The drivers and mechanisms of the decadal change in summer precipitation over eastern China have been widely investigated. In this context, the potential impact of vegetation changes on the Qinghai–Xizang Plateau (QXP) and its associated processes remain unclear. Using the GIMMS (Global Inventory Modeling and Mapping Studies) Leaf Area Index (LAI) dataset and gridded precipitation data over China, this study investigated the relationship between the decadal increase in spring vegetation cover in the QXP and the decadal change in summer precipitation over eastern China, and the associated mechanism. The results revealed that the late-spring (May) vegetation in the QXP has undergone notable decadal changes around 2010, with LAI anomalies turning from negative to positive. Concurrently, there has been a noticeable decadal change in the summer precipitation over eastern China after 2010, with a shift from wet to dry situations over the Huaihe River basin and South China, and precipitation anomalies turning from negative to positive over Northeast China, North China, and the middle and lower reaches of the Yangtze River basin. A composite analysis of the changes in the surface energy budget over the QXP and the East Asian summer monsoon circulation showed that, as the vegetation cover in the QXP increased in the late spring after 2010, the net surface shortwave radiation increased in response to the low-albedo characteristics of vegetation, leading to an increase in the surface temperature over the QXP and consequently enhancing the upward surface sensible heat flux. The stronger spring sensible heat flux provides a stronger atmospheric heating source in the subsequent summer, intensifying and extending the South Asian High eastward. This effect is accompanied by the strengthening and westward movement of the Western Pacific Subtropical High, thereby changing the decadal summer precipitation over eastern China around 2010. Thus, this study has elucidated the important modulating role of vegetation change in the QXP on the decadal change in summer precipitation over eastern China.

Using ground station observation data, ERA5 reanalysis data, and GPCP global precipitation data, this paper takes the regional heavy rainstorm caused by the coupling of the Qinghai–Xizang Plateau vortex and the Southwest China vortex on June 4–7, 2019, as an example to diagnose and analyze the coupling characteristics and formation mechanism of precipitation. The results show that (1) the weather process was caused by the combined action of the Qinghai–Xizang Plateau vortex and the Southwest China vortex, and the precipitation developed explosively from night to early morning. On June 4, the two vortices produced precipitation centers in the northwest and southwest, respectively, and the precipitation was located in the rear of the Qinghai–Xizang Plateau vortex and the center of the Southwest China vortex. The two vortices were coupled in the early morning of June 5–6, and the precipitation shifted to the front of the vortex, strengthening to its peak. The center of the heavy rainstorm was north of Hubei Province. On the night of 6, the two vortices moved eastward toward the sea, and the precipitation area disappeared on land. (2) During the evolution of the two vortices, when the plateau vortex moved eastward from the plateau, the divergence term and the vertical transport term enhanced the positive vorticity of the basin. At the same time, as the low-level convergence configuration strengthened, the southwest vortex was further induced to develop in the basin, and precipitation occurred in both vortex regions, providing a key dynamic condition for the subsequent coupling of heavy rain. (3) During the coupling development, the centers of the two cyclones at different heights merged to form a deep convective system, which generated strong positive vorticity columns and strong convective instability, accompanied by vertical upward motion and secondary circulation development. The southwest monsoon and the southerly airflow outside the subtropical high transported abundant water vapor to the front of the vortex, and under the synergistic effect of deep shear convergence and strong dynamic uplift, this process eventually led to heavy rain.

