Consistent response of European summers to the latitudinal temperature gradient over the Holocene
The drivers behind the current decadal trend toward longer and more extreme European summers are widely discussed. This is attributed to changes in the mid-latitude summer atmospheric circulation in response to Arctic Amplification and weakening of the latitudinal temperature gradients (LTGs), as well as to reduced aerosol emissions over Europe since the 1980s. However, causal links remain uncertain, limiting confidence in future projections. To gain statistical insights, evidence over periods longer than the instrumental record is necessary. Using seasonally resolved lake sediments, we reconstruct the evolution of the European summer-to-annual ratio over the last ten millennia. Our results indicate that summer weather dominated during the mid-Holocene, with an average of 195 summer days per year-falling within the extreme upper tail of summer distributions in the early-and late-Holocene. The Holocene variability in summer days aligns closely with simulated past changes in the LTG, supporting the hypothesis that dynamical processes influence midlatitude seasonal weather on decadal to millennial timescales. A 1 °C decrease in LTG would extend the summer season by ~6 days, potentially adding up to 42 summer days by 2100 under a business-as-usual scenario. These findings provide key observational constraints for understanding and projecting seasonal impacts on ecosystems and society.
The annual cycle of atmospheric circulation and associated weather regimes establishes the climate seasonal clock by determining the characteristic timing and recurrence of seasonal conditions. In the North Atlantic (NA)-European region, atmospheric circulation is prevailed by westerlies that transport heat and moisture from the ocean to the continent, thereby reducing the ocean-land surface temperature gradient and exerting a strong influence on the mean European climate. Additionally, a persistent high-pressure system over northern Europe, the so-called Scandinavian Blocking (SB), can divert the jet stream, creating a meandering, meridional flow and leading to prolonged climate extremes 1,2 (Fig. 1). According to the relationship between daily sea level pressure and continental temperature, the European climate seasonal clock is divided into two main seasons: summer and winter 2 . The onset of each season is determined by the
(Nature Communications. vol. 16, n° 2041-1723, 19/11/2025)
RHUL, UDC, GKT, ASTRAL, IMB, UB, Bordeaux INP, CNRS, Inria, UB, Bordeaux INP, EPOC, EPHE, PSL, UB, INSU - CNRS, CNRS
From Fields to Homes: How Agricultural Pesticides Contaminate Residential Environments? Findings from the PESTIPREV Study, France
Pesticide sprayings on crops can contaminate nearby homes, yet factors driving pesticide penetration indoors remain poorly understood. Our objective was to study the influence of factors related to air exchange and occupants track-in on agricultural pesticide surface loadings (SLs) measured in homes near vineyards. Indoor surface wipes were collected in 31 homes near vineyards during the peak pesticide application season in 2020 and 2021 and analyzed via LC/MS/MS or GC/MS/MS. Questionnaire data provided information on air exchange and track-in factors. Linear mixed models assessed their effects on SLs of seven fungicides across all surfaces (n = 667) and stratified by recently cleaned surfaces (n = 217), floors (n = 105), and high dusty surfaces (n = 130). Daytime airing reduced floor SLs by 14–28%, depending on the pesticide, but increased SLs by up to 38% on high surfaces. Nighttime airing was associated with a 20–65% increase in recently cleaned surface SLs. Double-flow ventilation increased high surface SLs by 21–126%, while ventilation grids reduced contamination, especially on floors (up to 51%). Home insulation tended to raise SLs. Pets and gardening were the main sources of pesticide track-in, contributing to an 11–87% increase in floor SLs. However, pets substantially reduced SLs on high surfaces. Direct yard access increased SLs in rooms by 5–10%. Room occupancy and outdoor activities showed mixed effects depending on the surface, whereas shoe removal and doormat use showed no significant impact. While ventilation effects remain complex to interpret for prevention, track-in findings suggest hygiene measures could help reduce indoor pesticide SLs.
