Sea ice and snow are crucial components of the cryosphere and the climate system. Both sea ice and spring snow in the Northern Hemisphere (NH) have been decreasing at an alarming rate in a changing climate. Changes in NH sea ice and snow have been linked with a variety of climate and weather extremes including cold spells, heatwaves, droughts and wildfires. Understanding of these linkages will benefit the predictions of climate and weather extremes. However, existing work on this has been largely fragmented and is subject to large uncertainties in physical pathways and methodologies. This has prevented further substantial progress in attributing climate and weather extremes to sea ice and snow change, and will potentially risk the loss of a critical window for effective climate change mitigation. In this review, we synthesize the current progress in attributing climate and weather extremes to sea ice and snow change by evaluating the observed linkages, their physical pathways and uncertainties in these pathways, and suggesting ways forward for future research efforts. By adopting the same framework for both sea ice and snow, we highlight their combined influence and the cryospheric feedback to the climate system. We suggest that future research will benefit from improving observational networks, addressing the causality and complexity of the linkages using multiple lines of evidence, adopting large-ensemble approaches and artificial intelligence, achieving synergy between different methodologies/disciplines, widening the context, and coordinated international collaboration.
There is debate about the use of wetlands as natural climate solutions due to their ability to act as a "double-edged sword" with respect to climate impacts by both sequestering CO2 while emitting CH4. Here, we used a process-based greenhouse gas (GHG) perturbation model to simulate wetland radiative forcing and temperature change associated with wetland state conversion over 500 years based on empirical carbon flux measurements, and CO2-equivalent (CO2-e.q.) metrics to assess the net flux of GHGs from wetlands on a comparable basis. Three CO2-e.q. metrics were used to describe the relative radiative impact of CO2 and CH4-the conventional global warming potential (GWP) that looks at pulse GHG emissions over a fixed timeframe, the sustained-flux GWP (SGWP) that looks at sustained GHG emissions over a fixed timeframe, and GWP* that explicitly accounts for changes in the radiative forcing of CH4 over time (initially more potent but then diminishing after about a decade)-against model-derived mean temperature profiles. GWP* most closely estimated the mean temperature profiles associated with net wetland GHG emissions. Using the GWP*, intact wetlands serve as net CO2-e.q. carbon sinks and deliver net cooling effects on the climate. Prioritizing the conservation of intact wetlands is a cost-effective approach with immediate climate benefits that align with the Paris Agreement and the Intergovernmental Panel on Climate Change timeline of net-zero GHG emissions by 2050. Restoration of wetlands also has immediate climate benefits (reduced warming), but with the majority of climate benefits (cooling) occurring over longer timescales, making it an effective short and long-term natural climate solution with additional co-benefits.
The Arctic-Boreal region is vulnerable to rapid climate change. Thawing of the permafrost and extended warm periods are expected to turn the region into a hotspot of enhanced CH4 emissions. We estimated CH4 fluxes by assimilating atmospheric CH4 mixing ratio data from a regional network into a global atmospheric inverse model, resulting in a mean uncertainty reduction of ~68% across the domain and improved agreement with observations. From 2010 to 2021, the Arctic-Boreal region emitted 45.4 ± 0.7 TgCH4 y-1 ( ~ 7% of global emissions), with no significant overall trend. However, on the regional scale a positive trend emerged in the Western Siberia Lowlands. Arctic-Boreal wetland emissions increased during warmer years, suggesting possible future increases as warming continues. Emissions varied regionally, with western Russia showing higher fluxes. Changes in winter hydroclimate significantly influenced emissions in the Western Siberian Lowlands, likely by enhancing the availability of soil moisture through snowmelt in spring. Our findings emphasize the importance of accounting for both temperature effects and changes in wetness, when assessing Arctic CH4 future emissions.
Statistical downscaling translates coarse-resolution climate model output into locally relevant information for climate services and impact assessment. Recent advances in artificial intelligence (AI) enable high-resolution, probabilistic, and computationally efficient approaches. This paper provides a perspective on the evolution from classical to AI-driven and hybrid downscaling approaches, assesses key challenges related to interpretability, uncertainty, data availability, and computational requirements, and outlines physically constrained and generative frameworks that support decision-making across sectors.
