| 摘要: |
| 【目的】研究给定设计标准下水文设计值置信区间的估算方法，评定设计值的不确定性。【方法】以陕西省关中地区4个气象测站年降水量为例，选用矩法（MOM）和极大似然法（ML）研究P-Ⅲ分布水文设计值的置信区间计算方法，计算出不同设计频率下水文设计值在95%置信水平下的置信区间，并对基于矩法和极大似然法估算的设计值置信区间的长度数据进行比较分析。【结果】在设计频率为0.2%时，周至、临潼、蓝田、户县4个测站采用极大似然法估算在95%置信水平下的设计值置信区间长度分别为414.28，338.88，360.65和371.24，采用矩法估算在95%置信水平下的设计值置信区间长度分别为546.61，435.64，384.91和403.92。在设计频率为99%时，上述4站采用极大似然法估算在95%置信水平下的设计值置信区间长度分别为144.44，124.93，177.67和160.73，采用矩法估算在95%置信水平下的设计值置信区间长度分别为183.28，151.26，187.31和178.62。由此可知，在低设计频率和高设计频率时，极大似然法估算的设计值置信区间长度均较矩法小。【结论】置信区间能够为设计值的估算提供相对可靠的变化范围，用极大似然法估算水文设计值的置信区间比矩法能更有效地降低水文设计值的不确定性。 |
| 关键词: 水文设计值 设计标准 标准差 置信区间 关中地区 |
| DOI： |
| 分类号: |
| 基金项目:国家自然科学基金项目(51179160，50879070，50579065)；高等学校博士学科点专项科研基金项目(20110204110017) |
|
| Estimation of confidence interval for hydrologic design values |
|
LIANG Jun,SONG Songbai
|
| Abstract: |
| 【Objective】This paper studied estimation methods for hydrologic quantile confidence interval under given design standard and assessed the uncertainty of design value.【Method】MOM and ML methods were used to study P-Ⅲ distribution hydrologic quantile confidence interval and figure out 95% confidence interval of hydrologic quantile estimation under different design frequencies at four meteorological stations in Guanzhong area.The results from MOM and ML were also compared and analyzed.【Result】At the design frequency of 0.2%,confidence interval lengths of hydrologic quantile under 95% confidence level based on ML for the 4 hydrologic stations of Zhouzhi,Lintong,Lantian and Huxian were 414.28,338.88,360.65,and 371.24,while those based on ML were 546.61,435.64,384.91,and 403.92,respectively.At the design frequency of 99%,confidence interval lengths of hydrologic quantile under 95% confidence level based on ML for Zhouzhi,Lintong,Lantian and Huxian were 144.44,124.93,177.67,and 160.73,while those based on ML were 183.28,151.26,187.31,and 178.62,respectively.Independent of design frequency,confidence interval lengths of hydrologic quantile based on ML were always smaller than based on MOM.【Conclusion】Confidence interval can provide relatively reliable range of variation for the hydrologic quantile estimation and the estimated confidence interval lengths based on ML were more effective in reducing the uncertainty of hydrologic quantile. |
| Key words: hydrologic quantile design standard standard error confidence interval Guanzhong area |
