MULTISPECTRAL REMOTE SENSING ALGORITHMS FOR PARTICULATE ORGANIC CARBON (POC) AND ITS TEMPORAL AND SPATIAL VARIATION

Hydrographic data including particulate organic carbon (POC) from the Northeastern Gulf of Mexico (NEGOM) study were used along with remotely sensed data obtained from NASA's Sea-viewing Wide Field-of-view Sensor (SeaWiFS) to develop POC algorithms to estimate POC concentration based on empirical and model-based principal component analysis (PCA) methods. In Case I and II waters empirical maximized simple ratio (MSR) and model-based PCA algorithms using full wavebands (blue, green and red wavelengths) provide more robust estimates of POC. The predicted POC concentrations matched well the spatial and seasonal distributions of POC measured in situ in the Gulf of Mexico. The ease in calculating the MSR algorithm compared to PCA analysis makes MSR the preferred algorithm for routine use. In order to determine the inter-annual variations of POC, MSR algorithms applied to calculate 100 monthly mean values of POC concentrations (September 1997-December 2005). The spatial and temporal variations of POC and sea surface temperature (SST) were analyzed with the empirical orthogonal function (EOF) method. POC estimates showed inter-annual variation in three different locations and may be affected by El $Ni{\tilde{n}}o/Southern$ Oscillation (ENSO) events.