Elucidating the relationship between aquifer heterogeneity and thermal dispersivity using stochastic heat transport simulations = Explication de la relation entre l’hétérogénéité de l’aquifère et la dispersivité thermale par des simulations stochastiques de transport de chaleur = Dilucidación de la relación entre la heterogeneidad del acuífero y la dispersividad térmica mediante simulaciones estocásticas de transporte de calor = Elucidando a relação entre heterogeneidade do aquífero e dispersividade térmica usando simulações estocásticas de transporte de calor

Homogeneous parametrization of aquifer properties is conventionally adopted due to data limitations and computational constraints, where the determination of effective model parameters and consideration of heterogeneity remains a key challenge. In this study, numerical simulations were developed to investigate the effect of heterogeneity on groundwater heat transport processes (conduction, dispersion and advection), and elucidate the relationship between aquifer heterogeneity and thermal dispersivity (αL). A simple finite element model was developed to simulate the operation of a geothermal system within a closed heterogeneous aquifer using stochastic permeability realizations. Sensitivity analysis revealed how the αL value is affected by (1) the scale of heterogeneity (R), (2) the distance between the well pair (L), and (3) the Darcy flux (qin). The effective permeability of the medium was found to continuously decrease as R increases; however, cold water reaches the outflow side faster through the high-conducting channels; thus, the thermal breakthrough time decreases in heterogeneous media. Calibrated αL values allowed homogeneous models to accurately fit the breakthrough curves obtained from the heterogeneous simulations. Increase in both R and L corresponded to higher αL, but R remained the dominant parameter. Modification of the Péclet number revealed that the thermal conduction weakens, yet does not remove, the effect of heterogeneity or αL. It is concluded that well-calibrated αL values are necessary for accurate predictions of heat transport processes in groundwater flow systems. This study provides guidance on the estimation of the longitudinal αL value for use in numerical modelling of practical geothermal problems.