论文部分内容阅读
随着气候变化、人类活动影响的加剧,水文序列一致性条件遭到破坏,本文选择下垫面变化剧烈的西大洋水库控制流域为研究区域,采用水文变异诊断系统,分析其年最大洪峰流量序列的变化趋势及变异点,确定序列变异形式,并对非一致性洪峰流量序列进行还原修正;根据贝叶斯理论将先验信息和样本信息有机结合,采用Gibbs-MCMC算法对修正前后的P-Ⅲ型频率分布曲线参数不确定性进行估计,给出了参数95%置信区间。将估计后的参数与适线法耦合进行水文频率分析计算,得到修正前后设计频率洪峰流量预估区间。选取覆盖率、平均带宽、平均偏移度等3个指标对修正前后预报区间优良性进行评价,其中覆盖率修正后比修正前提高17.10%,平均带宽修正后比修正前减少15.77%,平均偏移程度修正后比修正前减少3.27%,可见,对于人类活动影响较大的非一致性水文序列进行一致性修正可使得预报不确定性区间的可靠性得到提高。
With the aggravating influence of climate change and human activities, the consistency conditions of hydrological sequences are destroyed. In this paper, the control basin of the Western Ocean Reservoir, whose underlying surface changes drastically, is selected as the research area. Hydrological mutation diagnosis system is used to analyze the maximum peak flow series According to Bayesian theory, a priori information and sample information were combined organically, and the Gibbs-MCMC algorithm was used to analyze the variation of P- Ⅲ type frequency distribution curve parameter estimation uncertainty, given the parameter 95% confidence interval. Coupling the estimated parameters with the fitting method, we calculate the hydrological frequency and get the estimated range of the peak frequency of the design frequency before and after the correction. Three indicators, covering coverage, average bandwidth and average offset, were selected to evaluate the goodness of prediction before and after the correction. The coverage was improved by 17.10% after correction, the average bandwidth was reduced by 15.77% After the correction, the degree of migration was reduced by 3.27% compared with that before correction. Therefore, the consistency correction of non-consistent hydrological sequences with great influence on human activities could improve the reliability of the forecast uncertainty interval.