This study utilizes daily minimum temperature and relative humidity data from China’s CN05.1 dataset (1982–2022), along with ERA5 reanalysis data, to examine the spatiotemporal distribution characteristics and potential causes of two types of compound extreme low-temperature events in southern China during winter. The results indicate that cold–dry events are distributed across most of southern China, with high-frequency centers in the eastern and Jianghuai regions, whereas cold–wet events exhibit a south-to-north decreasing trend, occurring primarily in the southwestern and western South China regions. Over the past 41 years (1982–2022), cold–dry events have shown an increasing trend in most southern regions of China, whereas cold–wet events have significantly decreased, with notable interannual variability in both event types. Cold–dry events typically occur in early winter, while cold–wet events are more likely to occur in late winter. Further analysis of the different causes of these two event types reveals that early and late winter are associated with distinct atmospheric circulation systems and moisture conditions. The mid-to-high-latitude oblique ridge over Asia and the East Asian trough are key circulation features for cold–dry events, providing favorable conditions for the southward intrusion of cold air. During this period, pronounced moisture divergence over southern China promotes the occurrence of cold–dry events. In contrast, during cold–wet events, the northern branch of the wave pattern maintains a two-trough, one-ridge circulation over the middle and high latitudes of the Eastern Hemisphere. The southern branch of the wave pattern, propagating through West Asia and the Middle East, deepens the southern trough and interacts with the anomalous anticyclone over the western Pacific, transporting warm, moist air to southern China. This leads to moisture convergence in the region, providing abundant moisture for cold–wet events.

The characteristics of convective-scale updrafts in the outer core of tropical cyclones (TCs) simulated under lower- and upper-layer vertical wind shear are compared in this study. The results reveal notable differences between the two experiments. (1) The lower-layer shear experiment features a larger proportion of deep updrafts with top heights above 12 km; vertical mass transport and the maximum vertical velocity associated with outer core updrafts are statistically significantly larger in the lower-layer sheared TC than in the upper-layer sheared TC; this difference is related to the larger convective effective potential energy, equivalent temperature, and water vapor mixing ratio in the boundary layer. Furthermore, interactions among the lower-layer vertical wind shear, local wind shear induced by the circulation of the TC, and surface cold pools are conducive to the initiation and development of updrafts. (2) The radial tilts of updrafts are associated with the updraft top height and asymmetric radial flows. Deep-penetrating updrafts are prevalent in the lower-layer shear experiment, tilting radially outward with height in the downshear quadrants under upper-level radial outflow and inward in the upshear quadrants under upper-level radial inflow. However, in the upper-layer shear experiment, most updrafts tend to tilt radially inward with height, primarily driven by asymmetric radial inflow at upper levels in the downshear-right and upshear quadrants and at lower-to-middle levels in the downshear-left quadrant. (3) In the lower-layer shear experiment, relatively symmetric low-level radial inflow prevails, driving most updrafts to migrate inward toward the TC eye across all quadrants. However, in the upper-layer shear experiment, updrafts in the upshear quadrants tend to move radially outward under the influence of radial outflow in the lower levels, whereas those in the downshear quadrants tend to move radially inward under the radial inflow in the lower levels. (4) In the lower-layer sheared TC, most updrafts exhibit large positive thermal buoyancy, large negative water loading, and large negative dynamic buoyancy; hence, the total buoyancy ranges from negative to positive. In comparison, in the upper-layer sheared TC, most updrafts exhibit smaller positive thermal buoyancy, smaller positive dynamic buoyancy, and smaller negative water loading, resulting in smaller positive total buoyancy.

This study employs ERA5 reanalysis data, station precipitation data from the China Meteorological Administration, NOAA monthly average SST (sea surface temperature), and outgoing longwave radiation flux data from 1980 to 2020 to analyze the impact of zonal position of EASJ (East Asian subtropical jet) on summer precipitation in China, as well as its response to SST in the tropical central-eastern Pacific. The results indicate that in westward-shift years of the EASJ, upper-level convergence and lower-level divergence play an important role in YRV (Yangtze River valley). The eastward displacement and weakening of WPSH (western Pacific subtropical high) at 500 hPa suppress moisture transport to the YRV, leading to reduced precipitation. Meanwhile, upper-level divergence over the Hetao area promotes rising motion, accompanied by lower-level moisture convergence, increasing precipitation in the Hetao area. Meanwhile, in eastward-shift years of the EASJ, the WPSH strengthens and shifts westward. The convergence and divergence patterns in the upper and lower troposphere, along with moisture conditions and vertical motion in these two regions, exhibit reversed characteristics compared with those in westward-shift years of the EASJ. As a result, the precipitation anomalies of eastward-shift years of the EASJ display an opposite phase distribution compared with in westward-shift years of the EASJ. The EASJ’s zonal position is closely related to the summer SST in the tropical central-eastern Pacific. In this region, negative (positive) SST anomalies directly cool (warm) the atmosphere, suppressing (enhancing) local convective activity. These alterations modulate Walker circulation and lead to anomalous ascending (descending) motion in the tropical western Pacific. The vertical motion anomalies further affect East Asia’s mid-latitude region through meridional circulation, reducing (intensifying) convective activity near 30°N. Consequently, the meridional temperature gradient in the eastern part of the climatological jet region decreases (increases) due to diabatic heating, weakening (strengthening) the tropospheric westerlies and resulting in an anomalous westward (eastward) EASJ.

