基于DFA方法的珠江流域极端降水阈值研究

 流域极端降水事件发生频率低、样本少，其阈值定义存在定义方式不统一、方法多样等不确定性和无规则性问题。论文以珠江流域为研究区域，系统总结了绝对临界值法、百分位值法、极值分布拟合法在确定流域极端降水阈值存在的不确定性问题，并尝试运用DFA方法，研究探讨了该方法在确定流域极端降水阈值的物理机制及其适用条件。研究表明， DFA方法以系统动力学为基础，通过分析极端降水事件对降水时间序列长程相关性的影响，由此寻求流域极端降水阈值，能较好反映流域降水序列的统计效应和物理背景，比较适合流域极端降水阈值分析。

Researches on DFA-based Extreme Precipitation Threshold in the Pearl River Basin

Extreme precipitation events occur rarely and are difficult to collect samples. There are still uncertainty problems with the definition of extreme precipitation threshold, which is variable and not unified. This study summarizes the uncertainties of extreme precipitation threshold based on Fixed Critical Value Method, Special Percentile Value Method and the Goodness of Extreme Value Distribution Fitting Test. In order to overcome these problems, a new method,named as DFA, has been used to discuss the physical mechanism and applicable conditions. The system dynamics-based DFA, tracks extreme precipitation threshold by analyzing the characteristics of long range correlation of the precipitation series and can well reflect the statistical effects and scientific background of precipitation series.It is applicable to define extreme precipitation threshold within a basin.