Date of Award


Level of Access Assigned by Author

Campus-Only Thesis

Degree Name

Master of Science (MS)




Mary Jane Perry

Second Committee Member

Katja Fennel

Third Committee Member

Huijie Xue


The North Atlantic Spring Bloom is one of the main events that lead to carbon export to the deep ocean and drive oceanic uptake of CO2 from the atmosphere. However, the bloom and its associated carbon export are difficult to quantify, largely due to the lack of sufficient space and time-resolved observations. An experiment involving autonomous platforms was carried out south of Iceland from early April to late June 2008 in an attempt to overcome this obstacle. A Lagrangian float and four seagliders followed the evolution of a phytoplankton patch over the course of the spring bloom, providing high resolution measurements. The autonomously measured physical, chemical and bio-optical data were calibrated with in situ samples from four supporting cruises. During the experiment a major export event of diatoms was observed. A modelling study was carried out to implement this behavior in an ecosystem model. The model's physical framework is based on the 1-D General Ocean Turbulence Model (GOTM), implemented for the experiment location, following the Lagrangian float. It is forced with data on solar radiation, wind speed, air pressure, air temperature and humidity, and nudged to temperature and salinity observations in order to mimic the patch tracked by the float. This physical model is coupled to three biological model variants of different complexity. All include phytoplankton, zooplankton, detritus, dissolved inorganic nitrogen, silicic acid, chlorophyll and oxygen. The first variant includes only one phytoplankton group and the other two variants include two phytoplankton groups, representing diatoms and small phytoplankton. The third variant differs from the second in that diatoms are capable of turning into fast-sinking cysts under low nutrient conditions, in this case, silicate limitation. Variational data assimilation was applied to optimize the models and error analysis was conducted to verify the estimates of the biological parameters. The models were compared based on their performance in replicating the bloom and its associated carbon export. This study showed silica-controlled diatom cyst formation as an important process controlling carbon export during the North Atlantic Spring Bloom. The power of autonomous platforms for studying the oceans is demonstrated as more comprehensive observations allowed to constrain more parameters than in previous studies