Groundwater Level Changes Forecast Using Support Vector Machine and Ensemble Kalman Filtering Dual-Data Assimilation Technique
To improve the forecast ability of the groundwater level changes,recognize the dynamic change of groundwater level and improve the water resources management,this study applied support vector machines(SVM)and one Ensemble Kalman Filtering(EnKF)technique to construct the SVM and SVM-EnKF Dual-Data Assimilation(DDA)models to forecast the groundwater level changes at subsequent 1st to 3rd month using the in-situ meteorological variables(i.e.,air temperature,precipitation,solar radiation,ground temperature)and the groundwater level observation.The results indicate that:(1)the performance of SVM model is influenced by the meteorological forcing variables,which is only suit for the stations with strong relationship between groundwater level changes and meteorological variables.(2)the SVM-EnKF DDA model forced with limited meteorological forcing variables is more efficient than the SVM for the forecast of groundwater level changes at subsequent 1st to 3rd month.This study provides one efficient dual-data assimilation technique for the forecast of groundwater level changes at the regions with limited data.
万方数据
维普
