climatological parameters without capturing inter-annual variability. Here, we propose a new hierarchical Bayesian approach using partial pooling, which we coin 'climate-informed' cryospheric reanalysis. It jointly infers state trajectories, annual parameters, and local climatological hyperparameters using a nested hybrid particle smoothing workflow. We test the approach through three experiments by assimilating in situ snow water equivalent, satellite-based fractional snow-covered area, and glacier-wide mass balance data into a temperature index model at eight study areas, from the Swiss Alps to the High Arctic. Hierarchical reanalysis with partial pooling outperforms annually independent cryospheric reanalysis (no pooling) and static parameter calibration (complete pooling) for all experiments during calibration and for all but one during validation. Overall, it yields a 51% improvement in continuous ranked probability score relative to the prior, markedly exceeding that of annual reanalysis (38%) and static calibration (19%). Compared to a particle Markov chain Monte Carlo benchmark, our workflow reduces computational cost by several orders of magnitude through recycling and expectation maximization, with comparable or improved skill.

Abstract Seismic oceanography provides high-resolution imaging of thermohaline structure in the ocean interior, yet its application remains limited in the Southern Ocean, where direct oceanographic observations are sparse despite the region’s importance for ice-sheet stability and global overturning circulation. Here we integrate approximately 230 km of multichannel seismic reflection data acquired during the 2022 COLLAPSE expedition offshore George V Land (East Antarctica) with XBT/XCTD observations, vessel-mounted ADCP measurements, Argo profiles and the GLORYS ocean reanalysis to investigate the upper-ocean structure across the continental rise in East Antarctica. Seismic processing optimized for water-column imaging reveals laterally coherent reflections associated with thermohaline interfaces and delineates a sharp frontal transition corresponding to the southern boundary of the Antarctic Circumpolar Current (ACC). Quantitative comparison demonstrates that the reconstructed temperature field is consistent with the available oceanographic observations. The study identifies an energetically active frontal environment characterized by strong lateral thermohaline gradients, coherent undulating reflectors consistent with internal-wave-like deformation, and steeply inclined thermohaline interfaces associated with frontal strain and shear. This study provides the first integrated seismic and oceanographic characterization of the southern ACC boundary offshore George V Land and demonstrates how multichannel seismic reflection data can bridge the spatial limitations of conventional oceanographic sampling. These results establish seismic oceanography as an effective tool for investigating frontal structure and water-mass variability in data-sparse Antarctic regions and provide a new regional perspective on upper-ocean dynamics relevant to ocean heat transport toward the East Antarctic margin.

Snow accumulation on the Antarctic Ice Sheet could increase with greenhouse warming, storing water that would otherwise raise sea level. Employing multiple Earth system model simulations and reconstructions, we detect a warming-driven increase in accumulation consistent with greenhouse gas forcing, but observations can only be matched if wind and meltwater impacts on large-scale surface warming patterns are accounted for. This implies that predictions ignoring these effects are overconfident.

Subglacial topography controls how Antarctic ice flows towards the ocean and responds to climate change. In the Denman–Shackleton region, deep valleys connect the ice-sheet interior to the coast. Denman Glacier occupies one of the deepest and could contribute significantly to sea-level rise. Combining new ground-based gravity measurements with airborne data and statistical modelling reveals a more complex landscape than existing maps and a potential vulnerability to unstable retreat.

Preprint under review for ESSD (discussion: open, 0 comments) Mountain glacier history is hard to trace because field evidence is patchy and often erased by later ice advances. We used a fast computer model to simulate how glaciers grew and shrank over the last 130,000 years in nine mountain regions, and compared the results with mapped past and modern glaciers. By combining many plausible simulations, we map where glacier changes are reliable and where they are uncertain. The results provide a new baseline for studies of water, landscapes, and ecosystems.

Abstract The Miocene Climatic Optimum (MCO; 16.9–14.7 Ma) was a warm interval with atmospheric comparable to end‐of‐21st‐century projections under intermediate emission scenarios. Proxy records indicate the Oxygen Minimum Zone (OMZ) in the Eastern Tropical Pacific (ETP) was small or non‐existent during the MCO, expanding only when declined after 15 Ma. Conversely, the modern ETP OMZ is expanding under anthropogenic warming. The mechanisms behind these opposing trends remain unclear. Here, we use the Geophysical Fluid Dynamics Laboratory coupled model (GFDL CM2.1) with the offline biogeochemistry PCO2 model to simulate Miocene ocean oxygen across multiple levels and two paleogeographic configurations. Increasing produces widespread oxygen loss at shallow and intermediate depths. One configuration develops pervasive hypoxia throughout the water column in the Atlantic and Arctic, driven by stagnation and stratification associated with a collapsed Atlantic Meridional Overturning Circulation (AMOC). In the same configuration, the ETP OMZ contracts at higher due to weaker trade winds, reduced upwelling, and water mass transport changes through the open Central American Seaway (CAS) associated with the shutdown of North Atlantic Deep Water and Tethys Intermediate Water formation. The same circulation reorganization responsible for Atlantic deoxygenation thus contributes to ETP oxygenation, highlighting the spatial complexity of the marine oxygen response to warming. These results indicate Miocene ETP oxygen dynamics are not a perfect analog to future changes: the boundary conditions permitting ETP OMZ contraction (including an open CAS and collapsed AMOC) are absent from the modern ocean, limiting direct extrapolation from proxy evidence.

Climate change is driving the rapid retreat of glaciers across the sub-Antarctic zone. Using continuous lake sediment-based records of past glacier changes, the recent trends can be evaluated in a long-term context. Here, we analyse sediments from two glacier-fed lakes: Lake Athéna II and Lake Héra, located downstream the Chamonix Glacier on the sub-Antarctic Kerguelen Islands in the Indian Southern Ocean. Lake Héra is a threshold lake for the adjacent outlet glacier: the Agassiz Glacier, but has been solely receiving meltwater runoff from the Chamonix Glacier since the Agassiz Glacier retreated from its maximum extent. Based on distinct changes in grain-size and sediment density, we argue this transition occurred around 495 CE. After this, the glaciogenic sediments in both lakes are solely derived from the Chamonix Glacier. By tracking multi-decadal to centennial changes in incoming clastic sediments, we suggest the Chamonix Glacier retreated close to its current extent between 600 and 750 CE, primarily advanced between 750 and 1600 CE, before gradually retreating towards present-day. The glacier trends documented here are in line with mean annual air temperature and sea surface temperature records that indicate colder late Holocene temperatures, reflecting a negative Southern Annular Mode-like state in the Indian Southern Ocean. Conversely, phases of more positive Southern Annular Mode-like conditions, including gradually more positive anomalies after ∼1600 CE and until present-day, are in line with a decrease in activity at the Chamonix Glacier.

arXiv:2609.30420v1 Announce Type: new Abstract: Perturbed parameter ensembles (PPEs) reveal how physics parameters affect climate simulations, but interpreting parameter sensitivities across multivariate, spatially structured outputs remains challenging, particularly when calibrating models against observations. We develop an explainable contrastive learning model that maps 5 monthly cloud and radiation fields into a shared representation space. We train the model on the fields of two 100-member Community Atmosphere Model version 6 (CAM6) PPEs, spanning 34 parameters, that only differ in the warm rain microphysics scheme: KK2000, the default bulk microphysics scheme, and TAU-ML, a neural network emulator of a bin microphysics scheme. The learned representations separates two PPEs with over 94\% linear classification accuracy while preserving the seasonal variability and ensemble spread due to parameter perturbations. In the shared representation space, the representations of satellite observations occupy the same low-dimensional manifold as the PPEs but are displaced from them most strongly during boreal spring and autumn. TAU-ML PPE has a lower distance to observations compared to KK2000 in the representation space. Integrated Gradients attributions highlights the contributions in subtropical low-cloud regions, Northern and Southern Hemisphere storm track regions, and tropical convection regions to differences between PPEs and observations. Regional attributions correlate most strongly with parameters associated with cloud microphysics, boundary layer turbulence, and deep convection. These results demonstrate that explainable representations of climate fields can attribute model differences to specific variables, regions, seasons, and physical parameters.

Abstract Day-night compound hot days (CHDs) poses substantially greater risks to human health, ecosystems, and socioeconomic systems than that heat extremes occurring only during the day or night. Here, we investigate the spatiotemporal characteristics, physical mechanisms, and seasonal predictability of CHDs within the Yangtze River Basin (YRB), a recognized hotspot for CHDs in China. We show that CHDs are primarily associated with a local anomalous barotropic high-pressure/anticyclone, which is jointly influenced by a La Niña-like sea surface temperature pattern in the Pacific and a positive sea level pressure tendency over the North Atlantic through distinct atmospheric teleconnections. Based on these two physically interpretable ocean-atmosphere precursors, we develop a physics-based empirical model using data from 1980 to 2010. During the independent validation period of 2011–2022, the model exhibits substantial prediction skill, with a temporal correlation coefficient of 0.67. Nevertheless, it markedly underestimates the record-breaking CHDs in 2022. A comparison with 2013, when the second-highest CHDs was successfully predicted, revealed that land-atmosphere interactions, including the soil moisture-temperature positive feedback and the cloud-water vapor warming effects in the YRB, significantly amplified the CHDs in 2022. Because these land-atmosphere processes are not represented in the model, the exceptional severity of the 2022 CHDs is underestimated.

Strong winds can lift snow into the air, which influences the energy exchange between the surface and the atmosphere. We studied this phenomenon over the Greenland Ice Sheet using observations and a climate model. We found that blowing snow traps heat near the surface like a blanket and reflects incoming sunlight like a mirror. Simulating these processes increases the net energy at the surface and improves model accuracy. Therefore, we recommend including these effects in climate models.

In this work we summarize 15 years of research and developments of short-lived halogens (SLH) using the Community Earth System Model (CESM) and present a complete description of the implementation and capabilities achieved with the latest version v2. The CESM2-SLH release includes specific namelist options, input files and technical notes detailing the most important SLH updates that must be considered for the different model configurations and resolutions.

Journal of Advances in Modeling Earth Systems, Volume 18, Issue 10, October 2026.

Climate model based storylines of individual climate and weather events are increasingly used to quantify impacts, vulnerability, or stress-test infrastructure to inform adaptation decisions. Here, we present an approach to test the plausibility of unprecedented climate storylines based on physical conformity, internal consistency, and historical precedent. We apply this approach to assess the plausibility of month-long heatwaves in Western Europe that would exceed existing record temperatures by more than 5 K. These heatwaves are based on model simulations using ensemble boosting, a computationally efficient method to simulate unprecedented events. We compare these unprecedented heatwaves with historical heatwaves in a reanalysis data set, using standardised anomalies relative to a time-evolving climatology of relevant physical variables. We show that these unprecedented heatwaves are associated with well-known physical drivers that are similar to historical heatwaves, including strong anticyclones and low soil moisture. The anomalies of these drivers are more intense in magnitude, but similar in their spatial patterns and in their relationships with each other. However, we also reveal substantial discrepancies between historical and unprecedented heatwaves, which are related to atmospheric blocking, land-atmosphere interactions, and their temporal persistence. Despite these discrepancies, we conclude that these long-lasting, unprecedented heatwaves should not be ruled out as implausible and thus highlight the need to anticipate such events when planning adaptation measures. Similar approaches can be used to assess the plausibility of other types of extreme events.

