Gaining a Better Understanding of Surface-Subsurface Reactive Transport using a High-Order Advection Approach

Thursday, 18 December 2014
James Joseph Beisman III1, Reed M Maxwell1, Alexis Navarre-Sitchler1 and Carl I Steefel2, (1)Colorado School of Mines, Golden, CO, United States, (2)Lawrence Berkeley National Laboratory, Berkeley, CA, United States
Understanding the interactions between physical, geochemical, and biological processes in the shallow subsurface is prerequisite to the development of effective contamination remediation techniques, or the accurate quantification of nutrient fluxes and biogeochemical cycling. Here we present recent developments to the massively parallel reactive transport code ParCrunchFlow. This model, previously applicable only to steady-state, saturated subsurface flows, has been extended to transient, surface-subsurface systems. Proof-of-concept simulations detailing reactive transport processes in hillslope and floodplain settings will be presented. In order to reduce the numerical dispersion inherent in grid based advection schemes, which can lead to an over prediction of reaction rates, a weighted, essentially non-oscillatory (WENO) advection scheme has been implemented, providing formal fifth-order spatial and third-order temporal accuracy. We use a mass-conservative, positivity-preserving flux limiter while advecting solute concentrations to prevent non-physical solutions. The effects of advection schemes and their associated numerical dispersion on reaction rates are evaluated by comparing our scheme to a monotonic lower order scheme in a transverse mixing scenario. The work presented here allows a better understanding of nutrient cycling dynamics in watershed systems.