ADVANCES IN ATMOSPHERIC SCIENCES

Celebrating the Tenth Anniversary of Fudan University’s Discipline of Atmospheric Sciences: Research Highlights and Progress

<p<In April 2016, Fudan University formally established the discipline of Atmospheric Sciences and concurrently founded the Institute of Atmospheric Sciences. Two years later, the Department of Atmospheric and Oceanic Sciences was established. Since its inception, both the Institute and Department have successfully built an internationally competitive research team in the fields of weather, climate, climate change and environmental sciences by recruiting leading experts and systematically integrating existing academic resources within the university. Over the past decade, through deep interdisciplinary collaboration among atmospheric sciences, oceanography, mathematics, and physics, Fudan University has achieved a series of significant breakthroughs in four major research areas: climate change and its mechanisms; extreme weather and climate events and their impact assessments; interactions and effects of Earth’s multi-sphere systems; and observations and numerical simulations of the ocean–atmosphere system. These advancements have not only demonstrated scientific importance, but also provided critical scientific support for China’s strategies to address climate change and enhance disaster prevention and mitigation capabilities. The disciplines of atmospheric and oceanic sciences at Fudan University have continuously served major national strategic needs, such as ecological civilization construction, regional water and climate security development, as well as talent cultivation.</p<

ADVANCES IN ATMOSPHERIC SCIENCES. 2026 43(11): 2293-2319.

Artificial Intelligence for Atmospheric and Oceanic Forecasting and Research: A Fudan University Perspective

<p<Rapid advances in artificial intelligence (AI), together with the rapid growth of observational and reanalysis datasets, are reshaping the atmospheric and oceanic sciences. Based on representative studies from the Department of Atmospheric and Oceanic Sciences and the Artificial Intelligence Innovation and Incubation Institute at Fudan University, this paper reviews recent AI applications in atmospheric and oceanic sciences from a Fudan University perspective. We highlight the development of AI models for atmosphere, ocean and cloud systems, represented respectively by FuXi, Xihe, and DaYu, together with related advances in AI-based data assimilation, forecast utilization, hybrid modeling, and end-to-end forecasting systems. Beyond forecasting, we further discuss Fudan-related work that uses AI as a scientific tool, including studies on interpretability, predictability analysis, causality diagnosis, and emerging multimodal foundation models for Earth system science. Current challenges facing AI applications in the atmospheric and oceanic sciences, including physical consistency, extreme event prediction, generalization and long horizon stability, data limitations, and computational cost are discussed. Finally, we outline future directions for integrating AI with physics-based understanding in the atmospheric and oceanic sciences. Overall, this perspective suggests that AI should be viewed not only as a tool for improving prediction, but also as a framework for advancing scientific understanding of atmospheric and oceanic processes.</p<

ADVANCES IN ATMOSPHERIC SCIENCES. 2026 43(11): 2320-2338.

Responses of the Tropical Easterly Jet to Different Global Warming Patterns and Implications for Future Tropical Cyclone Activity

<p<The tropical easterly jet (TEJ) is a key upper-tropospheric circulation feature that influences regional climate and tropical cyclone (TC) activity. This study examines how the TEJ responds to distinct global warming patterns and the implications for future TC activity using high-resolution simulations from an atmospheric general circulation model. We find that La Niña-like warming strengthens the TEJ over the western Pacific but weakens it over the Indian Ocean, whereas El Niño-like warming weakens the TEJ in both regions. These contrasting responses are linked to pattern-dependent changes in the Walker circulation and associated convection heating patterns. Basin-separation experiments under El Niño-like warming further reveal that Indian Ocean warming exerts dominant control over the western Pacific TEJ (WP_TEJ), Pacific Ocean warming has the strongest influence on the Indian Ocean TEJ (IO_TEJ), while Atlantic SST anomalies provide a secondary influence on both jets through cross-basin teleconnections. A weakened Indian Ocean IO TEJ reduces easterly vertical wind shear, creating a more favorable environment for TC activity in the northern Indian Ocean during the transition seasons, and supporting more frequent intense storms. Over the western North Pacific, a weakened WP TEJ under El Niño-like warming is associated with reduced upper-level divergence, whereas La Niña-like warming strengthens this jet and is linked to a more favorable large-scale environment for TC formation. These findings emphasize that the spatial structure of sea surface temperature change, and the relative roles of individual ocean basins, must be accounted for in reliable projections of regional atmospheric circulation and future TC risk.</p<

ADVANCES IN ATMOSPHERIC SCIENCES. 2026 43(11): 2339-2349.