Indoor air quality (IAQ) is increasingly recognised as one of the most important aspects for public health, workplace safety and productivity. While indoor and outdoor factors both influence indoor pollutant levels, human presence and activity are key drivers of the emission of specific pollutants, including particulate matter (PM), total volatile organic compounds (TVOCs) and carbon dioxide (CO2). This study investigates the relationship between occupancy, physical activity measured by kinetic energy (KE), and air pollution concentrations in a real-world office setting, by combining data from air quality and radar motion sensors. Two exemplar office spaces were investigated, comprising an open-office area and a meeting room. PM, in the PM1 and PM2.5 size fractions, were found to be highly correlated with the outdoor conditions, whereas PM10 correlates more closely with indoor occupancy (up to r = 0.65). Even higher correlations, up to r = 0.74, were found between human activity, quantified as KE, and the PM10 concentrations. The TVOCs and CO2 showed even stronger correlations with KE (up to r = 0.83), suggesting that these metrics can be used as excellent proxies for estimating certain types of indoor air pollution. Notably, the impact of additional occupants varies depending on room characteristics and usage, underscoring the need for contextualised models of IAQ. By quantifying both outdoor infiltration and indoor emissions, this study offers a framework for disentangling pollutant sources and guiding interventions to optimise IAQ. These insights support evidence-based strategies to create healthier and more productive office environments.
The hydrological cycle in South America during austral summer, including extreme precipitation and floods, is significantly influenced by northerly low-level jets (LLJs) along the eastern Andes. These synoptic weather events have been associated with three different types of LLJs (Central, Northern, and Andes) and are sensitive to remote large-scale forcings. This study investigates how tropical forcings related to El Niño/Southern Oscillation (ENSO) and Pacific Decadal Oscillation (PDO) regulate the duration and frequency of each LLJ type and their impacts on extreme precipitation. Our analysis reveals that ENSO and PDO are important in driving the variability of LLJs over the past 65 years. Specifically, the Central LLJ type is more prevalent during El Niño and Warm/Neutral PDO phases, leading to heightened extreme precipitation in southern South America. Conversely, La Niña years during Cold PDO phases tend to favor the Northern and Andes LLJs, which are associated with increased precipitation extremes in the western Amazon and southeastern South America. Central and Andes LLJs tend to persist longer during these favored conditions, causing more pronounced precipitation events in the areas under their influence. This study enhances our understanding of the influence of large-scale atmospheric forcings on the regional precipitation dynamics in South America.
Earth system models inform water policy and interventions, but knowledge gaps in hydrologic representations limit the credibility of projections and impacts assessments. The literature does not provide conclusive evidence that incorporating higher resolutions, comprehensive process models, and latest parameterization schemes, will result in improvements. We compare hydroclimate representations and runoff projections across two generations of Coupled Modeling Intercomparison Project (CMIP) models, specifically, CMIP5 and CMIP6, with gridded runoff from Global Runoff Reconstruction (GRUN) and ECMWF Reanalysis V5 (ERA5) as benchmarks. Our results show that systematic embedding of the best available process models and parameterizations, together with finer resolutions, improve runoff projections with uncertainty characterizations in 30 of the largest rivers worldwide in a mechanistically explainable manner. The more skillful CMIP6 models suggest that, following the mid-range SSP370 emissions scenario, 40% of the rivers will exhibit decreased runoff by 2100, impacting 850 million people.
