End-to-end modeling for the Ross Sea Region Marine Protected Area: a review of available tools for conservation objectives

The challenge of protecting marine ecosystems, including biodiversity and manifold interactions in changing environments, has encouraged development of end-to-end models linking physical drivers through entire food webs to upper-level predators. Ideally such models should have temporal and spatial resolution adequate to capture important small-scale processes, while including scales large enough to encompass local food webs and movements of wide-ranging consumers. The need for component submodels to be manipulated and upgraded independently by scientists from different disciplines, and used selectively depending on questions addressed, favors modular structure allowing linkage of different submodels with varying strengths and weaknesses. As an example, this paper reviews modeling tools available, developments needed, and proposed integrative strategies to construct end-to-end models for the Ross Sea Region Marine Protected Area (MPA). Decisions there are especially needed on how to preserve biodiversity and ecosystem structure and function under potential fishing scenarios in the face of climate change. We find that Earth System Models currently lack adequate spatial resolution to address key management issues in the Ross Sea, whereas regional-scale physical models are available to drive biogeochemical models at scales of a few kilometers. Models of local food webs can be run concurrently in contiguous spatial cells with between-cell exchange of organisms and materials. Individual-based models that incorporate acoustic measures of the preyscape can account for dependence of upper-level predators on the patch structure of prey within food web cells. For the Ross Sea MPA, innovations important for developing end-to-end models include (1) accounting for differences between diatoms and Phaeocystis as the main primary producers, (2) incorporating the roles of sea ice and associated biota in pelagic processes, (3) considering the importance of benthic-pelagic coupling to food webs including species whose life stages span the water column, and (4) linking predator foraging on patches at scales of tens of meters to food webs at scales of multiple kilometers. Geographic Information Systems (GIS) have been advocated as common platforms to integrate observational data or submodel outputs into end-to-end simulations in a spatially explicit format. Such an approach would facilitate facultative use of different submodel types depending on the ecological and management questions of interest.