Abstract:Using conventional upper-air and surface observations, Doppler radar data, satellite nephogram and ERA5 reanalysis data, this paper analyzes drylines cases and drylines triggering convections in Shandong Province from 2001 to 2022. The results are as follows. The total number of days with drylines in Shandong from 2001 to 2022 is 451 d, of which there are 156 d with convective drylines， respectively. Convective drylines occur mainly from April to September. May and the first dekad of May are the month and dekad periods to have the highest occurrence frequency of drylines, respectively. Drylines mainly occur between 14:00 BT and 17:00 BT in the afternoon, distributed often in central and northwestern Shandong. The formation of drylines is related to the influence of mountain ranges and the distribution of land and sea. The widths of convective and non-convective drylines are 75-106 km and 80-114 km respectively, with dew point temperature gradients being 11.8-16.7 ℃·（100 km）-1 and 9.8-15.0 ℃·（100 km）-1 respectively. Temperatures on both dry and wet sides of drylines vary from 23℃ to 33℃, with a temperature gradient of 1-3 ℃·（100 km）-1, and temperature on the dry side is slightly higher. Compared to non-convective drylines, convective drylines exhibit higher temperatures and dew point temperatures on both the warm and cold sides, narrower widths, and greater dew point temperature gradients. This is because the circulation associated with the force-driven component becomes stronger when the horizontal dew point temperature gradient is large. The thresholds of the key environmental parameters for drylines triggering convections reveal that the convective parameters such as energy and stratification stability all exhibit a certain degree of distinguish ability. The vertical wind shear in the deep layer (0-6 km) is also distinguishable, whereas vertical wind shear in the lower layers (0-1 km and 0-3 km) is difficult to distinguish due to the small differences. The seasonal variation in convective parameters under unstable conditions is relatively small, but environmental parameters related to moisture conditions and 0-6 km vertical wind shear exhibit significant seasonal differences, so their thresholds should be differentiated by season. Drylines frequently develop under conditions of northwest airflow, ahead of westerly trough, and northeast low-pressure vortex at 500 hPa. At the lower 700 hPa and 850 hPa, drylines are primarily influenced by the shear line of the low-pressure vortex or the northwest airflow, with the boundary layer often accompanied by convergence lines. The surface primarily lies within the convergence flow field of continental warm low pressure or at the bottom or front of low-pressure systems.

Abstract:The northern part of Ningxia Hui Autonomous Region is situated in the temperate continental climate zone of Northwest China, where complex topography and the climatological characteristics of warm-season convective storm (CS) have not been fully understood. This study utilizes CINRAD/CD radar products and sounding data in Yinchuan from May to September in 2011-2016 to investigate the climatological characteristics of CS in northern Ningxia under different prevailing wind directions at 500 hPa and 700 hPa. In this research domain, we present a comprehensive analysis integrating the average circulation patterns, convective environments and CS intensity for the first time. The results demonstrate that solar radiative heating and topographic forcing from the Helan Mountains dominate the formation of CS climatological characteristics in northern Ningxia. Upon this foundation, dynamically varying prevailing winds induce more complex and diverse patterns in CS occurrence frequency, occurrence efficiency (defined as the ratio of CS occurrence frequency under a specific prevailing wind direction to the total frequency of that wind direction), diurnal variation, and spatial distribution. Notably, when southwesterly winds prevail at 500 hPa while easterly winds are blowing at 700 hPa, the influence of prevailing winds becomes particularly pronounced. The average 500 hPa circulation pattern associated with CS occurrence under a given prevailing wind direction characterizes the typical synoptic-scale systems and convective environments favorable for CS development under that flow regime, thereby directly determining CS occurrence efficiency and intensity. When northerly or northwesterly winds prevail at 500 hPa and northern Ningxia is positioned behind the trough, there is the lowest CS occurrence efficiency corresponding to the highest CAPE and CIN, and the CS intensity is the strongest. Under southerly or southwesterly wind regimes, northern Ningxia is situated between the ridge and trough, where favorable moisture transport conditions tend to cause maximum CS occurrence efficiency. During prevailing westerly flows, the mid-latitude circulation exhibits zonal characteristics with frequent upstream shortwave trough activity, causing the moderate levels of CS occurrence efficiency, CS intensity and environmental parameters. At 700 hPa, differences in CS occurrence efficiency and intensity across various prevailing wind directions are relatively smaller, with CS intensity showing a strong correspondence to the vertical temperature gradient.

