应用气象学报

2023, 34(6)

[FullText](97) [PDF 913KB](25)

Abstract:

A Probabilistic Forecast Experiment of Short-duration Heavy Rainfall in Beijing Based on CMA-BJ

Xing Nan, Zhong Jiqin, Lei Lei, Yang Yiya, Xu Luyang

2023, 34(6): 641-654. DOI: 10.11898/1001-7313.20230601

[FullText HTML] (83) [PDF 12149KB](122)

Abstract:
Based on numerical prediction products from China Meteorological Administration Beijing model (CMA-BJ), precipitation observation of ground weather stations in Beijing and ECMWF ERA5 dataset, the hourly rainfall samples from April to September during 2019-2021 are divided into short-duration heavy rainfall (SDHR, greater than 20 mm·h^-1), ordinary rainfall (between 0.1 and 19.9 mm·h^-1) and no rainfall (less than 0.1 mm·h^-1). The probability density distribution characteristics of physical parameters are comparatively analyzed, including moisture conditions, thermal and energy conditions, and dynamic conditions for three categories. Monthly predictors are selected from these parameters by comparing their ability to discriminate among SDHR, ordinary rainfall and no rainfall weather. It is found that the distributions and thresholds of physical parameters differ between months to some extent. Among that, moisture conditions, thermal and energy conditions, and dynamic conditions are relatively stronger but with less discrimination degrees among SDHR, ordinary rainfall and no rainfall weather in July and August. The background circulation and the distributions of physical parameters show obvious monthly differences, so the forecast model is established for each different period. After that, forecast model of SDHR for 0-12 h at 1 h intervals in different periods is established by using the ingredients-based method and fuzzy logic algorithms. When probabilistic and composite reflectivity thresholds are 0.6 dBZ and 15 dBZ, threat score (TS) and bias of SDHR are 0.14 and 1.14 in Beijing from April to September during 2019-2021, showing relatively better forecasts. Therefore, the probabilistic and composite reflectivity thresholds corresponding to the optimal TS and bias for 2019-2021 are taken as the forecast probability and eliminating false thresholds of SDHR, and 0-12 h hourly forecast products of SDHR four times a day are tested from April to September of 2022. Results show that TS and bias of SDHR are 0.104 and 1.341, respectively, indicating that the probability prediction products are better than that of CMA-BJ. SDHR products achieve greater improvement, and balance hit rate and false alarm rate well in the piedmont and plain areas with high SDHR frequency. But performances in mountainous areas are not as good as that in plain areas, which may be related to less stations in the mountainous areas of the forecast model. In addition, the result based on case analysis show the predicted area of the products is relatively larger, but high probability area has a good indication for SDHR in Beijing.

Application of Rain and Snow Phase Criterion Based on Ice-phase Particle Content Forecast by CMA-MESO

Wang Lei, Chen Qiying, Hu Jianglin, Xu Guoqiang

2023, 34(6): 655-667. DOI: 10.11898/1001-7313.20230602

[FullText HTML] (60) [PDF 8470KB](102)

