Scientific and media attention to extreme events has increased in recent years due to their severity and the damage they cause. Estimating the contribution of climate change to the probability of occurrence and intensity of these events has become a common practice, thanks to initiatives relying on climate science researchers, who diagnose the return period of the event and the contribution of the anthropogenic forcings in the event properties. These initiatives have succeeded in shedding light on the detection and attribution concept but still rely on the commitment of researchers to deal with an ever-increasing number of events. Here, we propose a proof of concept of an automatic attribution method which, in a few seconds, can attribute a heat wave that has occurred anywhere on the globe and estimate its statistical properties both in the past and in the future, using an existing method based on Bayesian statistics that combines past observations and climate models. Our web application currently covers extreme heat events over a period of 3 days and paves the way to a number of relevant applications, such as (i) enabling public or private organizations to become autonomous, i.e., processing the events they wish without waiting for an academic research investigation; (ii) handling events that have received less attention but have had a major impact, particularly in developing countries; and (iii) climate monitoring of extreme events over a specific region. SIGNIFICANCE STATEMENT: Quantifying the contribution of climate change to the frequency and intensity of extreme events (e.g., heat waves, floods) is an ad hoc exercise, often associated with an academic publication. Here, we propose a new paradigm via an automatic attribution method that, in a few seconds, provides key indicators regarding the contribution of anthropogenic climate change to the properties of hot extreme events that has occurred anywhere on the globe. This approach and its online application enable scientists and stakeholders trained in its use to study the events that interest them and to conduct climate monitoring of past events by refining the statistical indicators as more observations become available.

To ensure consistency in adaptation policies, the French government has adopted a Reference Warming Trajectory for Adaptation (TRACC). This trajectory is based on current international commitments to limit greenhouse gas emissions and translates them into global warming levels (1.5°C, 2°C, and 3°C), associated with three time horizons (2030, 2050, and 2100, respectively). To address the needs of adaptation stakeholders, these global warming levels have been expressed in terms of regional climate change over French territories, including both mainland France and overseas regions. Discrepancies over mainland France between regional climate projections from the CMIP5 generation and recent warming estimates derived from observational constraints motivated the development of a new methodology. This approach relies on regional warming levels to characterize future climate change consistently with the reference warming trajectory. This presentation outlines the principles of this methodology and its extension to overseas territories. It also describes how this method has been applied to describe future climate change in terms of averages, variability, extremes, and sectoral indicators. Perspectives for updating the description of the reference warming trajectory, based on the downscaling of CMIP6 simulations, are also discussed. They rely on the synthesis of a wide range of diverse and recently developed data sources, including kilometer-scale regional climate models, coupled regional climate models, AI-based emulators, very high-resolution global climate models, and observational constraints.

Convective self-aggregation is the spontaneous spatial organisation of deep-convective clouds into a limited region surrounded by drier, convectively inhibited regions, which occurs in many models. We propose a simple, piecewise linear model for self-aggregation based on primitive equations. In this model, each atmospheric column is in one of two possible thermodynamic regimes, deep convective or convectively inhibited, and the thermodynamics in each regime is linearised. The model simulates aggregated and non-aggregated steady states and reproduces many properties of self-aggregation as simulated by kilometre-resolution models. In particular, it exhibits a hysteresis with multiple, aggregated and non-aggregated equilibria and a similar sensitivity to convective inhibition, domain size, and boundary-layer radiative cooling in the dry region as in kilometre-resolution simulations. These results suggest that a self-aggregated state can be considered as a simple gravity wave with two phases: one convective and one convectively inhibited.

