国家级数值预报产品分发系统

随着９２１０工程系统建设的完成，Ｔ１０６中期数值预报产品分发采用了地面通信系统和卫星通信系统两种传输路径。地面通信系统由于受传输能力限制，国家气象中心根据区域、省，地市级的不同接收能力，采取了３种不同的传输方式以满足各级气象台站的需求；     

参考大气谱方法在国家气象中心谱模式上的应用

参考大气引入中期谱模式，能有效地减少截谱误差，提高预报精度。文章主要叙述了参考大气在国家气象中心业务预报谱模式上的进一步发展。计算结果表明，引入参考大气后，T63预报质量和降水的TS评分均有一定程度的提高。     

DOES WIND STRESS DEPEND ON SEA-STATE OR NOT? – A STATISTICAL ERROR ANALYSIS OF HEXMAX DATA

There have been several claims, either explicit or by implication, either based on experimental evidence or on theoretical reasoning, that the wind stress is modified by the stage of development of the wind sea. However, the overall evidence is weak, because theories are still incomplete and because it is questionable whether the sea-state effect, which is of the order of 10%, can be separated from experimental noise, which is of the order of 20%. In this paper a rigorous statistical analysis of HEXMAX data is pursued in order to establish the significance of sea-state effects. It appears that the enhanced drag, especially at high winds, which has already been established by previous analyses, cannot be attributed to the effect of young waves. The analysis provides no clues for the actual mechanism, which could be related to breaking or shoaling waves. As the effect of sea-state on wind stress is much smaller than the experimental noise level, it is hard to detect. Nevertheless, HEXMAX seems to contain a wave effect that is at the edge of statistical significance. It is, however, not the wave age itself that influences the drag, but a parameter involving wave height.Because the HEXMAX evidence is only indicative, we conclude that the issue set out in this paper cannot be answered on the basis of the HEXMAX data alone. It is recommended that error analyses are also carried out for other relevant observational data sets and that new measurements with suppressed noise will be taken up.     

Multiresolution Flux Decomposition

Geophysical variables are orthogonally decomposed by averaging timeseries using different averaging lengths, referred to as a (Haar)multiresolution decomposition. This simple and economic decomposition isassociated with cospectra that formally satisfy Reynolds averaging rules foreach averaging length. The multiresolution decomposition provides a naturalestimate of the random error in estimating a mean turbulent flux. The Fourierand multiresolution decompositions are compared using aircraft data fromBOREAS.     

用卡尔曼滤波预报南海热带气旋路径的试验

利用卡尔曼滤波方法预报南海热带气旋路径，发现采用卡尔曼滤波可以大大地降低预报误差，提高预测模型的预报能力。     

古洪水流量的误差计算

古洪水研究已在大型工程和长江三峡，黄河小浪底等的设计洪水中得到动用取得了令人虚心成功，但古洪水流量的推求与实测量的计算往往有大的差别，根据随机误差的传播给出了一般情况下的古洪不流量推误差计算公式，以小流底２３６０ａＢＰ古洪水注同误差的计算为例，对最终计算出的古洪水流量成果给出了一个误差范围。     

水库流域入库洪水预报误差分析

由于受水库枢纽工程的影响，使得反推计算的入库流量本身就存在较大的误差，从而增加了入库洪水预报模型研制的难度，影响了这时预报精度的考核。通过对全国诸多水电厂水库流域实时洪水预报的实践，对其入库洪水预报中存在的误差原因及解决方法进行了科学的分析和阐述，以指导实时预报。     

西南极菲尔德斯GPS形变网监测数据的误差特性分析与处理

本文对西南极菲尔德斯形变网ＧＰＳ监测数据的误差特性进行分析，利用数理统计原理检验了数据中的系统误差，提出了削弱这类系统误差的方法。该方法有３个特点：１．以监测网中两个稳定点为基准；２．对监测网数据进行尺度因子改正与坐标变换迭代计算；３．归算后监测网点位结构不变。经改化处理后的数据，基本上消除了系统误差的影响。     

灰色预测模型在径流长期预报中的应用

基于灰色系统的建模理论，利用河西地区有关河流的径流观测资料，建立了一个ＧＭ（１，１）残差序列周期修正径流预测模型，并用于春旱缺水期３～６月河流来水量的长期预报，经生产部门验证，准确率在８０％以上     

应用Kriging方法研究格尔木河流域地下水位动态观测网的优化配置

地下水动态是水文地质研究的一个重要方面，如何获满足一定精度要求全面而合理的地下水位动态资料具有现实经济效益。本文研究格尔木河流域地下水位动态测网，通过选取理想的变差函，运有Ｋｒｉｇｉｎｇ方法进行地下水位的线性无偏最优估计和计算估计误差的标准差，结合给定的允许误差限评判地下水位动态观测网的配置是否合适，并提出调整现有观测网的方案