Application of Grey Self-Memory Model Based on Smoothing Method Without Seasonal Fluctuations on Monthly Runoff Forecast

Abstract

Affected by the monsoon climate,there are abrupt changes,non-smoothing and variation in monthly runoff.However,good accuracy cannot be realized if the monthly runoff is simulated and predicted directly by the gray self-memory model.This paper proposes a new monthly runoff smoothing method (that is,smoothing method without seasonal fluctuations).First,the seasonal index is used to remove the “seasonal fluctuations” of monthly runoff,then,the three-point smoothing method is applied to smooth the monthly runoff without seasonal fluctuations,and finally the gray self-memory model is adopted to simulate and predict the smoothed data.After being applied to the Yanling Hydrological Station in Hunan Province,the results show that the average error of fitting and forecasting value of monthly runoff after seasonal fluctuation removing and smoothing is less than 20% compared with the measured value.At the same time,the short-term monthly runoff data is used to simulate and forecast monthly runoff by this method,which also meets the requirements of the specification.

Keywords

• monthly runoff;hydrological forecast;grey self-memory;seasonal index;three-point smoothing method