Stratospheric Harbingers of Boreal Summer Climate: The Role of Lower Stratospheric Processes in the Trans-seasonal Connection between Spring and Summer Northern Annular Modes

<p<Research over recent decades has fundamentally shifted our understanding of the polar stratosphere’s role in the boreal surface climate from a passive layer to an active driver of weather and climate extremes. The polar stratosphere’s influences often project onto major climate patterns such as the Northern Annular Mode (NAM), leading to predictable surface impacts such as cold air outbreaks. However, such top-down influences are considered contingent on the existence of the stratospheric polar vortex in the winter half-year. In this study, we report a significant negative correlation between the March and high summer (July and August) NAM during the past four decades. The lower stratosphere acts as a signal transmitter and converter: the negative NAM-like anomalies in the troposphere in March induce a stratospheric response in April that evolves into a positive NAM-like pattern in late spring propagating downward to the upper troposphere in June and maintaining through the summer. The nudged simulations with separate tropospheric and stratospheric states reveal the dominant role of the lower stratosphere in this trans-seasonal linkage. These results provide new insights into the seasonal prediction of the summer NAM and boreal surface climate.</p<

ADVANCES IN ATMOSPHERIC SCIENCES. 2026 43(11): 2350-2362.

Decadal Shift of Teleconnections Intensifies Summertime High-Ozone Episodes in China

<p<Summertime high-ozone pollution is a significant environmental concern for the North China Plain (NCP), affecting millions of residents. High-ozone episodes over the NCP often coincide with ridge formations, blocking the propagation of Rossby waves to the NCP, thereby creating favorable conditions for photochemical production of ozone. In recent years, more frequent ridge occurrences have intensified the emission-driven ozone problem over the NCP. We find that the climate factors driving the ridge formation have shifted over the decades. In the last two decades of the 20th century, wave trains induced by the low-latitude North Atlantic SST anomalies were the main driver of ridge frequencies. However, in more recent decades, the Eurasia teleconnection associated with Arctic sea ice variations over the Gakkel Ridge and warming anomalies in the central Indian Ocean had a greater influence. Based on CMIP6 projections, our study indicates that under the high-emission SSP5-8.5 scenario, the ridge frequency will likely increase, thereby amplifying meteorologically driven high-ozone pollution over the NCP.</p<

ADVANCES IN ATMOSPHERIC SCIENCES. 2026 43(11): 2363-2374.

Diagnosing Dominant Drivers of Sulfate Simulation Uncertainties with the EPICC Model: An Integrated Assessment of Chemical Mechanisms and Emission Inventories

<p<Sulfate (${\mathrm{SO}}_4^{2-} $) aerosol is a key driver of severe haze pollution in China, yet atmospheric chemical transport models (CTMs) typically underestimate its concentrations, and the dominant formation pathways remain highly controversial. Clarifying the specific model uncertainties responsible for this is critical for accurate haze forecasting and effective control measures. Here, we apply a diagnostic framework, integrating the EPICC (Emission and atmospheric Processes Integrated and Coupled Community) model with Bayesian Monte Carlo simulations and a Reduced-Form Model, to diagnose the sources of ${\mathrm{SO}}_4^{2-} $ aerosol simulations in China. Our results reveal a distinct seasonal shift in the dominant uncertainty sources of ${\mathrm{SO}}_4^{2-} $ simulations. Wintertime uncertainties are driven by the Mn-catalyzed heterogeneous pathway (MnHET, 45%–60%) and ammonia (NH<sub<3</sub<) emissions (11%–27%). These winter uncertainties are modulated by meteorology, peaking under high-RH and low-temperature conditions, explaining why models fail specifically during severe haze episodes. In contrast, summertime uncertainties associated with current model parameters are dominated by inaccuracies in emission inventories, particularly for condensable particulate matter (CPM, 25%–50%) and sulfur dioxide (SO<sub<2</sub<, 14%–36%), while an underestimated sulfur oxidation ratio also points to the contribution of missing chemical mechanisms. Incorporating the MnHET pathway and correcting for underestimated NH<sub<3</sub< and missing CPM emissions largely resolves the wintertime ${\mathrm{SO}}_4^{2-} $ simulation bias. We demonstrate that underestimating NH<sub<3</sub< causes a misattribution of the dominant ${\mathrm{SO}}_4^{2-} $ formation pathways by suppressing the Mn-catalyzed pathway. Furthermore, incorporating the MnHET pathway reveals that NH<sub<3</sub< emission reduction is a more effective haze mitigation strategy than conventional models suggest. This study highlights the critical need for accurate emission inventories and provides a robust scientific basis for developing the EPICC model and other CTMs.</p<

