Study On Mechanism Of Solute Transport In Heterogeneous Anisotropic Imbricated Aquifer

Taking the imbricate alluvial fan in the west of Erhai Basin as the study area,the solute transport characteristics and its influencing mechanism in the two-dimensional heterogeneous aquifer of alluvial fan in laboratory scale two dimension section were studied by seepage experiment and numerical simulation on the basis of literature investigation and field hydrogeological investigation.The effects of heterogeneous anisotropy on the reactive transport of PCE in imbricate aquifers were analyzed at laboratory and site scales by using the solute reactive transport numerical simulation program(RT3D).An efficient joint parameter estimation method in reactive transport model was established to solve the uncertainty in forward parameter estimation of stochastic model at site scale,and its effectiveness was verified.The main conclusions are as follows.(1)The heterogeneous anisotropy of the imbricated alluvial fan aquifer on the west of Erhai Lake is obvious.The alluvial fan aquifer in the west of Erhai Lake is deposited by eighteen cross-alluvial deposit from the Cangshan Mountain.Its sediment shows imbricated in macroscale and strong heterogeneity from the top to edge of fan due to lithology and sorting differences.The gravel imbricate sedimentary in a certain dipping angle along the longitudinal and vertical of riverbed due to the different sedimentary periods of hydrodynamic conditions,which affects or controls the anisotropy characteristics of aquifer.(2)The macroscopic flow of imbricated aquifer is controlled by heterogeneity,while the microscopic flow direction is altered by the anisotropy.In laboratory scale imbricated aquifers,the distribution and evolution of groundwater flow and determines the path and rang of contaminant reactive transport is controlled by the heterogeneity.The groundwater flow direction and velocity is altered by the anisotropy,and the flow retarding effect is increasing with the gravel inclination angle increase,which affects the transport velocity,concentration peak and arrival time of contaminants.(3)The horizontal transport of contaminants is retarded by the anisotropy of aquifer due to the increase of gravel dip angle.The laboratory flow experiment and numerical simulation results show that the retard effects of contaminant plume is increasing with the increase of gravel dip angle,and thus increase the residence time of contaminant plume on the top of clay lens.When the gravel dip angle changes from 10° to 20° and 20° to 30°,the peak concentration of contaminants in the same observation hole at the same time decreases by 38.6% and 36.1%,respectively.When the gravel dip angle changes from 10° to 30°,the retard of peak concentration arrival time of contaminants is about 200 s,and the peak concentration decreases by nearly 40%.(4)There is a certain degree of uncertainty in the forward modeling of hydrogeological parameters of aquifers by stochastic model.The study on solute reactive transport of random model at site scale shows that the maximum and minimum values of contaminant peak concentration in the breakthrough curve of 30 random realizations differ by nearly three times,indicating that the uncertainties of random parameters cause great differences to forward modelling results.The uncertainty of these parameters directly affects the accuracy of reactive transport simulation results of random anisotropic aquifers at site scale.(5)The Kriging adaptive Markov chain Monte Carlo method can improve the efficiency of parameter estimation.The Kriging model has no grid constraint,and the substitution fraction function adaptively selects new samples and integrates them to obtain the posterior distribution of model parameters,and the samples are concentrated in the high probability region to achieve the accurate PPDF estimation of the estimated parameters.Therefore,the number of calculations required by the alternative method is only 1.1% of that of the traditional MCMC model,which significantly reduces the computational cost of Bayesian estimation in the groundwater reactive transport model.The study results are expected to provide a basis for the transformation of hydrogeological parameters from laboratory scale to site scale.The proposed joint parameter estimation method can significantly reduce the cost and time of groundwater numerical simulation calculation,and is expected for the hydrogeological parameters inversion at site scale in alluvial fan aquifers with lower exploration accuracy.