Abstract:Based on the precipitation data of 155 meteorological stations in the eastern part of Northwest China from 1961 to 2022, as well as historical datasets including global atmosphere data, sea surface temperature (SST), and sea ice data, this study integrates the temporal convolutional network (TCN) module and the convolutional block attention module (CBAM) into the long short-term memory (LSTM) deep learning algorithm. A climate intelligent prediction model, CBAM-TCN-LSTM, based on the fusion of deep learning algorithms is thereby established for summer precipitation in the eastern part of Northwest China. The predictive performance of the model is verified, and its predictive capability is compared with that of other deep learning algorithms. The results show that the intelligent prediction model based on the fusion of multiple deep learning algorithms outperforms single-algorithm deep learning models. During the independent sample validation period (2018-2022), the prediction score (PS) for summer precipitation prediction ranged from 60 to 80, with an average value reaching 73.81. The anomaly correlation coefficient (ACC) was positive for all years except 2020, with an average value of 0.14, representing a significant improvement over other models. Compared with current seven mainstream models of machine learning algorithms, the CBAM-TCN-LSTM model presents superior performance across all five evaluation metrics: PS, percentage correct (PC), ACC, mean absolute error (MAE), and root mean square error (RMSE). In addition, the CBAM-TCN-LSTM model was successfully applied in the 2023 flood season prediction operation. It accurately predicted the characteristics of below-normal summer precipitation in most areas of the eastern part of Northwest China, with a PS as high as 81. By building a precipitation prediction model that combines the TCN module and the CBAM module on the basis of the LSTM model which has strong time-series predictive capability, this study provides a scientific basis and technical support for regional precipitation prediction, and the model holds a good prospect for popularization and application.

Abstract:This study develops a multi-model fusion quantitative precipitation forecast technique based on the 3D-Unet deep learning model and forecast products from the ECMWF-IFS global model （abbreviated as EC） and the CMA-SH9 regional model, and introduces the integrated gradient (IG) method to enhance model interpretability. The results indicate that the 3D-Unet can effectively improve the performance of raw precipitation forecasts, and the multi-model fusion deep learning model (DL) significantly outperforms the single-model correction model in terms of the TS score for 5 mm·h-1 precipitation. Real-time operational verification shows that the DL can produce substantially better precipitation forecasts than both the original model and the traditional optimal fusion method (OFM). Compared with that of OFM, the TS scores of precipitation forecasts of the DL model at the 5 mm·h-1 and 10 mm·h-1 thresholds are improved by approximately 70.3% and 26.5%, respectively. For the 24 h accumulated precipitation, the DL forecast achieves higher TS scores than a single model and OFM products for moderate rain, heavy rain and torrential rain categories. In particular, relative to that of the EC, at the longer lead time of 60 h and 72 h, TS scores of the DL for torrential rain forecasts are improved by about 42.7% and 24.3%, respectively. The DL can effectively correct the morphology and location errors of large-scale rain belts in raw forecasts, but it tends to overestimate the spatial extent of local extreme heavy precipitation. Feature importance analysis with the IG algorithm reveals that the DL can distinguish the different roles of the same factors in different models. Factors related to boundary layer processes and local dynamic-thermodynamic structures (e.g., CAPE and near-surface wind) possess higher importance in the CMA-SH9, while factors associated with synoptic-scale circulation background and air mass thermal state (e.g., surface pressure and 2 m temperature) contribute more prominently in the EC.

