SVM和Elman神经网络在乌鲁木齐污染物预报中的应用研究

选用2012年11月1日-2013年1月31日的逐6 h的空气污染物（SO2、NO2、PM10）和实况气象要素（温度、湿度、能见度、风速和气压）资料,利用支持向量机和Elman神经网络方法建立空气污染物预报模型。结果表明,支持向量机和Elman神经网络方法都可以得到较为理想的预测结果,支持向量机在泛化能力方面具有显著优势,预测结果更加准确。     

基于支持向量机的雷达地物回波识别研究

多普勒天气雷达探测过程中的非气象因子会显著影响雷达资料的定量化应用,在雷达基数据的应用前需对雷达资料进行抑制地物杂波、去距离折叠和退速度模糊等质量控制。本文在现有的自动识别地物回波方法的基础上,提出了基于支持向量机（Support Vector Machine,SVM）识别雷达地物杂波的方法,2013年6-8月对安庆和常州两地的CINRAD/SA雷达观测资料进行雷达地物回波识别,并将其与运用人工神经网络（Artificial Neural Networks,ANNs）识别的结果进行对比,结果表明支持向量机方法能够取得更好的效果。在地物、降水回波总样本识别和地物回波识别方面更为有效;降水回波的误判方面,神经网络略优于支持向量机,但两者差异不大,都将降水回波的误判率控制在了一个较小的范围内;另外支持向量机方法较之神经网络方法对于训练样本数目的依赖性较小,在训练样本较少时,支持向量机方法仍能保持有效的识别效果。     

支持向量机方法在天气预报中的应用

简要介绍了支持向量机方法(SVM)方法的基本原理和使用方法。用高空500hPa月平均高度、海洋温度以及地面资料作为因子，对西安6—9月份降水总量建立了SVM预报模型。     

Ensemble sensitivity analysis of an extreme rainfall event over the Himalayas in June 2013

Forecast Sensitivity of an extreme rainfall event over the Uttarakhand state located in the Western Himalayas is investigated through Ensemble-based Sensitivity Analysis (ESA). ESA enables the assessment of forecast errors and its relation to the flow fields through linear regression approach. The ensembles are initialized from an Ensemble Kalman Filter (EnKF) Data Assimilation in Weather Research and Forecast (WRF) model. ESA is then applied to evaluate the dynamics and predictability at two different days of the extreme precipitation episode. Results indicate that the precipitation forecast over Uttarakhand is sensitive to the mid-tropospheric trough and moisture fields for both the days, in general. The day 1 precipitation shows negative sensitivity to the trough over upstream regions of the storm location while in day 2, the sensitive region is found to be located over the southward intruded branch of the mid–tropospheric trough. Perturbations introduced in the initial conditions (IC) over the most sensitive region over the west of the storm location indicate significant variations in the forecast location of precipitation. IC perturbed experiments show that the perturbation amplitude is correlated linearly with predicted change in precipitation, which becomes nonlinear as the forecast length increases. ESA performed on convection-permitting ensembles show that precipitation over the Uttarakhand is mostly non-convective. However, when the location of the response function box is moved north-westward of the Uttarakhand, the sensitivity patterns show signs of convection.     

Comprehensive modeling of monthly mean soil temperature using multivariate adaptive regression splines and support vector machine

Soil temperature (T _s) and its thermal regime are the most important factors in plant growth, biological activities, and water movement in soil. Due to scarcity of the T _s data, estimation of soil temperature is an important issue in different fields of sciences. The main objective of the present study is to investigate the accuracy of multivariate adaptive regression splines (MARS) and support vector machine (SVM) methods for estimating the T _s. For this aim, the monthly mean data of the T _s (at depths of 5, 10, 50, and 100 cm) and meteorological parameters of 30 synoptic stations in Iran were utilized. To develop the MARS and SVM models, various combinations of minimum, maximum, and mean air temperatures (T _min, T _max, T); actual and maximum possible sunshine duration; sunshine duration ratio (n, N, n/N); actual, net, and extraterrestrial solar radiation data (R _s, R _n, R _a); precipitation (P); relative humidity (RH); wind speed at 2 m height (u ₂); and water vapor pressure (V_p) were used as input variables. Three error statistics including root-mean-square-error (RMSE), mean absolute error (MAE), and determination coefficient (R ²) were used to check the performance of MARS and SVM models. The results indicated that the MARS was superior to the SVM at different depths. In the test and validation phases, the most accurate estimations for the MARS were obtained at the depth of 10 cm for T _max, T _min, T inputs (RMSE = 0.71 °C, MAE = 0.54 °C, and R ² = 0.995) and for RH, V _p, P, and u ₂ inputs (RMSE = 0.80 °C, MAE = 0.61 °C, and R ² = 0.996), respectively.     

