The Study Of Groundwater Depth Prediction Methods In Baoding City

Water resources related problems are one type of the most serious worldwide resources problems now. Especially, the unreasonable development and utilization of groundwater problem has led to a series of negative effects. Therefore, sustainable development and utilization of groundwater resources has become an important direction in the field of water resources research.On the basis of the previous research on groundwater depth, this paper selected Baoding City as the study area and developed groundwater depth prediction models according to the collected meteorological hydrological materials and groundwater monitoring data, and analyzed the dynamic change characteristics of groundwater depth in the study area. The main achievements are as follows:(1) Analyze the dynamic change characteristics of groundwater depth based on precipitation, groundwater exploitation conditions, and groundwater depth data from 1990 to 2014. The main dynamic type of groundwater depth is rainfall infiltration-mining type. This paper selected three typical groundwater monitoring wells which have comprehensive data and materials, and analyzed the groundwater depth change rules of these three wells in high flow year, median flow year and low flow year.(2) Based on previous groundwater related studies in Baoding city, in this paper, we proposed to use the grey model and the BP neural network model to predict groundwater depth. Specifically, the grey model predicts the future groundwater depth by analyzing previous several years ’groundwater depth, which tries to capture the internal pattern of groundwater depth. The average relative error is 3.64% which is low. However, the BP neural network could analyze how external factors, e.g., precipitation, extraction, affect the groundwater depth. But its average relative error is 5.41%.(3) To address the limitation of both the grey model and the BP neural network model, in this paper, we proposed a new groundwater depth prediction model that combined the grey model and the BP neural network model, named grey-neural network model. Our experiments with this new model could generate better results than any of these two techniques. The average relative error of groundwater depth prediction is about 1.29%. This study finally could provide new methods to predict groundwater depth in Baoding City. The modified grey model is reasonable with only considering groundwater depth data, but the combined model is reasonable with knowing precipitation, exploitation quantity, etc.