Reconstructing past temperature variations is essential for predicting future trends under global warming. The recently developed biomarker, branched glycerol monoalkyl glycerol tetraethers (brGMGTs), has emerged as a promising proxy for continental paleotemperatures. However, lake water physicochemical conditions may influence brGMGT distributions, limiting their reliability as temperature indicators. Here, we analyzed brGMGT distributions in Sihailongwan Maar Lake (Lake SHL) in northeastern China. Results show that brGMGTs are abundant in downcore sediments and differ from those in catchment soils, indicating that they are primarily produced in situ. Further analyses demonstrate that, among the brGMGT-derived proxies, %brGMGTs exhibits a significant positive correlation with the mean lake water temperature reconstructed for Lake SHL since the last deglaciation (15.3 kyr BP), indicating that it effectively captures overall temperature variations. In contrast, the other proxies correlate significantly with precipitation, highlighting their susceptibility to precipitation influence in this lake. This divergence likely arises because brGMGT-producing bacteria may adjust their membrane composition to maintain stability in response to variations in ambient temperature—a signal effectively preserved by %brGMGTs—whereas the other proxies depend on weighted ratios among different brGMGT compounds, which can mask this temperature sensitivity when individual compounds are influenced by precipitation.

During the Pleistocene, the Alpine landscape and its surroundings were repeatedly reshaped by glacial cycles. Subglacial erosion carved out overdeepened basins into bedrock. Their sedimentary fillings are important archives for understanding the glacial impact on environmental transformation and the history of the Alpine region. Some of these infills are composed of multiple cycles representing separate glaciations; in typical cases, each is represented by a succession of lake deposits. While sediments of earlier glaciations are often eroded by subsequent glaciation cycles, these can be preserved in overdeepened troughs. The International Continental Scientific Drilling Program (ICDP) project “Drilling Overdeepened Alpine Valleys” (DOVE) investigates overdeepened structures on a larger scale by recovering sediment cores around the foreland of the European Alps. As part of the DOVE venture, we present two high-quality drill cores highlighting different basin generations from within the former Rhine Glacier domain of Upper Swabia in southern Germany. The first drilling ICDP 5068_6 near the village Gaisbeuren is located in the low-lying central part of the Rhine Glacier area. We subdivided the 144 m long ICDP 5068_6 core into four sedimentological lithotypes that, in seven stratigraphic units, build up the basin infill. The deposits reveal a glacially sheared bedrock horizon with overlying diamict layers, diamictic fines, and massive sands in alternating sequences. The sedimentary deposits show structureless sections attributed to melt-out, as well as presumed glacio-tectonic deformations. At the hamlet of Lichtenegg, a 76 m long core was recovered from a drill site on the Höchsten butte, targeting the correspondent small-scale Lichtenegg basin. Due to the basin's exceptional, elevated stratigraphic position underlying presumed Early Pleistocene (glacio-)fluvial gravel, it was included in the DOVE project as a bonus site. Based on petrographic and lithological features we divided the core contents into six sedimentological lithotypes and five lithostratigraphic units. The lowermost parts of the infill, which overlie the Molasse bedrock, consist of (sub-)glacial diamicts with glaciolacustrine to glaciodeltaic sediments followed by the (glacio-)fluvial gravels, with additional glacial diamicts on top. Therefore, both investigated cores differ strikingly from the above-mentioned, “typical” lacustrine-dominated sediment successions in overdeepened basins, which we attribute to a continuously ice-proximal setting during infilling. This study emphasizes the variability of overdeepening fills depending on the specific location within the basin and relative to the glacier front.

arXiv:2609.40140v1 Announce Type: new Abstract: Extreme heat is where urban adaptation needs kilometer-scale data the most, but the simulations training a downscaler can cost more than they save, and how much is needed has not been identified. We measured it with CASPER, a U-Net with a structure-preserving loss downscaling 32 km reanalysis to 1 km temperature, humidity and wind, across 24 configurations of one to eight months. Held-out error grows linearly with climatological distance to the training data, RMSE = 0.83 + 2.95 d, explaining 90% of its variance against 7% for volume and predicting unseen months in advance. On held-out extreme summer weeks CASPER preserves the fine-scale structure and cross-variable physics that matched-budget baselines degrade, and matches station observations during documented heat waves to within 1.8 K. Transfer to a new region degrades geographically; 11 days of local simulation cuts Vancouver's held-out error from 3.8 to 1.3 K. Training periods should span the target climate: the same accuracy for four times less simulation, putting kilometer-scale downscaling of extreme heat within reach of groups without large computing facilities.

Antarctic ice mass loss has intensified in recent decades. However, short-term glacier velocity variability in Victoria Land, East Antarctica, remains poorly documented. This study used Sentinel-1 SLC data and environmental records from 2017 to 2025 to quantify glacier-velocity variability in Victoria Land and examine its model-based statistical associations with selected atmospheric and oceanic variables using XGBoost–SHAP. The results show positive multi-year acceleration trends over the 2017–2025 observational period, superimposed on pronounced seasonal variability. The six selected glacier systems showed positive acceleration rates: Barnett Glacier (2.06 ± 0.92% a−1), Borchgrevink Glacier (2.39 ± 0.66% a−1), David Glacier–Drygalski Ice Tongue system (1.68 ± 0.23% a−1), Larsen Glacier (3.78 ± 0.69% a−1), McMahon Glacier (2.91 ± 0.73% a−1), and Oakley Glacier (2.54 ± 0.33% a−1). Spatially resolved velocity maps, centerline profiles, and terminus-region records revealed substantial differences in flow magnitude and variability among glaciers. Among the five environmental predictors examined, upper-ocean temperature had the largest overall model contribution, followed by near-surface air temperature, surface skin temperature, upper-ocean current speed, and runoff. This study provides a Sentinel-1-based record of recent glacier-velocity variability in Victoria Land and illustrates the utility of repeated, high-temporal-resolution SAR observations for characterizing glacier-flow variability in East Antarctica.

Reliable cryosphere information is limited in many mountain catchments of Central Asia. This paper develops a basin-scale evidence base for the Tupalang River Basin by combining MODIS snow observations, climate reanalysis, gridded water-balance products, a glacier inventory, and a 2025 Sentinel-2 snow and ice classification. Daily MODIS records for 2001 to 2025 were summarized as annual, winter, spring, and persistent-snow areas. Trend magnitude was estimated with the Sen slope, statistical significance was assessed with the Mann-Kendall test, and climate controls were examined through correlation analysis. Mean annual snow area was 1397.85 km2, whereas winter and spring means were 2182.48 and 1434.75 km2. The four snow indicators declined in slope terms, but none passed the 0.05 significance threshold. Annual mean air temperature increased by 0.041 degrees C per year and was the only significant monotonic climate trend reported in the study. Warmer conditions coincided with smaller snow areas in the annual, winter, and spring series. The GLIMS layer contained 618 mapped glacier polygons covering 78.61 km2. A Sentinel-2 classification for 2025 identified 526.35 km2 of snow and ice, of which 23.16 km2 intersected the mapped glacier layer. Tests with MODIS NDSI thresholds of 20, 30, and 40 produced different area totals but similar negative trend estimates. Comparisons with Landsat and Sentinel-2 showed that seasonal aggregation, image date, cloud screening, and spatial resolution materially affect area estimates. The results define a transparent monitoring baseline, while observed discharge and station measurements remain necessary for hydrological attribution.

The Southern Hemisphere Westerly Winds impact global climate and Antarctic ice sheet stability; however, there are few complete records over the past 12 000 years. We use a new mineral dust record from a South Pole ice core and identify a decrease in particle concentration and an increase in coarse particle percentage over the past ~ 11 000 years. Together with other records, our data suggests a southward shift in the winds starting ~ 6500 years ago related to warming in the Southern Hemisphere.

We present a global marine heatwave and cold spell dataset (1982–2024) based on satellite sea surface temperature. The dataset applies multiple definitions in parallel, varying baselines, thresholds, detrending, and event durations. It enables consistent comparisons of marine heatwave characterisation across methods and supports climate monitoring, and ecological impact studies.

Using Bennett Island (East Siberian Sea) as a case study, the role of shortwave radiation was examined in shaping the spatial distribution of the mass balance of Arctic ice caps. The aim of the study was to quantitatively assess the influence of spatial variability in insolation on glacier surface lowering. Changes in the surface elevation of four ice caps on Bennett Island, with a total area of approximately 56 km², over the period 2012–2022 were previously derived from multi-temporal ArcticDEM data. On average, the elevation change amounted to −6.27 ± 0.12 m, corresponding to a mass balance of −5.33 ± 0.39 m w. e. The mean incoming clear-sky shortwave radiation flux for a typical ablation season (76 frost-free days per year, based on ERA5-Land reanalysis data for 1994–2024) was modeled in SAGA GIS, taking into account both astronomical and terrain-related factors. The resulting distribution of values is characterized by spatial heterogeneity and direct dependencies on exposure and slope shading. The spatial variability of clear-sky insolation across the ice caps ranged from 360 to 520 W/m². For two of the ice caps, a statistically significant positive relationship between surface lowering and insolation was identified, with a linear trend of 0.013–0.021 m of additional melt per 1 W/m². Also, for two larger domes, a positive trend of ice melt was revealed, which equals 1.115–1.750 m per 100 m of altitude. It was shown that for smaller ice caps, the maximum differences in clear-sky insolation (160 W/m²) were equivalent to the effect of an elevation gradient of 100–200 m. For the largest ice cap, Toll, the influence of insolation could not be evaluated: due to horizontal ice motion, surface lowering derived using the geodetic method did not directly correspond to the actual melt. Therefore, further investigation required in situ mass balance measurements using the glaciological method.

Journal of Geophysical Research: Oceans, Volume 131, Issue 10, October 2026.

Selected multi-proxy data (NPP, palynofacies, Chironomidae, Cladocera, total organic carbon (TOC) and total inorganic carbon (TIC)) and a radiocarbon-supported chronological framework of the Holocene period in the Baćina Lakes, on the eastern Adriatic coast (southern Dalmatia), facilitated a comprehensive regional paleoenvironmental assessment. In this study, we analyzed an 8.4 m long lake sediment core BAJ-7 (extracted from the central part of Lake Crniševo). A high number of non-pollen palynomorphs (NPPs) were documented and results show that abundances of phytoplankton and zoobenthos NPPs seem to have comparable trends with other environmental factors such as organic matter, TOC, N and others over time. However, zoobenthos such as Chironomidae and Cladocera, as isolated NPPs, provide limited information on their own, and should be interpreted together with other NPPs and pollen data. The studied NPPs indicate that Lake Crniševo underwent multiple transitions from oligotrophic to eutrophic conditions. This research represents the first paleoecological study with zoological NPP-s as complementary bioindicators among others, within Croatian karst lakes, highlighting their strong potential for paleolimnological reconstructions.