Millimeter-wavelength cloud radar serves as a pivotal instrument for cloud microphysical characterization but is susceptible to clear-air echoes induced by insects, birds, and aerosols, leading to considerable retrieval biases. To enhance detection accuracy and generalizability, this study presents ATNN (Multi-Head Attention Neural Networks), a universal deep learning framework that exclusively employs the reflectivity factor, radial velocity, and relative height for adaptive learning of vertical structural features. Evaluated on observations from Baise, Guangxi, ATNN demonstrates robust performance during the rainy season, achieving a clear-air echo detection rate of 91.98% and a meteorological echo preservation rate of 94.28%; during the non-flood season transition, these metrics are 87.56% and 86.38%, respectively. In contrast, conventional neural networks (NN) achieve marginally higher detection rates but misclassify numerous low-altitude, weak meteorological echoes as clear-air signals, yielding a preservation rate of only 54.96% in non-flood periods. Nationwide validation across 15 operational radar sites further confirms that ATNN attains an average clear-air echo removal rate of 81.17% while preserving 93.89% of meteorological echoes, thereby demonstrating strong regional adaptability.

Based on subseasonal-to-seasonal (S2S) datasets obtained from four models (ECMWF, CMA, UKMO, and NCEP) and ground-observed precipitation data for April–September 1997–2022, an improved deep residual network (ResNet) model was developed for the extended-range probabilistic forecasting of persistent extreme precipitation (PEP). The deterministic and probabilistic forecasts of the improved ResNet model were compared with those of five other methods (single-model ensemble, multi-model ensemble, principal component regression, and Bayesian model averaging). Results demonstrated that the improved ResNet model outperformed other methods in probabilistic forecasting. The enhanced performance of the ResNet model is attributed to the improvement in spatial correlation coefficients, which stems from its ability to deeply extract PEP features from the ECMWF S2S data and the optimization of the ResNet structure. As a result, the improved ResNet model achieves a positive Brier skill score for forecast lead times up to 30 d, highlighting its considerable potential for extended-range probabilistic forecasting of PEP in eastern China.

Southwestern China experienced a typical compound heat–drought event between mid-August and mid-September 2024, which significantly impacted the local economy and livelihoods. In this region, such compound events are related to intraseasonal oscillations (ISOs) originating from different areas. However, conventional statistical methods fail to distinguish the specific contributions of various ISOs to these extreme conditions. Using atmospheric-circulation diagnostics and numerical simulations using the Weather Research and Forecasting model with varying-ISO-signal-driven boundary forcings, this study reveals the following: the ISO signal induced by a mid-latitude westerly wave train in the upper troposphere increased the regional average temperature by 1.331°C in Southwestern China, primarily affecting the northern part of the region. This ISO signal accounted for 62.8% of the overall compound heat–drought event. The ISO signal associated with a middle-tropospheric teleconnection wave train increased regional average temperatures by 0.768°C. This impact is mainly concentrated in the eastern Sichuan Basin and most of Chongqing, accounting for 37.2% of the event. The development of this compound event was predominantly influenced by a mid-latitude westerly wave train from the northwestern boundary. This promoted the significant expansion of the South Asian High over Southwestern China, resulting in strong positive geopotential height anomalies. Simultaneously, warm advection due to the westward extension and northward shift of the Western North Pacific Subtropical High, coupled with positive tropospheric geopotential height anomalies, contributed to regional temperature increases and decreased soil moisture. Overall, this study enhances the understanding of the mechanisms driving compound heat–drought events in Southwestern China, providing a foundation for improving subseasonal-prediction capabilities.

