Dowell, Mark D., Janet W. Campbell and Timothy S. Moore

Ocean Process Analysis Laboratory, University of New Hampshire, Durham, NH 03824, Tel: 603-862-1476, Fax: 603-862-0243, E-mail: mark.dowell@unh.edu

Dynamic Ocean Provinces: A biogeochemical and physiological template of the global ocean

The concept of oceanic provinces or domains has existed for well over a century. Such systems, whether real or only conceptual, provide a useful framework for understanding the mechanisms controlling biological, physical and chemical processes and their interactions. Criteria have been established for defining provinces based on physical forcings, availability of light and nutrients, complexity of the marine food web, and other factors. In general, such classification systems reflect the heterogeneous nature of the ocean environment, and the effort of scientists to comprehend the whole system by understanding its various homogeneous components. If provinces are defined strictly on the basis of geospatial or temporal criteria, the resulting maps exhibit discontinuities that are uncharacteristic of the ocean. While this may be useful for many purposes, it is unsatisfactory in that it does not capture the dynamic nature of fluid boundaries in the ocean. Boundaries fixed in time and space do not allow us to observe interannual or longer-term variability (e.g., regime shifts) that may result from climate change.

The current study illustrates the potential of using fuzzy logic as a means of classifying the ocean into objectively defined provinces using properties measurable from satellite sensors (MODIS and SeaWiFS). This approach accommodates the dynamic variability of provinces which can be updated as each image is processed. We adopt this classification as the basis for parameterizing specific primary production models for each of the classes. Once the class specific algorithms have been applied, retrievals are then recomposed into a single blended product based on the "weighted" fuzzy memberships. The provinces themselves are identified on the basis of global fields of chlorophyll, sea surface temperature and PAR which will also be subsequently used to parameterize primary production (PP) algorithms. We present a multi-year time series synthesizing the geographic and seasonal variability of specific variables relevant to primary production modeling as well as the global distribution of the estimated net primary production itself. The variability of photosynthetic parameters will also be characterized in each province, thus providing input to existing models for estimating primary production. In short the proposed approach will provide all of the oceanographic and ecological insight of traditional classification schemes whilst retaining the fluid boundaries and dynamic interaction of the different ocean biomes as perceived in global satellite imagery.