ADVANCES IN ATMOSPHERIC SCIENCES. 2026 43(11): 2375-2391.

Dominant Mechanisms of Sulfate Formation across Pollution Regimes in China: A GEOS-Chem Modeling Study

<p<Sulfate, a major component of secondary inorganic aerosols, substantially impacts environment and climate. However, atmospheric models often underestimate sulfate formation because key heterogeneous pathways remain incompletely represented, leading to large simulation biases and uncertainty in dominant mechanisms across multiple pollution regimes. We therefore incorporate dust heterogeneous chemistry and key reactions under haze conditions into GEOS-Chem, building a comprehensive framework to assess their impacts on sulfate. Compared to baseline, this framework markedly improves sulfate simulation for a 2018 dust event, 2016 haze episode, and the 2018 annual mean, reducing NMBs (normalized mean biases) from −62.76%, −41.97% and −35.01% to 0.44%, 14.85% and −16.52%, respectively, outperforming the simulation only including dust heterogeneous chemistry (−19.05%). Process diagnostics show that during the 2018 dust event, sulfate formation is dominated by SO<sub<2</sub<+OH oxidation on coarse-mode particles. Post-event, decreasing dust and OH weaken the dust chemistry and elevate the NH<sub<3</sub< by buffering aerosol acidity to strengthen particle-surface NO<sub<2</sub< oxidation. During the 2016 haze episode, heterogeneous chemistry dominates sulfate formation (67.41%), with Mn-catalyzed oxidation contributing 17.11%, lower than previous estimates. Sensitivity simulations indicate this reduction is driven by competition with dust chemistry through SO<sub<2</sub< consumption and lower simulated Mn concentrations. In terms of the 2018 annual mean, dominant sulfate formation pathways exhibit pronounced regional contrasts. In the Sichuan Basin, gas-/aqueous-phase oxidation (54.37%) and dust-phase SO<sub<2</sub<+O<sub<3</sub< oxidation (26.19%) prevail. In the Beijing–Tianjin–Hebei region, dust chemistry dominates (38.07%), followed by particle-surface NO<sub<2</sub< oxidation (6.96%) and black carbon–catalyzed oxidation (9.57%). Sulfate formation in the Yangtze River Delta region exhibits similar patterns but weaker heterogeneous contributions. Overall, the results show that explicitly representing the heterogeneous chemistry across dust and haze regimes improves sulfate simulation and provides a scientific basis for improving atmospheric chemistry parameterizations and informing region-specific sulfur mitigation strategies.</p<

ADVANCES IN ATMOSPHERIC SCIENCES. 2026 43(11): 2392-2412.

Dataset Construction of CO<sub<2</sub< Flux over Shifting Sand in the Taklimakan Desert and Assessment of Carbon Sequestration Capacity

<p<Desert ecosystems, which cover extensive global regions, play an essential yet often-overlooked role in sequestering atmospheric CO<sub<2</sub< at substantial yet gradual rates, primarily due to insufficient monitoring, data scarcity, and limited research attention. Current assessments of terrestrial carbon sinks often exclude desert areas, creating critical “data gaps” that undermine the completeness and accuracy of global carbon budget evaluations. Focusing on the Taklimakan Desert, we developed a high-resolution CO<sub<2</sub< flux dataset (2000–22) for shifting sand areas, featuring hourly temporal resolution and 0.25° × 0.25° spatial coverage, through an empirical estimation scheme incorporating internal carbon sequestration processes, ERA5 multi-layer soil temperature/moisture data, and extensive field-measured soil properties (soil organic carbon, pH). Our analysis reveals that spatiotemporal CO<sub<2</sub< flux patterns are jointly controlled by shifting sand temperature, moisture, and soil characteristics, with the shifting sand area of the entire desert sequestering approximately 1.05 × 10<sup<6</sup< t CO<sub<2</sub< annually. While two-thirds of the area functions as carbon sinks, the northwestern margins (one-third of the total area) exhibit carbon-source behavior (peaking at 64.49 g m<sup<−2</sup< yr<sup<−1</sup<) due to higher precipitation and organic carbon content. The ongoing warm–wetting trend in northwestern China is reducing both the stability and rate of carbon sequestration in these shifting sands—processes that may accelerate through climate change feedbacks. These findings provide critical scientific evidence for assessing the role of the desert in climate mitigation and inform policy-making for carbon neutrality strategies.</p<

