Upscaling multi-component reactive transport in presence of connected subsurface structures

Abstract:

Heterogeneity in hydraulic conductivity leads to incomplete mixing. Upscaling using the dispersion tensor in the advection-dispersion equation overestimates local mixing. Modelling multi-component reactive transport leads to an overestimation of reaction rates and overall reactions. Multi-rate mass transfer was shown previously to better represent mixing. But it is still unclear under what conditions this linear model is able to represent the underlying non-linear process.

We study explicit multi-component transport in heterogeneous aquifers for the example of calcite-dissolution. We compare different types of heterogeneity from intermediately well connected (multigaussian) fields to very well connected fields. The fundamental difference stems from their connectivity structure. We observe for the well connected field different dominating channels with an almost uniform advective velocity while the multigaussian fields show dominating channels with a varying advective velocity. Then, we compare our results with an effective reactive mass transfer model where the distribution of exchanges rates or the memory function are derived from information of the hydraulic conductivity field only.

We see that reactive multi-rate models show a good agreement for the well connected fields where the connected channels are more or less homogeneous and the immobile inclusions are of more or less equal size. We find connectivity important for upscaling reactive transport in highly heterogeneous conductivity fields.