Uncertainties in cloud simulations can affect key physical processes in climate modelling. Using satellite data from CloudSat/CALIPSO and CERES, together with ERA5 reanalysis data, this study analyzes the distribution characteristics of global climatological cloud fraction and cloud radiative effects (CREs) simulated by the atmospheric component FAMIL in the climate system model FGOALS-f3 at resolutions of 100, 25, and 12.5 km. Moreover, the study investigates the potential sources of simulation errors based on the Xu–Randall cloud diagnostic scheme. The results show that at different resolutions, cloud fraction is underestimated in low-latitude areas and overestimated in high-latitude areas and that the simulation performance changes nonlinearly with increasing resolution. On the global scale, the deviation in the 25 km model increases compared with the 100 km model, and the cloud fraction deviation of the 12.5 km model is approximately 4% higher than that of the 25 km model in regions such as the Western Pacific Warm Pool. As model resolution increases, the simulated CRE decreases in association with a reduction in cloud fraction. In the Xu–Randall scheme, relative humidity (RH) is the dominant factor controlling climatological cloud fraction. Resolution enhancement further changes RH by affecting variables such as temperature and vertical velocity, thereby indirectly affecting cloud fraction bias. The weak vertical velocity in the 100 km model indirectly leads to underestimation of middle- and high-level cloud fractions in low-latitude areas. The higher temperature in the 25 km model compared with the 100 km model increases the cloud fraction bias. Furthermore, the lower temperature in the 12.5 km model compared with the 25 km model decreases the cloud fraction bias.

The El Niño–Southern Oscillation (ENSO) represents the most prominent interannual signal within the global ocean–atmosphere system and stands as one of the crucial factors influencing the interannual variability of summer precipitation in the Beijing–Tianjin–Hebei (BTH) region. Based on data such as spring SST (Sea Surface Temperature), summer precipitation in the BTH region, and atmospheric circulation from 1961 to 2023, this study analyzes the interdecadal variation in the relationship between spring ENSO and summer precipitation in the BTH region as well as its possible causes. Results show that the key spring ENSO region significantly impacting the interannual summer-precipitation anomalies in the BTH region underwent an interdecadal shift in the late 1990s, migrating from the eastern Pacific in the earlier period to the central Pacific in the later period. Although SST anomalies (SSTAs) in the key spring ENSO region exhibit a strong negative correlation with the summer precipitation anomalies in the BTH region before and after the shift, the underlying mechanisms differ significantly. When SSTAs exist in the key spring ENSO region, the summer SSTAs in the North Pacific, Indian Ocean, and Atlantic Ocean show different characteristics, which in turn cause circulation anomalies in the tropics and mid-high latitudes through the synergistic effect of SSTAs in different regions. When the key spring ENSO region exhibits warm SSTAs, a distinct anticyclonic circulation is observed in the subtropical northwestern Pacific. After the interdecadal shift, this anticyclonic circulation strengthens, but it still fails to transport water vapor to the BTH region. The BTH region is controlled by a cyclonic circulation. In the earlier period, this cyclonic circulation is associated with SSTAs in the North Pacific, and in the later period, it is caused by the zonal wave train over the Eurasian continent, excited by SSTAs in the North Atlantic. The control of anomalous northerly winds over the BTH region reduces precipitation.

Dual-satellite observations of FY-4A and FY-4B exhibit continuity and complementarity, but differences in the satellites’ positions and observation paths may make their actual detection results inconsistent. This study, which is based on the AGRI full-disk Level 1 data and cloud top height products from dual-satellite imagers during June–August 2022, conducted parallax correction on the dual-satellite data and analyzed differences in thermal infrared channel brightness temperatures (BTs) under various cloud height scenarios and underlying surface conditions. The results showed that within the common observation area, the BT differences between FY-4A and FY-4B gradually turned positive to negative from west to east. In the equatorial region, the BT differences reached 8 K. As cloud height increases, the root mean square error of the BT differences increases from 1.63 K to 3.06 K. The differences were larger over land than over the ocean. Additionally, this study investigated the theoretical effects of satellite zenith angle on dual-satellite BT differences using the RTTOV fast radiative transfer model. Simulated BT difference trends between the satellites show broad consistency with observed empirical trends. Therefore, future research can perform physical correction on the dual-satellite observations according to these patterns, enabling joint applications of the dual-satellite data.