ADVANCES IN ATMOSPHERIC SCIENCES. 2026 43(11): 2413-2425.

In Situ Measurement of Dust Particle Transport to the Tropopause Layer over the Northern Tibetan Plateau

<p<Previous satellite observations and simulations have shown that dust particles can be transported to the upper troposphere, but there is very little evidence from in situ observations. This study provides the first balloon-borne measurements of dust particle transport to the tropopause layer during a dust event over the northern Tibetan Plateau (TP). Vertical profiles of size-resolved aerosol number distributions were obtained before and during the dust event, which revealed that the number of dust particles larger than 0.416 μm increased significantly throughout the troposphere and above the tropopause at the sounding site. The backward trajectories of these enhanced dust particles suggest that the particles at different altitudes exhibit different transport pathways and dust emission sources. In the lower troposphere, the dust particles originated from the Taklamakan Desert following flow to the northern TP. The particles above the tropopause originated from the Thar Desert. They experienced both synoptic-scale lifting associated with the trough and jet stream over the Thar Desert and forced uplift due to the high terrain of the TP.</p<

ADVANCES IN ATMOSPHERIC SCIENCES. 2026 43(11): 2426-2435.

Warming-induced Decrease in Snowfall Fraction Shifts Runoff Seasonality over the Tibetan Plateau

<p<Climate warming facilitates a transition from snowfall to rainfall, altering snowpack dynamics and affecting runoff seasonality. The Tibetan Plateau, located at the westerly–monsoon transition zone, exhibits varied impacts of snowmelt on runoff across different climatic regions. Using monthly runoff data from nine snow-affected river basins, the runoff seasonality response to snowfall fraction changes is investigated during the historical (1998–2017) and future (2021–60) periods. Distinct seasonal runoff responses to decreasing snowfall fraction are observed across basins dominated by different climate regimes, with a critical snowfall fraction threshold of 0.5. Specifically, the center timing of runoff mass in westerly dominated basins shifts earlier, while it is delayed in monsoon-dominated basins. Seasonal runoff variability decreases with reduced snowfall fraction in westerly dominated basins, but increases in monsoon-dominated basins, primarily due to asymmetric cold–warm season runoff responses. In westerly basins, declining snowfall reduces both warm- and cold-season runoff, dampening the variability. In monsoon basins, warm-season runoff increases while cold-season runoff declines, amplifying the seasonality. This shift is driven by snowfall fraction effects on snowpack dynamics. Below the threshold, rapid spring snowmelt occurs with sensible heat flux dominance. Above the threshold, the persistent snowpack shifts energy partitioning toward latent heat flux, delaying and moderating peak runoff. Future projections indicate that runoff timing will advance by 22–29 days with reduced variability in westerly dominated basins, but delay by 18–20 days with increased variability in monsoon-dominated basins. Warming-induced earlier snowmelt will increase spring runoff, while rising evapotranspiration reduces summer and autumn water availability, exacerbating seasonal water shortages.</p<

ADVANCES IN ATMOSPHERIC SCIENCES. 2026 43(11): 2436-2454.