Abstract:
The forecast of rain and snow phase is one of the difficulties in precipitation forecast, which is of great significance for disaster prevention and reduction. Rain or snow phase is mainly discriminated according to the traditional temperature-thickness criterion, or the combination of numerical model results with the judgment of forecasters on environmental conditions in present operatorial forecast. However, the determination of temperature-thickness criterions is subjective, complicated and various in different regions. The precipitation phase product of numerical model is based on temperature, humidity and liquid water content forecasts, resulting in errors of other variables besides microphysics introduced. Therefore, many uncertainties exist in the forecast of the transition of rain and snow, especially in the mixed phase. China Meteorological Administration mesoscale model (CMA-MESO) is a regional numerical model and has been applied to national operatorial weather forecast. Its precipitation phase is diagnosed using temperature, humidity and other basic atmospheric variables, including only rain, snow, freezing rain and hail, excluding mixed phase. Therefore, it is urgent to study on a more effective method for rain and snow, especially for the sleet forecast.A criterion for discriminating rain and snow phase is determined using the ice-phase particle content directly output from microphysics scheme of CMA-MESO, and applied to discriminate the range and transition of rain and snow in a widespread precipitation process in China during 14-15 January 2023. The proportion threshold of ice particles is firstly determined by the statistical threat scores. Results show that problems of larger range of sleet and underreporting scattered sleet in eastern China discriminated by traditional thickness criterion are obviously improved by ice-phase criterion. Threat scores for 6-18 h forecast of sleet increase by 75%-100%, and those for 24 h forecast of snow increase by 67% using ice-phase criterion compared with those using thickness criterion, respectively. Threat scores of 3-36 h forecast for rain, snow and sleet are 0.76 to 0.62, 0.69 to 0.63 and 0.11 to 0.08. There are false alarm and missing for rain and snow, respectively, and obvious false alarm for sleet within 24 h. The ice-phase criterion performs well on discriminating the transition process of rain and snow. The forecast error of phase transition start time at representative stations is about 1-2 h using ice-phase criterion, better than thickness criterion. Besides, the ice-phase criterion performs better in discriminating the duration of sleet for the representative station in eastern China too, while the thickness criterion will make forecast results longer than observations. These results could provide a more reliable and objective forecast product for the rain and snow phase forecast in operation.

Correction of Gust Estimation in the Yellow and Bohai Seas and Adjacent Areas

Hu Haichuan, Qian Chuanhai, Qu Hongyu

2023, 34(6): 668-680. DOI: 10.11898/1001-7313.20230603

[FullText HTML] (65) [PDF 6196KB](56)

Abstract:
Due to the influence of local factors such as underlying surface, the fitting performance of the gust factor method can vary among different observation stations in the Yellow and Bohai Seas and adjacent areas. Therefore, based on observations from January 2021 to December 2022, causes for different performances are analyzed, and a gust estimation correction (GECO) method is proposed based on the coefficient of difference and frequency matching to correct fitting results of the gust factor method. It's concluded that, as the gust wind speed increases, the fitting results tend to be biased towards the gust at the whole observation time rather than the maximum gust within the hour. The difference coefficient calculated from the maximum gust within the hour and the gust at the whole hour observation time can quantitatively characterize the influence of local characteristics of the observation station on fitting results. For stations with large difference coefficients, the gust factor method has a larger negative bias in fitting the strong gusts, so it needs to be corrected to a greater extent. Conversely, for stations with small differences, the gust factor method has a smaller negative bias in fitting the strong gusts, so it only needs to be corrected to a smaller extent. GECO method constructed based on statistical results of 12 benchmark observation stations can also be applied to 364 stations in the Yellow and Bohai Seas and adjacent areas, demonstrating the feasibility of quantitatively characterizing the influence of local factors such as underlying surface on fitting results of gusts, as well as the stability of GECO method. In the comparative test of fitting errors for 364 stations in the Yellow and Bohai Seas and adjacent areas, after correction by GECO method, the root mean square error of gusts above 12 m·s^-1 and 16 m·s^-1 is reduced by 12.3% and 11.5%, respectively, compared to the gust factor method. Although GECO method may slightly increase the fitting error of weak gusts, it can significantly improve the fitting performance of strong gusts. In the verification during the impact of Typhoon Muifa from 14 September to 16 September in 2022, the fitting performance of strong gusts is also significantly improved after correction by GECO method. The improvement of fitting performance is the basis for improving forecasting skill. Combining GECO method with other objective gust forecasting methods can further enhance the forecasting service capability of gusts.