A lack of meaningful action in reducing greenhouse gas emissions has fuelled debate on whether global warming can still be limited to 1.5°C. In 2022, the Working Group III contribution to the Sixth Assessment Report found 97 scenarios that could limit global warming to 1.5°C above pre-industrial levels with no or limited overshoot, while 9 kept median warming projections under 1.5°C throughout the 21st century. These pathways relied on emissions reductions by 2025 that have not materialized. We show that updating the emissions and climate model calibrations to 2023, provides additional constraints on the projected warming outcomes. The window to keep warming below 1.5°C with a greater than 50% likelihood has closed. We use several approaches to show that the objective of limiting warming to 1.5°C needs to be pursued from above after a temporary overshoot. Immediate, rapid deep reductions in CO 2 emissions and non-CO 2 forcers, can still limit peak warming to around 1.7°C, and returning below 1.5°C before 2100 is attainable. Keeping peak warming as low as possible and pursuing to reverse it below 1.5°C thereafter will contain the adverse consequences of temperature overshoot and limit the scale of negative emissions required.

Abstract Quantifying the role of the atmospheric circulation in the temperature increase in Europe over the recent decades remains an open question. Here, we present an innovative dynamical adjustment framework based on a convolutional neural network (UNET), trained on CMIP6 simulations and fine‐tuned on reanalysis, to estimate the circulation‐induced temperature at the daily timescale and the subsequent trends over 1979–2024. Following this methodology, which provides high reconstruction scores at the daily timescale, we find temperature trends induced by the combination of horizontal winds at 850 hPa of 0.05 (−0.03, 0.14)°C/decade annually, 0.08 (−0.00, 0.17)°C/decade in summer and 0.09 (−0.11, 0.29)°C/decade in winter, accounting for between 10% and 20% of the total trends. In addition, we show that using sea‐level pressure rather than upper‐level winds would result in slightly higher trends, especially in summer.

We describe an improved method and the associated package for estimating the statistics of temperature extremes in a Bayesian framework. Building on previous work, this method uses a range of climate model simulations to provide a prior of the real-world changes, and then considers observations to derive a posterior estimate of past and future changes. The new version described in this study makes it possible to process several scenarios simultaneously, while keeping one single counterfactual world (i.e., the world without human influence). We offer a free licensed, easy-to-use command-line tool called ANKIALE (ANalysis of Klimate with bayesian Inference: AppLication to extreme Events), which can be used to reproduce the analyses presented here, as well as to process user-defined events. ANKIALE is based on a python code, but is designed to be used from the command line interface. ANKIALE is natively parallel, enabling it to be used on a personal computer as well as on a supercomputer. To derive the posterior, ANKIALE uses state of art MCMC-methods to sample the posterior distribution. The potential of this method and tool is illustrated via an application to maximum temperature over Europe between 1850 and 2100 (the posterior is derived from ERA5, covering the period from 1940 to 2024), at a 0.25 degrees resolution, for a range of four emission scenarios, including a particular focus on the city of Paris (France).

In a rapidly changing climate, evidence-based decision-making benefits from up-to-date and timely information. We track twelve key sets of indicators of the state of the climate system, closely following Intergovernmental Panel on Climate Change (IPCC) Sixth Assessment report (AR6) methods, to produce our fourth annual publication. One of the indicators, the Earth's energy imbalance (EEI) provides a crucial integrative measure of the overall heating of the planet and the pace of climate change - this has more than doubled since the 1976-1995 period. A newly added indicator of temperature extremes, the number of days experiencing marine heatwaves, has more than tripled between 1991 and 2025.For the 2016-2025 decade average, observed warming relative to 1850-1900 was 1.26 [1.13 to 1.36] degrees C, of which 1.24 [1.0 to 1.5] degrees C was human-induced. Human-induced warming reached 1.37 degrees C relative to 1850-1900 in the year 2025, increasing at a rate of 0.27 [0.2-0.4] degrees C per decade over 2016-2025. This high rate of warming, which matches the all-time high seen last year in the instrumental record, was caused by a combination of greenhouse gas emissions being at an all-time high of 54.6 +/- 5.5 GtCO2e yr-1 over the last decade (2015-2024), as well as reductions in the strength of aerosol cooling. Despite this, there is evidence that CO2 emission growth is slowing. The continuation of these annual updates could track decreases or increases in the rate of human influence and climatic changes presented here, reflecting the outcomes of societal choices during the critical 2020s decade.The data presented herein can provide a useful reference point for the drafting of the IPCC seventh assessment report. In total, we employ analysis from over 40 global datasets (10.5281/zenodo.20499280, Smith et al., 2026a). Future monitoring of these indicators, such as ocean and satellite measurements of the Earth's energy imbalance, are threatened by geopolitical and public funding decisions. Our ability to consistently track many of the indicators requires the continuity of observation programs and coordination mechanisms, including the Global Climate Observing System (GCOS) program, that enable their effective integration and use.

