Study On Optimization Of Soil Salt Sampling Layout In Yongji Irrigation Area Of Hetao Irrigation Area

Soil salinization is the main restrictive factor for the sustainable development of agriculture and ecological environment in arid,semi-arid areas and coastal areas.Understanding the temporal and spatial distribution law of soil salt and its causes is an important prerequisite for the prevention and control of soil salinization.Drawing the distribution map of soil salt and groundwater depth is an important means to understand the temporal and spatial distribution law of soil salt and its causes.The mapping accuracy of soil salinity and groundwater table depth and the disclosure of their relationship largely depend on the number and layout of monitoring points.How to obtain the highest mapping accuracy at the lowest cost has always been a hot issue,which is of great significance to improve the efficiency of soil salinization control and promote the sustainable development of agriculture in arid and semi-arid areas of North China.Hetao Irrigation Area in Inner Mongolia is an important commercial grain production base in China Affected by natural and human factors,the problem of soil salinization has always been prominent.Based on the sampling data of a test area(with an area of 1574 hm2)in Yongji irrigation area of Hetao Irrigation Area in 2020,and combined with the 10 sampling data of the expanded area(with an area of 5670 hm2,referred to as test area 0 in the text)from 2018 to 2020,this paper optimizes the sampling scheme of soil salt single variable,soil salt and groundwater buried depth respectively based on Geostatistics and spatial simulated annealing algorithm,and evaluates the optimization results,The results can provide reference for the layout of saline alkali land monitoring points in Hetao irrigation area.The main results and conclusions are as follows:(1)Through the regression Kriging method with the shortest distance from wasteland as the covariate(referred to as the shortest distance from wasteland-RK method),the regression Kriging method with sand content as the covariate(referred to as the sand content-RK method)and the regression Kriging method with comprehensive index as the covariate(referred to as the comprehensive index-RK method),the soil salt sampling scheme is optimized.The estimated variances are 0.0807(g·kg-1)2,0.0851(g·kg-1)2 and0.0791(g·kg-1)2 respectively,Compared with the ordinary Kriging method(hereinafter referred to as OK method),the estimated variance(0.1143(g·kg-1)2)is reduced by 29.39%,25.55%and 30.8%respectively.(2)From the three aspects of global optimality,geospatial representativeness and attribute spatial representativeness,the optimization results of the shortest distance from Wasteland-RK method,sand content-RK method,comprehensive index-RK method and OK method perform well in three aspects:global optimality,representativeness in geospatial space(the assurance rate of spatial interpolation neighborhood is more than 93%)and representativeness in attribute space(both reach a significant level).Four methods are comprehensively evaluated from three aspects.It is recommended to optimize the salt sampling scheme based on the comprehensive index-RK method and the shortest distance from wasteland-RK method.(3)The salinity and groundwater depth are optimized one by one(taking soil salinity as the first variable and groundwater depth as the second variable)and jointly.The soil salinity sampling points and groundwater depth monitoring points obtained from the joint optimization are more balanced in accuracy,geographical spatial representation and attribute spatial representation.Compared with optimizing the two variable sampling layout one by one(without considering the existing groundwater depth monitoring log,the sampling number is fixed),the estimated variance of soil salinity increases by 8.7%,and the estimated variance of groundwater depth decreases by 1.9%;The representativeness of soil salinity in geographical space decreased(the assurance rate of spatial interpolation neighborhood decreased by 7.2%),and the representativeness of groundwater table depth in geographical space increased slightly(the assurance rate of spatial interpolation neighborhood increased by 2.8%);The representativeness of soil salt sampling points in the attribute space expressed by the shortest distance from wasteland and sand content is greatly weakened.In general,for this study area,the effect of optimizing two variables one by one is better.