arXiv:2610.03314v1 Announce Type: cross Abstract: Flow- and diffusion-based generative models have recently emerged as flexible and highly efficient forecasting models for dynamical systems. When combined with inference-time guidance, they offer a promising route to high-dimensional non-Gaussian data assimilation (DA), the problem of combining forecasts with observations to estimate latent system states. Existing filters, however, condition on a fixed history and assimilate only the most recent observation, leaving them unable to revise past states when new observations arrive. Estimates then stay tethered to a history that later observations may contradict, and errors accumulate over the assimilation run. To this end, we introduce **DAWIS**, a unified DA method covering filtering, fixed-lag smoothing, and block smoothing within a single framework. DAWIS replaces the single flow time of a state-level prior with a multitask stochastic interpolant over a window of consecutive states, assigning a separate flow time to each. An assimilation cycle inverts the window to a vector of per-state turning points and regenerates it under observation guidance, with the turning points controlling how strongly each state is held fixed, revised, or generated from scratch. The same construction can also absorb the forecast into the assimilation cycle, removing the need for a separate forecasting model. Experiments on challenging nonlinear systems show that DAWIS improves on both filtering and smoothing baselines under sparse, noisy, and nonlinear observations. The code for DAWIS is available at https://github.com/Erik-Wikingsson/DAWIS

arXiv:2609.30746v1 Announce Type: cross Abstract: Extreme events in chaotic systems are difficult to learn from short trajectories because they are controlled by transient finite-time instability rather than by frequently observed bulk dynamics. We propose a mechanism-aware conditioning plug-in framework that turns a nudged coarse ensemble into a non-intrusive sensor of local instability geometry. In the small-noise regime, the ensemble covariance aggregates the same finite-time deformation kernels that govern local instability, providing a Jacobian-free proxy for the local amplification structure around a synchronized coarse trajectory. A small FiLM module injects statistics of this ensemble geometry into an otherwise unchanged backbone while leaving the coarse simulator unchanged. We demonstrate this interface in two distinct pipelines: a Transformer-style residual-attention corrector for a controlled low-dimensional chaotic system and a probabilistic recurrent STORN corrector for topographic two-layer quasi-geostrophic (QG) flow. In the low-dimensional benchmark, ensemble covariance directions co-activate with OTD modes and FiLM conditioning improves 99th-percentile exceedance-frequency errors over an identical no-context Transformer baseline. In QG, a fixed ensemble-conditioned FiLM-STORN model trained on only \(50\) time units substantially improves long-horizon rare-event statistics in the data-limited regime, including density-tail errors, exceedance frequencies, and spatial exceedance-area distributions relative to an unconditioned STORN trained on the same data; on averaged high-threshold exceedance diagnostics, it also outperforms the baseline STORN trained with $20$ times more high-resolution data. These results show that local instability geometry is not merely interpretable post hoc, but an actionable conditioning signal for data-efficient rare-event emulation.

Journal of Geophysical Research: Oceans, Volume 131, Issue 10, October 2026.

The Arctic is warming rapidly, and winter rain-on-snow events are becoming more common. These events can alter snow layers and disturb climate information stored in glaciers. We studied how small rain-on-snow events affect snow structure and water isotopes near Ny-Ålesund, Svalbard. Field experiments show that meltwater flow creates complex features, but isotope changes remain confined to them, meaning seasonal climate signals can still be preserved where enough snow accumulates.

Journal of Geophysical Research: Earth Surface, Volume 131, Issue 10, October 2026.

Reconstructing Holocene surface-water hydrography in the North Atlantic is complicated by the different seasonal and ecological sensitivities of commonly used biological proxies. To address this, we introduce the Coccolithophore Hydrography Index (CHI), a new coccolith assemblage-based environmental proxy. Calibrated against modern coccolith species distribution, CHI condenses assemblage polarity into a single, front-sensitive signal of upper ocean hydrography, enabling continuous and structurally explicit reconstructions of Atlantic-Polar water alternations. Applied to a Holocene record from the Iceland Basin (core DA12-11/2-ST1-GC01), the CHI demonstrates strong ecological reliability and systematic coherence: elevated CHI values correspond to warm, saline, stratified water, while low CHI values mark cool, fresh, mixing regimes linked to East Greenland/Iceland Current influence. Short-lived fluctuations between 4.5 and 4.0 ka BP further show CHI’s potential for resolving transient stratification events. Combining the CHI with dinoflagellate cyst (dinocysts) and diatom records, a multi-dimensional view of Holocene hydrography, yields a coherent reconstruction of frontal dynamics in the Iceland-Irminger region.

In recent years, global warming and frequent extreme weather events have prompted attention to the impact of land cover changes on surface temperature, especially the role of forest cover changes. While changes in forest cover influence local mean and extreme land surface temperature (LST) through radiative and non-radiative pathways, the mechanisms vary by latitude and remain contentious in mid-latitude regions. Focusing on the mid-latitude Yangtze River Delta (YRD) urban agglomeration, this study employs a space-and-time analysis and a corrected surface energy balance model to quantify the impact of forest change on mean and extreme LST (ΔTsm). Here, ΔTsm is treated as a diagnostic indicator of the overall thermal response to forest conversion. At the same time, the Non-Radiative Forcing Index (NRFI) serves as a complementary decomposition indicator of the relative non-radiative contribution. We find that afforestation was generally associated with cooling of daytime mean LST (−0.21 ± 0.11 °C), whereas deforestation was associated with warming (0.17 ± 0.07 °C), with effects amplified in spring and summer. At night, these effects reversed slightly. For extreme-high LST, afforestation was typically associated with lower values (−0.13 ± 0.14 °C), except for cropland conversion, which was associated with higher summer values. Conversely, deforestation was associated with higher extreme-high LST (0.20 ± 0.09 °C). Non-radiative effects associated with energy redistribution contribute substantially to the modeled LST responses under the reference correction factor c = 25%, with an overall mean NRFI of approximately 82.16%. Sensitivity tests across c = 10–25% show that the principal annual response patterns and the predominance of non-radiative effects are robust, although the magnitude of the modeled response varies with c . These findings highlight the seasonal and transition-specific effects of forest-cover change on regional LST and provide a scientific basis for sustainable land management and climate adaptation in urban agglomerations and similar regions.

This study represents the first application of CorrDiff (Corrective Diffusion), a diffusion-based model, to achieve 3 km resolution statistical downscaling over China by integrating various variables at multiple levels. Through structural improvements to CorrDiff, our methods produce forecasts that outperform the China Meteorological Administration Mesoscale Model for almost all target variables. Our results show that for radar reflectivity prediction, CorrDiff yields more realistic outputs than regression models.

Nature, Published online: 29 September 2026; doi:10.1038/d41586-026-03035-w A heatwave in 2025 would have been one-third of a degree cooler without the greenhouse gases released since 2015.

arXiv:2609.35949v1 Announce Type: new Abstract: Extreme temperatures are the leading cause of climate-related mortality world-wide. Climate-health research and operational weather forecasting require accurate estimates of human thermal stress. The Universal Thermal Climate Index (UTCI) is among the most sophisticated and widely used feels-like temperature metrics. However, its ubiquitous polynomial approximation does not generalize well to extreme weather conditions. Here, we introduce Neural-UTCI, a neural network that calculates UTCI with substantially higher accuracy across global conditions at a lower computational cost for operational use. Neural-UTCI reduces the polynomial approximation RMSE from 2.78{\deg}C to 0.36 {\deg}C, an 87% improvement, and lowers thermal stress misclassification rates from 5.3% to 1.7%, with consistent performance across resampling experiments. These differences affect thermal exposure metrics. For example, during the 2003 European heatwave summer in Rome, Italy, the number of very strong heat stress days increases from 15 to 35 days when using Neural-UTCI compared to operational products like ERA5-HEAT. Simultaneously, Neural-UTCI reliably classifies extreme cold stress conditions, allowing continuous global application. By improving UTCI accuracy, Neural-UTCI can strengthen climate-health risk assessments and public weather warning systems, especially as global warming increases the incidence of extreme events.

This study investigates the dramatic global cooling during the Late Ordovician, 440 million years ago. Our simulations reveal a powerful feedback loop: as the ice sheet grew, it generated strong, cold winds flowing down its slopes. These winds further cooled the continents, causing the ice sheet to expand even more. The super ice sheet led to a substantial global temperature drop of about 1.5 °C. Interestingly, while continents froze, the oceans warmed.

The fauna of the Emperor Seamount Chain (ESC) is unique in composition and endemic to a certain degree. The isolation of the ESC allows considering them a model region for studying long-term climate changes in the world oceans. We studied planktonic foraminifera (PF) in the surface sediments of the Nintoku, Jingy, Ojin, and Koko guyots located in the southern the ESC (35°–42° N). This area is influenced by the warm North Pacific Current and the transition zone between sub-Arctic and tropical water masses. We identified twenty-three species of PF in the assemblages that include representatives of various climatic groups: polar, subpolar, transitional, subtropical, and tropical. Two discrete upwelling-influenced sites (Nintoku LV86-21 and Koko LV86-5) exhibit exceptionally high PF shell concentrations, indicating elevated biogenic sediment flux and enhanced carbonate preservation under high-productivity conditions. In contrast, station LV86-6 shows advanced carbonate dissolution features, including elevated fragment content and selective enrichment of dissolution-resistant species, confirmed by scanning electron microscopy. These findings demonstrate that sedimentary PF assemblages in the ESC are not simple climate proxies but rather taphonomically modified records shaped by the interplay of biogenic sediment input, hydrodynamic focusing, topography-induced upwelling, and depth-dependent carbonate dissolution.

ABSTRACT Hydroclimate in Portugal is influenced by the North Atlantic subtropical high‐pressure system (the Azores High) and by internal climate modes, most prominently the North Atlantic Oscillation, which may shift in position and/or intensity due to anthropogenic climate change. The development of high‐resolution paleoclimate records from sites in Portugal is critical for providing baseline climate estimates, including hydroclimate variability, for better constraining climate model predictions under future warming scenarios. Here, we present a near‐annually resolved composite stable carbon isotope (δ 13 C) and oxygen isotope (δ 18 O) time series from three stalagmites from two proximally located caves in the Algarve region of southern Portugal. U/Th dating indicates that our record spans the last millennium, with a short hiatus at ~1450 CE. During the last millennium, the stalagmite record shows coherence with regional sea surface temperature, Northern Hemisphere air temperature, and the North Atlantic Oscillation, demonstrating that the stalagmite proxies broadly reflect hydroclimate. Hydroclimate during the Medieval Climate Anomaly was relatively dry/warm and generally more variable than modern conditions and transitioned to an overall wetter/cooler early Little Ice Age with similar hydroclimate variability. From ~1650 to 1700 CE, conditions in southern Portugal became increasingly dry/warm, with hydroclimate remaining fairly stable from 1700 CE to the present.

SHELDA is sub-hourly European quality controlled sea level dataset with 257 records in NetCDF format, each representing a quality-controlled time series sampled at intervals between 1 and 15 min, along with residual derived by removing tidal components. It provides reliable, high-quality data essential for accurately analysing short-term sea level variations and extreme events, which are often overlooked due to limited resolution or insufficient quality control in existing datasets.

Central Asia is highly vulnerable to a wide range of natural hazards due to its complex mountainous terrain, glacierized catchments, and highly variable climate. Hazards such as glacial lake outburst floods (GLOFs), mudflows, snow avalanches, river floods, and droughts pose significant risks to communities, infrastructure, and water resources across the region. Many of these hazards originate in the high mountain systems of the Tien Shan and Pamir ranges, where climate change is accelerating glacier retreat, altering hydrological regimes, and increasing the likelihood of extreme events. Effective monitoring and management of these hazards require timely and spatially comprehensive information, which is often limited by sparse ground-based observations in remote mountain environments. Earth Observation (EO) technologies provide powerful tools for monitoring environmental conditions and supporting disaster risk management in such regions. Satellite remote sensing enables the detection of glacial lake expansion, monitoring of snow and ice dynamics, mapping of flood and mudflow events, and assessment of drought conditions over large areas. Radar and optical satellite data allow near-real-time hazard detection and provide critical information for early warning and emergency response. This short commentary article reviews the applications of EO technologies for monitoring major natural hazards in Central Asia and highlights their role in improving disaster preparedness and response. The study also discusses operational satellite platforms and emerging technologies that can support hazard assessment and risk reduction across the region.