To describe regional climate change, climate services typically rely on an ensemble of climate model simulations. The development and arrival of observational constraints at regional scales are questioning this approach, as some simulations may not align with warming trajectories estimated by these techniques. This study proposes a methodology for describing future regional changes that combines multiple sources of information: global and regional observational constraints applied to the CMIP6 ensemble, along with existing regional climate model simulations driven by CMIP5. This approach uses Regional Warming Levels (RWLs), mirroring the use of Global Warming Levels (GWLs) in the IPCC AR6. We apply it to mainland France, a region with discrepancies in warming projections between global models, regional models, and observational constraints. Results show that the standard GWL approach produces unrealistically low warming estimates due to overly low regional-to-global warming ratios in some models. Using RWLs allows separation of the annual mean warming estimation (based on observational constraints) from the detailed climate change characteristics (based on regional models). We explore ways to link RWLs and GWLs and assess associated uncertainties. This methodology has been selected to describe future climate change in mainland France, as part of the definition of a reference trajectory for adaptation set by the French government. It can be replicated in other regions and applied to existing or upcoming climate projections to express them in terms of regional warming levels at the national scale. Practical implications: The French government has recently chosen to adopt a reference trajectory for adaptation to climate change in France, known as the TRACC (Trajectoire de Rechauffement de reference pour l'Adaptation au Changement Climatique). This trajectory defines 3 levels to which the country needs to prepare for, corresponding to +1.5 degrees C global warming in 2030, +2 degrees C in 2050 and +3 degrees C in 2100 compared to 1850-1900. The aim is to establish a single framework for climate change impact studies including climate services, the definition and analysis of adaptation actions, standardizing practices nationwide and facilitating a coherent response to climate challenges. This article describes the methodological choices associated with this trajectory, based on a description of future changes at a fixed regional warming level (RWL) consistent with the chosen global trajectory. For mainland France, the 3 TRACC levels are expressed as an average warming over the country of 2 degrees C in 2030, 2.7 degrees C in 2050 and 4 degrees C in 2100 compared to 1850-1900. These are derived from observational constraints, combining models and observations. The subsequent description of local scale climate change is based on existing regional climate model simulations. The article finally provides a description of some of the changes associated with these 3 regional warming levels.

In a rapidly changing climate, evidence-based decision-making benefits from up-to-date and timely information. Here we compile monitoring datasets (published at https://doi.org/10.5281/zenodo.15639576; Smith et al., 2025a) to produce updated estimates for key indicators of the state of the climate system: net emissions of greenhouse gases and short-lived climate forcers, greenhouse gas concentrations, radiative forcing, the Earth's energy imbalance, surface temperature changes, warming attributed to human activities, the remaining carbon budget, and estimates of global temperature extremes. This year, we additionally include indicators for sea-level rise and land precipitation change. We follow methods as closely as possible to those used in the IPCC Sixth Assessment Report (AR6) Working Group One report. The indicators show that human activities are increasing the Earth's energy imbalance and driving faster sea-level rise compared to the AR6 assessment. For the 2015–2024 decade average, observed warming relative to 1850–1900 was 1.24 [1.11 to 1.35] °C, of which 1.22 [1.0 to 1.5] °C was human-induced. The 2024-observed best estimate of global surface temperature (1.52 °C) is well above the best estimate of human-caused warming (1.36 °C). However, the 2024 observed warming can still be regarded as a typical year, considering the human-induced warming level and the state of internal variability associated with the phase of El Niño and Atlantic variability. Human-induced warming has been increasing at a rate that is unprecedented in the instrumental record, reaching 0.27 [0.2–0.4] °C per decade over 2015–2024. This high rate of warming is caused by a combination of greenhouse gas emissions being at an all-time high of 53.6±5.2 Gt CO2e yr−1 over the last decade (2014–2023), as well as reductions in the strength of aerosol cooling. Despite this, there is evidence that the rate of increase in CO2 emissions over the last decade has slowed compared to the 2000s, and depending on societal choices, a continued series of these annual updates over the critical 2020s decade could track decreases or increases in the rate of the climatic changes presented here.