During July-August 2024, East Antarctica experienced the most intense winter heatwave in the 46-year satellite era, with regional mean surface air temperatures across Dronning Maud Land exceeding the climatological mean by more than 9°C for 17 consecutive days. To explore the physical drivers and quantify the anthropogenic contribution to this unprecedented event, we propose a multi-model, multi-method attribution framework integrating regional climate model-based storyline attribution, circulation analogues, and large-ensemble probabilistic attribution. The results show that a pronounced weakening of the stratospheric polar vortex initiated a quasi-barotropic high-pressure anomaly, which enhanced meridional heat and moisture transport and accounted for approximately 50% of the observed surface warming. Across different models and attribution methods, synthesis of the attribution results indicates that anthropogenic warming intensified the event by approximately 0.7°C and more than doubled the likelihood of such exceptional winter heatwaves in the current climate. Probabilistic attribution further indicates that, compared to a natural climate without human influence, the likelihood of such events increases from 2-3 times today to ~6 times under moderate emissions and up to 26 times under high emissions by 2100. These findings reveal how human-induced warming is transforming even the coldest regions, with implications for ice shelf stability and predictability of future Antarctic extremes.
Paleoclimate records provide a critical long-term perspective on natural climate variability, essential for understanding contemporary climate variations. However, existing paleoclimate proxies lack sufficient spatial-temporal coverage for studying high-impact weather extremes like tropical cyclones (TCs). Here we introduce a multi-source framework that contextualizes the contemporary TC landfalls in East Asia against a multi-century baseline (1368-1911) reconstructed from historical documents. Leveraging pre-industrial and contemporary-era data, the analysis reveals that the relatively small shift toward earlier landfalls in the contemporary era (1946-2020) falls well within the range of fluctuations documented historically (1651-1900). Rather than indicating detectable anthropogenic changes, these results suggest the dominance of natural variability in modulating landfall timing. Our work also suggests consistent natural controls of TC timing in contemporary and pre-industrial eras. This consistency lends credibility to pre-industrial observational datasets and climate simulations, providing a robust template for assessing changes in the seasonality of high-impact extremes.
Earth's climate sensitivity quantifies the ultimate change in global mean surface air temperature in response to a doubling of atmospheric CO2 concentrations. Recent assessments estimate that Earth's climate sensitivity very likely lies between 2.3 °C and 4.7 °C, with the representation of clouds in climate models accounting for a large portion of its uncertainty. Here, we adjust the climate sensitivity of individual contemporary climate models after using satellite observations to alleviate biases in their representation of mixed-phase clouds. A resulting moderate average climate sensitivity of 3.63 ± 0.98(1σ) °C arises due to opposing responses of clouds. While increasing the proportion of liquid within cold clouds prior to CO2 doubling increases climate sensitivity via transitions from solid to liquid hydrometeors, a strongly opposing increase in reflective cloud cover decreases climate sensitivity. This emphasizes the need to reconsider the role of mixed-phase cloud cover changes in climate sensitivity assessments.
Climate records have been broken with alarming regularity in recent years, but the events of 2023-2024 were exceptional even when accounting for recent climatic trends. Here we quantify these events across multiple variables and show how excess energy accumulation in the Earth system drove the exceptional conditions. Key factors were the positive decadal trend in Earth's Energy Imbalance (EEI), persistent La Niña conditions beginning in 2020, and the switch to El Niño in 2023. Between 2022 and 2023, the heating from EEI was over 75% larger than during the onset of similar recent El Niño events. We show further how regional processes shaped distinct patterns of record-breaking sea surface temperatures in individual ocean basins. If the recent trend in EEI is maintained, we argue that natural fluctuations such as ENSO cycles will increasingly lead to amplified, record-breaking impacts, with 2023-2024 serving as a glimpse of future climate extremes.
Internal climate variability (ICV) remains a major source of uncertainty in climate projections, complicating impact assessments across critical sectors, especially at stakeholder-relevant scales. Given that ICV emerges from the nonlinear interactions of the climate system, we argue that nonlinear dynamical (NLD) approaches can improve its characterization, providing physically interpretable insights that strengthen adaptation strategies and support multisector decision-making. However, despite their suitability for such problems, NLD approaches remain largely underutilized in the analysis of initial condition large ensembles (LEs). We argue that a diverse suite of NLD approaches offers a promising pathway for systematically extracting robust insights from LEs. If effectively applied and systematically integrated, these methods could fully harness the potential of LEs, uncovering underlying patterns and variability across ensemble members to refine fundamental insights from climate projections. This will help bridge the gap between complex climate dynamics and practical resilience strategies, ensuring that decision-makers, resource managers, and infrastructure planners have a more reliable foundation for navigating irreducible uncertainty.
