Abstract:Aiming at the problems that the existing evaluation methods of the Shenyang vertical observation system are imperfect and mostly rely on single-equipment data, and the insufficient exploration of the synergistic application value of multi-source vertical observation data in the monitoring and early warning of severe convective weather under cold vortex conditions. This study intends to establish a multi-element data quality evaluation system suitable for the cold vortex circulation background. Meanwhile, it explores the monitoring performance of multi-device collaborative observations during cold-vortex severe convective rainstorms, so as to further enhance the regional three-dimensional vertical monitoring and nowcasting capabilities for cold vortex weather. Based on the observation data from the Shenyang Ground-based Remote Sensing Vertical Observation Comprehensive Station, this paper takes the extremely heavy regional rainstorm process in Liaoning from July 23 to 28, 2024 as an example, combines radiosonde observation data, establishes a multi-element data quality evaluation system for vertical observation suitable for the cold vortex background, and carries out systematic evaluation and application analysis. The results show that: (1) The operational availability of all five types of equipment in the vertical observation system reaches 100%, and the data quality meets the requirements of operational application and analysis; (2) Multi-element fusion analysis reveals that affected by the Northeast cold vortex, water vapor transport from the periphery of Typhoon Gaemi, and the subtropical high, the rapid accumulation of water vapor and the vigorous vertical development of cloud systems, combined with vertical temperature difference, low-level convergence and upward movement, as well as mid-level jet momentum transport, jointly trigger the severe convection process; (3) In terms of forecast and early warning, the wind profile radar can indicate changes in dynamic conditions 6-12 hours in advance; the microwave radiometer combined with GNSS/MET observation can capture water vapor and precursor signals 1.5 hours in advance; the millimeter-wave cloud radar can monitor the imminent development characteristics of convection 30-60 minutes in advance. The study shows that the established multi-element quality evaluation system can effectively support the data quality evaluation and application analysis of the vertical observation system under the cold vortex background, and the multi-equipment collaborative observation shows significant advantages in the identification of dynamic triggers, tracking of water vapor evolution, and imminent early warning of severe convective weather, providing important technical support for the refined monitoring and short-term forecasting of severe convective weather in Liaoning.

Abstract:Reanalysis products hold significant application value in extreme weather risk assessment, renewable energy planning, climate change research, and artificial intelligence. Systematic evaluation of the accuracy and applicability of 2m temperature reanalysis products is a prerequisite for their reliable utilization in scientific investigations. This study employs hourly 2m temperature observations from Chinese automatic weather stations (2021–2022) to evaluate three reanalysis products—CMA Regional Re-Analysis (CMA-RRA), ECMWF Reanalysis v5 (ERA5), and Modern-Era Retrospective Analysis for Research and Applications, v2 (MERRA2). Results indicate that: Summer high-temperatures are concentrated in the Yangtze-Huai River Basin (June) and South China and Sichuan Basin regions (July–August),with 2022 high-temperature intensity surpassing 2021 levels. CMA-RRA exhibiting the smallest RMSE (<2°C) and highest correlation coefficient (R≥0.9). For Sichuan Basin extreme high-temperatures, CMA-RRA effectively captures diurnal variations and spatial patterns, achieving the highest Critical Success Index. While ERA5 shows premature heat onset timing and spatial overestimation, and MERRA2 systematically underestimates both intensity and duration of high-temperature events.

Abstract:Based on daily precipitation data from 126 national meteorological stations in Yunnan Province from June 2023 to June 2025, this study evaluated the applicability of Fengyun-4B (FY-4B) quantitative precipitation estimation (QPE) daily precipitation products over Yunnan Province using the correlation coefficient (CC), root mean square error (RMSE), relative bias (RB), probability of detection (POD), false alarm ratio (FAR), and critical success index (CSI), and compared FY-4B QPE with Global Precipitation Measurement Integrated Multi-Satellite Retrievals (GPM IMERG) and climate hazards group infrared precipitation with station data (CHIRPS) products. The results show that the precipitation detection capability of FY-4B QPE is superior to that of GPM IMERG and CHIRPS, with the highest CSI (0.44) and the lowest FAR (0.44). However, it exhibits a certain degree of precipitation underestimation (RB=?0.15). In comparison, GPM IMERG has the highest FAR (0.53), and CHIRPS shows a relatively low POD (0.41). FY-4B QPE exhibits a pattern of higher precipitation over the southwest and lower precipitation over the east of Yunnan Province. The precipitation detection capability in southwestern and southern Yunnan is superior to that in the central and eastern regions. The southwestern and southern areas, covering approximately 31% of the province, have an average CSI that is 0.14 higher than that of the central and eastern regions. The 500 to 1000 m elevation range is the optimal application area for FY-4B QPE. FY-4B QPE performs best under light rain, with pronounced underestimation under heavy precipitation. FY-4B QPE is sensitive to seasonal variations, with strong detection capability but a tendency toward underestimation in summer, while overall accuracy decreases somewhat in winter. The evaluation results can provide a scientific reference for the research and operational applications of FY-4B QPE for precipitation monitoring in complex terrain areas of Yunnan Province.