(Water, Air, and Soil Pollution. vol. 237, n° 0049-6979, pp. article no 150, 19/11/2025)
BPH, UB, INSERM, CHU Bordeaux, ARS NA, EPOC, EPHE, PSL, UB, INSU - CNRS, CNRS, ECOSYS, INRAE
One-year high-frequency environmental and behavioral data from ALAN experience in a French coastal area
Despite the widespread exposure to artificial light at night (ALAN) on coastal ecosystems, its effects are poorly studied, and data remain scarce compared to terrestrial ecosystems. Coastal areas are critical for ecosystem services, providing biodiversity, social, and commercial benefits. To acquire high-quality data on ALAN impacts on coastal ecosystems, we conducted a one-year in situ study in the “Ile aux Oiseaux” site, part of the Arcachon Bay (France), from December 2023 to November 2024. The experimental platforms consisted of two oyster tables: one for the control condition exposed to natural light and one exposed to a skyglow ALAN intensity using underwater LEDs. The impacts of ALAN were assessed using two oyster species (Crassostrea gigas and Ostrea edulis) as sentinel organisms, recording continuously their valve behavior using high-frequency non-invasive valvometer biosensors. To characterize the local environment, the experimental platform was also equipped with multiple sensors for the long-term measurement of light irradiance and additional physical parameters (temperature, water level, salinity, turbidity, conductivity, and low-frequency sound).
(Scientific Data. vol. 12, n° 2052-4463, pp. 1808, 18/11/2025)
EPOC, EPHE, PSL, UB, INSU - CNRS, CNRS, GEODE, UT2J, Comue de Toulouse, CNRS
Shallow benthic invertebrate communities in relation to substrate types in coastal environments of the sub-Antarctic Crozet archipelago.
Coastal ecosystems of sub-Antarctic islands are threatened by increasing climate-driven changes and direct anthropogenic pressures. Significant effects on marine communities are expected, but benthic ecosystems of these isolated islands remain largely under-explored. Effective preservation of these nearshore environments requires deeper ecological assessments and comprehensive biodiversity knowledge. In this regard, this study reports findings from a survey carried out in 2021 at two sites – Baie du Marin and Crique du Sphinx – located on the eastern coast of Ile de la Possession (sub-Antarctic Crozet archipelago, Southern Ocean). We investigated the composition and structure of nearshore benthic faunal communities using a quantitative fieldwork protocol and an integrative molecular- and morphology-based taxonomic approach. A total of 124 morphotypes were identified, including a high proportion (72%) of rare species. Both sites exhibited similar benthic invertebrate communities. Structurally complex habitats such as hard substrates or areas dominated by macroalgae exhibited higher species richness and diversity. The investigated benthic invertebrate communities are typical of the sub-Antarctic area but featured unique structures, including dense tube-dwelling polychaete colonies. This study will provide a baseline for future monitoring programs and for the preservation of sub-Antarctic coastal benthic ecosystems.
(Frontiers in Marine Science. vol. 12, n° 2296-7745, pp. 1692217, 14/11/2025)
BGS, EPHE, PSL, CNRS, UBE, LBM, ULB, LECOB, SU, CNRS, OOB, SU, CNRS, ULiège, ULiège, UMAG, CHIC, BASE, EPOC, EPHE, PSL, UB, INSU - CNRS, CNRS, IMR, UiB, UIB, UAM
Développement de nouveaux algorithmes d’apprentissage statistique pour coupler projections climatiques et observations passées en vue de réduire les incertitudes du changement climatique à venir
Le climat change à cause des émissions humaines de gaz à effet de serre. Les modèles climatiques permettent de projeter son évolution selon des scénarios futurs d’émissions humaines (SSP). Cependant, les résultats des différents modèles climatiques développés dans le monde sont quantitativement divergents, générant une grande incertitude appelée par la suite incertitude modèle.Cette incertitude modèle est particulièrement forte pour la circulation méridienne de retournement en Atlantique (AMOC), un système de courants océaniques qui affecte les conditions climatiques de nombreuses régions du globe. En effet, même si tous les modèles climatiques CMIP6 prévoient un ralentissement de l’AMOC, l’ampleur varie fortement : en moyenne, ils projettent un ralentissement AMOC de 33 ± 36% d’ici 2100 (confiance 90%, scénario SSP2-4.5).Pour réduire l’incertitude modèle, les méthodes dites de contraintes observationnelles (COs) ont été récemment développées. Elles contraignent les projections des modèles par l'observation réelle d'une (univarié) ou de plusieurs (multivarié) variables. Ces approches sont cependant difficiles à utiliser car dans la littérature (i) les fondements statistiques sur lesquels elles reposent ne sont pas toujours identiques et manquent souvent de clarté, (ii) les approches sont très diversifiées et souvent univariées, et (iii) le choix des variables observables est subjectif. Cette thèse se compose de trois articles qui traitent chacune de ces trois problématiques.