Abstract This paper presents the results of a new radiocarbon dating program for the Upper Palaeolithic site of Badanj, Bosnia and Herzegovina. After putting together older dates originating from the 1970s and 1980s with recent ones acquired as part of new fieldwork, we then combine the available radiocarbon record with the present understanding of the stratigraphy to create two alternative Bayesian models. Both models are in general agreement and demonstrate that the site was in use from the Late Glacial until the end of the Pleistocene, which correspond in the local typo-chronological terminology to the Early and Late Epigravettian. Implications of this unique documentrary resource for our understanding of the Late Upper Palaeolithic in the Adriatic basin are briefly considered.

Glaciers of the Hindu Kush Himalaya (HKH) are critical to regional hydrology and climate systems, yet the rapidly expanding body of related research lacks a comprehensive bibliometric synthesis. This study systematically analyses glacier-related publications from 1980 to 2020 using the Scopus database with 24,086 articles examined through advanced bibliometric and network visualization techniques implemented in Python. Results indicate a pronounced and sustained growth in scientific output, with a compound annual growth rate of 13.98%, accelerating markedly after the mid-1990s in response to advances in remote sensing and geospatial technologies. The analysis reveals a strong dominance of Chinese institutions and authors, alongside key publication outlets such as Geomorphology and Earth and Planetary Science Letters. Keyword co-occurrence patterns identify “climate change,” “Tibetan Plateau,” and “remote sensing” as central research themes, reflecting the interdisciplinary evolution of HKH cryospheric studies. Collaboration networks further highlight clustered research communities and limited cross-group integration. By synthesizing four decades of scientific output, this study provides critical insights into research trajectories, dominant contributors, and emerging gaps, offering a robust foundation for future investigations and policy frameworks in high mountain cryosphere research.

arXiv:2610.03355v1 Announce Type: new Abstract: Climate oscillations are intrinsically out-of-equilibrium phenomena, yet the connection between their predictability and nonequilibrium thermodynamic properties remains unexplored. We develop a coarse-grained stochastic thermodynamic framework for Ornstein-Uhlenbeck (OU) processes and apply this framework to climate data via linear inverse models (LIMs), linking predictability, entropy production, dynamical activity, and path action. Climate oscillations are often studied in a 2d phase space of climate indices, which are even under time-reversal, motivating our focus on 2d OU dynamics with even variables. We find that coordinate-invariant properties of 2d OU processes can be reduced to a minimal three-parameter description: two diffusivities and a phase-space rotation. We show that these quantities fundamentally characterize distinct aspects of the dynamics. We find that higher irreversibility and dynamical activity do not necessarily imply reduced predictability. Rather, it is the path action which is most closely connected to a commonly used measure of predictability, the anomaly correlation coefficient (ACC). As an illustrative example, we apply this framework to the Madden-Julian Oscillation (MJO) over the twentieth century. We find that the MJO has become simultaneously more predictable, as seen previously, while at the same time becoming more irreversible and more active. Predictability is driven predominantly by the reduced mean diffusion. Entropy production reflects the combined evolution of all three parameters. These results suggest that stochastic thermodynamics provides a powerful lens for studying nonequilibrium climate phenomena.

arXiv:2609.30485v1 Announce Type: new Abstract: Mesoscale convective systems (MCSs) produce much of the extreme precipitation, floods, and landslides that affect northern South America (NOSA), yet remain difficult to anticipate because they are strongly governed by local processes. On the night of 31 March 2017, an MCS triggered a catastrophic flash flood in Mocoa, Colombia ---an event whose severity was not foreseen with sufficient lead time for emergency action, which typically requires several hours to a couple of days. Motivated by this disaster, this study investigates whether such events emerge within recurrent environments that provide a source of large-scale predictability, rather than as isolated and inherently unpredictable occurrences. We combine a 21-year (2001--2021) MCS tracking dataset, ERA5 reanalysis, equatorial Rossby (ER) wave diagnostics, and a Self-Organizing Map (SOM) to identify recurrent environments associated with MCS occurrence in the Orinoco--Amazon basin. The Mocoa environment maps onto one of nine recurrent configurations characterized by enhanced low-level moisture, neutral Orinoco low-level jet conditions, favorable upper-level circulation, and strong ER-wave activity. We further trace convective precursors to an upstream source region over the Guiana Highlands up to 24~h before arrival in the Mocoa region. Together, these results show that the Mocoa event was not climatologically exceptional but a high-impact expression of a recurrent, partly predictable large-scale regime, demonstrating that portion of MCS predictability arise from recurring large-scale conditions that can be monitored hours to days in advance, and support hazardous weather forecasting in NOSA and other tropical regions.

Journal of Geophysical Research: Solid Earth, Volume 131, Issue 10, October 2026.

Journal of Geophysical Research: Solid Earth, Volume 131, Issue 10, October 2026.

This study investigates how Quaternary colluvial deposits preserved within rock shelters record palaeoenvironmental changes and potential evidence of human occupation in Serra da Capivara, Northeast Brazil (NEB). It integrates morphostratigraphic and granulometric analyses, optically stimulated luminescence (OSL) chronology, phytolith assemblages and indices, and geochemical and mineralogical characterisation (EDXRF, XRD, FTIR, and UV–Vis spectrophotometry) across four sites. Morphostratigraphic and chronological evidence provided the strongest constraints on depositional architecture, sediment accumulation dynamics, and stratigraphic discontinuities, whereas geochemical and phytolith proxies offered complementary insights into environmental conditions and site-use patterns, while remaining more strongly influenced by lithology, taphonomy, and post-depositional processes. Results spanning approximately 37.5–6.0 ka, from the Late Pleistocene to the Middle Holocene, demonstrate that stratigraphic discontinuities and buried palaeosurfaces are associated with both coherent and contrasting responses among proxies, reflecting their differential sensitivity to environmental change, depositional stability, and post-depositional transformation throughout this interval. Phosphorus enrichment was observed in Quaternary deposits relative to the weathered bedrock, indicating allochthonous inputs and providing a useful chemical marker when assessed alongside stratigraphic and archaeological contexts. These findings suggest that the integration of morphostratigraphic, chronological, phytolith, and geochemical approaches strengthens the interpretation of sedimentary archives in rock shelters and provides an integrated basis for evaluating relationships between environmental dynamics and potential pulses of human activity.

Melting alpine snowfields develop cup-shaped hollows that reduce how much sunlight the snow reflects, accelerating melt. Over three seasons in the Swiss Alps, we monitored how these hollows form and grow with a laser scanner. Two processes reduce reflectivity at the same time: hollows trap light through internal reflections, and meltwater washes dark particles into them. Because they are difficult to separate, surface roughness alone emerges as a proxy for overall reflectivity loss.

Earth's Future, Volume 14, Issue 10, October 2026.

Accurate runoff simulation in glacier-fed alpine basins is essential for sustainable water resource management and water security in downstream arid regions, particularly under climate warming and accelerating glacier retreat. To address the limited representation of glacier processes in the Soil and Water Assessment Tool (SWAT) and the challenges of runoff modelling in data-scarce alpine basins, this study compared monthly runoff simulations from SWAT, SWAT_Glacier, Light Gradient Boosting Machine (LightGBM), and Extreme Gradient Boosting (XGBoost) in the Yurungkash and Karakash River Basins, two glacier-fed tributaries of the upper Hotan River Basin, Xinjiang, China. Pearson correlation analysis was used to screen runoff-related predictors, and SHapley Additive exPlanations (SHAP) was applied to quantify model-specific feature attribution in runoff simulations. SWAT_Glacier improved validation-period Nash–Sutcliffe efficiency (NSE) from 0.50–0.57 to 0.80–0.81 and modified Kling–Gupta efficiency (KGE) from 0.57–0.66 to 0.76–0.82 relative to SWAT, while reducing root mean square error (RMSE) by 32.4–37.0%. LightGBM and XGBoost achieved comparable testing performance, with NSE and R2 of 0.81–0.82 and KGE of 0.80–0.86. Across the two basins, the relative performance of SWAT_Glacier and the machine-learning models varied by evaluation metric and basin characteristics, rather than showing a consistent model ranking. Temperature-related variables showed the strongest statistical associations and model attributions in both basins. These findings demonstrate that model selection should consider glacier cover, process representation, and data availability to improve runoff simulation reliability and support sustainable water resources management in glacier-fed basins.

Science Advances, Volume 12, Issue 40, October 2026.

Science, Volume 394, Issue 6819, Page 87-90, October 2026.

Journal of Geophysical Research: Atmospheres, Volume 131, Issue 19, 16 October 2026.

arXiv:2610.02069v1 Announce Type: new Abstract: Rare weather regime transitions pose a challenge for data-driven modeling due to class imbalance. In this study, we develop a probabilistic deep learning emulator for a prototypical system with regime transitions, the stochastic Holton--Mass model of stratospheric variability, and analyze the structure of its learned latent space. The Holton--Mass model exhibits two metastable regimes, a strong and a weak polar vortex, maintained by nonlinear wave--mean flow interactions, with weak stochastic forcing intermittently triggering rare transitions between these regimes that qualitatively represent SSW events. We employ a ResNet-inspired Conditional Variational Autoencoder with six-layer encoder and decoder layers and explicit current-state conditioning to model the distribution of the system's state at the next time step (one day). The emulator accurately reproduces short-term dynamics, steady-state probability distributions, regime persistence statistics, rare transition rates, the transition committor function, and the transition expected lead time of the physical model. Beyond emulation fidelity, we interrogate the learned latent representation to understand how the model internalizes the underlying metastable structure of the dynamics. Principal Component Analysis of the 32-dimensional latent space reveals a clear and unsupervised separation into four physically interpretable clusters corresponding to strong versus weak vortex regimes and stable versus transition-prone configurations. Such emergent regime separation in latent space is hard to identify for deep generative models applied to high-dimensional stochastic systems. Our results show that carefully designed probabilistic emulators can uncover physically meaningful manifolds governing extreme-event dynamics, potentially aiding the development of improved operational advanced warning systems.

arXiv:2605.03997v2 Announce Type: replace-cross Abstract: Skill scores, which measure the relative improvement of a forecasting method over a benchmark via consistent scoring functions and proper scoring rules, are a standard tool in forecast evaluation, yet their sampling uncertainty is rarely rigorously quantified. With modern forecasting applications being increasingly multivariate and involving evaluations across multiple horizons, variables, spatial locations, and forecasting methods, standard tools like the pairwise Diebold-Mariano forecast accuracy test or pointwise confidence intervals fail to account for the multiple comparison problem, leading to inflated Type I error rates and invalid joint inference. To address the lack of a coherent, statistically rigorous framework for quantifying uncertainty across these multi-dimensional evaluation problems, we introduce simultaneous confidence bands for expected scores and skill scores. Our framework provides a versatile tool for joint inference that is applicable to any forecast type from mean and quantile to full distributional forecasts. We develop a bootstrap implementation and show that our bands are valid under multivariate extensions of the classical Diebold-Mariano assumptions. We demonstrate the practical utility of the approach in two case studies by quantifying the benefits of time-varying parameter models for macroeconomic forecasting, and by comparing data-driven and physics-based models in probabilistic weather forecasting.