Abstract:Based on the C-band dual-polarization radar data from the Wutaishan Cloud Physics Experimental Base in Shanxi Province and the operational data of artificial rain enhancement, a fuzzy-logic hydrometeor classification algorithm was used to analyze the full-stage microphysical characteristics of a ground-based rocket rain enhancement operation during a mixed rain event on 22 May 2025. The results show that after the seeding of silver iodide (AgI) ice nuclei, ice crystal nuclei formed through heterogeneous nucleation, and the growth of ice-phase particles was promoted through the Bergeron process and collision-coalescence process. During the main impact stage of rain enhancement, ice crystals and wet snow particles showed an inverse relationship in number variation. Snow aggregate particles were mainly produced by the collisional growth of ice crystals, accounting for 38.9% of the total number and becoming the dominant ice-phase hydrometeors, while high-density graupel accounted for only 3.19%. Wet snow particles were mainly derived from aggregate particles above the 0degrees Celsius level, and their generation region was located at 3.6-4.1 km altitude. They were also an important source of the bright band near the 0degrees Celsius level. Based on the differences before and after AgI entered the cloud, a three-stage conceptual model of cloud microphysical evolution was established, including AgI ice nuclei seeded at the supercooled water accumulation zone, ice formation and aggregation growth, and mixed-phase transition near the 0degrees Celsius level with the formation of raindrop.

Abstract:Based on the observation data from 126 national meteorological stations in Yunnan Province from 2016 to 2025 and and ERA5 reanalysis data for the same period, the spatial distribution characteristics of regional heavy fog in Yunnan were diagnosed and analyzed using Rotated Empirical Orthogonal Function (REOF), and fog zoning was carried out combined with topographic factors. Furthermore, machine learning methods such as random forest were employed to establish zonal heavy fog forecasting models, and the results were compared and verified with SCMOC and ECMWF numerical forecast products. The results show that: 1)Heavy fog in Yunnan Province exhibits four dominant spatial modes: an out-of-phase pattern between the southwest and the rest of the region, southeastern Yunnan, northeastern Yunnan, and northwestern Yunnan. Combined with topography, the province is finally divided into four fog zones: Zone 1 (western and northwestern Yunnan), Zone 2 (northeastern and central Yunnan), Zone 3 (southwestern and southern Yunnan), and Zone 4 (southeastern Yunnan). 2)The random forest model achieves an accuracy of 0.82~0.84 for modeling samples and 0.80~0.83 for test samples in each zone, which is significantly superior to decision tree, K-nearest neighbor, logistic regression, and Bayesian methods. 3)There are regional differences in the importance of forecast factors in each fog zone. High-altitude areas (Zone 1 and Zone 2) are strongly dependent on water vapor conditions at 600—700 hPa. Regions significantly affected by eastern industrial activities (Zone 2 and Zone 4) are largely contributed by initial visibility. In relatively low-altitude areas (Zone 3 and Zone 4), the importance of sea level pressure and temperature factors within the boundary layer is more prominent. 4)The comprehensive performance of the short-range heavy fog forecast model established based on random forest is superior to those of two mainstream forecast products, SCMOC and ECMWF. The improvement in forecast effect is more significant in autumn and winter, and Regions 3 and 4 exhibit higher forecast accuracy.. For a regional heavy fog event that lasted two consecutive days, the TS scores of the random forest model are 0.38 and 0.41, which are significantly higher than those of SCMOC (both < 0.1) and ECMWF (0.26 and 0.11, respectively).

