The purpose of this three-year NASA funded effort to objectively map and detect changes in large ocean ecosystems. We are using algorithms developed in the field of bioinformatics. These large ocean ecosystems are called biogeographic provinces. Biogeographic provinces provide useful categories for comparing and contrasting important ocean processes such as primary production, carbon flux, and species distribution and diversity. Climatological provinces have been identified using a priori expert knowledge. Discerning temporal trends and fine scale structures require objective automatic methods. We use objective classification on global remote sensing data to automatically produce time and space resolved province distributions. Our results show seasonal patterns in province geography reflect well known ocean processes. These province locations are verified by independent in-situ data from Canada’s Federal Department of Fisheries and Oceans, the Coriolis project for operational oceanography and the National Oceanic and Atmospheric Administration Atlantic Oceanographic and Meterological Laboratory.

Project Result

• We have computed the mean state of the biome locations based on the entire MODIS-Aqua dataset. There are approximately 20 different types of biomes (depending on which time period is analyzed). Several of them show remarkable similarity to the biomes/provinces that Longhurst (1998) first described. However, because our analysis is not dependent on long-term climatological analysis to define the regions, we can examine the evolution of these provinces over time, as well as what the interactions are along the boundaries in high detail.
• We have verified our predicted locations of province boundaries with in-situ ship tracks and well known climate indices. We have verified boundary locations in the North Atlantic, the South Pacific, the Equatorial Upwelling Region, and the Mediterranean Sea. These five regions are hydrographically dissimilar, so we are confident that our approach will work world wide.
• We have made two time series of province distributions through time, one for MODIS-Aqua and one for a SeaWiFS/AVHRR combination. The MODIS-Aqua time series from Jan-2002 through Dec 2006 shows the seasonal evolution of these provinces over a five year period.

An example of province distributions computed from an annual average of MODIS-Aqua. The different colors represent the different classes of biomes or ecosystems.

Increase in the total area of oligotrophic biomes (purple areas in top figure) over a five year period. This shows both the seasonal and secular changes in province distribution in the global ocean.

Project Resources & Results

• Oliver, M. J., Moline, M. A., Mobley, C. D., Sundman, Schofield, O. M. E. 2007.  Bioluminescence in a Complex Coastal Environment II: Prediction of Bioluminescent Source Depth from Spectral Water Leaving Radiance. Journal of Geophysical Research. 112, C11017, doi:10.1029/2007JC004136
• Longhusrt, A. R. 1998. Ecological geography of the sea. Academic, San Diego, 398pp.