预测10年尺度地震危险性的综合模型

本文依据信息量的定量判定方法，从分析地震地质和前兆因子与强震发生的关系出发，对各种因子所能提供的长期或１０年尺度强震危险的信息量作了估计。     

中央气象台短期降雨预报水平初步分析

该文利用1988～1993年资料，对中央气象台降雨预报水平状况进行了初步分析。结果表明:最近6年，中央气象台的有雨预报、小雨预报、中雨预报和等级综合预报的准确率和技巧水平均有不同程度的提高；暴雨和大暴雨两个等级的落点预报的准确率均超过气候概率，具有正技巧（技巧评分＞0），但技巧水平提高不明显。     

多准则模糊决策方法在旱涝趋势预报中的应用

应用“多准则模糊决策”方法，把多种预报方法以偏少，正常，偏多及总的预报准确率作为评价条件，对每种预报方法均做出总的评价，进行最佳预报方法排序，作为最终预报的依据。     

风预报的数值产品释用研究

根据地转风和偏差风原理，建立了风预报的数值产品释用方案。通过一年试验预报和３年业务预报检验，证明此方案是一个有效的实用的方案。     

一次寒潮低温过程的低频振荡特征

从大气存在低应振荡这一事实出发，阐述了１９９３年１１月１７－２１日寒潮低温过程的环流及低频振荡特征，对提高寒潮中期预报能力，有积极的意义。     

兰州盛夏大降水短期预报

根据预报员的思路，运用专家系统匹配方法，依次对大降水的环流场、影响系统、高低空物理量配置等建立推理规则，然后输入计算机算出结果。     

Systematic errors in the medium range prediction of the Asian summer monsoon circulation

The present study describes an analysis of Asian summer monsoon forecasts with an operational general circulation model (GCM) of the European Centre for Medium Range Weather Forecasts (ECMWF), U.K. An attempt is made to examine the influence of improved treatment of physical processes on the reduction of systematic errors. As some of the major changes in the parameterization of physical processes, such as modification to the infrared radiation scheme, deep cumulus convection scheme, introduction of the shallow convection scheme etc., were introduced during 1985–88, a thorough systematic error analysis of the ECMWF monsoon forecasts is carried out for a period prior to the incorporation of such changes i.e. summer monsoon season (June–August) of 1984, and for the corresponding period after relevant changes were implemented (summer monsoon season of 1988). Monsoon forecasts of the ECMWF demonstrate an increasing trend of forecast skill after the implementation of the major changes in parameterizations of radiation, convection and land-surface processes. Further, the upper level flow is found to be more predictable than that of the lower level and wind forecasts display a better skill than temperature. Apart from this, a notable increase in the magnitudes of persistence error statistics indicates that the monsoon circulation in the analysed fields became more intense with the introduction of changes in the operational forecasting system. Although, considerable reduction in systematic errors of the Asian summer monsoon forecasts is observed (up to day-5) with the introduction of major changes in the treatment of physical processes, the nature of errors remain unchanged (by day-10). The forecast errors of temperature and moisture in the middle troposphere are also reduced due to the changes in treatment of longwave radiation. Moreover, the introduction of shallow convection helped it further by enhancing the vertical transports of heat and moisture from the lower troposphere. Though, the hydrological cycle in the operational forecasts appears to have enhanced with the major modifications and improvements to the physical parameterization schemes, certain regional peculiarities have developed in the simulated rainfall distribution over the monsoon region. Hence, this study suggests further attempts to improve the formulations of physical processes for further reduction of systematic forecast errors.     

新亚欧大陆桥新疆段暴雨灾变事件的灰色预测

通过对新亚欧大陆桥新疆段水害严重区段暴雨灾变事件的分析,确定暴雨灾变事件的临界雨强,并运用灰色灾变预测方法,根据暴雨灾变序列确立相应的灾变年份序列,运用ＧＭ（１,１）模型,分别求出灾变年份序列的ＧＭ（１,１）序号响应式和灾变年份预测式。     

基于集合预报的台风“利奇马”登陆后远距离暴雨预报的敏感性分析

2019年第9号台风“利奇马”在8月10日登陆后引发了远距离大范围的暴雨,本文利用ECMWF(EC)和GRAPES全球集合预报模式等资料对暴雨短期预报的误差及原因进行了分析。此次台风远距离暴雨主要集中在8月10日夜间的山东中部地区,EC集合预报对该区域的降水量预报效果总体优于GRAPES集合预报。集合敏感性分析可以识别出和预报变量高相关(敏感)的天气系统,结果表明山东区域平均降水量对同期500 hPa副高、台风西北侧海平面气压和山东北部低层温度较为敏感,而对流层高层的高度及经向风存在更大范围的敏感区。根据暴雨预报TS评分选取EC集合预报成员作为优势组和劣势组,结果表明优势组预报成员表现为山东上空300 hPa低槽前倾,北侧高空偏南急流更强,同时配合低层台风外围偏东风急流,形成高层辐散、低层辐合的有利条件。另外,优势组预报的中纬度低层冷空气和斜压锋区更强,导致优势组在山东中部预报出暴雨,更加接近于实况。     

基于关键对流参数分级的强对流潜势预报

通过对上海地区1998—2009年4—9月各类强对流天气的统计分析,选取42个对流参数及其时间变量,采用逐步回归方法建立了针对各类强对流天气的0～12h潜势预报方程。在此基础上,提出了基于关键对流参数进行分级的强对流潜势预报方法,选取K指数、SI指数、PWV(大气可降水含量)指数和θsedif85(500hPa和850hPa假相当位温差)等反映大气热力和水汽条件的关键对流参数,根据对流分布情况将各对流参数分别分为3个等级,并分级建立了针对不同强对流天气的潜势预报方程。与未分级方程对比表明:基于关键对流参数分级的预报方程对雷雨大风、强雷电和所有对流等预报效果上有明显提升,采用如下组合评分更佳:雷雨大风的预报采用SI分类方程,强雷电和所有对流采用PWV分类方程。将基于关键对流参数分级的强对流潜势预报方法在数值预报模式中进行了业务应用,取得了较好效果。