(i) Le premier article propose un modèle statistique rigoureux pour une CO linéaire multivariée, appelé ClimLoco1.0. Il prend en compte l’incertitude sur l'observation réelle et l'incertitude provenant du nombre limité de modèles climatiques, toutes deux souvent négligées ou mal considérées dans la littérature. Dans un souci de pédagogie, l’article détaille progressivement la construction de ClimLoco1.0 et fournit des interprétations graphiques des résultats mathématiques obtenus.(ii) Le second article compare les performances de quatre approches de CO, par validation croisée. L’AMOC future y est contrainte soit par l’AMOC passée (univarié), soit par l’AMOC passée, la température et la salinité de surface de différentes régions océaniques choisies subjectivement (multivarié). L’approche la plus performante est la régression Ridge dans le cas multivarié, qui est particulièrement adaptée à notre étude. Elle fournit une estimation du ralentissement AMOC d’ici 2100 de 51 ± 8% (SSP2-4.5). Ce ralentissement plus important que celui estimé par la moyenne multi-modèle (33%) résulte principalement de la considération du biais des modèles climatiques concernant la salinité de surface en Atlantique sud. Cela est en accord avec la littérature reliant ce biais à la rétroaction positive d’advection de salinité, un mécanisme clé pour la stabilité de l’AMOC.(iii) Le troisième article utilise une méthode spécifique de clustering pour créer des groupes de régions associées à de nombreuses variables climatiques. Ces groupes sont interprétés pour identifier les mécanismes qui expliquent l’incertitude modèle de l’AMOC future. Ils sont également résumés par des variables synthétiques qui sont utilisées dans la méthode ClimLoco1.0 pour contraindre cette incertitude modèle. Cette méthode suggère que l’incertitude modèle est notamment expliquée par la rétroaction positive d’advection de salinité et les remontées d’eaux profondes, en accord avec la littérature. Contrainte par l’observation de ces variables synthétiques, l'estimation du ralentissement AMOC d'ici 2100 est de 45 ± 10% (SSP2-4.5).Les deux derniers articles estiment un ralentissement AMOC similaire, qui est plus important et moins incertain que celui estimé par la simple moyenne multi-modèles. Il semble donc opportun que ce résultat soit pris en compte afin d’améliorer les projections climatiques et les stratégies d'adaptation associées.
(14/11/2025)
ASTRAL, IMB, UB, Bordeaux INP, CNRS, Inria, UB, Bordeaux INP, EPOC, EPHE, PSL, UB, INSU - CNRS, CNRS
Global Carbon Budget 2025
Accurate assessment of anthropogenic carbon dioxide (CO2) emissions and their redistribution among the atmosphere, ocean, and terrestrial biosphere in a changing climate is critical to better understand the global carbon cycle, support the development of climate policies, and project future climate change. Here we describe and synthesise datasets and methodologies to quantify the five major components of the global carbon budget and their uncertainties. Fossil CO2 emissions (EFOS) are based on energy statistics and cement production data. Emissions from land-use change (ELUC) are estimated by bookkeeping models based on land-use and land-use change data. Atmospheric CO2 concentration is measured at surface stations, and the global atmospheric CO2 growth rate (GATM) is computed from the annual changes in concentration. The global net uptake of CO2 by the ocean (SOCEAN, called the ocean sink) is estimated with global ocean biogeochemistry models and observation-based fCO2products. The global net uptake of CO2 by the land (SLAND, called the land sink) is estimated with dynamic global vegetation models. Additional lines of evidence on land and ocean sinks are provided by atmospheric inversions, atmospheric oxygen measurements, ocean interior observation-based estimates, and Earth System Models. The sum of all sources and sinks results in the carbon budget imbalance (BIM), a measure of imperfect data and incomplete understanding of the contemporary of 2023-2024 in the Northern Hemisphere. The ocean CO2 sink was 3.2 ± 0.4 GtC yr -1 during