Enhanced Internal Moisture Redistribution as the Primary Driver for Summer Wetting over the Inner Tibetan Plateau since the 1990s

<p<Summer precipitation over the inner Tibetan Plateau (TP) has increased markedly since the 1990s, leading to widespread lake expansion and exerting profound impacts on local ecosystems and infrastructure. By integrating an Eulerian moisture back-tracking model (Water Accounting Model: WAM-2layers) with large-scale circulation diagnostics for 1979–2021, we find that intensified internal moisture redistribution plays a more prominent role in driving the wetting trend compared to changes in external moisture transport or local evaporation. Specifically, the eastern TP (ETP) has emerged as the dominant contributor, with its moisture supply increasing by 7.01 mm mon<sup<–1</sup< (22.4%) since 1999, accounting for 36.7% of the total precipitation increment (19.10 mm mon<sup<–1</sup<). This contribution exceeds those from both southern and western external sources. Notably, this enhanced ETP contribution is driven by anomalous easterly winds rather than local thermodynamic factors. Such a shift coincides precisely with the interdecadal regime transition of the Silk Road Pattern (SRP) from a positive to a negative phase around 1998/1999. The negative SRP phase anchored an anomalous anticyclone to the north flank of the TP, triggering the robust low-level easterly anomalies that redirected moisture from the ETP westward into the inner TP. These results highlight that intensified internal water recycling is a major mechanism reshaping the regional hydrological balance and accelerating lake expansion in the inner TP.</p<

ADVANCES IN ATMOSPHERIC SCIENCES. 2026 43(11): 2455-2467.

Turbulence Intermittency Characteristics and Energy Balance over Three Distinct Land Surfaces

<p<Turbulence intermittency is a frontier challenge in atmospheric turbulence research, affecting turbulent flux estimation and surface energy balance nonclosure. This study uses atmospheric boundary layer observations from three typical underlying surfaces in northern China to quantitatively investigate the characteristics of turbulence intermittency and its impact on energy balance closure. Particular attention is given to the consistency and variability of these characteristics across different surface types. The results indicate that turbulence intermittency exhibits broadly similar characteristics and effects over the three underlying surfaces. During burst periods of turbulence intermittency events, turbulent flux transport is enhanced, whereas sub-mesoscale motions contribute less to transport than turbulent motions. The opposite tendency is observed during quiescent periods. Stronger intermittency generally weakens turbulent flux transport. Empirical relationships are established between turbulence intermittency intensity, mean turbulent transport capacity, and background meteorological fields. The contribution of turbulence intermittency to surface energy balance closure appears to be a common feature across different surfaces. During burst periods, surface energy residuals decrease by 1.91–44.86 W m<sup<−2</sup<, and the mean closure ratios improve by 3%–39%. Tazhong station, which represents an arid region with strongly heterogeneous and thermally forced underlying surfaces, exhibits the strongest and most frequent turbulence intermittency. It also shows that more frequent co-occurrence of dynamical and thermal intermittency is associated with greater reductions in residuals and more pronounced improvements in closure. These findings may provide a basis for numerical simulations of turbulence intermittency and for refining turbulence parameterization schemes in atmospheric models.</p<

ADVANCES IN ATMOSPHERIC SCIENCES. 2026 43(11): 2468-2483.

Climate Responses to Vegetation Greening in China and India and Their Combined Effects

<p<Global vegetation greening is a prominent trend, with China and India as major contributors. However, the individual and combined climatic impacts of greening in these two regions remain insufficiently understood. Based on satellite data from 1982 to 2020, we selected regions with significant greening in China and India as sensitive areas, and used a coupled land–atmosphere model (CAM6.0-CLM5.0) to simulate the biophysical climate effects under three scenarios: individual China greening; individual India greening; synergistic vegetation greening in both regions. Results show that vegetation greening in China induces a local cooling effect, accompanied by regionally heterogeneous precipitation responses, with increased precipitation over southern and northeastern China and reduced precipitation over central eastern China. In contrast, greening in India leads to local warming and a widespread reduction in precipitation. When greening occurs simultaneously in both regions, more than half of the land area exhibits climate responses consistent with the sign-consistent superposition of the individual effects. Notably, the synergistic greening produces a robust cooling and wetting response over southern China and a consistent cooling over western Eurasia, while warming dominates over northern North America. Further analysis indicates that these climate responses are associated with changes in surface energy fluxes, cloud cover, diabatic heating, and large-scale atmospheric circulation, including modifications to Rossby wave propagation and the upper-level jet. Our findings highlight the role of cross-regional vegetation–atmosphere interactions and reveal the influence of low- to mid-latitude vegetation change on climate variability in northern high latitudes.</p<

ADVANCES IN ATMOSPHERIC SCIENCES. 2026 43(11): 2484-2499.

