Abstract:

In order to discuss the application of changeable weight combination forecasting model in groundwater levels forecasting, based on the date of groundwater levels from 1997 to 2008 in eastern suburbs of Tongliao City, Inner Mongolia Autonomous Region, the changeable weight combination forecasting method which is based on index model, grey model and linear regression model was used to simulate and forecast groundwater levels of the research area. It was calculated that the sum squares of errors of various forecasting were 5.116 1,5.080 1,4.914 8,4.672 3 respectively. The results indicated that the changeable weight combination forecasting method was more precise than one single forecasting method. From the predictive value, the depth of groundwater level in this research area will decrease to 15.39 m by the end of 2015, and compared with the depth of groundwater level in 2008, the decreasing range will come up to 20%.

Key words: combination forecasting, changeable weight coefficients, groundwater level, groundwater