Abstract:IN the spring of 2026 (March to May), the temperature in most part of China was higher than normal, with a national mean temperature of 12.0°C, ranking as the fourth warmest since 1961. The national average precipitation was 166.3 mm, 15.8% above the normal. Spatially, precipitation in central and eastern China presented two main rainfall belts in the North and South: The northern rainfall belt is mainly located from the southern part of North China to the Huanghuai region, while the southern rainfall belt is located from the central-eastern part of Southwest China to the western part of Jiangnan. At 500hPa geopotential height field over the middle and high latitudes of Eurasia exhibited a “two ridges and one trough” pattern. Positive height anomalies prevail over the region from the Ural Mountains to Lake Balkhash and over Northeast Asia, whereas negative height anomalies were center over Lake Baikal and the central Siberia area to its north, a circulation pattern that facilitating the southward intrusion of cold air along the central and western paths into China. In the lower- tropospheric wind field, Northeast Asia was dominated by an anomalous anticyclonic circulation, which favors the transport of warm and moist water vapor from the western North Pacific to northern China. Combined with the southward-moving cold air, this led to significantly above-normal spring precipitation over areas from the southern part of North China to the Huanghuai region. The positive phase of the spring North Atlantic Tripole (NAT) server as an important external forcing signal responsible for the formation of the northern rainy belt over central and eastern China in spring.

Abstract:The May Day and National Day holidays are China""s golden tourism week with favorable climate conditions in most regions. However, the sudden and destructive feature of tourism meteorological disasters, combined with the concentrated travel during these periods, may have severe impact on the tourism industry. Daily observational data from national meteorological stations and information of landfall tropical cyclones over China since 1961 were used to reveal the spatiotemporal patterns of tourism meteorological disasters during the May Day and National Day holidays and further analyze long-term trends. The results indicate that the number of meteorological disaster days during the National Day holiday is generally less than that during the May Day holiday. Except for fog and cold wave events, all types of tourism meteorological disasters exert weaker impacts in the National Day holiday than in the May Day holiday. The number of meteorological disaster days decreases in the later stages of both the May Day and National Day holidays, which is favorable for tourism. There are obvious spatial differences in various tourism meteorological disasters across China. North China is primarily impacted by strong winds, dust storms and cold waves. Southwest China is greatly affected by severe convection and fog. Tourism meteorological disasters in South China are mainly dominated by rainstorm and high temperature. The number of national average meteorological disaster days during the two holidays shows a significant decreasing trend. The downward trends in strong winds and dust storm days are especially notable, while high temperature days increase significantly. Overall, the probability of tourism meteorological disaster is higher in northeastern Tibet, southern Qinghai, northwestern Sichuan and southern Yunnan, while it is low in Northeast China, North China, Huang-Huai, the southeastern part of Northwest China, and northern Xinjiang. A total of 18 tropical cyclones made landfall in China during the National Day holiday from 1961 to 2024, with the highest landfall frequency in Guangdong Province. Under the background of global climate warming, more attentions should be paid to the impact of high temperature and landfall tropical cyclones during the May Day and National Day holiday.

Abstract:In response to the issue of insufficient accuracy in identifying tornadoes using conventional threshold methods,this study proposes a machine learning-based tornado identification method driven by multiple radar-derived storm features. The method first calculates and constructs a storm feature dataset based on radar base data from tornado cases in China between 2003 and 2023. The dataset includes three categories: (1) tornadic storms, (2) non-tornadic storms with significant three-dimensional vortex signature(3DVS), and (3) non-tornadic storms without significant 3DVS. Feature importance ranking is then performed using XGBoost, and the top 14 features are selected as model inputs after combining with PCA. A TFM+CNN deep learning model for tornado recognition is constructed by using the encoder part of the TFM architecture and utilizing CNN to increase the input vector dimension and adding multiple fully connected layers at the end of the model. Three binary classification experiments are conducted using this model: (1) tornado storms vs. non-tornadic storms without significant 3DVS, (2) tornado storms vs. non-tornadic storms with significant 3DVS, and (3) tornado storms vs. non-tornadic storms. Comparative experiments with the XGBoost model are also performed. The results show that the TFM+CNN model achieves CSI values of 84%, 71%, and 73%; POD values of 94%, 83%, and 83%; and FAR values of 11%, 16%, and 14% in the three binary classification experiments, respectively. Compared to the XGBoost model, the TFM+CNN model performs better in most metrics, except for a 5% higher FAR in experiments 1 and 3. Both models exhibit the worst performance in experiment 2. Both models performed worst in experiment 2, indicating that tornadic storms and non-tornadic storms with significant 3DVS are the most challenging to distinguish. Furthermore, the TFM+CNN model has a larger AUC under ROC, indicating stronger generalization ability. It can be seen that the TFM+CNN model has strong tornado recognition ability.

