Evaluation of a power averaging technique of effective parameters on simulation of atrazine leached in soil at field scale

The heterogeneity of agricultural soils leads to the complexity of flow and solute transport.To improve the efficiency of numerical simulation,some parameters must be upscaled.In this study,we take the numerical simulation of atrazine leaching through unsaturated soils at field scale as an example,by taking 100 soil samples from a field plot with an area of 27 m×27 m,located in the eastern suburb of Beijing,and analyze their main physicochemical properties.The hydraulic parameters of the van Genuchten‘s type for all soil samples are generated by using pedotransfer function,and the dispersivities of atrazine transport are indirectly estimated by using the hydraulic information.The adsorption coefficients of atrazine on the soil texture are indirectly calculated,too.On the assumption of column model,the atrazine leachs in the soil on the condition that the rainfall infiltrat ionredistribution is simulated.Firstly,the dynamics of the soil water matric potential and the atrazine concentration in individual soil column are simulated by using 100 simulat ions of HYDRUS-1D software,and then their mean spatial distribution is obtained from the results.Secondly,the dynamics of soil water matric potential and atrazine concentration in an equivalent porous media is simulated by using the effective parameters and one simulation of HYDRUS-1D.In this paper,the emphasis is put on discussing the effect of the equivalent effective parameters,being calculated by the power law averaging technique,on the numerical simulation of the mean distribution of soil water matric potentia and atrazine concentration at field scale.It is shown that the upscaled numerical simulation of high accuracy can be obtained,when the geometry mean or harmonic mean values of hydraulic parameters,transport parameters and adsorption parameters are taken as their effective parameters,while the arithmetic mean of soil water matric potential of normal distribution under initial condition is taken as its equivalent quantity.