Groundwater Quality Assessment Of Kezuozhong Banner Based On BP Neural Network-Fuzzy Mathematical Series Model

Groundwater quality is the main factor for the growth and development of plants and the impact on the ecological environment.It is a core indicator for the assessment of groundwater environment.Agricultural irrigation in Tongliao region of Inner Mongolia is mainly dependent on groundwater.At present,the area is undergoing a large-scale "water-saving and grain-increasing operation",which is bound to cause changes in the groundwater environment.This article took Kezuozhongqi Wanmu Demonstrative Field as the research object,carried out groundwater environment monitoring,revealed the change law of groundwater environment in the study area,and applied the constructed BP neural network-subordination serial model to evaluate the groundwater environment of Kezuozhong Banner.It will provide a basis for the effective treatment of farmland groundwater environment,and it has significant significance for maintaining the sustainable development of local agriculture.The main research results are as follows:(1)During the growing period of crops,the groundwater level decreased,and the groundwater depth fluctuates more obviously.Groundwater depth in 2016 showed a“U” change during the growth period;in July 2017,due to intensive rainfall,the water level rose,and the depth of groundwater showed a “W” change during the growth period.(2)Changes in the groundwater quality during the growth period are relatively large,accompanied by the elevation of the groundwater depth,the replenishment of rainfall,and the application of fertilization: July,groundwater pH increase;salinity,total salt,ammonia nitrogen,nitrate,sulfuric acid Salt,sodium ions,magnesium ions,chlorides and total hardness produced maximal values in June and August,and in July,they produced minimum values with more pronounced fluctuations.Bicarbonate and total alkalinity showed undulating changes.There is a very significant positive correlation between salt content and salinity,bicarbonate and total alkalinity.At the end of August,the concentration of each ion gradually decreased,and the ion concentration values were similar before and after the growth period.The local groundwater quality in April and October was superior to other months.(3)The established BP neural network-ranking model of membership degree is consistent with the actual monitoring data.This model is applicable to the assessment of farmland groundwater quality in Kezuozhongqi area.(4)Based on the groundwater IV criteria,a BP neural network-subordination series model evaluation results show that the groundwater quality of the study area is generally above Class IV water,and the groundwater quality is in good condition.The groundwater quality in 2017 is better than that of 2016.