Dual Polarization Radar Characteristics of Severe Downburst Occurred in Weak Vertical Wind Shear

Guo Feiyan, Diao Xiuguang, Chu Yingjia, Ma Yan

2023, 34(6): 681-693. DOI: 10.11898/1001-7313.20230604

[FullText HTML] (56) [PDF 11386KB](82)

Abstract:
Based on S-band Dual polarization doppler radar data and conventional observations, characteristics of 3 severe downbursts occurred under the background of weak vertical wind shear are analyzed, and their possible physical formation mechanisms are studied. It shows that they all occur with high convective available potential energy, but the vertical wind shear (less than 10 m·s^-1) of the environmental atmosphere is weak. 6·30 storm and 7·2 storm are featured by wetter air at low level and drier air at middle (or high) level, while the atmosphere air for 6·26 storm is dry from low level to high level except for the near surface layer. With weak wind vertical shear and higher 0℃ layer height, this type of severe downburst is mostly along with high intensity precipitation (over 3 mm per minute). Before the severe downbursts touch down, the storms grow intensively (over 60 dBZ) and expand over 10 km height, the differential reflectivity (Z_DR) and specific differential phase shift (K_DP) columns are higher over -10℃ layer, and K_DP columns' area at -10℃ layer are larger than Z_DR columns'. For dual polarization radar, the appearance of high K_DP region (more than 3.0°·km^-1) around or above 0℃ layer can be regard as a criterion for identifying downbursts. The high K_DP region with high concentrations of liquid partials or small melting ice particles around or above 0℃ layer can be treated as the overhanging quality roll of liquid particles, which is similar to the overhanging and descending reflectivity core. The appearance and descending of high K_DP region initiates the development of severe downburst along with short-time high intensity precipitation. Due to weak entrainmental zone mean wind speed, the contribution of downward momentum transportation mechanism on the surface gale wind is probably weak. If the environmental atmosphere is wet, the dominant formation mechanism for severe downburst is the gravity dragging effect by abundant liquid (or small melting ice) particles and a tiny quantity of big hail, and the subordinate formation mechanism is the melt cooling effect by ice phase particles. If the environmental atmosphere is dry especially at middle or high level, the entrainment effect and evaporative cooling effect by dry air also contribute to the maintenance and acceleration of downdrafts.

Indentification of Weather Radar Abnormal Data Based on Deep Learning

Zhang Lin, Wu Lei, Li Feng, Li Yan, Shi Lijuan, Sun Kangyuan

2023, 34(6): 694-705. DOI: 10.11898/1001-7313.20230605

[FullText HTML] (70) [PDF 2470KB](86)

Abstract:
The world's largest weather radar observation network which consists of 236 weather radars is built up in China. The quality control of weather radar data becomes an indispensable part in operation as data grow. In real-time operation of CMA Meterological Observation Center, the abnormal data caused by radar hardware fault or calibration problem usually leads to a bad image, and the problem directly affects the quality of weather radar data, quantitative estimation of precipitation and analysis of the weather system. At present, the abnormal data are processed by artificial corrigendum in real-time operation, so that it does not affect the application of follow-up data. In recent years, artificial intelligence technology has developed rapidly, and deep learning algorithms are used to build convolutional neural network models to extract image features from abnormal data such as radar hardware failures and electromagnetic interference. According to characteristics of the abnormal data, two kinds of abnormal dataset are established. The first kind is labeled data for those samples with whole abnormal picture or precipitation. The second kind is the electromagnetic interference in fixed direction, which is easily mixed with precipitation, and it is used to label the samples for pixels. Based on the convolution pre-training network model (ResNet), the R-ResNet model is built with the first kind of abnormal data. The model evaluation shows that the R-ResNet model achieves more than 99% accuracy in both training and test datasets, and the prediction results of the model in test datasets are all consistent with the label data. For the second kind of abnormal data, based on the image semantic segmentation network model LinkNet, a group of encoders and decoders are extended, and a hop-link structure is added, and ResNet50 is used as the encoder structure. The R-LinkNet model is constructed and the accuracy is over 98% on the training and test datasets, with intersection over union of 83.4% and 83.2%, respectively. Two models can be used to monitor the abnormal data of the national radar in real-time operation, so that abnormal data can be corrected automatically and the workload of the manual on duty can be reduced greatly.