To investigate the mobility of the Southwest China Vortex (SWV) and its associated rainstorm regions, this study examined 25 SWV-related rainstorm events that affected Chongqing during 2019–2023. A dynamic compositing method that tracks the movement of the vortex center was employed to analyze the dynamic characteristics and genesis of the SWV and the distribution of heavy rainfall relative to the vortex center. The conclusions are as follows: SWV rainstorms can be categorized into three types: southeastern, eastern, and multi-quadrant types. The distribution of heavy rainfall for each SWV type is closely related to the position of the upper-level jet, the intensity and meridional extent of the trough, the exit region of low-level jets, and the distribution of warm-air and cold-air advection, while the thermodynamic and dynamic characteristics exhibit distinct similarities and differences. The southeastern type is influenced by the short-lived SWV with weaker upward motion, shallow vertical development, and a pronounced northwestward tilt. The eastern type involves the long-lived, intense SWV with deeper vertical development and a low-level northwestward tilt. The multi-quadrant type is influenced by the SWV with sustained cold-air intrusion from the rear and strong uplift movement. The configuration features symmetric winds in the front and rear, pronounced cold- and warm-air convergence, the deepest vertical development among the three types, and a low-level northward tilt.

Aircraft turbulence is primarily caused by atmospheric turbulence, which is closely related to changes in wind speed, temperature, and topography. This phenomenon is especially pronounced in the southwestern region, where the complex terrain and the significant uplifting effect of the plateau on air currents make aircraft highly susceptible to turbulence. The turbulence event experienced by flight MU5954, operating from Tengchong to Kunming on 4 April 2024, was analyzed using reanalysis data, topography data, satellite data, and aircraft in-flight reports. The results showed that the turbulence occurred near the boundary between the windward and leeward slopes of the mountains, where the air currents were unstable. The alternating positive and negative distributions of vertical velocity, divergence, and vorticity in the upper-level jet stream, combined with the convergence of cold and warm airflows, jointly triggered atmospheric turbulence, making the aircraft susceptible to turbulence during flight.

The China Meteorological Administration Mesoscale Model (CMA-MESO) serves as a critical operational system for precipitation forecasting. However, its eight-times-daily rolling updates present new challenges for forecast evaluation. A comprehensive understanding of mean bias characteristics, variations across different forecast lead times, and error sources is key to model improvement. This study systematically evaluated the precipitation forecast performance of the CMA-MESO over East China during the summer of 2023. The key findings are as follows: (1) While the averaged multiple forecasts generally reproduced the spatial distribution patterns of observed precipitation, systematic biases existed in precipitation intensity and frequency. (2) The model significantly overestimated the precipitation amount for moderate-and-above rainfall events, with 3.7% underestimation of the frequency of no-rainfall events but 2.6% overestimation of the frequency of light rain events, along with elevated false alarm rates for heavy-and-above rainfall events. (3) The forecast performance showed a distinct lead-time dependence, with precipitation biases largest in 3～6 h nowcasting, gradually decreasing to a minimum in 21～24 h forecasts, and then increasing again in the 24～36 h range. (4) Physical analysis revealed that the nowcasting overestimation stemmed from excessive hydrometeor introduction during cloud analysis, and the overestimation in the forecast after 24 h resulted from excessive low-level wind speeds that promoted moisture transport to the target region.

