This project will investigate scalable parallel algorithms for the solution of the hydrodynamic flow and transport equations. The algorithms studies will be implemented on a variety of platforms, including clusters of workstations and massively parallel systems. The mathematical model will be based on the generalized wave continuity equation for surface elevation, the non-conservative form of the momentum equations for the velocity field, and the advection-diffusion equation for transport. Alternative numerical solution techniques, such as mixed finite element methods, will be investigated. There are five phases to the research. The first phase will use domain decomposition in conjunction with an existing hydrodynamic model to achieve parallelism. In the second phase, parallelism will be realized by exploring new algorithms specifically designed to take full advantage of parallel hardware, e.g., domain decomposition techniques. In conjunction with the second phase, a third phase will expand the codes capabilities by incorporating more physics into the governing equations. For example, baroclinic terms, which account for driving forces due to density variations, would be added so that simulations of, say fresh water/salt water interactions could be modeled. The fourth phase will rigorously test and validate the model against field data and analytical solutions. In the fifth phase, the code will be disseminated to users in academia, consulting, and industry, where it will be used to research topics, such as freshwater flow requirements in sensitive coastal ecosystems.
