扩展经验正交函数(EEOF)及其在月、季降水预测中的应用

本文提出一种降水长期预测的新方案，用扩展经验正交函数(EEOF)展开连续月组成的月（季）降水分布场，求取各月的特征向量场和对应的时间权重系数，分析各场的天气学意义及前后承替的相互关系，用前期出现的特征向量场的特征来预测后期的降水场分布趋势，同时利用特征向量场所对应的时间系数作二维点聚图，估算预报月（季）份的降水总趋势。     

运用经验正交函数分解建模制作麦播期降水定量预报

针对鹤壁市麦播期（９月２１日￣１０月３１日）降水量的高频振荡性，利用均值生成函数的延柘函数构成的序列，提取时间函数，作为主分量分析方法的时间函数场，制作鹤壁市麦播期降水定量预报方程并进行试报。     

均生函数模型在降水预测中的应用

采用均生函数模型对单一的要素序列作气候预测试验。首先计算出要素的均生函数，经逐步回归，筛选出周期性预报因子，以期建立历史拟合和预测效果好的模型。实例计算分析表明，均生函数模型预测的结果具有一定的可信度。     

经验正交函数在解释数值预报产品中的应用

本文用经验正交函数展开的分析方法，确定了各种数值预报产品物理场中和对应时效的运城日降水量显著相关的物理量要素及关键区域，对关键区域的物理量场使用典型相关分析进行处理，得到了数量较少，而又能代表该关键区物理量分布状况主要信息的新因子，从而建立使用数值预报产品制作短期单站降水预报模式，进行预报效果较好。     

短期气候(月、季、年)逐月降水预测系统研究

描述了短期气候逐月降永预测系统的概念及其实现。首先讨论了把气候模式与人工神经网络结合起来开发一个气候预测支持系统的必要性和可能性。然后，给出了系统设计的框图，并对系统结构作了详细的讨论。最后，提出了有关预测结果分析和进一步完善的看法。     

均生函数模型在高原汛期降水预报中的应用

以均生函数为基函数,建立了正交化、主成分和最优子集回归三类预测模型。对云南昆明、滇中（包括昆明、玉溪和楚雄）及西藏林芝的汛期（６￣８）降水进行了模拟和预测、其拟合效果,特别是对极值的拟合十分理想。结果表明：在高原地区,均生函数模型预报降水具有较高的精度,拟合报准率和实际报准率分别为９７％￣１００％和６０％￣８０％,且具有多步预测能力。     

利用经验正交函数方法（EOF）浅析中国夏季降水时空分布特征

根据1951～2000年中国夏季月降水资料，用经验正交函数方法(EOF)分析了中国夏季降水分布特征和旱涝情况。     

均生函数预测方法在青藏高原冬季降水预报中的应用及改进

均生函数预测方法在天气预报等领域得到了越来越广泛的应用，但预测方法没有涉及到图形处理。在对方法、原理及过程进行深入细致的研究后，重新编写了程序的头部、中部、尾部语句及子过程调用函数。经多次调试试验，实现了人口资料调用灵活、多变、计算模拟结果屏幕显示等，提高了预测方法的整体实用性和预测信息的可视化程度。检验得知：该方法对青藏高原冬季降水的预报具有较高的预报精度。     

均生函数的改进模型在重庆地区主汛期降水预测中的应用试验

本文应用均生函数模型及其改进模型对重庆地区 主汛期(6～8月 )降水进行预测试验，结果表明：针对均生函数固有的缺点提出的改进方案所建立的数学预 报模型具有更好的预测效果，是一种具有较高使用价值的长期预报方法。     