As the climate warms, cold waves are expected to become less intense and less frequent. Is there still a risk of reliving events comparable to the most intense cold spells we can remember? We analyze four remarkable cold spells that have occurred since 2010 in different regions: Western Europe, Texas, China, Brazil. We show that all these recent events have a moderate to high probability of not happening again by 2100 – typically 50% to 90% in an intermediate emissions scenario, depending on the event. The probabilities are even higher for iconic events of the 20th century or earlier. Our results suggest that the most intense cold snaps, and their associated icy landscapes in mid-latitude regions, are disappearing or have already disappeared due to anthropogenic climate change.

This work focuses on inferring design life levels for extreme events under non-stationary conditions. Its objectives are twofold. The first one is to provide a single indicator that summarizes relevant and interpretable information about large values in time series, even when stationarity cannot be assumed. Classical risk indicators such as the 100-year return level become difficult to interpret in a non-stationary framework. To address this, we leverage the existing concept of the equivalent reliability (ER) level. Under stationarity, the ER level coincides with the classical return level, but it differs otherwise. More precisely, the ER level ensures that the probability of having all observations below the ER level during a specified design period is controlled. This definition ensures interpretability in terms of safety or failure risk. A second objective is to capture stochastic and estimation uncertainty, a key aspect in any risk analysis, as uncertainties due to inference schemes can grow with extreme intensities. We incorporate both by using the Bayesian predictive distribution. Although well known in Bayesian statistics, the predictive distribution has rarely been applied to climatological time series risk analysis. Our approach is demonstrated on simulated data and on a case study of annual maxima of temperatures at a site in Southern France. To do so, a non-stationary Bayesian hierarchical extreme value model is used to combine data from 26 CMIP6 general circulation model simulations (SSP2-4.5, 1850-2100) with observations. The resulting predictive ER levels clearly indicate that non-stationarity over a design period of interest, as well as sampling and estimation uncertainty, have to be taken into account for risk assessment. For example, the 1000-year posterior predictive ER level for 2050-2100 is higher than any non-stationary 1000-year return level median estimate over the same period, reflecting the increasing risk due to the non-stationarity of the SSP 2-4.5 pathway.

High-resolution climate information is critical for the Vulnerability, Impacts, Adaptation, and Climate Services (VIACS) communities. Coordinated ensembles generated by initiatives like the Coordinated Regional Climate Downscaling Experiment (CORDEX) provide consistent and comparable information for the present and future over all land areas of the globe. This manuscript focuses on the European CORDEX initiative (EURO-CORDEX) and its coordinated effort to build regional climate ensembles for the years to come. In its first phase, EURO-CORDEX produced a rich ensemble of regional climate simulations under different representative concentration pathway scenarios. The EURO-CORDEX dataset is openly available and was fed into the Regional Atlas of the IPCC Sixth Assessment Report. However, this ensemble suffered from several shortcomings, which the community seeks to address in the next phase of production. Chief among these is the oft-cited criticism that the selection of GCMs that provide input to the regional climate models was not rigorous and that the resulting ensemble represents an "ensemble of opportunity." The present paper provides a description of how the community has addressed these shortcomings. We present a comprehensive, flexible, and traceable evaluation framework and toolkit for assessing the suitability of GCMs for downscaling, using EURO-CORDEX as an example. Its value lies in its explicit recognition of subjectivity and mechanisms implemented to transparently track decision-making. Further, the utility of the framework extends well beyond predownscaling decisions to also include postdownscaling investigations performed by the VIACS communities and beyond, to include researchers investigating such topics as model biases, future constraints, and exploring future storylines. SIGNIFICANCE STATEMENT: The European Coordinated Regional Climate Downscaling Experiment initiative (EURO-CORDEX) community has created a comprehensive evaluation framework and open-source toolkit for global climate model assessment. These approaches provide a robust, transparent, and traceable approach to both pre-and postdownscaling ensemble designs whose utility extends well beyond the immediate regional climate community. This will enable improved and more nuanced assessments of regional climate change and its impacts.