This study investigates the distribution and sources of Polycyclic Aromatic Hydrocarbons (PAH) across the Eocene-Oligocene Transition (EOT) in the Caltavuturo Formation at Portella Colla (northern Sicily, Italy). The EOT represents a critical shift from greenhouse to icehouse conditions, and this research evaluates the utility of PAHs as geochemical proxies for paleoenvironmental and paleoclimatic changes. Nineteen clay samples were analyzed using Gas Chromatography–Mass Spectrometry (GC–MS) to determine the concentration of 14 PAH compounds and assess their origin through diagnostic. Results reveal a pronounced spike in total PAH concentrations at the EOT, primarily driven by high-molecular-weight pyrogenic compounds. This peak correlates with a dramatic reduction in deep-water agglutinated foraminifera (DWAF) diversity and the dominance of the opportunistic species Repmanina charoides. Diagnostic ratios indicate that while late Eocene samples reflect a background of biomass burning, the EOT interval is characterized by a significant increase in pyrogenic activity (wildfires), likely triggered by cooler, more seasonal, or more arid continental conditions. Subsequent Oligocene intervals show a transition toward mixed or petrogenic-dominated sources, alongside fluctuations in the biogenic marker perylene, which indicates shifts in the sediment redox boundary and organic matter flux. The integration of geochemical and micropaleontological data suggests that climate-driven changes in continental processes, such as enhanced erosion and wildfire frequency, directly influenced deep-marine environmental stability and benthic faunal ecology during the EOT.

Geophysical Research Letters, Volume 53, Issue 19, 16 October 2026.

Journal of Geophysical Research: Oceans, Volume 131, Issue 10, October 2026.

arXiv:2501.03135v3 Announce Type: replace-cross Abstract: Geodesic vortex detection is a tool in nonlinear dynamical systems to objectively identify transient vortices with flow-invariant boundaries that defy the typical deformation found in 2-d turbulence. Initially formulated for flows on the Euclidean plane with Cartesian coordinates, we have extended this technique to flows on 2-d Riemannian manifolds with arbitrary coordinates. This extension required the further formulation of the concept of objectivity on manifolds. Moreover, a recently proposed birth-and-death vortex framing algorithm, based on geodesic detection, has been adapted to address the limited temporal validity of 2-d motion in otherwise 3-d flows, like those found in the Earth's stratosphere. With these adaptations, we focused on the Lagrangian, i.e., kinematic, aspects of the austral stratospheric polar vortex during the exceptional sudden warming event of 2002, which resulted in the vortex splitting. This study involved applying geodesic vortex detection to isentropic winds from reanalysis data. We provide a detailed analysis of the vortex's life cycle, covering its birth, the splitting process, and its eventual death. In addition, we offer new kinematic insights into ozone depletion within the vortex.

This paper describes simulations of the present climate using the NASA Goddard Institute for Space Studies Earth System model. Major highlight is the capability to include complex interactions of the different components of the Earth System, including biology, and compare well with a wide variety of observations. We demonstrate that this model can be used to address present, past and future changes in the Earth system and their impacts and interactions with ecosystems.

The History of Anthropogenic Phosphorus Inputs (HaPi) provides a dataset of global human-driven phosphorus fluxes to terrestrial biosphere from 1860–2020 at a resolution of 5 arcmin. This comprehensive dataset consists of phosphorus fertilizer/manure application to cropland, manure application/deposition to pasture, manure deposition to rangeland, and atmospheric phosphorus deposition. It supports analyses of nutrient budgets and provides essential forcing data for Earth system models.

During the Late Paleozoic, subaerial exposures at the top of cyclothems were important markers of sea-level lowstands linked to glacioeustasy. Although subaerial exposure horizons carry valuable information on atmospheric conditions, they still tend to be underutilized in climate reconstructions. This study focuses on subaerial exposures in the isolated Pennsylvanian Valdorria carbonate platform situated in the Cantabrian Mountains, Spain. Previous studies have reported three major subaerial exposure horizons exhibiting distinct calcrete and/or paleokarst profiles, which were identified to have developed over a 2-My time interval. This research applies various analytical techniques, including scanning electron microscopy and X-ray diffraction, combined with major and trace element and 87Sr/86Sr isotope mass spectroscopy to further characterize the horizons. The results shed light on paleoclimatic conditions along the active margins of southeastern Laurentia–northern Gondwana at paleoequatorial latitude (ca. 5-15º south latitude) during the Late Paleozoic. As a result of this integrated approach, the oldest emersion surface displays a thick calcrete profile characterized by the presence of illite-smectite and lack of kaolinite, and a minor karstic system indicating a semi-arid climate. The second emersion surface exhibits a thin calcrete profile that displays a distinct red matrix with similar clay mineralogy to that of the older profile and lacks evidence of karstification, suggesting a shift to an arid climate. Finally, the youngest emersion surface at the platform-top exhibits an extensive karstic system with abundant kaolinite throughout, which indicates that the platform was exposed during a hot and wet climate prior to burial of the carbonate succession. Geochemistry of the subaerial exposure horizons combined with presence of diagnostic trace minerals indicate that siliciclastic sediment was derived from an intermediate volcanic arc source through aeolian transport. These results demonstrate that combining a variety of analytical techniques is a powerful tool for the study of paleosols beyond the investigated deposystem. The high-resolution record of climate variations, partly deviating from long-term trends in overall climate evolution, highlights the value of using an integrated methodology to better constrain changes documented in underexplored regions, such as the western Paleotethyan realm during the Late Paleozoic.

ABSTRACT The Fiorano quarry paleosol developed within the Colli Albani volcanic district (Italy) in the time interval bracketed by the deposition of the Pozzolanelle pyroclastic‐flow unit (365 ± 4 ka) and the emplacement of the Capo di Bove lava flow (277 ± 2 ka). The paleosol is characterized by an accretionary feature produced by the progressive accumulation of clastic material eroded from the surrounding volcanic landscape. The δ 13 C composition of soil organic matter (δ 13 C OM ) shows limited but systematic variations with stratigraphic age (from −20.77‰ to −27.29‰), reflecting changes in plants' contribution and organic matter input during pedogenesis. The δ 13 C OM record also matches global and local proxies over the 365–277 ka interval, spanning Marine Isotopic Stage 10 (MIS) to MIS 8.5. These data suggest that δ 13 C OM values were primarily controlled by variation in vegetation density and cover at the paleosol surface. The Fiorano paleosol thus preserves a well‐resolved continental stable isotope record for MIS 9 in the central Mediterranean area, showing the potential of δ 13 C OM variations in paleosols to constrain paleo‐environmental and paleo‐climatological conditions in the central Mediterranean area during the middle Pleistocene.

We installed six automatic weather stations from 3,810 m to 8,810 m on Mount Everest region to better understand weather conditions at high altitudes. These measurements reveal how temperature, humidity, and wind changes with height and season. Our results show differences between observations and widely used ERA5 datasets. This improved knowledge helps scientists better understand mountain weather and climate, climate change and supports water resource planning for communities downstream.

Seasonal snow is a critical water resource in Northern Kazakhstan, where spring snowmelt dominates surface runoff and agricultural water supply. However, traditional rainwater harvesting frameworks often neglect snow water equivalent (SWE), limiting their applicability in cold, snow-dominated steppe environments. This study develops an integrated remote sensing and GIS-based approach to identify potential sites for spring runoff harvesting. Using Sentinel-2 imagery, snow cover fraction (SCF) was derived and combined with a digital elevation model (DEM) to estimate snow depth through a quadratic regression model, validated against in situ data from 23 meteorological stations. Snow water equivalent was then calculated by incorporating field-measured snow density. Alongside SWE, six additional parameters—slope, drainage density, land use/land cover (LULC), and soil texture (sand, silt, clay)—were integrated within an Analytical Hierarchy Process (AHP) multi-criteria decision analysis. Results show that SWE, drainage density, and LULC collectively contributed 75% of the weight in site suitability, underscoring their dominant role in snowmelt-driven hydrology. The final suitability map indicates that 28.03% of the region falls into the “Very High” class, concentrated in central and northwestern districts. Developing just 30% of these potential zones could yield approximately 4.2 × 108 m3 of usable meltwater, supporting up to 105,000 ha under full irrigation or 280,000 ha under deficit irrigation. The study demonstrates that integrating high-resolution SWE with MCDA provides a scalable framework for enhancing agricultural resilience, mitigating floods, and strengthening water security in cold, semi-arid landscapes.

Geophysical Research Letters, Volume 53, Issue 18, 28 September 2026.

arXiv:2509.19384v2 Announce Type: replace-cross Abstract: Reconstructing high-resolution regional significant wave height (SWH) fields from sparse buoy observations is a critical challenge for ocean monitoring. We introduce AUWave, a hybrid deep learning framework that fuses a station-wise encoder with a multi-scale U-Net enhanced by self-attention to recover regional SWH fields. Trained and validated using NDBC buoy observations and ERA5 reanalysis over the Hawaii region, AUWave achieves high accuracy. It consistently outperforms a representative baseline, especially in configurations with more than a single buoy, demonstrating the benefit of its multi-scale architecture. Spatial error analysis shows performance is highest near observation sites, as expected. Further, buoy ablation studies identify critical anchor stations whose removal disproportionately degrades performance, offering actionable guidance for observational network design. AUWave provides a scalable pathway for gap-filling, creating high-resolution priors for data assimilation, and contingency reconstruction. Cross-basin evaluations in the Atlantic and Pacific confirm the model robustness and portability, highlighting its potential for operational use across diverse ocean regimes.

Geophysical Research Letters, Volume 53, Issue 19, 16 October 2026.

Journal of Geophysical Research: Oceans, Volume 131, Issue 10, October 2026.

arXiv:2512.18289v3 Announce Type: replace Abstract: In three-dimensional variational data assimilation (3DVar) for numerical weather prediction (NWP), the observation operator $\mathcal{H}$ plays a central role by mapping model state variables to an observation equivalent. For weather radar, however, specifying $\mathcal{H}$ is particularly challenging: reflectivity is a nonlinear, microphysics-dependent diagnostic quantity that only indirectly relates to the model's prognostic variables, making traditional parameterised radar operators complex, regime-dependent and difficult to tune. In this study, we propose a neural-network (NN)-based observation operator for radar reflectivity and apply it within a 3DVar framework. Using five years (2019-2023) of radar reflectivity data from the Lisca radar and 4.4 km-resolution short-range forecasts from ALADIN model over Slovenia, we train a convolutional encoder-decoder neural network to map model temperature, humidity, horizontal wind components and surface pressure fields to radar reflectivity. Across independent test cases spanning clear-sky, stratiform, and convective regimes, the NN-based operator accurately reproduces the spatial structure and intensity of observed reflectivity, relying primarily on the model state near the observation point. In the extreme precipitation case, which caused widespread floods in Slovenia on August 4, 2023, assimilating the full radar disc reduces the domain-averaged reflectivity root-mean-square error from 5.99 dBZ to 3.47 dBZ and improves the alignment between the analysed and observed convective bands. Embedded within 3DVar, the Jacobian of the NN observation operator allows radar reflectivity observations to inform model state variables, producing corresponding analysis increments. The proposed NN radar observation operator offers a flexible alternative to traditional parameterised radar operators for improving convective-storm forecasts.

arXiv:2606.19302v2 Announce Type: replace Abstract: As deep learning for physical systems continues to grow in popularity, efforts to improve generalizability have primarily focused on designing architectures that embed physical constraints. However, for machine-learning surrogate climate models (emulators), we show that the low structural diversity in existing scenarios commonly used to generate training data places a ceiling on predictive skill. Here, we examine whether training datasets themselves can be optimized to improve generalization. We introduce a method to create datasets that produce emulators capable of generalizing to new, structurally different scenarios absent from the training data. We use a differentiable Simple Climate Model (SCM) to calculate the sensitivity of emulator loss to perturbations in the training data, iteratively updating the training data to maximize emulator skill. For an SCM, training on one scenario optimized in this fashion outperforms an emulator trained on six standard ScenarioMIP pathways. We achieve this higher predictive skill despite training on a smaller dataset, finding that our emulator successfully isolates distinct physical behaviors of different climate forcing agents (e.g., greenhouse gases vs. aerosols) without single-forcing runs. We then demonstrate that scenarios optimized using an SCM, when used to drive an intermediate-complexity climate model, produce a training dataset that yields a more skillful emulator than training on ScenarioMIP outputs. Our results suggest that, in the compute-constrained environment of running full-scale climate models, generating a small number of dynamically rich scenarios provides greater marginal value for emulation and characterizing system responses than expanding the suite of traditional emissions pathways.