Previous modeling studies have indicated that Atmospheric rivers (ARs) will become more frequent in the warming climate. However, whether we have experienced more ARs during historical period is less studied. Here, we show that winter AR frequency has significantly increased over the mid-latitude Northern Hemisphere from 1950-2022. Using station-based observations, we confirm that ARs have driven coherent long-term trends in both total and extreme precipitation over land. The warming-induced rise in atmospheric moisture alone accounts for an AR frequency increase of ~0.6-0.8% per decade. AR trends exhibit meridional dipolar patterns over western North America and Europe, governed by positive trends of the Pacific-North America Pattern and North Atlantic Oscillation. The "Pineapple Express" ARs have been suppressed, declining of AR landfalling over the Pacific Northwest. Through atmospheric model analyses, we demonstrate that observed sea surface temperature changes dominate Atlantic AR trends, while exerting minor effects on Pacific AR trends.
Heatwaves are one of the leading causes of climate-induced mortality. Using the examples of recent heatwaves in Europe, the United States and Asia, we illustrate how the communication of dangerous conditions based on temperature maps alone can lead to insufficient societal perception of health risks. Comparison of maximum daily values of temperature with physiological heat stress indices accounting for impacts of both temperature and humidity, illustrates substantial differences in geographical extent and timing of their respective peak values during these recent events. This signals the need to revisit how meteorological heatwaves and their expected impacts are communicated. Close collaboration between climate and medical communities is needed to select the best heat stress indicators, establish them operationally, and introduce them to the public. npj Climate and Atmospheric Science (2023) 6:33.
Previous studies suggested that atmospheric water vapor, as a greenhouse gas, can amplify surface warming by enhancing downwelling longwave radiation (DLR), thereby potentially intensifying extreme heatwaves. However, the atmospheric water vapor anomalies associated with heatwaves and their radiative effects remain poorly quantified. Here, using radiative kernels to decompose DLR anomalies, we demonstrate that atmospheric water vapor exerts a substantial influence on summertime interannual-scale heatwave-related DLR across the Northern Hemisphere, despite DLR being largely dominated by near-surface air warming. This influence exhibits pronounced regional contrasts: mid‑ to high‑latitude heatwaves are typically accompanied by moist air, which enhances DLR, whereas heatwaves in India and western North America often coincide with drier conditions that reduce DLR and thereby partially offset the positive DLR anomalies caused by air warming. The dryness over India is linked to an interannual weakening of the Indian summer monsoon, while that over western North America is associated with soil‑moisture deficits. Our findings provide a quantitative assessment of the contribution of atmospheric moisture to extreme heatwaves.
Glaciers in Alaska contribute greatly to sea-level rise and are losing mass at a faster rate than any other region. Yet, our understanding of ongoing changes and ability to model them are hindered by a lack of observations, particularly at high spatiotemporal resolution. Here, we leverage Sentinel-1 synthetic aperture radar (SAR) data to produce temporally-varying glacier melt extents and snowlines from mid-2016 to 2024 for 99% of glaciers in Alaska greater than 2 km2. The melt extents are strongly correlated with temperatures, revealing that each 1°C increase in summer temperature causes up to 3 additional weeks of glacier melt. The high spatiotemporal resolution also captures subseasonal changes such as the 2019 heat wave, which caused subregional snowlines to retreat up to 105 m higher and exposed up to 28% more of the underlying glacier compared to typical years. Our snowlines agree well with optical datasets (r2 up to 0.94), thus providing unprecedented reliable data unencumbered by clouds or lighting conditions. Moving forward, our automated, open-source workflow can easily be applied to other regions. These data also present unique opportunities to calibrate and validate large-scale glacier evolution models, a critical step for improving projections of glacier changes and their impacts.
