Temporal And Spatial Distribution Analysis And Prediction Of Shallow Groundwater Level In Shijiazhuang Plain

Shijiazhuang plains region is a typical shallow groundwater funnel area.The analysis on Shijiazhuang plains area's present situation and the dynamic state of shallow groundwater level,as well as effective simulation of the underground water level.It had scientific guiding significance to reflect the proposed response characteristics of the buried depth of groundwater for water saving measures.The article analyse on Shijiazhuang shallow groundwater system evolution characteristics and the main driving factors of dynamic of shallow groundwater table.The downward trend of groundwater level and its spatial dynamic characteristics were clarified,and a saving water concept had been put forward.Then,Based on the observed data of groundwater level,the main recharge and discharge factors and the water balance theory,extreme learning machine(ELM),mind evolutionary neural networks(MEANN),genetic(GANN),wavelet(WNN),fuzzy(FNN)and generalized regression neural network(GRNN)model were constructed.the error of spatial distribution trend was analyzed by using ArcGIS software.Their performance was evaluated comprehensively by using Matlab software.The main obtained results were as follows:(1)The main recharge and discharge of shallow groundwater in Shijiazhuang plain were precipitation and exploitation.In the past 25 years,the average level of shallow groundwater had decreased with 0.78m/a,and less than 20m buried depth of the areas were less than 1%of the total study areas,and nearly 52%of the areas groundwater levels were between 30-40m.There was a great relationship between the groundwater level and the water requirement of crops in Shijiazhuang plain.In spring and summer,the water level decreased faster than that in autumn and winter.To slow down the downward trend of groundwater level in the Groundwater flow field dominated by agricultural exploitation,the surface coverage and soil moisture content should be taken into consideration for agricultural water-saving measures.(2)The simulation results showed that the accuracy,stability,and space uniformity of the ELM were optimal.The ELM model could fit nonlinear dynamic characteristics of groundwater under the dual influence of natural and man-made accurately,and it realized the high precision prediction under the complex condition of shallow groundwater depth.What's more,it could use data close to the site to simulate the lack wells with high precision.(3)ELM model could identify the underground water level dynamic depth of the main driving factors.There were the high accuracy with only exploitation and precipitation as the input parameters.It could be used as recommended simulation of shallow groundwater forecast model when the investigation material lacked fulfilling discharge data.(4)If maintaining the present mining strength in mostly years,the average water table would be a downward trend in the next 3 years,which surpassed the average of the previous 25 years.If adjust the planting structure in the next 3 years,the winter wheat areas should be reduced by 77.30%in mostly years the mining intensity of the present condition could be correspondingly reduced by 50%.Similarly,In wet year,a 71.45%in the reduction area would reduce the mining strength by 40%.In dry year,a 71.45%reduction in area would reduce the mining strength by 40%.Finally,the Shijiazhuang plain's groundwater level changed from continuous decline to gradual rebound.