Characterizing Pelagic Habitats within Gulf of Mexico Coastal Waters Using Remote Sensing Data

We have applied the non parametric, non linear, neural network classification known as Self Organizing Maps(SOM) to classify the waters in the Gulf of Mexico and establish the correspondence between the SOM method and the Coastal and Marine Ecological Classification Standard (CMECS) scheme.

The following data products were used for the SOM technique: Sea surface temperature (SST), Salinity,ChlorophyllEuphotic depth, and the bottom to surface temperature differential (Delta T).

For each SOM class we have computed the spatial extent in square km and how this varies with time. In the time series the annual and biannual signals are clearly evident. A good example of a clear annual cycle in the areal extent is Class 4 in the Eastern Gulf of Mexico. A good example of a clear biannual cycle in the areal extent is Class 61 in the Eastern Gulf of Mexico.

The classifications are available monthly for a five year period 2005-2009 for three parts of the Gulf of Mexico:

We have also compared a SOM classification to fish catch data, this is available as a zip archive of google earth files here.