Geophysical Research Letters, Volume 53, Issue 19, 16 October 2026.

by Claudia R. Binder, Alejandro Gomez, Jair Campfens, Maria Anna Hecher, Simón Ladino Cano, Léonard Léchot, Hanbit Lee, Simon Montfort The world is on track to exceed 1.5 °C of warming. To avoid further temperature increase, accelerated action to decarbonize our societies is needed. Scholars claim that this transition requires surpassing social tipping points (STPs). STPs are defined as non-linear processes of transformative change in social systems, where a small perturbation triggers a response that generates self-amplifying changes in behaviors. While research efforts have focused on analyzing triggers to tipping, we have little understanding of the structures and mechanisms that stabilize a new state after the tipping phase. We argue that post-tipping trajectories rarely converge to a single stable state. Instead, they converge to multiple coexisting (including reversed) states that can exist as spatial patterns and need to be designed-in for stability. Combining insights from ecological systems, where phenomena that occur after tipping have been researched, with empirical evidence on social tipping in transitions towards net zero, we shed light on states after tipping and on how they may be stabilized. Four phenomena shape post-tipping trajectories: (i) lock-in through co-adoption of technologies or feedback loops; (ii) the co-existence of states; (iii) the reversibility of sustainable behaviors; and (iv) infrastructure as a stabilizer of co-existing or new states. We found that a systemic approach that combines technological and social levers to create positive feedback is needed to stabilize a future state. We propose three policy design principles to achieve this: (1) prioritize infrastructure that creates positive lock-ins to new technologies; (2) combine deep (social norms) and shallow (co-adoption of technologies) levers; and (3) target network brokers to bridge adopter and non-adopter groups and establish diffusion structures.

In Argentina, and specifically in the Province of San Juan, glaciers and periglacial landforms are protected under National Law No. 26,639 and Provincial Law No. 1,076-L, respectively. Both laws, in force since 2010, establish the protection of glaciers and periglacial environment landforms, recognizing them as strategic water reserves in solid state. At the local level, the provincial government is responsible for assembling and keeping the provincial inventory up to date as a management tool. This study presents the results of the first update of the Provincial Glacier Inventory for the San Juan River Hydrological System for the period 2015-2020. In this update, 685 glacial landforms (clean-ice glaciers, snow patches-glacierets, debris-covered glaciers, and debris-covered glaciers with rock glaciers) and 3,227 periglacial landforms (mostly rock glaciers) were identified, which collectively cover an area of 471.58±48.22 km² and represent 1.24% of the total area of the San Juan River Hydrological System. The estimated water equivalent volumes were 6.204±0.29 km³ (stored in glacial landforms) and 2.177±0.42 km³ (stored in periglacial landforms). By comparing these results with those obtained in 2017, it is observed that the area occupied by clean-ice glaciers decreased 18.74%, and the area covered by snow patches-glacierets decreased 65.15%. For the assessment of inventory quality, guidelines established by ISO Standard 19,157 were applied, with the omission, commission, and uncertainty indices determined during the mapping and digitization processes. This work primarily contributes to compliance with the provincial regulatory framework, enhancing our understanding of the current state of the cryosphere and its evolution in recent years.

The Lautaro volcano is the northernmost and highest volcanic center of the Austral Volcanic Zone (AVZ) of the Andes. It is in the northwestern part of the plateau of the Southern Patagonian Ice Field along the northern continuity of the Pío XI chain. The volcano is almost entirely ice-covered, nourishes numerous large outlet glaciers, and has been the object of only very limited field-based geological studies. It represents a key location within the Southern Patagonian Ice Field due to its potential influence on the advancement trend cycles of the Pío XI glacier. The present contribution uses the Red Relief Image Map technique to identify volcanic landforms and structures in a glacial context, providing the first quantitative morphological assessment of this volcano. The results show plateau-like lava flows at the base of the volcano, possibly concentrically emplaced from a single emission center during the first stages of volcano construction. Subsequent stages comprise two main edifices formed by overlapped lavas and coulées on top, with multiple emission centers aligned along a NE-SW volcanic structure. The early evolution of the volcano seems to have been influenced by the regional paleotopography. The approximate pre-eruptive surface is relatively flat, with a depression in the northern and eastern sectors of the volcano and a positive relief towards the southern and western parts. Furthermore, a review of the historical activity and tephra records from lakes in Argentina was performed to reassess the eruptive activity of Lautaro. This review reveals a cycle of intermittent historical volcanic activity, with at least three major eruptions during the 1870s and 1880s, and classifies Lautaro as the most active volcanic center in the AVZ, since the late 19th century. These antecedents advance our knowledge about the interactions between volcanoes and glaciers in southern Patagonia, providing new perspectives on the evolution of this remote, ice-capped volcano.

في الدراسة الحالية، تم اخذ عينات سطحية وتحت سطحية من عمق (50) سم باستخدام أداة الجرافة إلى عمق (21) مترًا بواسطة الغواص لدراسة الاختلافات في مجموعة الأوستراكودا والفورامينيفرا من (15) عينة، (5) منها ضحلة و(10) تحت سطحية على طول الخليج العربي في الفاو (كاسر الأمواج) في مدينة البصرة جنوب العراق لدراسة اختلافات الرواسب السطحية وتحت السطحية في مجموعة الأوستراكودا والفورامينيفيرا. تم دراسة (15) عينة تحت المجهر الثنائي. تم تحديد اربعة وعشرون نوعًا من تصنيفات الأوستراكودا تنتمي إلى (21) جنسًا وأيضًا واحد وخمسون نوعًا من انواع الفورامينيفيرا تنتمي إلى (19) جنسًا. توصلت الدراسة الحديثة الى ان هناك الأوستراكودا مع العمق عن وجود وفير من Cypreidis و Tanella و Actinocythereis و Neocytheromorpha كذلك ازدياد في الأنواع العائدة للأجناس التالية من الفورامنفرا مع ازدياد العمق Ammonia و Elphidium و Quingloculina و Spiroloculina في جميع العينات وهي السائدة بينما تتوفر بقية الأجناس والأنواع بنسب متفاوتة في جميع العينات. تشير الأجناس السائدة في العينة السطحية إلى بيئات المد والجزر والمسطحات الطينية البحرية ذات الملوحة المنخفضة وحالة الطين المسطح بينما تشير باقي الأجناس الأخرى في العينات بأكملها إلى البيئة البحرية وظروف الملوحة العالية إضافة إلى تأثير التغاير في مستوى سطح البحر. تم أخذ عينات من الرواسب السطحية وتحت السطحية من الجرف البحري والمنحدر العلوي لفهم تجمعات الأوستراكودا والفورانيفيرا الحديثة وتصنيفها. كما تم إجراء وقياس تحليل حجم الحبيبات والمادة العضوية.

Coffee production and trade are globally important, and Brazil is the world’s largest producer and second-largest consumer. However, climate change poses a major challenge to coffee production because the crop is highly sensitive to irregular rainfall distribution and temperature fluctuations. We evaluated temperature and precipitation fluctuations from 1984 to 2023 at two coffee-producing units in southern Minas Gerais, Brazil, to characterize regional climate dynamics and their associations with coffee yield and total production. We used the Mann–Kendall methodological approach, linear regression and Sen’s slope estimator to identify trends and estimate rates of change in maximum, minimum, and mean air temperature and in total, dry-season, and wet-season precipitation. We used Pearson’s and Spearman’s correlation coefficients to assess relationships between climatic and agronomic variables, selecting the appropriate coefficient based on prior Shapiro–Wilk normality tests. Most temperature variables showed warming trends in September, during the transition from the dry to the wet season. Minimum temperature showed more warming trends in the first months of wet season. Cooling trends were also detected in January, February, March, April, May, and December. Precipitation declined in May and increased in June but showed no trend in either the dry or wet season. Relationships between climatic and agronomic variables differed between production units: correlations were positive at one unit and negative at the other. Locally, these trends indicate vulnerability during critical phenological stages that coincide with these seasonal windows, resulting in pollen tube desiccation, greater consumption of photo assimilates during physiological rest, and poor bean development under reduced dry-season rainfall. The contrasting correlation patterns were attributed to local climate fluctuations and to the microclimatic, land-use, and land-cover characteristics of each production unit. Under this scenario, mitigating these adverse conditions is essential to minimize damage to coffee crops. Recommended strategies include the implementation of agroforestry systems, rigorous monitoring of pests favored by climate fluctuations, and the adoption of more climate-resilient cultivars. Ultimately, this study highlights that crop management and mitigation strategies designed to buffer these climate variations must be tailored to the distinct conditions of each production unit.

Journal of Geophysical Research: Oceans, Volume 131, Issue 10, October 2026.

arXiv:2610.01095v1 Announce Type: new Abstract: An important and unresolved problem in the physical sciences is explaining the predictions made by neural networks. Several explainable artificial intelligence (XAI) methods have been proposed to address this problem, including gradient XAI, Integrated Gradients, and GradientSHAP. We evaluate the baseline XAI methods according to four scores: sensitivity (XAI patterns strongly affect predictions), attribution (XAI patterns reproduce the change in prediction relative to a baseline), robustness (XAI patterns remain stable for nearby inputs), and coherence (XAI patterns are spatially smooth). We also introduce a new method, average gradient outer product (AGOP) XAI, that uses global gradient information to identify an important direction for a specific input. We apply XAI to neural network predictions of the El Ni\~no-Southern Oscillation (ENSO) based on data from the Zebiak-Cane model. AGOP XAI achieves the highest attribution, robustness, and coherence scores in the architecture and lead-time comparisons reported here. Its sensitivity is surpassed by gradient XAI, which is maximally sensitive by definition. Beyond diagnosing neural-network behavior, AGOP XAI can generate candidate hypotheses about physical mechanisms. The method highlights an equatorial thermocline-depth signal consistent with recharge oscillator physics, together with a southeastern-Pacific lobe that may be specific to the Zebiak-Cane model. Finally, we test the physical relevance of AGOP using optimized perturbations that move the Zebiak-Cane model along AGOP explanation coordinates. Such perturbations can suppress the selected extreme events or, from a near-neutral ensemble, generate strong El Ni\~no or La Ni\~na events 10 months later.

arXiv:2610.01677v1 Announce Type: cross Abstract: In this paper we explore and extend the state of the art for understanding the skill and utility of early warnings of critical transitions (tipping points) in forced nonlinear systems. We highlight some of the challenges and opportunities of transforming estimates of a dynamics-based early warning system, based on trends in stability or resilience into forecasts with high skill. We highlight the importance of (a) quantifying consequences of possible actions in response to warnings that may be false negatives or positives (b) considering finite time horizon predictions to give verifiable predictions (c) assumptions necessary for valid extrapolations of trends. We evaluate approaches to improving the skill and utility of the forecast.

