High-temporal resolution in situ imaging and machine learning to observe copepod-parasite interactions

Eric Coughlin Orenstein1, Christian BriseƱo-Avena2, Paul Roberts1, Jules S Jaffe3 and Peter J. S. Franks4, (1)Monterey Bay Aquarium Research Institute, Moss Landing, United States, (2)Oregon State University, Hatfield Marine Science Center, Newport, United States, (3)Scripps Institution of Oceanography, La Jolla, CA, United States, (4)Univ California San Diego, La Jolla, United States
Abstract:
Zooplankton play a critical role in virtually all aquatic ecosystems, forming the link between photosynthetic microorganisms and higher trophic levels. Biotic controls on zooplankton populations are often modeled as top-down, via predation, or bottom-up, via preferred prey availability. There is, however, a growing body of evidence suggesting that parasitism exerts an influence on zooplankton condition (e.g., growth, reproduction, nutritional content). Observing and quantifying these interactions remains difficult due to limited spatial-temporal resolution of traditional sampling techniques and challenges associated with creating realistic environments in the laboratory. In situ microscopy systems alleviate some of these issues and provide a window into these cryptic relationships. Here, we outline the use of the Scripps Plankton Camera to monitor covariations in the relative abundance of the copepod Oithona sp. and Paradinium sp., a parasitic Rhizarian. With human annotators working in concert with a machine classifier, we constructed a two-year long record at hourly resolution from images captured at the Scripps Pier in La Jolla, California. Preliminary time series analysis suggests a coupling between the presence of the parasite and female egg bearing Oithona.