Mpi-Based Parallelization Of Ecosystem Modeling and Services Assessment For High-Resolution Regional and Global Sustainability.

MPI-based parallelization of ecosystem modeling and services assessment for high-resolution regional and global sustainability analysiss

Agroecosystem models that can incorporate management practices and quantify environmental effects are necessary to assess sustainability associated food and bioenergy production across spatial scales. The models would also be useful for guiding decision and policy makings through providing evaluations of potential opportunities and trade-offs among options. However, most developed agroecosystem models are at a point or field scale. Tremendous work on simulations and datasets is needed when large scale of high-resolution spatial simulations are conducted. Parallelism design of modeling and database management presents unique strategy of handling these challenges. We used the message passing interface (MPI) technique and developed a master-slave scheme for an agroecosystem model, EPIC on global food and bioenergy studies. On the supercomputer, Titan at Oak Ridge National Laboratory, we successfully shortened the running time from days to 30 minutes for a global switchgrass modeling effort with the message passing interface based EPIC (mpi_EPIC). This study would be beneficial to explore high-resolution assessment on regional, national and global efforts on the critical issues of food, bioenergy and sustainability.