High Mountain Asia is undergoing rapid warming, with consequences for snow, ground thermal conditions, runoff, sediment availability, and water-related mass-movement hazards. This study combines watershed-scale hydroclimatic and cryospheric diagnostics with an event-scale reconstruction of the 17 July 2018 Badswat debris flow–river blockage in Gilgit–Baltistan, Pakistan. Snow-cover area (MODIS MOD10A1; 2000–2025), surface frost-number terrain (MODIS MOD11A2; 2001–2025), air temperature (ERA5-Land; 1950–2025), and precipitation (CHIRPS; 1981–2025) were analyzed in Google Earth Engine. The event was reconstructed in RAMMS-DF with Voellmy–Salm rheology (μ = 0.09, ξ = 500 m s−2) on a 12.5 m grid. Air temperature increased by 0.13 °C decade−1 (p < 0.001). Snow-cover and permafrost-favorable extents were negatively associated with temperature (r = −0.74 and −0.78), but neither showed a significant temporal trend over the shorter satellite record. CHIRPS precipitation increased by 1.7 mm yr−1 (p < 0.001), mainly in summer, although its annual variability agreed poorly with ERA5-Land and the trend is therefore product-specific. The calibrated scenario reproduced the mapped runout path, river-blockage location, and depositional zone, and yielded maxima of 15.17 m flow height, 33.82 m s−1 velocity, 2544 kPa impact pressure, and 81 kPa basal shear stress, with about 352,540 m3 of deposition. Because dynamic outputs lack independent measurements and joint parameter uncertainty was not quantified, they are scenario estimates rather than design values. The study provides an uncertainty-explicit event reconstruction and separates long-term environmental context from the unresolved event-scale trigger.

Grass (Poaceae) pollen is largely overlooked in investigations of grassland evolution because the pollen of most species cannot be differentiated using traditional optical microscopy. However, the combination of superresolution microscopy and deep learning enables the capture and quantification of distinct morphological variation across the pollen of grass species. Using a semisupervised deep-learning strategy, we trained convolutional neural networks (CNNs) on superresolution images of known modern and unlabeled fossil grass pollen. We used learned CNN features to estimate taxonomic diversity and discriminate between C 3 and C 4 grass species—applying Shannon entropy as a measure of morphological variability and a gradient-boosted decision tree as a classifier of photosynthetic type. We validated the method on modern grass pollen assemblages and then applied our trained models to fossil assemblages from a 25,000-y lake-sediment record from Mt. Kenya, correlating shifts in grass diversity through time with changes in atmospheric CO 2 concentration and proxy records of local temperature, precipitation, and fire occurrence. Our data show that the species diversity of these montane grasslands declined substantially between 21,500 and 16,000 years ago, coincident with the most severe regional cooling of the last ice age, and that the fraction of C 4 grasses gradually decreased after the late-glacial–Holocene transition, associated with higher temperatures and decreasing fire activity. Our results demonstrate that CNN features of pollen morphology can advance palynological analysis and enable robust estimation of changes in grass diversity and C 4 grass abundance in ancient grassland ecosystems.

Abstract Climate change poses a significant threat to agricultural productivity and food security in sub-Saharan Africa (SSA), where many people rely heavily on rain-fed agriculture for their livelihoods. Rising temperatures, erratic rainfall and more frequent extreme weather events have caused substantial fluctuations in crop yields. However, the extent to which socio-economic, institutional and adaptive-capacity factors moderate these climate effects across different agro-ecological settings remains insufficiently understood. Using panel data from the World Development Indicators (WDI), FAOSTAT and the Climate Research Unit (CRU) datasets for 2000–2022, this study applied panel regression and dynamic GMM estimators to examine the effects of temperature and precipitation on agricultural outcomes and to assess how resilience mechanisms shape productivity, agricultural trade and food price stability. The results show that the Climate Resilience Productivity Index (CrPI) is positively associated with agricultural trade, with coefficients ranging from 0.23 under fixed effects to 0.87 under pooled OLS and random effects models, while higher CrPI values are associated with lower food prices in the random and fixed effects specifications. Temperature increases generally constrain agricultural outcomes, whereas resilience-enhancing mechanisms help stabilize markets against climate-induced volatility. These findings are consistent with the Land Maxing framework, which shows that closing yield gaps through the integration of indigenous food trees, agroforestry and community-driven value chains can address food security, non-timber forest product (NTFP) provision and rural livelihoods simultaneously. Evidence-based policies targeting yield gap closure and the strengthening of community food systems represent a pathway towards sustainable agricultural resilience in Sub-Saharan Africa.

Tajikistan and the Kyrgyz Republic are among the most hydropower-dependent countries in Central Asia, with electricity systems strongly influenced by mountain snow and glacier-fed river flows. Hydropower provides important benefits for energy security, economic development, and regional electricity trade, but this dependence also creates vulnerability to changes in water availability and seasonal hydrological conditions. This review examines the energy systems of Tajikistan and the Kyrgyz Republic, focusing on hydropower development, energy-sector challenges, climate-related changes in glaciers and hydrology, renewable-energy diversification, and transboundary water–energy management. The review synthesizes evidence from scientific literature, international organizations, government sources, and other publicly available materials. The findings indicate that rising temperatures, glacier retreat, changes in snow and melt dynamics, and increasing hydrological variability may increase uncertainty in river flows and reservoir inflows, creating additional challenges for hydropower planning and long-term energy security. Although hydropower is expected to remain the principal source of electricity in both countries, complementary development of solar, wind, geothermal, and biomass resources could reduce the concentration of hydrological risk. The review also highlights the importance of climate-informed energy planning, infrastructure modernization, improved hydrological and climate information, and stronger transboundary water–energy coordination. Overall, long-term energy resilience in Tajikistan and the Kyrgyz Republic will depend on integrating hydropower development with renewable-energy diversification, climate adaptation, and regional cooperation. Further quantitative research is needed to assess climate impacts on hydropower generation and the technical and economic feasibility of alternative energy pathways under future climate conditions.

Preprint under review for ESSD (discussion: open, 0 comments) We present PhyEddy-Net, a global mesoscale eddy detection dataset at 0.125° resolution for 2010–2021. It combines satellite altimetry, geostrophic vorticity, and sea surface temperature using a multimodal deep-learning model. The catalogue contains 36.7 million detections with polarity, area, radius, and probability. Independent Argo validation shows an 87.1 % hit rate and 92.1 % type agreement. PhyEddy-Net supports studies of eddy occurrence, spatial distribution, and ocean model validation.

Nature Geoscience, Published online: 28 September 2026; doi:10.1038/s41561-026-02114-x Arctic observations over 2003–2024 identify three aerosol regimes with contrasting responses to warming and cloud condensation nuclei trends. Model projections suggest that the aerosol–cloud feedbacks will evolve differently across the Arctic by 2100.

Nature Reviews Earth & Environment, Published online: 28 September 2026; doi:10.1038/s43017-026-00839-2 Joseph Boctor explains how machine learning can be used in combination with automated biochemical analyses to identify viable natural biocatalysts to remediate environmental pollution.

Nature Climate Change, Published online: 29 September 2026; doi:10.1038/s41558-026-02736-4 Security analysts and government economists increasingly present policymakers with different pictures of the same climate and nature risk. We argue that governments need a shared way to weigh these assessments, particularly given that under-preparing for a severe, irreversible outcome costs far more than over-preparing.

Journal of Geophysical Research: Solid Earth, Volume 131, Issue 10, October 2026.

Journal of Geophysical Research: Solid Earth, Volume 131, Issue 10, October 2026.

Journal of Geophysical Research: Solid Earth, Volume 131, Issue 10, October 2026.

Journal of Geophysical Research: Solid Earth, Volume 131, Issue 10, October 2026.

Journal of Geophysical Research: Solid Earth, Volume 131, Issue 10, October 2026.

Journal of Geophysical Research: Solid Earth, Volume 131, Issue 10, October 2026.

Journal of Geophysical Research: Solid Earth, Volume 131, Issue 10, October 2026.

Journal of Geophysical Research: Solid Earth, Volume 131, Issue 10, October 2026.

Journal of Geophysical Research: Solid Earth, Volume 131, Issue 10, October 2026.

Mountain glaciers are important freshwater reservoirs. Quantifying their loss is essential for forecasting water availability. While glacier retreat is well studied in many regions, glacier loss on the Nevados de Chillán volcanic complex in south-central Chile remains uncertain. Using satellite and drone images, we find all 28 glaciers on Nevados de Chillán lost volume over the past 7 decades, with melt rates accelerating markedly since 2000.

arXiv:2609.36381v1 Announce Type: new Abstract: We present Su\^etes, a fully differentiable, non-hydrostatic, limited-area atmospheric dynamical core implemented in JAX. Su\^etes includes both a semi-implicit semi-Lagrangian scheme and a Split-Explicit Runge-Kutta scheme with acoustic substepping, enabling applications from coarse regional simulations to convection-permitting dynamics. Its terrain-following geometry, spatial discretization, time integration, initial conditions, lateral boundary treatment, and included physical parameterizations are all compatible with reverse-mode automatic differentiation. A single reverse-mode pass can therefore compute sensitivities of forecast diagnostics to selected initial and boundary fields, physical parameters, topography, and computational geometry. We provide a series of numerical experiments---including a tracer-source inverse problem, a sensitivity analysis of a three-dimensional squall-line, adversarial perturbation construction an ERA5-driven downslope hurricane force wind event, topography optimization, and terrain-following coordinate optimization---demonstrating the versatility of Su\^etes as a framework for gradient-based diagnosis and model development at regional and convection-permitting scale.

Nature Reviews Earth & Environment, Published online: 30 September 2026; doi:10.1038/s43017-026-00836-5 Comprehensive environmental codes, exemplified by China’s new Ecological and Environmental Code, seek to integrate governance across climate, biodiversity and pollution, yet remain misaligned with the dynamic, interconnected nature of Earth system processes. Aligning legal frameworks with these dynamics requires adaptive, process-based governance, predictive monitoring and modelling, and cross-scale coordination.

by Pono Pono, Alan M. Jones Climate change and air pollution affect every major organ system, altering both the disease burden and patient responses to pharmaceutical treatments. However, current knowledge on how health systems, patients, healthcare professionals, and governments should prepare for these challenges remains fragmented. This review synthesises evidence on the potential impact of climate change on human health and pharmaceuticals, with the aim of identifying the diseases, populations, and medications most vulnerable, and assessing potential implications for pharmacotherapy. Emerging evidence shows that climate-related stressors, particularly extreme heat, contribute to increased morbidity and mortality and exacerbation of diseases affecting cardiovascular, cerebrovascular, gastrointestinal, renal, and respiratory systems. Furthermore, the review found that high temperature and humidity can affect drug stability, bioavailability, and pharmacokinetics by degrading active pharmaceutical ingredients and altering pharmaceutical formulations, potentially reducing efficacy and increasing the risk of adverse drug events. Older adults, patients with chronic diseases, and those taking medications that impair the thermoregulatory system are particularly at higher risk. In addition, extreme weather events may also disrupt pharmaceutical supply chains, further threatening access to essential medicines. This review finds that climate change has implications for healthcare through increasing disease burden and affecting the effectiveness and safety of medicines. Therefore, there is an urgent need for integrated research to explore the interplay between climate change, therapeutic response, and drug safety to support resilient, climate‑ready healthcare systems.

