大渡河流域超阈值降雨样本模拟及不确定性分析

发布时间：2018-06-10 02:12

本文选题：GPD模型 + 轮廓似然函数　；参考：《人民长江》2016年07期

【摘要】：大渡河流域地形十分复杂,流域内气候随高程变化差异较大。通过超阈值模型模拟大渡河流域范围内9个雨量站的逐日降水资料序列,利用极大似然估计法计算模型参数,采用概率图、分位数图、重现水平图、密度函数图4种较直观的诊断图形对模型的合理性进行了全面评估,并借助轮廓似然方法估计模型关键参数及设计强降雨的置信区间。研究结果表明,各站点超阈值降雨样本均服从Pareto分布,可以选择GPD模型作为大渡河流域强降雨统计推断的分布函数类型,轮廓似然法能反映重现期长短对设计降雨置信区间的影响。可为大渡河流域降雨不确定性的定量评估及梯级水库群洪水预报提供依据。
[Abstract]:The topography of the Dadu River basin is very complex, and the climate varies greatly with elevation. The daily precipitation data series of 9 rainfall stations in the Dadu River Basin are simulated by the model of over-threshold. The model parameters are calculated by using the maximum likelihood estimation method, and the probability map, quantile map and horizontal map are used to reproduce the model parameters. The rationality of the model is evaluated by four kinds of intuitive diagnostic graphs of the density function, and the key parameters of the model and the confidence interval of heavy rainfall are estimated by using the contour likelihood method. The results show that the samples of all stations are distributed from Pareto, and the GPD model can be selected as the distribution function type of statistical inference of heavy rainfall in Dadu River Basin. The contour likelihood method can reflect the influence of the recurrence period on the design rainfall confidence interval. It can provide basis for quantitative evaluation of rainfall uncertainty and flood forecast of cascade reservoirs in Dadu River Basin.
【作者单位】：中国水利水电科学研究院水资源研究所;中国矿业大学资源与地球科学学院;
【基金】：国家重点基础研究发展“973”计划项目“梯级水库群风险预警与应急处置机理”(2013CB036406) 国家自然科学基金项目“风险规避视角下跨流域调水工程可调水量概率预报的基础研究”(51109224)
【分类号】：P333

【相似文献】