Physical Inspection of Randomized Trial for the Artificial Rain Enhancement Experiment at Gutian from 2014 to 2022

Hu Shuping, Lin Wen, Lin Changcheng, Li Dan, Jiang Shanci, Feng Hongfang

2023, 34(6): 706-716. DOI: 10.11898/1001-7313.20230606

[FullText HTML] (63) [PDF 1588KB](65)

Abstract:
The verification of the effectiveness of artificial rain enhancement is a worldwide challenge. Based on random experiments of ground-based rocket artificial rain enhancement at Gutian from 2014 to 2022, physical verification of stage samples is carried out using radar macro parameters such as radar echo intensity, echo top height, and thickness of the negative temperature layer. The evolution characteristic analysis of dual-polarization parameters Z_DR and K_DP is conducted. The physical characteristic response of seeded cases is studied by combining ground raindrop size distribution data. The analysis results show that the non-seeded samples have a small increase in echo intensity, echo top height, and negative temperature layer thickness after hypothetical artificial precipitation enhancement operation, but then quickly decrease. Seeded samples generally show an increase in echo intensity after operation, reaching a peak after 36 minutes. There are 52.6%, 21.1% and 7.9% of seeded samples with maximum growth rates of 0-20%(excluding 0), 20%-50%(excluding 20%) and above 50%, respectively. Half of seeded samples show an increase in echo top height and negative temperature layer thickness, with the former showing a significant increase after 30 minutes and maintaining stability, while the latter increase significantly after 12 minutes. Among them, 36.8%, 13.2% and 2.6% of seeded samples increase with maximum growth rates of 0-20%(excluding 0), 20%-50%(excluding 20%) and over 50%, respectively. At the same time, the dual-polarization parameters Z_DR and K_DP show sustained enhancement after operation. From the physical response characteristics of seeded individual case, Z_DR column appearing in the cloud after operation indicates that the upward airflow in the cloud has increased, causing the echo top height to rise and the echo intensity to increase or maintain. At the same time, K_DP column also indicates that there are more large raindrops or partially melted ice particles in the cloud after operation. The ground rainfall size distribution shows a characteristic of increasing first with small and medium drops and then with large drops, resulting in an increase in spectral width. The maximum minute rainfall intensity generated by the cloud system within 50 minutes after the operation increases significantly, and the cumulative rainfall has increased by 49% compared to the operation time. In summary, cloud seeding helps the development, enhancement, and maintenance of clouds, not only increasing the number and size of precipitation particles, but also to some extent prolonging the life of the cloud.

Accuracy of Atmospheric Profiles Retrieved from Microwave Radiometer and Its Application to Precipitation Forecast

Zhou Bingxue, Zhu Langfeng, Wu Hao, Dong Zipeng, Wang Xuan, Luo Yuyan

2023, 34(6): 717-728. DOI: 10.11898/1001-7313.20230607

[FullText HTML] (41) [PDF 40844KB](95)

