| Groundwater arsenic with high content is one of the serious water resources problems facing the international society.The alluvial plain of lower Yellow River in northern Henan Province is an important grain producing area of Henan Province.There is a large amount of arsenic enrichment in groundwater used for agricultural irrigation and drinking water,which will have adverse effects on human health.In order to better understand the range of arsenic pollution in groundwater in this region,this paper selects the optimal variable combination by recursive feature elimination algorithm based on the measured arsenic concentration values of1081 shallow groundwater in the study area and combined with environmental factors such as climate,soil physical and chemical characteristics,sedimentary environment,human activities,hydrogeology.The Stacking integrated algorithm model with RF,XGBoost,SVM as the base learner and LDA as the meta-learner is constructed.The paper predicts the spatial distribution of high arsenic groundwater in the alluvial plain of the lower Yellow River and identifies the key environmental variables affecting the spatial distribution of high arsenic groundwater in the study area.The following research results were obtained:1.In the groundwater samples of the study area,the measured arsenic concentration ranges from 0.01 to 190μg/L,arsenic concentration>10μg/L(exceeding the international standard)accounted for 16.76%of the total samples,and the average arsenic concentration in the groundwater samples is 7.06μg/L,and there is a spatial variation.The high-arsenic groundwater area is distributed in the direction of northeast to southwest,and gradually expands to the south.2.The Stacking prediction model of high-arsenic groundwater with RF,XGBoost and SVM as learning tools and LDA as meta-learning tools is constructed and the spatial distribution of high-arsenic groundwater in northern Henan is well predicted.The integrated model of Stacking had the maximum AUC,Accuracy,Specificity and sensitivity compared with RF,XGBoost and SVM models.The spatial distribution of arsenic in shallow groundwater in northern Henan Province is very uneven.The area with arsenic content exceeding 10μg/L in shallow groundwater is 6673.25km~2,accounting for 33.78%of the total area of the study area.It is mainly distributed in the depression in front of the alluvial and diluvial fan of Taihang Mountain and the Yellow River sill fan.Specifically,some areas of Fan County and Puyang County in Puyang City,the north of Yanjin County,Yuanyang County,and Fengqiu County in Xinxiang City,and the southwest of Hua County and Neihuang County in Anyang City are seriously polluted by high arsenic water.3.The key environmental factors affecting arsenic enrichment in groundwater in northern Henan are identified.The situation of the Yellow River breach,annual average temperature,annual precipitation,elevation,and hydraulic gradient are the environmental variables that rank first in terms of their impact on high arsenic groundwater in northern Henan.The sedimentary environment(number of Yellow River crevasses,Quaternary geomorphic type,elevation,clay sand ratio)has a significant impact on high arsenic groundwater.The probability of occurrence of high arsenic groundwater gradually decreases with the increase of the number of Yellow River crevices,precipitation,permeability coefficient,water yield,elevation,slope,hydraulic gradient,and prediction variables of distance from the river,but the probability of occurrence of high arsenic groundwater gradually increases with the increase of surface soil silt content,temperature,and evaporation. |
