Remote Sensing of Harmful Algal Blooms in the Mississippi Sound and Mobile Bay: Modelling and Algorithm Formation

Date of Award


Degree Type


Degree Name

Doctor of Philosophy (PhD)


Coastal Sciences, Gulf Coast Research Laboratory

First Advisor

William E. Hawkins

Advisor Department

Coastal Sciences, Gulf Coast Research Laboratory


The incidence and severity of harmful algal blooms have increased in recent decades, as have the economic effects of their occurrence. The diatom Pseudo-nitzschia spp. caused fisheries closures in Mobile Bay during 2005 due to elevated levels of domoic acid. In the previous 4 years Karenia brevis counts of >5,000 cells L-1 have occurred in Mobile Bay and the Mississippi Sound. Population levels of this magnitude had previously been recorded only in 1996. Increases in human populations, urban sprawl, development of shoreline properties, sewage effluent and resultant changes in N-P ratios of discharge waters, and decline in forest and marsh lands, will potentially increase future harmful algal bloom occurrences in the northern Gulf of Mexico. Due to this trend in occurrence of harmful algal populations, there has been an increasing awareness of the need for development of monitoring systems in this region. Traditional methods of sampling have proven costly in terms of time and resources, and increasing attention has been turned toward use of satellite data in phytoplankton monitoring and prediction. This study shows that remote sensing does have utility in monitoring and predicting locations of phytoplankton blooms in this region. It has described the composition and spatial and temporal relationships of these populations, inferring salinity, total nitrogen and total phosphorous as the primary variables driving phytoplankton populations in Mobile Bay and the Mississippi Sound. Diatoms, chlorophytes, cryptophytes, and dinoflagellates were most abundant in collections. Correlations between SeaWiFS, MODIS and in situ data have shown relationships between R rs reflectance and phytoplankton populations. These data were used in formation of a decision tree model predicting environmental conditions conducive to the formation of phytoplankton blooms that is driven completely by satellite data. Empirical algorithms were developed for prediction of salinity, based on R rs ratios of 510 nm/555 nm, creating a new data product for use in harmful algal bloom prediction. The capacity of satellite data for rapid, synoptic coverage shows great promise in supplementing future efforts to monitor and predict harmful algal bloom events in the increasingly eutrophic waters of Mobile Bay and the Mississippi Sound.