Abstract:
Real-time and effective detection of atmospheric profiles is of great significance in understanding the evolution of climate system. Ground-based microwave radiometers can provide atmospheric temperature and humidity profiles with extremely high temporal and spatial resolution. Domestic MWP967KV microwave radiometer has effectively made up for problems of imported microwave radiometers, but there are relatively few studies on the performance evaluation and application of this microwave radiometer. In order to better apply data and products of MWP967KV microwave radiometer, inversion data from June 2018 to July 2021 at Jinghe Station of Xi'an are compared with L-band radiosonde observation. The accuracy of atmospheric temperature, relative humidity and vapor density retrieved from microwave radiometer under clear skies and different cloudy skies (classified as low cloud, middle cloud and high cloud, respectively) are analyzed, and the applicability of the related products in precipitation is further explored. Results show that correlation coefficients of temperature between microwave radiometer and radiosonde are 0.99, correlation coefficients of vapor density are 0.97, and correlation coefficients of relative humidity are less than 0.50 under clear skies and cloudy skies, all passing 0.01 significant test. The difference of temperature between clear and cloudy skies is small, but root mean square error of relative humidity in cloudy skies is more than 25%, which is significantly larger than that in clear skies. It indicates that the presence of clouds reduces the accuracy of the humidity inversion, causing large errors, and the inversion accuracy is higher near the ground. Under different cloud types, the temperature difference is small, while root mean square error and bias of relative humidity in low cloud are the largest, which are 26.85% and 9.51%, respectively. In addition, a case analysis shows that relative humidity, liquid water content, atmospheric precipitable water vapor and liquid water path increase significantly before the occurrence of precipitation, which can be used as indicators of the possible occurrence of precipitation. Statistic results show that the atmospheric precipitable water vapor reaches 4 cm and liquid water path reaches 0.2 mm during several precipitation cases, and these indexes can be used as the reference threshold for judging the precipitation of Xi'an.

Effects of Covering and Uncovering Date on Seedling Emergence of Overwintering Ginseng and American Ginseng

Yuan Fuxiang, Mu Jia, Gao Yan, Liu Wei, Liu Gang, Song Minghai, Zhan Weidong, Li Xianqiang, Zhang Jing

2023, 34(6): 729-738. DOI: 10.11898/1001-7313.20230608

[FullText HTML] (35) [PDF 1232KB](35)

Abstract:
Ginseng and American ginseng need to be covered and protected against cold in winter to survive, and covering and uncovering date has great influences on safe overwintering and seedling emergence. To determine the optimal time, field experiments are conducted at Fusong of Jilin Province from November 2021 to June 2022. Effects of different covering dates on the soil temperature in the overwintering layer and seedling emergence rate of ginseng and American ginseng are studied through different experiment themes. For some cases they are covered in stages in early winter and simultaneously uncovered in spring, and there are also other cases when the ginseng and American ginseng are covered simultaneously in early winter and uncovered step by step in spring. Results show that the emergence rate of ginseng and American ginseng decreases with the delay of coverage date. When 5 cm soil temperature is 0℃, the coverage date for cold prevention is the best, with the highest emergence rate of ginseng and American ginseng. Below -12℃, most or all of American ginseng is frozen to death. When the instantaneous 5 cm soil temperature drops to -14℃, the emergence rate of ginseng still reaches 75%. The emergence rate of ginseng seedlings in other treatments above this temperature coverage is mostly over 86%. When the soil temperature is -7℃ or above, the emergence rate of ginseng seedlings is above 97%. The emergence rate of ginseng seedlings covered by snow is 91%. The soil temperature fluctuates between -14℃ and -8℃, and all uncovered ginseng with an extreme minimum of -16℃ is frozen to death. When ginseng seedlings emerge, the soil temperature of 5-20 cm is around 8-9℃, slightly higher than that of American ginseng. Covering with high insulation fibers to prevent cold, the later the cover is removed, the lower the soil temperature becomes, and the later the seedlings emerge. The date of uncovering affects the progress of seedling emergence, but has little correlation with the emergence rate. The optimal date of uncovering needs to be analyzed further. On frost prone areas in early spring, combined with climate prediction, the emergence progress can be adjusted by uncovering date to avoid spring freezing damage.

