Analysis on runoff time series dynamics character based on approximate entropy

摘要

Runoff time series dynamic character and runoff forecasting are studied by approximate entropy measure of complexity theory in this paper. Relative change of measured runoff approximate entropy is analyzed at the different hydrologic station in the main Yellow River, and some dynamic character of runoff change is disclosed. The results show that abnormal change character of the low value (or high value) after the high value (or low value) of approximate entropy is looked upon as change character of runoff before its peak value (or valley value), which explain various physical characteristics of all kinds of affecting factor before peak value (or valley value) of runoff, and cause significant change law of runoff series space-time structure that is from disordered (or ordered) state to ordered (or disordered) one. This phenomenon between abnormal change character and corresponding runoff change character seems natural one between lightning and rumbling, which provide the base for runoff peak and valley forecasting.