Atmospheric rivers (ARs) are key agents of poleward heat and moisture transport, yet the extent to which tropical moisture directly feeds AR precipitation at higher latitudes remains debated. Here we present a forty-year global climatology of moisture sources for AR-related precipitation, combining deep learning-based AR detection with Lagrangian moisture tracking. We reveal that AR precipitation worldwide organizes into two coherent and contrasting moisture transport regimes. Along western continental boundaries, AR precipitation is predominantly sustained by extratropical moisture convergence, with tropical contributions typically limited to 30-40%, whereas eastern boundaries exhibit shorter transport pathways with substantially larger tropical contributions, averaging 60-70%. These persistent west-east asymmetries reflect systematic shifts in moisture origin as ARs evolve from early subtropical acquisition to mature extratropical transport. We further show a strong negative correlation between tropical contribution and latitude, demonstrating that direct tropical moisture delivery is not a requirement for AR precipitation at high latitudes. Instead, tropical moisture primarily preconditions AR formation, while the water mass that ultimately precipitates is progressively replaced along the poleward pathway. Taken together, our findings reconcile divergent assessments of tropical moisture influence on AR precipitation and thus deepen our understanding of how Earth's hydrological engine redistributes atmospheric water across the planet.
As the ocean absorbs over 90% of excess radiative heat, recent ocean warming is shaped by a combination of anthropogenic surface heat gain and dynamical processes redistributing heat. To distinguish these contributions, we introduce a novel framework that decomposes temperature changes into three components: 'spice' (density-compensated variability) and 'heave' (density-contributing variability), with heave further divided into 'passive' (net warming) and 'dynamic' (redistribution) contributions. Passive heave captures anthropogenic warming subducted along isopycnals, while spice and dynamic heave, which globally sum to zero, represent heat redistribution. Observations and climate models demonstrate general agreement on passive heave, establishing it as a key oceanic fingerprint of anthropogenic climate change. In contrast, dynamic heave, driven by interannual-to-decadal variability, exhibits significant spatial heterogeneity, with notable discrepancies between models and observations. This framework links ocean heat uptake to sea-level change, with passive heave driving global thermosteric rise and dynamic heave contributing to regional dynamic sea level changes.
While it is widely believed that the intense rainfall in summer 2022 over Pakistan was substantially exacerbated by anthropogenic climate change1,2, climate models struggled to confirm this3,4. Using a high-resolution operational seasonal forecasting system that successfully predicted the extreme wet conditions, we perform counterfactual experiments simulating pre-industrial and future conditions. Both experiments also exhibit strong anomalous rainfall, indicating a limited role of CO2-induced forcing. We attribute 10% of the total rainfall to historical increases in CO2 and ocean temperature. However, further increases in the future suggest a weak mean precipitation reduction but with increased variability. By decomposing rainfall and large-scale circulation into CO2 and SST-related signals, we illustrate a tendency for these signals to compensate each other in future scenarios. This suggests that historical CO2 impacts may not reliably predict future responses. Accurately capturing local dynamics is therefore essential for regional climate adaptation planning and for informing loss and damage discussions.
Organic aerosol particles (OA) can absorb solar radiation with varying efficiencies depending on their chemical composition and physical properties. This light-absorbing fraction of OA, commonly referred to as brown carbon (BrC), is difficult to accurately represent in climate models due to the inherent diversity of its optical properties. This variability arises from differences in emission sources and atmospheric processing, as well as from variations in experimental design and the analytical methods used to quantify BrC absorption. As a result, the climate effect of BrC remains uncertain. Here, we studied the light absorption properties of surface ambient OA using measurements from 17 sites across Europe. Combining multi-wavelength absorption measurements from filter-based photometers with OA mass concentrations and source apportionment derived from ACSM/AMS data, we derive empirical estimates of the OA mass absorption cross section (MACOA), its wavelength dependence (AAEOA), the OA density (⍴OA), and the MAC associated with different primary and secondary OA sources. We further develop parameterizations that relate MACOA, AAEOA and ⍴OA to the ambient black carbon-to-organic aerosol ratio (eBC/OA) and propose a corresponding parameterization for the imaginary refractive index (kOA). Given the widespread availability of eBC and OA measurements in global monitoring networks, the framework presented here provides a practical approach for estimating the absorptive properties of surface OA particles under real-world conditions.