Abstract:This study develops a multi-model fusion quantitative precipitation forecast technique based on the 3D-Unet deep learning model. The method integrates upper-air and surface variables from the ECMWF-IFS global model and the CMA-SH9 regional model for the period 2019–2023. Furthermore, the Integrated Gradients (IG) method is introduced to enhance the interpretability of the model. The results indicate: (1) The 3D-Unet model can effectively enhance the performance of the original precipitation forecast of the model, and the multi-model fusion deep learning forecast model (DL) significantly outperforms the single-model correction model in terms of TS score for 5mm/h precipitation forecast. (2) Real-time operational verification results show that the precipitation forecast performance of the DL model is significantly better than that of the original model forecast and the traditional optimal weight fusion method (OFM). The TS scores for hourly precipitation forecasts of 5mm/h and 10mm/h have increased by approximately 70% and 26.8%, respectively, compared to OFM. For 24-hour cumulative precipitation forecasts, the TS scores of the DL model for moderate rain, heavy rain, and rainstorm levels are superior to those of single models and OFM products; especially for longer forecast horizons of 60 hours and 72 hours, its TS scores for rainstorms have increased by approximately 42.7% and 24.3% compared to the EC model, respectively. (3) The DL model can effectively correct the shape and location of the original forecast rain belt of the model (such as the southwest vortex rainstorm), but it tends to have a wider range of local extreme heavy precipitation. (4) Using the IG algorithm to analyze the importance of feature factors, it is found that the DL model can identify the differences in the effects of the same factors across different models. Factors related to boundary layer processes and local dynamic and thermal structures (such as CAPE and near-surface wind field) have higher importance in the CMA-SH9 regional model, while factors related to synoptic-scale circulation background and air mass thermal state (such as surface pressure and 2m temperature) contribute more prominently in the EC global model.

Abstract:Sub-seasonal to Seasonal(S2S) hydrological prediction is a core supporting technology for watershed water resource scheduling and flood control & disaster reduction. A semi-distributed hydrological modeling system for the Huaihe River Basin was established based on the SWAT2012 model. Runoff prediction was carried out by driving the SWAT model with historical hindcast data from three sub-seasonal to seasonal (S2S) climate models (CMA-CPSv3, CMME-S2S, and AH-EDDS) from 2006 to 2024 (AH-EDDS from 2009 to 2024). The simulation accuracy of the hydrological model and the prediction skills of the climate models were systematically evaluated. The results show that: the calibrated SWAT model exhibited excellent simulation accuracy during both the calibration period (1981-2010) and validation period (2011-2020), with a coefficient of determination (R2)≥0.88 and Nash-Sutcliffe efficiency (NSE)≥0.85. Four parameters, including groundwater delay time, are the most sensitive parameters. Runoff prediction skills driven by S2S models show significant seasonal differences: the effective lead time in the dry season can reach 40 days, while the effective lead time of CMA-CPSv3 and CMME-S2S is only 1–5 days and that of AH-EDDS is 6–10 days in the wet season. The overall prediction skill in the dry season is higher than that in the wet season. All three models show low hit rate, high false alarm rate and systematic underestimation for extreme precipitation, among which AH-EDDS performs the best in spatial precipitation prediction and comprehensive runoff prediction. The tercile probabilistic prediction and extreme event threshold method can effectively capture watershed hydrological risks, successfully capturing the flood signals in the wet seasons of 2008, 2016, 2020 and the drought signals in the dry seasons of 2009, 2011, 2014. The framework of “multi-model S2S climate prediction driving hydrological model” established in this study can effectively improve the reliability of sub-seasonal to seasonal runoff prediction in the Huaihe River Basin, which has been verified in real time during the 2025 wet season and shows potential for operational application.