Abstract:Doppler weather radar observations play a vital role in hail identification, but the spatio-temporal coverage of single-source radar data is limited and the timeliness and accuracy of traditional identification methods are not satisfying. To address this issue, this paper proposes a hail identification method based on satellite and radar data fusion. Utilizing the spatio-temporal complementarity between satellite and radar data, the method combines threshold characteristics of satellite and radar observations before and after hailfall to achieve efficient multi-source data fusion and identification through deep learning algorithms. Experiment results indicate that the proposed method effectively integrated satellite and radar data, with the mean average precision of the YOLO v7 model is 0.9083. It successfully identified hail-affected regions, providing important references for early warning of hail weather. Notably, in areas where radar data were susceptible to terrain occlusion, the method significantly improved the problem of inaccurate identification of hail areas caused by poor data quality, demonstrating high practical application values.

Abstract:To understand the new variation features of acid rain pollution deeply in typical northern regions of China, the precipitation pH value and conductivity at Taishan Station during 1993-2021 and Tai’an Station during 2007-2021 are analyzed. On this basis, synchronous intensive observation data of precipitation chemistry which was performed from May 2018 to April 2019 are analyzed to get the contributions of below-cloud processes and long-distance transport to acid precipitation in the Mount Taishan Area. Combined with the results of earlier studies, the new situation and causes of acid rain in the Mount Taishan Area in recent years are investigated. The study results show that the annual precipitation pH value at Taishan Station varied in the range from 4.36 to 6.47 during 1993-2021, while that at Tai’an Station ranged from 4.73 to 6.66 during 2007-2021. Owing to the strong below-cloud scavenging effect at Tai’an Station, the neutralization effect of alkaline pollutants was more pronounced, leading to a significant reduction in precipitation acidity compared with that at Taishan Station. During 1993-2021, the precipitation acidity at Taishan Station experienced a phased variation process, i.e., weakening, intensifying, moderately weakening, and relatively stabilizing or slightly intensifying. The neutralization effect of alkaline pollutants exerted an important influence on the long-term variation trend of acid rain pollution. The concentration of SO2-4 in precipitation in the Mount Taishan Area decreased to approximately 50% of that from 2004 to 2008 and the SO2-4/NO-3 at Taishan Station dropped from 2.84-3.98 to 1.21, which indicates that the source of acid rain had shifted from SO2 dominated to SO2 and NOx dominated. It is noteworthy that the NO-3 concentration in precipitation at Taishan Station and Tai’an Station increased by approximately 47% and 20%, respectively, which was contrary to the decreasing trend of NOx emissions in the whole China and Shandong Province during the same period. Backward trajectory air mass analysis shows that the precipitation characteristics at Taishan Station were mainly affected by the pollution emissions from Shandong Province and its surrounding provinces, including the Beijing-Tianjin-Hebei Region and Inner Mongolia.

Abstract:The May Day and National Day holidays are China’s golden tourism festivals with favorable climate conditions in most regions. However, the sudden and destructive feature of tourism-related meteorological disasters, combined with the concentrated travel of people during these periods, may have a serious impact on the tourism industry. In this study, daily observation data from national meteorological stations and information of landfall tropical cyclones in China since 1961 are used to reveal the spatio-temporal patterns of tourism-related meteorological disasters during the May Day and National Day holidays. Moreover, the disasters’ long-term variation trends are analyzed. The results indicate that the number of me-teorological disaster days during the National Day holiday is generally less than that during the May Day holiday. Except for fog, cold wave events and tropical cyclones, all types of tourism-related 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 tourism-related meteorological disasters across China. Northern China is primarily impacted by strong wind, sand-dust and cold wave events. Southwest China is greatly affected by severe convection and fog. Tourism-related meteorological disasters in South China are mainly dominated by torrential rains and high temperature. The number of national-average meteorological disaster days during the two holidays shows a significant decreasing trend, and the downward trends in strong winds and sand-dust days are especially notable. However, the number of high temperature days shows a significant increasing trend. Overall, the occurrence probability of tourism-related meteorological disasters is higher in northeastern Xizang, southern Qinghai, northwestern Sichuan and southern Yunnan than in Northeast China, North China, Huang-Huai Region, the southeastern part of Northwest China, and northern Xinjiang. From 1961 to 2024, there were a total of 18 tropical cyclones having made landfall in China during the National Day holiday with the highest landfall frequency in Guangdong Province. Under the background of global climate warming, more attention should be paid to the impact of high temperature and landfall tropical cyclones during the May Day and National Day holidays.

