Illuminating the Twilight Zone: In-Situ Imaging Facilitates Discovery of Mesopelagic Plankton Dynamics and Community Structure

Ocean ecosystems provide both food and regulation of climate. Much of ocean ecosystem functioning occurs through processes in the pelagic, where plankton are the dominant organisms. While plankton have been researched for several centuries, recent advances in technology are greatly improving the ability to better understand plankton and how their communities are structured. Specifically, in-situ imaging allows oceanographers and plankton ecologists to better sample fragile taxa and record organisms’ true traits at high frequencies. These tools can facilitate investigation into new topics in plankton ecology and better understand critically understudied habitats, like the mesopelagic of open-ocean systems. However, in-situ imaging devices are not a plankton ecologist’s panacea; there are many drawbacks and limitations which need to be fully understood when utilizing these novel tools. In this dissertation, I present four chapters which investigate in-situ imaging for plankton ecology and provides new discovery into plankton communities in the oligotrophic ocean and throughout the water column. First, I investigate how well an in-situ imaging device compares to traditional, net-based sampling. Ultimately, I show there are some taxa which are better sampled while others are under sampled. Additionally, I provide a framework for many aspects of processing data associated with in-situ imaging. Then, using in-situ images of copepods, I investigate how traits influence diel vertical migration patterns. I also analyze Rhizaria populations across time and vertical space in the Sargasso Sea. I show through this research that Rhizaria occupy different portions of the water column. Finally, I present a literature review of studies which utilize in-situ imaging to research plankton ecology. I provide an evaluation of the statistical tools commonly used in ecological investigations and highlight the need to consider the sampling probabilities associated with these devices.