经验正交函数分解质量控制法在地面观测资料变分同化中的个例研究与应用

同化地面观测资料能够获得丰富的地面大气信息,这对于大气边界层的准确模拟尤为重要。由于地面观测资料同化一直受到地面观测资料质量较差的影响,因此,地面观测资料的质量控制是提高地面资料同化效果的重要方法之一。为了分析基于经验正交函数分解质量控制方法(Empirical Orthogonal Function quality control,EOF-QC)对地面资料同化效果的影响,并进一步检验该方法在实际同化试验中的应用效果,在WRF的三维变分同化系统中引入了经验正交函数分解质量控制法,同时通过一系列同化试验比较了经验正交函数分解质量控制法与原系统自带的基于观测与模拟偏差质量控制法(Observation Minus Background quality control,OMB-QC)的差异。2008年1和7月的多个强降水预报试验结果表明,经验正交函数分解质量控制法能够保留更多天气系统的有效观测信息,更为客观准确地反映大气真实状态;同化经过经验正交函数分解质量控制法后的观测资料,模式预报的温度降低,在北部形成一个气旋性环流,该环流底部的偏西气流带动北部冷空气东移入海,同时冷空气南下也削弱了带有丰富水汽的西南气流,从而使模式预报的降水范围和强度更加合理。降水的空间分布对比结果也表明,经验正交函数质量控制法改善了模式对降水落区和强度的预报能力,各个量级的降水评分有明显提高,模拟结果更接近于实况。各组数值模拟试验结果表明,经验正交函数分解质量控制法在WRF-3DVAR中具有较高的应用潜力。     

广东省降水的多尺度时空投影预测方法

采用多尺度时空投影(MSTP)预测思路建立广东月降水和季节降水预测方法。通过EOF分解、小波分析和Lanczos滤波方法进行周期分解, 采用MSTP方法进行预测。借鉴年际增量法, 对预报结果用最小二乘法进行误差订正, 得到降水预测结果。PS预测评分和均方根误差10年独立样本检验(2006—2015年)结果显示:订正后, PS预测评分起伏较小, 68.8%的月降水和季节降水PS预测评分明显提高的年份超过6年, 且有87.5%的月降水和季节降水PS预测平均分达到70以上; 在±0.5个标准差范围内, 订正后均方根误差在40%以上的概率分布明显高于订正前, 订正后的月和季节降水占81.3%, 订正前占31.3%;在±1个标准差范围内, 概率分布在70%以上的月季降水订正前后相差不多, 订正后占56.3%, 订正前占50%。     

动力延伸(月)数值天气预报中的信息提取和减小误差试验

用一个全球谱模式通过较多个例的月数值天气预报试验,研究了预报结果的有用信息提取问题.模式预报误差的谱分析表明,纬向平均(零波)场误差占很大比例,试验了两种用气候倾向改善纬向平均(零波)场误差的方案,一是对逐日预报结果进行订正,二是在积分过程中进行订正,两种方案都取得了一定成效.     

Seasonal Prediction of Monthly Precipitation in China Using Large-Scale Climate Indices

In this study, seasonal predictions were applied to precipitation in China on a monthly basis based on a multivariate linear regression with an adaptive choice of predictors drawn from regularly updated climate indices with a two to twelve month lead time.A leave-one-out cross validation was applied to obtain hindcast skill at a 1% significance level.The skill of forecast models at a monthly scale and their significance levels were evaluated using Anomaly Correlation Coefficients(ACC) and Coefficients Of De...     

Empirical Orthogonal Function Analysis of Daily Rainfall in the Upper Reaches of the Huai River Basin, China

Summary  The design and operation of hydro-structures for flood control and water conservation bring a need for improved characterization of precipitation patterns. A 73 000 km² study area in East Central China is situated in the East Asian monsoon region and experiences a strong seasonality in the rainfall regime. The characteristics of daily rainfall from 230 gauges during 1967–1986 were investigated for four periods in the summer monsoon season using empirical orthogonal function analysis (EOF) and extended empirical orthogonal function analysis (EEOF). The EOF analysis showed that for all four periods most of the variance was explained by an elongated spatial rainfall pattern. The pattern varied in direction, from roughly west-east to southwest-northeast, in the different periods. The zonally oriented patterns were interpreted as being caused by the stationary Mei-Yu front and the southwest-northeast patterns interpreted as cold fronts in cyclones that were developing over the study area. The latitude of the rain belt described by the first mode moved slightly northward with the advance of the East Asian monsoon from the first period, 9–22 June, to the third period, 23 July–5 August, and then withdrew southward again in accordance with the known seasonal movement of the Mei-Yu front. The EEOF analysis was used to show the development of the rainfall area over sequences of three days. During all four periods rainfall intensified on the second day, compared to the first and third days. During the first and last periods, 9–22 June and 5 August–30 September, respectively, there appeared to be little movement in the rainfall. During the second and third periods, the patterns were interpreted as a cold front in a developing cyclone. The results show the connection between the temporal variation in rainfall intensity and the temporal succession of spatial patterns over three day periods and should be used in the construction of design rainfalls for the study area. Received February 10, 1998 Revised June 23, 1998     