基于公众天气预报预测塑料大棚逐日极端气温

利用浙江省慈溪市的公众天气预报和草莓大棚内极端气温的观测数据,构建一个以室外日最高气温、最低气温、相对湿度、最大风级、白天和夜间天空状况作为输入变量,棚内日最高气温和日最低气温作为输出变量的BP神经网络预测模型。用以预测草莓大棚室内日最高气温和日最低气温。结果表明,该模型对大棚内日最高气温、日最低气温的训练值和实测值之间的均方根误差分别为4.0℃和1.3℃,绝对误差则分别为3.2℃和1.0℃;日最高气温和日最低气温的预测值和实测值之间的均方根误差分别为3.6℃和1.2℃,绝对误差为3.0℃和1.0℃。该模型数据获取方便,实用性强,模拟精度较高,可以较准确的预测未来温室内的极端气温,为温室管理和调控提供依据。     

基于EC细网格产品在乌鲁木齐机场低能见度预测中的释用

利用2015—2018年乌鲁木齐机场航空例行天气报告(METAR报)、ECMWF(European Centre for Medium-Range Weather Forecasting)细网格数值预报产品对影响能见度的主要因子进行分析,提取与低能见度相关性高的物理量作为预报因子,采用SVM方法,分别基于Poly、RBF核函数建立乌鲁木齐机场未来21 h能见度预报模型。结果表明:(1)基于预报因子区间分类的SVM模型物理意义明确,试验结果较好;以RBF为函数建立的SVM模型(SVM-RBF)预报能力更好,其训练样本预测的TS评分0.84,准确率89.20%。(2)SVM-RBF模型的检验样本中,预报准确样本的预报误差整体偏小;在漏报样本中则有能见度越低、预报误差越大的特点,模型的振荡性明显。(3)结合NCEP/NCAR再分析资料研究SVM-RBF模型对天气过程的预报表现,发现模型对于特定天气形势下引发的低能见度天气,预报误差较小且预报提前量较大。     

Evaluation of radar and automatic weather station data assimilation for a heavy rainfall event in southern China

To improve the accuracy of short-term(0–12 h) forecasts of severe weather in southern China, a real-time storm-scale forecasting system, the Hourly Assimilation and Prediction System(HAPS), has been implemented in Shenzhen, China. The forecasting system is characterized by combining the Advanced Research Weather Research and Forecasting(WRF-ARW)model and the Advanced Regional Prediction System(ARPS) three-dimensional variational data assimilation(3DVAR) package. It is capable of assimilating radar reflectivity and radial velocity data from multiple Doppler radars as well as surface automatic weather station(AWS) data. Experiments are designed to evaluate the impacts of data assimilation on quantitative precipitation forecasting(QPF) by studying a heavy rainfall event in southern China. The forecasts from these experiments are verified against radar, surface, and precipitation observations. Comparison of echo structure and accumulated precipitation suggests that radar data assimilation is useful in improving the short-term forecast by capturing the location and orientation of the band of accumulated rainfall. The assimilation of radar data improves the short-term precipitation forecast skill by up to9 hours by producing more convection. The slight but generally positive impact that surface AWS data has on the forecast of near-surface variables can last up to 6–9 hours. The assimilation of AWS observations alone has some benefit for improving the Fractions Skill Score(FSS) and bias scores; when radar data are assimilated, the additional AWS data may increase the degree of rainfall overprediction.     

Simulation of extreme rainfall and projection of future changes using the GLIMCLIM model

Theoretical and Applied Climatology - In this study, the performance of the Generalized LInear Modelling of daily CLImate sequence (GLIMCLIM) statistical downscaling model was assessed to simulate...     

