Impact Of Hydraulic Conductivity Distribution Mapped By Five Geostatistical Methods On The Fate Of Groundwater Contaminant Transport At The Field-scale

Numerical simulation of groundwater flow and solute transport based on a large amount of regional measured data,which satisfy large scale and multi-precision requirements,can provide theoretical support for solving regional groundwater resources issues and environmental problems.However,due to the spatial variation of exploration data and the heterogeneity of aquifer system during regional model building process,groundwater flow and solute transport simulation results tend to display larger deviation with reality.Looking for a method which can be used to take full advantage of limited exploration data for the reasonable valuation of aquifer parameters in unknown areas,is now the key problem in large regional groundwater flow simulation.At present,the method for identification of parameters based on the traditional groundwater flow and solute transport model and classical probability statistical method has not been able to analysis and evaluate the spatial variation of migration parameters objectively.In this paper,geostatistics promotion to large scale migration of groundwater pollution.Based on the measured field data of Massachusetts Military reserves(Massachusetts he Reservation,MMR),ordinary kriging,quantile kriging and indicator kriging method,sequential Gaussian simulation method and sequence indicator simulation method are used to generate a random hydraulic conductivity field combined with specific geological conditions.The generated parameter values are used in the concept model and mathematical model of the study area,which is solved by using the MODFLOW and MT3DMS,to compare the effects of the different geostatistics methods on contaminant transportation.The results show that,regionalized variable theory can reflect the random and structural characteristics of the spatial distribution of variables;variogram is a basic tool of geostatistics,and one of the most effective method to estimate the distribution features of variables;the plume centroid coordinates of the four geostatistical methods are basically the same,which proved effective centroid position is determined by the average hydraulic conductivity to determine the theoretical;plume in space spread range is mainly affected by the spatial variability of hydraulic conductivity of variance.The simulation methods overcome the smoothing effect of estimation methods,better reproduce the volatility of real curves.With the increase in number of simulations,hydraulic conductivity(lnK)plume space with second-order moment estimated variance showed a smaller trend,and the degree of sequential indicator simulation is more apparent.