(13/11/2025)
CICERO, UiO, UEA, GKSS, VLIZ, MPI-M, WUR, LMU, LMU, LMU, CSIRO, LSCE, UVSQ, INSU - CNRS, CNRS, DRF (CEA), CEA, BIOGEO, LSCE, UVSQ, INSU - CNRS, CNRS, DRF (CEA), CEA, UiO, PMEL, NOAA, KIT, CIMAS, RSMAS, BIOS, UOR, MFRI, BSC-CNS, LMD, INSU - CNRS, X, IP Paris, SU, CNRS, ENS-PSL, PSL, ENPC, IP Paris, CSIRO, JAMSTEC, SATINV, LSCE, UVSQ, INSU - CNRS, CNRS, DRF (CEA), CEA, JMA, LGENS, INSU - CNRS, CNRS, ENS-PSL, PSL, EPOC, EPHE, PSL, UB, INSU - CNRS, CNRS
Investigation of the combined influence of salinity and particle concentration on the adsorption of anionic and zwitterionic PFAS onto estuarine sediment using the RSM modelling approach
Salinity (S) and suspended particulate matter (SPM) are key factors influencing the sorption of micropollutants in estuaries, due to strong gradients in these ecosystems. Previous laboratory or field-based studies have typically investigated the impact of S or SPM separately. Thus, the combined effects of S and SPM as well as their interactions on the sorption of micropollutants such as per- and polyfluoroalkyl substances (PFAS) in estuarine environments still remain poorly understood. We initially investigated the adsorption kinetics of 11 anionic and zwitterionic PFAS onto estuarine sediment under one S/SPM combination in laboratory-controlled conditions, as well as their adsorption isotherms under two S/SPM combinations. We also determined their distribution coefficients (K d ) across 35 S/SPM combinations covering a wide range of estuarine conditions. The adsorption kinetics of PFAS could be described by a pseudo-second-order model (equilibrium time <24h). Sorption isotherms were fitted by both linear and Freundlich models; the linear sorption range was in the range 0.12–1.31 nM and K d varied between 0.6 and 55271 L/kg. Based on response surface modelling, both S and SPM were significant factors, i.e. K d was positively related to S (salting-out effect), while it was negatively related to SPM concentration (third-phase effect). SPM had a stronger effect than S for short-chain carboxylates, whereas S was the dominant factor for most other compounds. We also present, for the first time, evidence of a significant negative interaction between these two factors. This study therefore provides a new perspective to model the fate of PFAS at the land-sea interface.
(Peer Community Journal. vol. 5, n° 2804-3871, pp. e124, 12/11/2025)
EPOC, EPHE, PSL, UB, INSU - CNRS, CNRS
The ongoing spread of the Asian date mussel (Arcuatula senhousia) within the French Atlantic coast: colonisation dynamics and associated drivers in a historically invaded ecosystem (Arcachon Bay)
Arcuatula senhousia is a non-indigenous species first observed in Arcachon Bay in 2002. At that time, the species’ distribution was restricted to the northern part of this coastal lagoon. During the following 7–8 years, the species also started to colonise its south-eastern parts. Then, two surveys conducted in 2018 and 2021 showed that the species was observed over most of the investigated tidal flats within the lagoon. Between those two periods, there was a threefold increase in both frequencies of occurrence and average densities. The highest average densities and frequencies were observed in areas colonised in 2002. It suggests that this area is either the main area of introduction/settlement for the species or the area where it could find the most suitable ecological conditions. However, the use of Species Distribution Modelling showed that, considering habitat features, most of the intertidal flats in Arcachon Bay were a highly suitable habitat for A. senhousia . Further colonisation of the lagoon during the coming years appears very likely. The study of its habitat in this area suggested that the presence of meadows favoured the settlement of A. senhousia individuals. Furthermore, lower bottom current velocity and the erosion potential it induces seemed to be the principal environmental factors driving the distribution pattern of this species within the bay. However, other factors that could regulate the spread of A. senhousia must be considered. Understanding the dynamics of A. senhousia colonisation and identifying its drivers aimed to characterise possible areas to be colonised by this species. This is a crucial point in determining how to manage its distribution and assess the risk of spreading within other ecosystems.
