基于非线性RSH方法的泥石流灾害预测预报

Prediction and Evaluation of Debris Flow Disaster Based on Nonlinear RSH Method

泥石流灾害系统是一种自组织临界性系统,这种系统的演化总是处在混沌的边缘,系统的演化过程是可以预测的。将改进的重标极差分析方法(RSH)引入泥石流灾害前兆观测资料的处理和分析中,并在计算赫斯特指数H时,定义1个涵盖序列历史和序列长时记忆信息的时间窗。应用该方法对云南蒋家沟泥石流总径流量观测数据进行研究表明:云南蒋家沟泥石流总径流量时间序列不仅具有持久相关性,而且在激增前赫斯特指数值要经历1个先降维,然后升维的过程。根据对云南蒋家沟泥石流总径流量时间序列赫斯特指数值及其变化的分析,提出了预测预报泥石流灾害的3种定性判据。

Debris flow disaster is a self-organized critical phenomenon. Evolution of the system is always on the boundary strip of chaos and the evolution process can be predicted. The precursor monitoring data of the debris flow disaster was processed and analyzed with the improved rescaled range analysis （RSH algorithm）. And, a time window which contains sequences history and longtime memory information was defined when calculating Hurst index H. Monitoring data on total runoff of the Jiangjiagou debris flow in Yunnan province was analyzed with the method. It was shown that the time sequence of total runoff of the Jiangjiagou debris flow was long time dependent and Hurst index experienced fractal descending initially, ascending subsequently before the rapid rise of total runoff. Based on the analyses results, three kinds of quantitative criterions for debris flow forecasts were presented.