A Technique of Rainfall Assimilation for Dynamic Extended Range Monsoon Prediction

Summary ?A methodology has been developed to assimilate satellite-measured rainfall during the initial phase of model integration for extended range monsoon prediction. The vertical profiles of latent heating corresponding to different rain rates have been derived from the model statistics. These heating rates have been assimilated through nudging in the thermodynamics equation of the model. This procedure of assimilating observed heating has corrected the simulation of heating location in the model and consequently removed the anomalous sinking motion over Indian landmass. With the correction of vertical circulation, both mean July rainfall over India and the distribution have improved. Interannual variability has been brought out for the years 1987 and 1988. In view of the availability of rainfall profile from Tropical Rain Measuring Mission (TRMM) there is a scope of adopting this method of assimilating observed rainfall, for extended range monsoon prediction. Received February 12, 1999/Revised May 4, 1999     

Dynamical and Machine Learning Hybrid Seasonal Prediction of Summer Rainfall in China

Seasonal prediction of summer rainfall is crucial to reduction of regional disasters, but currently it has a low prediction skill. We developed a dynamical and machine learning hybrid(MLD) seasonal prediction method for summer rainfall in China based on circulation fields from the Chinese Academy of Sciences(CAS) Flexible Global Ocean–Atmosphere–Land System Model finite volume version 2(FGOALS-f2) operational dynamical prediction model. Through selecting optimum hyperparameters for three machine learning methods to obtain the best fit and least overfitting, an ensemble mean of the random forest and gradient boosting regression tree methods was shown to have the highest prediction skill measured by the anomalous correlation coefficient. The skill has an average value of 0.34 in the historical cross-validation period(1981–2010) and 0.20 in the 10-yr period(2011–2020) of independent prediction, which significantly improves the dynamical prediction skill by 400%. Both reducing overfitting and using the best dynamical prediction are important in applications of the MLD method and in-depth analysis of these warrants a further investigation.     

提高月降水量预报精度的一种尝试

用原始数据建立的多元自回归预报模型,虽然对气压,气温,绝对湿度的预报精度很高,但对月降水量的预报粗度较低,改用自典型相关因子建模,大大提高了月降水量的预报精度,对气压,气温,绝对湿度的预报亦有所改进。     

Monthly Extended Predicting Experiments with Nonlinear Regional Prediction. Part I: Prediction of Zonal Mean Flow

Systematic errors have recently been founded to be distinct in the zonal mean component forecasts,which account for a large portion of the total monthly-mean forecast errors. To overcome the difficulty of numerical model, the monthly pentad-mean nonlinear dynamic regional prediction models of the zonal mean geopotential height at 200, 300, 500, and 700 hPa based on a large number of historical data (NCEP/NCAR reanalysis data) were constituted by employing the local approximation of the phase space reconstruction theory and nonlinear spatio-temporal series prediction method. The 12-month forecast experiments of 1996 indicated that the results of the nonlinear model are better than those of the persistent, climatic prediction,and T42L9 model either over the high- and mid-latitude areas of the Northern and Southern Hemispheres or the tropical area. The root-mean-square of the monthly-mean height of T42L9 model was considerably decreased with a change of 30.4%, 26.6%, 82.6%, and 39.4%, respectively, over the high- and mid-latitudes of the Northern Hemisphere, over the high- and mid-latitudes of the Southern Hemisphere, over the tropics and over the globe, and also the corresponding anomaly correlation coe cients over the four areas were respectively increased by 0.306-0.312, 0.304-0.429, 0.739-0.746, and 0.360-0.400 (averagely a relative change of 11.0% over the globe) by nonlinear correction after integration, implying that the forecasts given by nonlinear model include more useful information than those of T42L9 model.