Abstract:This paper assesses the accuracy of operational position and intensity estimation as well as the accuracy of track and intensity forecasts for the typhoons over the Northwest Pacific and the South China Sea in 2025. It has been found that, in 2025, the mean error of typhoon position estimation and the mean absolute error (MAE) of typhoon intensity estimation made by the National Meteorological Centre (NMC) were 13.5 km and 1.1 m·s-1, respectively. The accuracies of position and intensity estimations by the NMC were superior to those of other official typhoon forecasting agencies, including the Japan Meteorological Agency and the Joint Typhoon Warning Center. The mean errors of 24 h and 48 h typhoon track forecasts by the NMC were 56.6 km and 95.8 km, respectively, and the MAEs of 24 h and 48 h typhoon intensity forecasts were 3.6 m·s-1 and 4.9 m·s-1, respectively. Thus, the performance of NMC’s typhoon track and intensity forecasts with lead times of 24 h and 48 h reached the highest level on record. For AI-based weather prediction (AIWP) models, however, the 95th percentile track forecast errors were generally larger than those of numerical weather prediction (NWP) models at short lead times, but smaller than those of NWP models at long lead times. Correcting the typhoon intensity forecasts of AIWP models with an adaptation model can reduce the underestimation of AIWP models for typhoon intensity and improve the accuracy of typhoon intensity forecasts.

Abstract:In the spring (March to May) of 2026, the temperature in most parts of China was higher than normal, with average temperature over China reaching 12.0℃, making this spring the fourth warmest since 1961. The average precipitation over China was 166.3 mm, 15.8% above normal. Spatially, precipitation in central and eastern China presented two main rainfall belts in the north and south. The northern rainfall belt was mainly located from southern North China to the Huanghuai Region, while the southern rainfall belt was from the central-eastern part of Southwest China to the western part of the Jiangnan Region. The 500 hPa geopotential height field over the middle and high latitudes of Eurasia exhibited a “two ridges and one trough” pattern. Positive height anomalies controlled the region from the Ural Mountains to Lake Balkhash and over Northeast Asia, whereas negative height anomalies were centered over Lake Baikal and the central Siberia Area to its north. This circulation pattern facilitated 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, favoring the transport of warm and moist water vapor from the western North Pacific to northern China, which combined with the southward-moving cold air, led to significantly above-normal spring precipitation in the area from southern North China to the Huanghuai Region. In addition, the positive phase of the spring North Atlantic tripole served as an important external forcing signal for the formation of the northern rainy belt in the central and eastern parts of China in spring.

Abstract:In July 2026, the mid- and high-latitude circulation in the Northern Hemisphere exhibited a predominantly four-wave pattern. The Northwest Pacific subtropical high was located farther north than normal and extended westward in stages, and the rainy season in North China began earlier than normal. The national average precipitation was 122.8 mm, 1.0% more than normal. The above-normal precipitation occurred mainly in eastern Northwest China, Northeast China, eastern and southern North China, central regions south of the Yangtze River, and South China. Six widespread precipitation events and four severe convective weather events occurred during this month. Extreme daily precipitation was recorded at 88 national meteorological observation stations, among which the daily precipitation at 22 stations exceeded the historical records. The national average temperature was 23.5℃, 1.3℃ warmer than normal and ranked as the second highest for the same period since 1961. The national average high-temperature days were 5.7 d, the third most for the same period on record. Daily maximum temperatures at 225 stations reached the extreme high-temperature threshold, with 40 stations seeing the temperature reach or exceed their historical records. Specifically, the maximum temperature at Dongkan Station in Turpan of Xinjiang was up to 49.8℃. In addition, five typhoons formed successively over the Northwest Pacific and the South China Sea, three of which made landfall in China bringing strong winds and heavy rainfall to parts of the country. Severe convective activities were relatively intense. From the evening to the night of 6 July, three tornadoes occurred in eastern Hubei, two of which were of EF3 level and one was of EF1 level.