As urbanization accelerates, hazards including flash floods, debris flows, and urban inundation are becoming more frequent and severe, with escalating social disruption and economic losses. A scientific and effective radar quantitative precipitation estimation (QPE) product is essential for enhancing the monitoring, forecasting, and early warning of extreme precipitation, which is of great value to current disaster prevention and mitigation efforts. This study used data from a single S-band dual-polarization radar and six X-band dual-polarization phased array radars in Jiaxing to develop S-band and X-band radar QPE products for the region. After preprocessing the radar observation data, including removing nonmeteorological echoes and performing quality control of the differential propagation phase, the effective minimum hybrid elevation for each radar was calculated. Then, the rain rate was estimated using single radar observations and raindrop size distribution parameters provided by the Dual-frequency Precipitation Radar onboard the Global Precipitation Measurement satellite (GPM-DPR). Finally, the rain rate was mosaicked for multiple radars, forming a gridded QPE product with a temporal resolution of 1 min and a spatial resolution of 50 m. To verify the accuracy of the proposed method, two heavy precipitation events that occurred on 10 and 16 September 2024, in the Jiaxing area were selected to evaluate the radar QPE products. The evaluated products included the original Jiaxing X-band QPE product and the X-band, S-band, and mosaic products developed in this study. The results show that the proposed QPE products outperform the original X-band product in terms of accuracy and stability, with the mosaic product yielding the best estimation result. For localized heavy rainfall with horizontally inhomogeneous distribution, the mosaicking product accurately reflects the precipitation distribution across various towns (subdistricts), contributing to the early issuance of accurate warnings.

Two large-hail supercells that occurred before spring in Zhejiang, on February 21, 2024 (“24·2·21”), and March 3, 2025 (“25·3·3”), were analyzed using the reanalysis data, automatic weather station observations, lightning location data, and S-band dual-polarization radar data. The results indicate that both convection events occurred after the passage of a cold wave front and were elevated convection systems. Above the inversion layer, conditionally unstable stratification and positive convective available potential energy were present. Deep convection was rapidly triggered under the influence of the 850 hPa shear line. Strong, deep-layer vertical wind shear favored the development of supercells. The dry layer in the “25·3·3” case was deeper than that in the “24·2·21” case. In both supercell hail processes, hailfall was accompanied by a decrease in storm centroid height and in the maximum shear height of the mesocyclone. The “24·2·21” supercell exhibited a long maintenance time, a mid-level double vortex structure, and differential reflectivity (Z_DR) and specific differential phase (K_DP) columns extending above 6 km. The Z_DR and K_DP values within the overhanging echo were small, and the cross-correlation coefficient (C_C) decreased with decreasing height, indicating the presence of melting small ice phase particles, which favored moist hail growth. In contrast, the “25·3·3” supercell had a shorter maintenance time, a deep single vortex (mesocyclone) structure, and a pronounced rear inflow. The Z_DR and K_DP columns were confined below 5 km. Small Z_DR and K_DP values accompanied by large C_C values at upper levels indicated a greater abundance of dry ice phase particles. This was manifested by an increased frequency of intracloud lightning flashes, which may be associated with accelerated freezing caused by rear-inflow dry cold air and was conducive to dry hail growth in the upper cold cloud region.

Using multi-level observation data of the near-surface layer on the Weizhou Island, Beihai, Guangxi, this study examined the periods influenced by typhoons Yagi, Sanba, and Talim during 2023–2024. First, quality control and correction were conducted on the observation data for each event. Subsequently, the structural characteristics of the typhoon boundary layer under the influence of the complex underlying surface of the Beibu Gulf islands were analyzed. The analysis included near-surface wind field evolution, wind profile characteristics, and variations in aerodynamic parameters, namely, friction velocity (u_*), roughness length (z₀), and drag coefficient (C_d), along with near-surface wind speed and direction. The results indicate that the typhoon boundary layer within 10–32 m exhibits near-neutral stability and the mean wind speed profile follows a logarithmic distribution. Under the influence of the complex underlying surface of the Beibu Gulf islands, the variations in u_*, z₀, and C_d with near-surface wind speed exhibit characteristics typical of a land underlying surface, showing obvious nonmonotonicity, and cannot be simply categorized as those typical of a homogeneous marine underlying surface. Specifically, u_* ranges between 0.5 and 1.5 m s⁻¹; when the 10 m wind speed exceeds approximately 15 m s⁻¹, the magnitudes of z₀ and C_d are on the order of 10⁻² m and 10⁻³, respectively. The u_*, z₀, and C_d values are generally higher under easterly winds than under westerly winds. The findings contribute to a deeper understanding of typhoon boundary layer structures over the complex underlying surface of an island and hold significance for optimizing numerical simulations and parameterization schemes of typhoons in the Beibu Gulf.