Luminosity and Current Characteristics of Metal-vaporized Channel of an Artificially Triggered Lightning

Xu Weiqun, Lü Weitao, Qi Qi, Fan Yanfeng, Chen Lüwen, Wu Bin, Wang Xuejuan, Ma Ying, Xia Dengcheng

2023, 34(6): 739-748. DOI: 10.11898/1001-7313.20230609

[FullText HTML] (49) [PDF 2272KB](27)

Abstract:
Channel current is an important parameter of the lightning discharge, but it's difficult to be directly measured due to the randomness and instantaneity of the natural lightning. The channel luminosity, however, is relatively easier to obtain. If there is a definite relationship between channel current and luminosity, the channel current can be estimated based on its luminosity. The correlation between the current and luminosity of lightning channel can be obtained through artificially triggered lightning experiments, during which the channel current can be directly measured and close-range optical observations of the lightning channel can be carried out.Base on observations of an artificially triggered lightning obtained at the Field Experiment Base on Lightning Sciences, China Meteorological Administration (CMA_FEBLS) in 2022, characteristic parameters of the channel-base current, the luminosity of the metal-vaporized channel, and their correlation are analyzed. The spatial movement of the metal-vaporized channel during the multiple return stroke processes is distinguished using still image with high spatial resolution. Combined with the high-speed video camera images and the channel current data, the correlation between the luminosity of the metal-vaporized channel and the channel current in the process of return strokes, and M components are studied. The results show that compared with the peak current, its squared value has stronger correlation with the peak luminosity for 13 return strokes. For the return stroke followed by a long continuing current, as well as the return stroke decay stage, both the current and M components superimposed on it show good linear correlations with the channel luminosity, with correlation coefficients of 0.981 and 0.988, respectively. However, the slope values of correlation fitting lines for the channel current versus the channel luminosity of the return stroke decay stage and the subsequent continuing current are obviously different. For M components superimposed on the long continuing current, a time delay for the peak luminosity relative to the peak current is revealed, and it is found that a smaller pulse peak current corresponds to a larger delay time.

Three-dimensional Numerical Simulation of the Protective Effect of Tall Building on Short Building

Wu Meng, Tan Yongbo, Lin Yuhe, Wang Xuewen

2023, 34(6): 749-758. DOI: 10.11898/1001-7313.20230610

[FullText HTML] (24) [PDF 3130KB](25)

Abstract:
Tall buildings distort the electric field of the surrounding environment, resulting in a relatively strong electric field at top corners, which affects the lightning strike process, and the protection effect that short buildings receive from tall buildings varies with the distance between them. A three-dimensional fine-resolution lightning attachment model with multiple upward leaders (LAMM) is used to simulate protection effect of a tall building and a short building with different height and distance, with an isolated building set as a control group. Experimental results show that when there is a tall building with a short building in the space and two buildings are close to each other, the distortion range of the tall building almost completely contains the distortion range of the short building, the development of downward leader is completely affected by the tall building, and the shielding effect of the tall building on the short building is significant. With the increase of the building distance, the shielding effect of the tall building on the short building decreases exponentially. When tall building is 250 m high and the short building is 150 m high, probabilities of lightning strikes on short building with distance from 50 m to 600 m with interval of 50 m are 8.3%, 15.0%, 26.5%, 36.7%, 39.5%, 47.5%, 58.9%, 57.0%, 56.0%, 57.2%, 61.0%, and 62.5%. When there is a cut-off point where the increasing trend of the probability of lightning strikes on short buildings appears to slow down significantly, the probability of lightning strikes on short building differs from its probability of lightning strikes in isolation by only 3.6%. Comparing the lightning strike results for different building distance and isolated building, the difference in lightning strike results caused by the presence of tall building decreases from 44.5% to 22.7% when the horizontal distance between tall and short buildings is increased from 400 m to 600 m. Given the height of tall building, the probability of lightning strikes on short building with different heights follows a similar trend from a large increase to a flat increase with an inflection point. Corresponding to short building height from 50 m to 200 m, the horizontal distances reaching cut-off point are 300, 450, 550 m and 600 m, respectively. When two buildings are far separated, the effect of tall building on the probability of lightning strikes on short building is weaker, and it can be assumed that there is a state when tall building have no effect on the lightning attachment process.

Vol.34, NO.6, 2023