Modelling Korean extreme rainfall using a Kappa distribution and maximum likelihood estimate

Summary Attempts to use the 4-parameter Kappa distribution (K4D) with the maximum likelihood estimates (MLE) on the summer extreme daily rainfall data at 61 gauging stations over South Korea have been made to obtain reliable quantile estimates for several return periods. A numerical algorithm for searching MLE of K4D by minimizing the negative log-likelihood function with penalty method has been described. The isopluvial maps of estimated design values corresponding to selected return periods have been presented. The highest return values are centered at sites in the south-western part of the Korean peninsula. The distribution of return values for annual maxima of 2-day precipitation (AMP2) is more similar to the climatological features of annual total precipitation of Korea than that of annual maxima of daily precipitation (AMP1). Our results of return values delineate well the horizontal patterns of the heavy precipitation over the Korean peninsula. Received January 15, 2001 Revised October 8, 2001     

Comparison of classification and clustering methods in spatial rainfall pattern recognition at Northern Iran

Pattern recognition is the science of data structure and its classification. There are many classification and clustering methods prevalent in pattern recognition area. In this research, rainfall data in a region in Northern Iran are classified with natural breaks classification method and with a revised fuzzy c-means (FCM) algorithm as a clustering approach. To compare these two methods, the results of the FCM method are hardened. Comparison proved overall coincidence of natural breaks classification and FCM clustering methods. The differences arise from nature of these two methods. In the FCM, the boundaries between adjacent clusters are not sharp while they are abrupt in natural breaks method. The sensitivity of both methods with respect to rain gauge density was also analyzed. For each rain gauge density, percentage of boundary region and hardening error are at a minimum in the first cluster while the second cluster has the maximum error. Moreover, the number of clusters was sensitive to the number of stations. Since the optimum number of classes is not apparent in the classification methods and the boundary between adjacent classes is abrupt, use of clustering methods such as the FCM method, overcome such deficiencies. The methods were also applied for mapping an aridity index in the study region where the results revealed good coincidence between the FCM clustering and natural breaks classification methods.     

最小二乘支持向量机在云量预报中的应用

基于2003-2006年逐年1、8月WRF区域数值预报产品和单站观测资料,采用最小二乘支持向量机回归方法,结合选取合适的参数和核函数,分别按月通过不同长度样本序列建立了台北和厦门站总云量和低云量短期释用预报模型,利用2007年1、8月样本资料对模型进行了预报和检验,并与神经网络方法进行了对比.结果表明:最小二乘支持向量机回归方法的预报效果要好于神经网络方法;两站不同长度样本的总云量和低云量预报模型,预报效果较好,其预报准确率不会因为训练样本的减少而降低.可见,最小二乘支持向量机回归在云量等气象要素释用预报方面,具有较好的应用前景.     

Analysis of spatial and temporal extreme monsoonal rainfall over South Asia using complex networks

We present a detailed analysis of summer monsoon rainfall over the Indian peninsular using nonlinear spatial correlations. This analysis is carried out employing the tools of complex networks and a measure of nonlinear correlation for point processes such as rainfall, called event synchronization. This study provides valuable insights into the spatial organization, scales, and structure of the 90th and 94th percentile rainfall events during the Indian summer monsoon (June–September). We furthermore analyse the influence of different critical synoptic atmospheric systems and the impact of the steep Himalayan topography on rainfall patterns. The presented method not only helps us in visualising the structure of the extreme-event rainfall fields, but also identifies the water vapor pathways and decadal-scale moisture sinks over the region. Furthermore a simple scheme based on complex networks is presented to decipher the spatial intricacies and temporal evolution of monsoonal rainfall patterns over the last 6 decades.     

Role of unusual moisture transport across Equatorial Indian Ocean on the extreme rainfall event during Kerala flood 2018

Southwestern Indian state, Kerala experienced extreme devastating statewide flood event of the century during 2018 monsoon season. In this study, an attempt has been made to bring out the salient dynamical factors contributed to the Kerala flood. There were 3 active spells over Kerala during 2018 Monsoon season. All the three spells were accustomed with the intrinsic factors of low frequency components of the active spells such as strength of monsoon Low Level Jet (LLJ), Monsoon depressions in the Bay of Bengal, favorable Madden-Julian oscillation (MJO) phases and Western Pacific systems. Though all the common ingredients remain same, the third spell is distinct with the less evaporation flux over Western and Central Arabian Sea and unusual moisture transport from maritime continent through South Equatorial Indian ocean (SEIO) towards the Kerala coast across Equator. Strong meridional pressure gradient force created by the combined effect of high pressure anomaly oriented Northwest-Southeast direction across tropical Indian ocean and anomalous low pressure along monsoon trough might have contributed to this unusual moisture transport across SEIO originating from west of Australia. The anomalous high pressure in South Indian ocean was greatly influenced by the position of the Mascarene high. Subtropical Indian ocean dipole (SIOD) also exhibits an influential role by altering tropical Southern Indian ocean dynamics in favor of the unusual moisture transport. The position of the monsoon depression and presence of typhoons in Western Pacific might have aided to this moisture transport. However, the normal moisture transport from Central Arabian sea towards Kerala coast by virtue of the strong LLJ along with additional moisture transport directly from South of maritime continent through SEIO across the Equator towards Kerala coast might have played a dominant role in the historical flood event over entire Kerala state.     