The first version of the Detection and Attribution Model Intercomparison Project (DAMIP v1.0) coordinated key simulations exploring the role of individual forcings in past, current and future climate as part of the Coupled Model Intercomparison Project, Phase 6 (CMIP6). The simulations have been used extensively in the literature for detection and attribution of long-term changes, constraining projections of climate change, attributing extreme events and understanding drivers of past and future simulated climate changes. Attribution studies using DAMIP v1.0 simulations underpinned prominent assessments of human-induced warming in the Intergovernmental Panel on Climate Change (IPCC) Sixth Assessment Report. Here, we describe the set of DAMIP v2.0 simulations, proposed for the next phase of CMIP, CMIP7. Detection and attribution studies rely on pre-industrial control simulations and historical simulations, which will be part of the Diagnostic, Evaluation and Characterization of Klima (DECK) set of simulations for CMIP7. In addition, we identify the three highest-priority single-forcing experiments for CMIP7 to be run as “Assessment Fast Track” simulations in support of the Seventh Assessment Report of the IPCC: simulations with natural forcings only, anthropogenic well-mixed greenhouse gases only and anthropogenic aerosols only. Beyond this, the DAMIP v2.0 experimental design includes full-column ozone-only simulations and land-use-only simulations, such that the set of individual forcing experiments, when these are considered together, represents the full set of historical forcings. While concentration-driven simulations are prioritised for attribution, emissions-driven versions of the DAMIP experiments are also proposed to support understanding of the influence of carbon-cycle feedbacks on the simulated responses to individual forcings.

State-of-the-art climate models project a substantial decline in precipitation for the Mediterranean region in the future1. Supporting this notion, several studies based on observed precipitation data spanning recent decades have suggested a decrease in Mediterranean precipitation2–4, with some attributing a large fraction of this change to anthropogenic influences3,5. Conversely, certain researchers have underlined that Mediterranean precipitation exhibits considerable spatiotemporal variability driven by atmospheric circulation patterns6,7 maintaining stationarity over the long term8,9. These conflicting perspectives underscore the need for a comprehensive assessment of precipitation changes in this region, given the profound social, economic and environmental implications. Here we show that Mediterranean precipitation has largely remained stationary from 1871 to 2020, albeit with significant multi-decadal and interannual variability. This conclusion is based on the most comprehensive dataset available for the region, encompassing over 23,000 stations across 27 countries. While trends can be identified for some periods and subregions, our findings attribute these trends primarily to atmospheric dynamics, which would be mostly linked to internal variability. Furthermore, our assessment reconciles the observed precipitation trends with Coupled Model Intercomparison Project Phase 6 model simulations, neither of which indicate a prevailing past precipitation trend in the region. The implications of our results extend to environmental, agricultural and water resources planning in one of the world’s prominent climate change hotspots10. Our assessment of a 27-country weather station dataset in the Mediterranean region revealed long-term stability in precipitation over 150 years, along with substantial short-term variability on annual to decadal scales driven by atmospheric circulation; these findings align with the precipitation trends seen in CMIP6 models.