Poleward Redistribution of Tropical Cyclone Hazards over East Asia

<p<Tropical cyclones (TCs) in the western North Pacific have undergone notable changes over recent decades alongside ocean warming and large-scale atmospheric circulation variability. Using TC best-track data from 1980 to 2025, we identify a pronounced poleward redistribution of TC occurrence across East Asia. This redistribution is characterized by decreasing landfall likelihood at lower latitudes and increasing likelihood at higher latitudes, despite a decline in basin-wide TC genesis frequency. Consistent with this pattern, TC activity exhibits a poleward shift, with reduced occurrence south of 30°N and enhanced activity at higher latitudes, accompanied by more landward- and poleward-oriented tracks. The ratio of landfalling TCs decreases south of 20°N but increases north of 30°N, providing direct quantitative evidence for this redistribution. In addition, TC translation speed has increased modestly, leading to shorter oceanic durations prior to landfall, particularly during the most recent decade, which further modulates coastal hazard patterns. In contrast, no significant long-term trend is found in lifetime maximum intensity, and mean intensity shows a slight weakening. These results suggest that recent changes in TC hazards are primarily driven by shifts in TC tracks and translation characteristics rather than systematic intensification. Overall, our findings highlight a coherent poleward redistribution of TC hazards along the East Asian coast, rather than a uniform increase, with important implications for regional coastal hazard patterns.</p<

ADVANCES IN ATMOSPHERIC SCIENCES. 2026 43(11): 2500-2517.

Rad3DGFM: A Deep Learning Network for Completing Data Gaps in 3D Radar Reflectivity Mosaics

<p<Three-dimensional (3D) radar reflectivity mosaic products generated from weather radar networks, such as the Severe Weather Automatic Nowcasting (SWAN) system of the China Meteorological Administration, provide essential information for severe convective weather nowcasting, short-term forecasting, and quantitative precipitation estimation. However, these products often contain lower-level missing-data gaps caused mainly by Earth’s curvature and terrain blockage of radar beams. To address this problem, a Radar 3D Gap-Filling Model (Rad3DGFM) was developed in this study to reconstruct missing reflectivity information. Rad3DGFM is built upon a 3D U-Net architecture integrated with an attention mechanism and an enhanced conditional generative adversarial network (CGAN). The CGAN is improved by using a least-squares adversarial loss, label smoothing, and Gaussian-noise-perturbed discriminator training, together with MAE (mean absolute error) and MSE (mean squared error) reconstruction constraints in the gap regions. This design improves training stability and helps the model generate radar reflectivity fields that are both structurally realistic and quantitatively accurate. The experimental results based on 6562 test samples demonstrate that Rad3DGFM effectively reconstructs the missing data for all precipitation intensities. For reflectivity values exceeding 35 dB<italic<Z</italic<, the model achieves POD (probability of detection), FAR (false alarm ratio), TS (threat score), and FSS (fraction skill score) values of 0.71, 0.13, 0.67, and 0.81, respectively. Furthermore, the model performs robustly when applied to regions with different terrain conditions and data gaps from the training area, suggesting that Rad3DGFM has the potential to be applied in different regions without additional training.</p<

ADVANCES IN ATMOSPHERIC SCIENCES. 2026 43(11): 2518-2535.

Sensible Heat and Momentum Flux in the Sea-Ice-Covered Southern Ocean Based on a Shipborne Observation Experiment

<p<In the Southern Ocean’s marginal ice zone (MIZ), sea ice complicates accurate estimation of air–sea fluxes, limiting the predictive skill of climate models. From December 2020 to January 2021, the 37th Chinese National Antarctic Research Expedition conducted ship-based measurements in the MIZ using an EC (eddy covariance) system, focusing on sensible heat and momentum fluxes between the air and the sea. A dataset has been compiled and made publicly available. It includes meteorological variables, energy and momentum fluxes, and turbulence variables (friction velocity and atmospheric stability), offering valuable insights for studies of ocean–ice–atmosphere interactions and climate modeling. This study also details refinements to the COARE3.5 algorithm informed by the observations, leading to improved flux estimates.</p<

ADVANCES IN ATMOSPHERIC SCIENCES. 2026 43(11): 2536-2547.