印江极端降水和极端强降水日数统计特征

利用印江气象站的逐日降水资料,在对该站点极端降水和极端强降水过程阈值进行科学界定的基础上,对50 a来极端降水和极端强降水过程进行了常规统计。结果表明:印江极端降水和极端强降水过程有弱的增加趋势,极端(强)降水天数与降水量成正相关,20世纪60年代极端降水天数和极端强降水的离散程度最大,21世纪前8 a极端强降水的离散程度最小。极端降水和极端强降水日数变化均达不到气候突变的标准。     

Least squares support vector machine for short-term prediction of meteorological time series

The prediction of meteorological time series plays very important role in several fields. In this paper, an application of least squares support vector machine (LS-SVM) for short-term prediction of meteorological time series (e.g. solar irradiation, air temperature, relative humidity, wind speed, wind direction and pressure) is presented. In order to check the generalization capability of the LS-SVM approach, a K-fold cross-validation and Kolmogorov–Smirnov test have been carried out. A comparison between LS-SVM and different artificial neural network (ANN) architectures (recurrent neural network, multi-layered perceptron, radial basis function and probabilistic neural network) is presented and discussed. The comparison showed that the LS-SVM produced significantly better results than ANN architectures. It also indicates that LS-SVM provides promising results for short-term prediction of meteorological data.     

鲁西北持续性暴雨非地转湿Q矢量分析

利用常规观测资料、NECP/NCAR提供的1°×1°FNL全球再分析资料,对2012年鲁西北一次持续性暴雨进行了湿Q矢量方法诊断分析。结果表明：此次持续暴雨出现在有利的环流背景下,降水区域集中并有明显的中尺度特征,湿Q矢量方法是分析强降水落区很好的工具;925～850 hPa湿Q矢量散度与强降水落区有较好的对应关系,但暴雨并不总是出现在湿Q矢量散度负值区中心,有时出现在湿Q矢量散度梯度大的负值区一侧;700 hPa湿Q矢量涡度正值中心与散度负值重叠的区域是中尺度低值系统发展有利的区域,与未来6～12 h暴雨落区有很好的对应;湿Q矢量锋生函数差值预报强降水落区明显优于锋生函数。     

基于数字摄像能见度仪的北京地区降雨和雾霾天气能见度对比分析

数字摄像能见度仪（Digital Photographic Visibility System,DPVS）仿照人工目测能见度的原理测量大气能见度。本文应用2017年3—8月北京地区DPVS、前向散射仪（PWD22）、大气透射仪（LT31）三种观测仪器在降雨天气和雾霾天气观测数据进行了对比。结果表明：能见度观测数据与相对湿度、颗粒物浓度、降水粒子等要素之间有明显的负相关性;在低能见度天气条件下,三种仪器观测数据变化趋势一致,但存在一定的差异;DPVS在中雨天气、大雨天气、暴雨天气和中度雾霾天气中,观测数据离散性更小,稳定性更好。但DPVS在白天和夜间的交替过渡期观测值不够稳定,这也是今后算法优化的重点方向。     

Impact of land-use changes on the genesis and evolution of extreme rainfall event: a case study over Uttarakhand,India

In recent years, hot summers (HS) have played an important role in affecting people’s health and causing natural disasters. However, it is not very clear what HS should be attributed to. In order to investigate that, the anomalous of anticyclone associated with HS in the Mediterranean and North China is examined and compared by using data during the time period 1949~2018. Statistical analysis shows that summer temperature in the Mediterranean and North China is revealing a good correlation when removing the global warming trend. The composite results indicate that the anomalous warming during HS over different regions is both dominantly controlled by an anomalous anticyclone, which enhances the subsidence adiabatic heating of the temperature. Furthermore, the subtropical high (STH) variations faithfully represent the fluctuations in summer temperature over the Mediterranean (cor = 0.78) and North China (cor = 0.80). The anticyclonic anomalies over both focus areas are associated with North Atlantic and Northwest Pacific sea surface temperature (sst), respectively. These results indicate that the total influence of the STH position anomaly should be taken into consideration in different places during HS. However, whether such atmosphere-ocean feedback can be improved by numerical experiments is worthy to be further studied.