In the context of global warming, the spatiotemporal correlation between extreme drought and heatwave events has intensified over time. Their compounding effects not only amplified their individual impacts but also exacerbated the pressure on water resources, agriculture, and ecological environments. Based on the fifth-generation reanalysis data from the European Centre for Medium-Range Weather Forecasts (ERA5), this study calculated the standardized precipitation evapotranspiration index (SPEI) and the standardized temperature index (STI) and employed the Frank Copula function to construct the standardized compound dry and hot index (SCDHI). This approach systematically revealed the spatiotemporal evolution characteristics of compound hot–dry events in the Mongolian Plateau between 1982 and 2023. Additionally, the study integrated a random forest regression model to identify and quantify the key climatic factors that influenced compound hot–dry events in the region. The results indicated that over the past four decades, the intensity, the time proportion, and the affected area proportion of compound hot–dry events in the Mongolian Plateau exhibited a significant decreasing trend. Among the climatic variables, STI was the most influential factor affecting the compound events, with a feature importance value of approximately 0.5. This was followed by SPEI, temperature, and potential evapotranspiration (PET), with feature importance values of 0.4, 0.1, and 0.1, respectively, while precipitation had a relatively weak impact. This study underscored the need to incorporate the long-term impacts of compound hot–dry events into the protection and management of the grassland ecosystem of the Mongolian Plateau. It emphasized the importance of developing adaptive management strategies based on scientific predictions and risk assessments to enhance the resilience and sustainability of the ecosystem.

Using surface dense automatic precipitation observations, ERA5 reanalysis data, and radar composite reflectivity data, this study analyzes the causes of a heavy rainstorm event over the western Sichuan Basin on July 11, 2023, focusing on the impacts of inertial oscillation on wind fields and terrain–airflow interactions on rainfall. A physical conceptual model of this rainstorm event is also developed. The results show that this precipitation event exhibited three prominent characteristics: a linear distribution of heavy rainfall along mountainous terrain, pronounced nocturnal development, and strong convective activity. Before the heavy rainstorm, the Sichuan region was under the control of the South Asian High, while the basin was influenced by southerly airflow on the periphery of the Western Pacific Subtropical High, with no considerable weather systems affecting the area. Quantitative analysis revealed that the Coriolis force and geostrophic deviation terms showed large variations. While atmospheric advection and vertical transport were weak, inertial oscillation dominated the wind speed increase in the central–eastern Sichuan Basin and eastern Yunnan–Guizhou Plateau. This process led to nocturnal boundary layer jets and clockwise rotation of the ageostrophic wind. The development of southeastward ageostrophic winds at night enhanced the southeasterly flow over the basin. The superposition of jet transport from the eastern Yunnan–Guizhou Plateau and wind speed growth in the central–eastern basin enhanced wind convergence in the basin. The combined effects of the divergence term and vertical transport term promoted the development of cyclonic wind fields. Under the combined influence of airflow in the northern sector of the cyclonic wind field and ageostrophic winds, southeasterly winds persisted over the northwestern basin, playing a crucial role in the formation of nocturnal rainfall. The southeasterly winds transported warm, moist air westward, leading to considerable wind and moisture convergence along the mountains of the western basin. The intrusion of dry air at mid-to-upper levels combined with low-level warm and moist airflow enhanced convective instability. The perpendicular intersection of southeasterly winds with the mountains of the western basin produced strong upslope flow. Orographic lifting on windward slopes and the formation of secondary vertical circulations triggered convection ahead of the mountains. As nocturnal southeasterlies intensified, convergence ahead of the mountains strengthened, and convection continued to develop. The persistent southeasterlies combined with topographic blocking and cyclonic wind field development, generated northerly winds along the western basin margin and easterly winds along the northern edge. Their convergence with easterly/southeasterly flows formed wind shear lines that promoted convective organization. The steep terrain transition from the basin to the mountains and the perpendicular intersection of the southeasterlies with the northeast–southwest-oriented mountains in the western basin favored strong upward motion on windward slopes. Sensitivity experiments conducted using a multisource observation data assimilation system and the WRF model demonstrated that halving terrain slopes reduced precipitation intensity by one order of magnitude. This decrease clearly indicates the enhancement effect of steep topography on rainfall. Under the combined forcing of low-level shear lines and steep terrain, a quasi-linear intense convective system developed along the mountains, forming a linear downpour zone along the western basin’s foothills. With the interaction of persistent southeasterly flow and the unique topography, efficient precipitation persisted, ultimately causing the extreme rainfall event.