Climate change monitoring and adaptation policies require reliable and updated indicators about global warming. Reports of the Intergovernmental Panel on Climate Change (IPCC) provide updated indicators every 5 to 7 years, based in particular on the latest available data and evidence. In the last report, projections were derived from climate model simulations constrained by observations. Current warming was derived from observations averaged over the last 10 years, or again a combination of model and observations in attribution statements. Here we explore the possibility of updating present and future warming estimates on a yearly basis, by incorporating new observations to observational constraints. First, we show that adding the latest temperature observations each year leads to a continuous improvement in estimating warming projections. In particular, warming estimates are not affected by year-to-year internal variability. Second, regarding current warming, we show that observational constraints can be used to derive an estimate of the forced warming for the current year, without having to average over the last 10 years - as the IPCC does in its latest report. This provides an unbiased estimate of current warming, without adding further variance. Third, we show that updating model data can lead to a gap in the estimate of current and future warming, but this remains within the uncertainties we estimate. Finally, we argue that annual updates of current and future warming estimates provide accurate, robust, and reliable information about climate change, while remaining consistent with previous years' estimates.

In Europe, temperature variations are mainly driven by the North Atlantic atmospheric circulation. Here, with data from the MIROC6 large ensemble, we investigate a convolutional neural network (a UNET) for reconstructing daily temperature anomalies in Europe from Sea Level Pressure (SLP) as a proxy of the atmospheric circulation, and we compare the results with a traditional analogs approach. We show an excellent ability of the UNET to estimate temperature variations given information from SLP only. This novel method outperforms the analogs method, at both daily and inter‐annual time scales. Our study also shows that during the training, the UNET learns information such as the seasonal cycle of the relationship between sea‐level pressure and temperature anomalies, which could explain part of its excellent scores. This exploratory work opens up promising prospects for estimating the contribution of atmospheric variability to observed temperature variations.

The Detection and Attribution Model Intercomparison Project (DAMIP) coordinates single forcing climate model simulations for detection and attribution analysis and other applications. DAMIP simulations were carried out with fifteen climate models as part of CMIP6, and these simulations were used in at least 270 published articles. These simulations were also used directly in at least five chapters of the IPCC Sixth Assessment Working Group I Report, and they underpinned the estimate of anthropogenic attributable warming highlighted in the Summary for Policymakers of that report, and quoted directly in the UNFCCC Glasgow Climate Pact. For CMIP7, natural-only, well-mixed greenhouse gas-only, and aerosol-only simulations have been proposed as fast track DAMIP simulations, and planning of a broader set of experiments is currently underway. This talk will highlight key DAMIP results from CMIP6, and will discuss plans for the CMIP7 version of DAMIP. Comments and suggestions regarding the CMIP7 DAMIP experimental design will be welcomed.

Europe is one of the fastest-warming region of the world and temperatures of the recent years have been systematically higher than best estimates of the forced response. This may be due to internal variability in favor of warmer situations, or it may indicate that the forced response is underestimated.In Europe, inter-annual temperature variations are primarily linked to the variability of North Atlantic atmospheric dynamics. The temperature T for a given day and year can be written as the sum of the forced response µ (the non-stationary climate normal) and the internal variability D + ε (D the atmospheric dynamics and ε the residual).We investigate two methods for estimating the forced response in transient simulations, via a denoising of the temperature T from the dynamical term D. To test these methods, we use a perfect model framework, here the large ensemble of 50 MIROC6 transient simulations. The mean of the large ensemble provides an accurate estimate of the forced response (the « truth »), to which estimates from individual members can be compared.The contribution of the D term is first estimated with a circulation analogues method. We reconstruct the temperatures of one individual member from similar atmospheric situations of the other 49 members. The analogues are calculated on the T-µ series, rather than on the T series directly.Second, we reconstruct temperatures using deep neural networks. Using a U-NET, we estimate the function f in the equation T=µ+f(X)+ ε, where X is the sea level pressure. Our network is trained on the different members of the large ensemble.