arXiv:2609.38763v1 Announce Type: new Abstract: Summertime synoptic weather regulates Arctic land-ocean exchange through meteorological tides during the ice-free period, when the predominantly easterly regime of the Beaufort High is punctuated by episodic Arctic cyclones, changing coastal winds to westerlies. We analyze wind-driven exchange flows in a northern Alaska lagoon using a 26-year reconstruction (2000-2025) validated by new in situ observations and FVCOM simulations. Phase-space EOF analysis of ERA5 data links deep cyclonic anomalies to 89% of Extreme Eastward Throughflow (EET) events (24 of 27). Coastal easterly occurrence frequencies show a consistent decrease (up to 38.3% at > 10 m/s) while low-to-moderate westerly occurrence frequencies increase by up to 30.9% but decrease by 40.6% at extreme westerly wind speeds (>10 m/s). Offshore ERA5 wind occurrence frequencies increased, with extreme offshore events (> 10 m/s) increasing by 82.4%. Local ASOS observations reveal a near-monotonic ~88% reduction in atmospheric forcing variability (East-West Wind Excursion variability drops from 2.98X10^4 to 0.35X10^4 km). The In-Out Flow Excursion variability declines steadily by 70% from 1.40X10^3 to <0.40X10^3 km, while the positive (outward) flow ratio variability contracts steadily from 7.4% to below 2.2%. Extreme Westward Throughflow (EWT) events were infrequent, but were concentrated in the most recent period, with five of six events occurring in 2022-2024 and one in 2013. These results indicate that the summertime coastal regime remains predominantly easterly, but exhibits progressively lower variability, while the wind distribution shifts toward more frequent low-to-moderate westerlies. This evolution of weather-forcing state may impact flushing and biogeochemical connectivity of Arctic lagoon ecosystems.

arXiv:2609.39379v1 Announce Type: new Abstract: A paradox recently emerged in artificial intelligence (AI) weather prediction research. While some claim AI weather models cannot simulate atmospheric butterfly effect, this conflicts with AI models' limited predictability and advances in AI ensemble forecasting. This study demonstrates via counterexamples that the butterfly effect does exist in AI weather predictions. For Super Typhoon Khanun, AI predictions are constrained by a double-attractor system. Minor initial perturbations confined to two regions trigger state transitions between two local attractors, causing a 1006-km difference in the predicted storm position on Day 7. This behavior is consistent with numerical weather prediction models and observed in ~12% of tropical cyclones in the past 5 years. These findings verify AI's ability to capture atmospheric chaos and provide the physical basis for AI ensemble forecasting.

arXiv:2609.39737v1 Announce Type: cross Abstract: The two-dimensional sphere embedded in three-dimensional Euclidean space S2, plays a central role in a variety of scientific and engineering domains, including geophysics, planetary science, geodesy, atmospheric physics, quantum chemistry, cosmology, and virtual reality, among many others. As machine learning increasingly permeates these fields, the demand grows for robust tools that process and model functions on the sphere, while respecting the inherent topological and symmetry properties of the domain. We present torch-harmonics, a comprehensive library that offers efficient, differentiable implementations of advanced signal processing and machine learning (ML) methods for spherical data. These include the spherical harmonic transform (SHT), the spherical analogue of the Fourier transform, vector spherical harmonics, discrete-continuous and spectral convolutions, as well as both global and neighborhood spherical attention mechanisms. Beyond traditional representations, torch-harmonics provides the building blocks for state-of-the-art spherical ML architectures such as spherical transformers in order to enable scalable, rotationally-aware learning and inference in modern scientific and engineering applications.

Journal of Geophysical Research: Solid Earth, Volume 131, Issue 10, October 2026.

by Lauren D’Mello-Guyett, Neelina Jaggarwal, Jane Wilbur, Belen Torondel, Meghna Ranganathan Menstrual health and hygiene (MHH) encompasses access to menstrual materials; water, sanitation, and disposal facilities; privacy and safety; accurate information; supportive social environments; access to care for menstrual discomfort and disorders; enabling people to manage menstruation with dignity and choice. Climate change is disrupting these interlinked domains, disproportionately affecting women, girls, and people who menstruate. This review synthesises evidence on how climate change affects MHH across diverse climate hazard contexts. We conducted a global mixed-methods systematic review of studies published in English between 1947 and 2025, examining associations between climate-related hazards and MHH. We conducted searches in Medline, Scopus, Embase, and Web of Science. Quantitative, qualitative, and mixed-methods studies were eligible. Given the heterogeneity in exposures, outcomes, and designs, findings were synthesised narratively. From 3,890 records, 25 studies met inclusion criteria, covering floods (n = 7), air pollution (n = 6), cyclones (n = 3), droughts (n = 3), saltwater intrusion (n = 3), and multiple hazards (n = 4). Flood- and cyclone-related studies reported disrupted access to menstrual materials, damaged water and sanitation infrastructure, reduced privacy and safety, limited access to care for menstrual disorders, and unsupportive social environments characterised by stigma and restricted mobility. Drought and salinity intrusion constrained hygiene practices through reduced water availability and quality, affecting comfort, dignity, and infection risk. Across hazards, studies documented adverse physical outcomes, including dermatological irritation and reproductive or urinary tract infections, alongside psychosocial consequences such as shame, anxiety, and harassment. Air pollution studies linked exposures with earlier menarche and menstrual cycle irregularities. Evidence on extreme heat and several MHH domains, including access to information, was limited. Climate-related hazards undermine multiple, interconnected domains of MHH. The evidence base remains fragmented across hazards, geographies, methods, underscoring the need to integrate MHH into climate adaptation, disaster preparedness, and humanitarian response frameworks and to address critical research gaps related to extreme heat, slow-onset hazards, and equity.

Proceedings of the National Academy of Sciences, Volume 123, Issue 39, September 2026. SignificanceAgriculture produces greenhouse gas emissions from on-farm activities and through conversion of natural ecosystems to agricultural use. Agricultural practices that increase yields can reduce demand for agricultural land, reducing emissions ...

Proceedings of the National Academy of Sciences, Volume 123, Issue 39, September 2026. SignificanceThis study presents a computational framework to mechanistically simulate neuronal lesions and interventions for the treatment of disorders of consciousness (DoC). Through a neural network with biologically plausible features, we modeled ...

Abstract Long multi-decadal transient convection-permitting climate simulations are crucial for reliable regional climate projections. However, the large computational resources required are often not available, particularly in the context of African climate change studies. This poses challenges for regions like the Lake Victoria basin, where extreme weather events, such as thunderstorms and associated wind gusts, result in high waves and flash floods with devastating impacts. Here, we configure a mini-ensemble of convection-permitting pseudo-global warming (PGW) regional climate simulations at 2.8 km for the Lake Victoria basin, providing a computationally efficient approach that allows us to assess model uncertainties related to changes in large-scale dynamics in the driving GCMs and identify robust climate change signals across the members. Our results reveal that, while extreme precipitation scaling is robust among the members (~5%/K), primarily driven by thermodynamic controls, the mean precipitation response is sensitive to uncertainties in large-scale dynamics, resulting in significant ensemble spread. Our projections indicate consistent drying over Lake Victoria during 18-24h UTC (21-03h local time) across the three ensemble members (up to −36% in the simulation with the strongest reduction in easterly trade winds), and a consistent shift from low and intermediate precipitation intensities to high values. Overall, despite uncertainties in large-scale dynamics, our ensemble projections provide critical insights for future climate resilience planning in the Lake Victoria basin.

Journal of Geophysical Research: Solid Earth, Volume 131, Issue 10, October 2026.

Journal of Geophysical Research: Solid Earth, Volume 131, Issue 10, October 2026.

arXiv:2610.03106v1 Announce Type: new Abstract: The subseasonal-to-seasonal (S2S) timescale, roughly from two weeks to two months ahead, is a critical forecast window for sectors such as agriculture, energy, and water management. Yet, it is widely known as the `predictability desert'. Recent AI weather models excel up to two weeks ahead but deteriorate beyond, largely because they are trained to predict fine-scale details that are neither predictable nor essential at S2S timescales. We argue that a more physically grounded objective is to forecast only the slowly varying components that remain predictable. Computer vision reached the same conclusion with the Joint-Embedding Predictive Architecture (JEPA), which predicts in latent space, discarding unpredictable details. In this work, we introduce S2S-JEPA, which brings the JEPA paradigm to S2S forecasting. It is tailored to this task through design elements from state-of-the-art AI weather models. S2S-JEPA achieves comparable skill to the gold-standard ECMWF physics-based ensemble and surpasses it on multiple metrics at weeks 5 to 6.

It is challenging to assimilate land surface temperature (LST) owing to its fast temporally varying nature. This study proposes a scheme by jointly updating the soil temperature and soil moisture. Results show marginal enhancement in LST, yet soil temperature bias over Northeast Asia (NA) drops sharply. Snow temperature and snow depth over NA, and soil moisture in the humid tropics also improve significantly. These consistent improvements demonstrate the effectiveness of the proposed scheme.

Coastal seas are strongly shaped by complex coastlines, shallow depths, and inputs from rivers, which influence marine life and water quality. This study introduces a new high-resolution modelling system that combines ocean circulation and marine ecosystem processes on unstructured grids. Applied to the northern Adriatic Sea, the model realistically captures seasonal changes in key biogeochemical variables, offering improved tools to support coastal environmental management.

Planning for climate change requires understanding how climate and weather interact to drive possible impacts. We developed a tool that simulates realistic monthly climate patterns and translates them into useful measures like energy needs. It captures seasonal cycles and year-to-year variations to bridges the gap between global climate predictions and local planning, making climate risk assessment faster and more accessible.

The spatial variability in the land surface properties is often not captured by the resolution of land surface models. To overcome this limitation, most models subdivide the grid cells into fractions with homogeneous characteristics, for which the land processes are calculated separately. In reality, the fractions interact via the lateral exchange of water and heat, and the present manuscript details an approach to include these fluxes in the land component of the ICON modeling framework.

Surface hoar crystals occasionally grow on the top of snowpacks overnight. Little work has focused on climatic conditions leading to surface hoar. We use data from three field campaigns in the Colorado Rockies to investigate. We show that surface hoar events decline as the climate warms, and that there may be 14 % fewer events per year in the future. There is still work needed to reconcile measured amounts of surface hoar crystals, models, and the relationship with turbulent air.

This study examines how high-energy particles develop and propagate in thunderstorm electric fields. Simulations at four high-altitude sites show that accounting for air density improves the description of particle avalanche growth across different atmospheric conditions. The study also examines how electrons and gamma rays propagate after leaving the accelerating field, providing a framework for interpreting high-energy atmospheric observations.

Earth's Future, Volume 14, Issue 10, October 2026.

Earth's Future, Volume 14, Issue 10, October 2026.

We applied a consistent modeling approach for both field and regional scales of multi-pollutants to evaluate the air-surface exchange processes contributing to regional air quality modeling biases when evaluated against observed network and satellite ammonia concentrations. This multi-resolution approach will serve the modeling and measurement community in their future development and generalization of air-surface exchange models utilizing flux, routine network and satellite observations.

Nature, Published online: 05 October 2026; doi:10.1038/d41586-026-03091-2 Karl Deisseroth, Peter Hegemann and Georg Nagel win the 2026 Nobel Prize in Physiology or Medicine for their work on optogenetics, which has revolutionized neuroscience.

Preprint under review for ESSD (discussion: open, 0 comments) China has built the world’s largest reservoir system, yet modeling of its climate and environmental impacts lacks attribute values. We developed a database of 5,142 reservoirs across China with over 20 key attributes, covering 87.5 % of total storage and 62 % of hydropower capacity. It enables more realistic identification of reservoir-served irrigation areas and improves streamflow simulations, offering a valuable resource for hydrology, agriculture, energy, climate, and environmental research.