An extratropical cyclone is an important weather system in mid-latitude regions and has a profound impact on global weather and climate. The frontal system is a key component of the cyclone, and its accurate identification is crucial for improving the accuracy of weather forecasts. In recent years, deep learning methods have been applied to automatic frontal recognition, but research on frontal identification over the Eurasian continent has been limited, with a focus on detecting single-type fronts. Achieving simultaneous recognition and classification of cold and warm fronts remains a major challenge. Therefore, this paper proposes a method for the synchronous recognition of cold and warm fronts based on the Detection Transformer (DETR) model called DETR-FRONT (DETR-FRO). This method constructs RGB feature images using meteorological variables, such as sea level pressure, temperature advection, temperature, and relative humidity at 850 hPa. The DETR model is trained using an artificially created cold and warm front label set, enabling automatic synchronous recognition and classification of both types of fronts. DETR-FRO improves the original DETR model by enhancing its transmission channel and incorporating multiple meteorological variables to construct RGB feature images. This modification strengthens the ability of the model to capture and express frontal features, leading to improved recognition performance for cold and warm fronts. Multidimensional validation results show that DETR-FRO can simultaneously identify cold and warm fronts in extratropical cyclones over the Eurasian continent with high accuracy and stability. In a typical snowstorm case study, the DETR-FRO model accurately reconstructs the cold and warm fronts and their evolution processes. The recognition results are consistent with dynamic and thermodynamic fields, further confirming the model’s generalization ability and meteorological interpretability. This improvement provides new technical support for weather forecasting and the study of extratropical cyclones. In addition, the long-term frontal dataset created using DETR-FRO, covering the period from 1995 to 2024, reveals the seasonal distribution characteristics of cold and warm front frequencies in extratropical cyclones over the Eurasian continent at the climate scale. This information contributes to a deeper understanding of the climatic characteristics of cold and warm fronts in extratropical cyclones.

The downslope-enhanced convective storms in the Beijing urban area triggered destructive winds that reached 12 on the Beaufort scale from 1300 to 1600 (Beijing time. the same below) on May 30, 2024, causing severe damage, including the toppling of utility poles and large trees with a diameter at breast height exceeding 1 m. This study combines damage surveys with polarimetric Doppler weather radar observations to investigate the characteristics of the disaster and the physical processes associated with the severe winds. The results show that the destructive winds exhibited westward divergent patterns, with extreme winds generated by microbursts and downbursts on the sub-100-meter scale, reaching at least EF1 intensity and, locally, reaching EF2 intensity. Observational analyses reveal that enhanced low-level southwesterly winds and sustained convergence zones along the mountainous areas facilitated moisture accumulation in Beijing’s western foothills 1–2 h before the storms moved downslope, with the Global Positioning System observing increased precipitable water vapor. The downslope movement of the storms essentially resulted from gust fronts interacting with mountain-foot convergence zones, triggering new convective storms that rapidly developed into a squall line with reflectivity levels exceeding 55 dBZ within 18 min. Physical process analysis indicates that damaging winds in areas such as Bei’anhe were caused by downbursts embedded within cold pool-driven gust fronts. Extreme winds near Yuanmingyuan additionally involved a low-level γ-mesoscale vortex, whose path closely matched the severely damaged areas, demonstrating the direct impact of the vortex on localized wind intensification. The radar detected radial velocities reaching 53 m s⁻¹ (at an altitude of 0.38 km) following velocity de-aliasing at 1430 BJT, together with features resembling tornadic debris, albeit with a relatively weak echo intensity. Quantitative estimation of different physical processes suggests that cold pool outflow contributed approximately 20 m s⁻¹, whereas the low-level γ-mesoscale vortex and downbursts each enhanced near-surface winds by 15–18 m s⁻¹.