This multidisciplinary research project will explore and expand the frontier of Information Technology capabilities by incorporating variability across a large range of spatial scales (through innovative embedded griding and advanced code architecture) and generalized adaptive ecological food webs (through multiple planktonic size classes, functional groups, and limiting nutrients) in realistic model configurations for the Pacific Ocean. This proposed combination of elements is unprecedented and thus pushes against the limits of simulating reality. The simulations of oceanic circulation and biogeochemistry will be based on the continued development of the Regional Oceanic Modeling System (ROMS), a relatively new computational code with innovative space-time discretization, sub-gridscale transport parameterizations, open-boundary conditions, and embedded griding capabilities. The computational research will focus on: 2-way embedding and flux-consistency across embedding levels, parallelization performance and portability, and software protocols for code architecture. The biological component will derive from a new global ecosystem model that explicitly treats iron limitation, nitrogen fixation, size structure, and diatom blooms. The strategy will be to simulate the entire Pacific Ocean circulation and biology at coarse spatial resolution (i.e., not resolving the eddies) and then investigate mesoscale influences at fine resolution in three productive, regional subdomains along the eastern boundary-tropical, subtropical and subpolar-with distinctive biomes. The coupled solutions will be analyzed to determine the rectification of mesoscale biological-physical variability to larger scales, the emergence of distinct biomes from a generalized set of ecological rules, and the level of biological complexity required to match observations based on information theoretic model-selection techniques.
Broader Impacts: Although these studies are focused on particular places with particular oceanic behaviors, they will provide a generic prototype for oceanic processes globally and for the important challenge of downscaling planetary-scale climate simulations to determine their regional-scale consequences. This project is a multi-disciplinary collaboration between UCLA and Woods Hole Oceanographic Institution. The personnel have expertise in physical, biological, and chemical oceanography and in the algorithms and procedures required for high-end computational modeling at the national supercomputing centers and they will train a post-doc and a graduate student in using these complex tools.