Date Awarded

Spring 2017

Document Type


Degree Name

Doctor of Philosophy (Ph.D.)


Virginia Institute of Marine Science


Marjorie A. M. Friedrichs

Committee Member

Walker O. Smith

Committee Member

Elizabeth A. Canuel

Committee Member

Elizabeth H. Shadwick

Committee Member

Eileen E. Hofmann


As Earth’s climate changes, polar environments experience a disproportionate share of extreme shifts. Because the Ross Sea shelf has the highest annual productivity of any Antarctic continental shelf, this region is of particular interest when striving to characterize current and future changes in Antarctic systems. However, understanding of mesoscale variability of biogeochemical patterns in the Ross Sea and how this variability affects assemblage dynamics is incomplete. Furthermore, it is unknown how the Ross Sea may respond to projected warming, reduced summer sea ice concentrations, and shallower mixed layers during the next century. To investigate these dynamics and explore their consequences over the next century, high-resolution glider observations were analyzed and used in conjunction with a one-dimensional, data-assimilative biogeochemical-modeling framework. An analysis of glider observations from two latitudinal sections in the Ross Sea characterized mesoscale variability associated with the phytoplankton bloom and highlighted potential mechanisms driving change in the assemblage. In particular, an observed increase in the ratio of carbon to chlorophyll (C:Chl) suggested a marked transition from a phytoplankton assemblage dominated by Phaeocystis antarctica- to one dominated by diatoms. The expected control of phytoplankton variability by Modified Circumpolar Deep Water and mixed layer depth were shown to be insignificant relative to the effects of wind and sea surface temperature on the temporal/spatial scales measured by the glider. Additional glider measurements were used to force the Model of Ecosystem Dynamics, nutrient Utilisation, Sequestration and Acidification, which was adapted for use in the Ross Sea (MEDUSA-RS) to include both solitary and colonial forms of Phaeocystis antarctica. The impacts of climate-induced changes on Ross Sea phytoplankton were investigated with MEDUSA-RS using projections of physical drivers for mid- and late-21st century, and these experiments indicated increases of primary productivity and carbon export flux. Additional scenario experiments demonstrated that earlier availability of low light due to reduction of sea ice early in the growing season was the primary driver of simulated productivity increases over the next century; shallower mixed layer depths additionally contributed to changes of phytoplankton composition and export. Glider data were assimilated into MEDUSA-RS using the Marine Model Optimization Testbed (MarMOT) to optimize eight phytoplankton model parameters. Assimilation experiments that used different data subsets suggest that assimilating observations at the surface alone, as are typically available from remote-sensing platforms, may underestimate carbon export to depth and overestimate primary production. Experiments assimilating observations characteristic of a cruise-based sampling frequency produced a wide range of solutions, depending on which days were sampled, suggesting the potential for large errors in productivity and export. Finally, assimilating data from different spatial areas resulted in less variation of optimal solutions than assimilating data from different time periods in the bloom progression; these temporal differences are primarily driven by decreasing colonial P. antarctica growth rates, increasing colonial P. antarctica C:Chl, and faster sinking of colonies as the bloom progresses from the accumulation stage through dissipation. Overall, this dissertation research demonstrates the value of using bio-optical glider observations in conjunction with modeling to characterize phytoplankton dynamics in a remote marine ecosystem. High-resolution glider data are better able to resolve mesoscale physical-biological relationships, which are typically not discernible from lower frequency data, but it can be difficult to identify mechanistic relationships from in situ measurements alone. In addition, biogeochemical models can be used to extend insights gained by empirical observation, but application is often limited by the quantity and type of in situ data appropriate for evaluation and forcing. The use of gliders for facilitating development and operation of a lower trophic level model demonstrated the effectiveness of a synthetic approach that partly overcomes the individual limitations of these otherwise distinct approaches. Finally, the combination of these approaches is especially useful for gaining a better understanding of ecosystem dynamics in regions similar to the Ross Sea that are undergoing substantive climate-induced changes and where harsh conditions make other means of access difficult.




© The Author