(Aquatic Invasions. vol. 20, n° 1798-6540, pp. 427-450, 12/11/2025)
EPOC, EPHE, PSL, UB, INSU - CNRS, CNRS, PatriNat, MNHN, IRD, CNRS, OFB - DSUED, OFB, LERAR, COAST, IFREMER
Deep learning for carbonate rocks petrography
This project aims to investigate deep learning capacity to enhance the identification of factors that may impact the dynamic behavior of carbonate reservoirs. This is achieved through the integration of comprehensive datasets comprising images acquired over decades from hydrocarbon carbonate reservoirs and identified outcrop analogues. Convolutional neural networks represent natural candidates for analyzing the thousands of available thin sections. This thesis developed a suite of complementary methods for their quantitative and automated description. An optimal subdivision strategy for high-resolution images was established to enable multi-scale characterization of this data type, closely approximating traditional microscopic analysis methods. The approach combines representation learning through a rotation-invariant variational autoencoder with automatic clustering, achieving a two-fold acceleration in labeling time while forming geologically coherent groups based primarily on R. Dunham's classifications for limestones, R. Folk's classifications for dolomites, and the incorporation of secondary minerals such as anhydrite facies. The model can then be refined for these subsidiary classification tasks, enabling semantic description of the samples studied (>88% mAP). To address the heterogeneity of existing petrographic descriptions and the lack of precise quantification on real data, two methods were developed: SynSection procedurally generates training data by simulating depositional processes (+20% mAP) in grain limestones, while SynSection2 unifies synthetic-to-realistic transformation with segmentation (+5% additional mAP). These approaches jointly enable automatic segmentation of elements (bioclasts, intraclasts, ooids, and peloids) and the creation of a homogeneous quantified database. Deep semantic segmentation applied to granular limestones provides access to quantification of relative proportions and grain size distribution analysis. The developed metrics achieve determination coefficients (R²) of 0.78 for mean grain size and 0.72 for sorting. This integrated methodology enables standardized and reproducible description following established petrographic nomenclatures and can be superimposed on pore network analysis. Automated quantification of pore types reveals heterogeneities that facilitate comparative analysis between samples. This approach constitutes a powerful tool for quantitative petrographic analysis, capable of incorporating local refinements to fully exploit deep learning's potential in characterizing carbonate systems and understanding associated petrophysical responses.
(10/11/2025)
EPOC, EPHE, PSL, UB, INSU - CNRS, CNRS
Rising atmospheric CO2 concentrations
Across the last deglaciation, the atmospheric CO 2 concentration (CO 2 ) increased substantially from ∼ 180 to ∼ 280 ppm, yet its impact on vegetation dynamics across this major climatic transition remains insufficiently understood. In particular, Iberian pollen records reveal an intriguing feature that can be related to an often overlooked role of CO 2 in shaping vegetation responses during the last deglaciation. These records reveal the near disappearance of forests during the cold Last Glacial Maximum (LGM) and Heinrich Stadial 1 (HS1) phases and an unexpected recovery during the Younger Dryas (YD) cold phase when CO 2 increased. Here, we present high-resolution tracers of terrestrial (pollen, C 29 : C 31 organic biomarker) and marine (alkenone-derived Sea Surface Temperature, C 37 : 4 %, and long-chain n-alkanes ratios) conditions from the southwestern (SW) Iberian margin Integrated Ocean Drilling Program Site U1385 ("Shackleton site") for the last 22 cal kyr BP. This direct land-sea comparison approach allows us to investigate how the Iberian Peninsula vegetation responded to major global CO 2 changes during the last deglaciation.
Our results show that cool and moderately humid conditions of the LGM supported a grassland-heathland mosaic ecosystem, but low CO 2 likely caused physiological drought and suppressed forest development. HS1, the coldest and most arid period, combined with sustained low CO 2 values, almost suppressed forest growth in favour of Mediterranean steppe. In contrast, the warmer Bølling-Allerød, characterised by a temperature optimum and variable but generally wetter conditions, along with the rise of CO 2 above 225 ppm at ∼ 15 cal kyr BP, contributed to substantial forest development. During the YD, sufficient moisture combined with increasing CO 2 enabled the persistence of a mixed grassland-forest mosaic despite cooler temperatures. Our study suggests that during cool and humid periods (LGM and YD) different pCO 2 values led to contrasting SW Iberian vegetation responses. In contrast, during periods of relatively high CO 2 , temperature and precipitation played the main role in shaping the distribution and composition of the vegetation.
(Biogeosciences. vol. 22, n° 1726-4170, pp. 6631 - 6650, 07/11/2025)
IPMA, CCMAR, UAlg, CCMAR, UAlg, CIMAR/CCMAR, EPOC, EPHE, PSL, UB, INSU - CNRS, CNRS