A sequential-parallel approach for coupling chemistry and transport in groundwater quality modelling

Abstract

Groundwater quality modelling requires taking into account coupled mechanisms such as chemistry and hydrodynamics, for which numerical treatment is almost always necessary. However, when attempting to simulate natural systems, the computation time on conventional computers limits the use of such calculation codes. A numerical code (CAT//), in which the Transport module and the Chemistry module are successively called at each time step, was developed by the BRGM. We show that such a sequential algorithm is readily adaptable to parallel computing with relatively little programming effort. A test of this code on four different parallel computers showed the required CPU time to be reduced by a factor close to the number of processors available. This approach seems well adapted to modelling coupled phenomena since coupling of major mechanisms can be approximated by a sequential algorithm.