Solute transport in a heterogeneous aquifer: a search for nonlinear deterministic dynamics

Abstract. The potential use of a nonlinear deterministic framework for understanding the dynamic nature of solute transport processes in subsurface formations is investigated. Time series of solute particle transport in a heterogeneous aquifer medium, simulated using an integrated probability/Markov chain (TP/MC) model, groundwater flow model, and particle transport model, are studied. The correlation dimension method, a popular nonlinear time series analysis technique, is used to identify nonlinear determinism. Sensitivity of the solute transport dynamics to the four hydrostratigraphic parameters involved in the TP/MC model: (1) number of facies; (2) volume proportions of facies; (3) mean lengths (and thereby anisotropy ratio of mean length) of facies; and (4) juxtapositional tendencies (i.e. degree of entropy) among the facies is also studied. The western San Joaquin Valley aquifer system in California is considered as a reference system. The results indicate, in general, the nonlinear deterministic nature of solute transport dynamics (dominantly governed by only a very few variables, on the order of 3), even though more complex behavior is possible under certain (extreme) hydrostratigraphic conditions. The sensitivity analysis reveals: (1) the importance of the hydrostratigraphic parameters (in particular, volume proportions of facies and mean lengths) in representing aquifer heterogeneity; and (2) the ability of the correlation dimension method in capturing the (extent of) complexity of the underlying dynamics. Verification and confirmation of the present results through use of other nonlinear deterministic techniques and assessment of their reliability for a wide range of solute transport scenarios are recommended.