A discrete, stochastic model of colonial phytoplankton population size structure: Development and application to in situ imaging-in-flow cytometer observations of Dinobryon.

The scientific community lacks models for the dynamic changes in population size structure that occur in colonial phytoplankton. This is surprising, as size is a key trait affecting many aspects of phytoplankton ecology, and colonial forms are very common. We aim to fill this gap with a new discrete, stochastic model of dynamic changes in phytoplankton colonies' population size structure. We use the colonial phytoplankton Dinobryon as a proof-of-concept organism. The model includes four stochastic functions-division, stomatocyst production, colony breakage, and colony loss-to determine Dinobryon population size structure and populations counts. Although the functions presented here are tailored to Dinobryon, the model is readily adaptable to represent other colonial taxa. We demonstrate how fitting our model to in situ observations of colony population size structure can provide a powerful approach to explore colony size dynamics. Here, we have (1) collected high-frequency in situ observations of Dinobryon in Lac (Lake) Montjoie (Quebec, Canada) in 2013 with a moored Imaging FlowCytobot (IFCB) and (2) fit the model to those observations with a genetic algorithm solver that extracts parameter estimates for each of the four stochastic functions. As an example of the power of this model-data integration, we also highlight ecological insights into Dinobryon colony size and stomatocyst production. The Dinobryon population was enriched in larger, flagellate-rich colonies near bloom initiation and shifted to smaller and emptier colonies toward bloom decline.

The private life of Cystodinium: in situ observation of its attachments and population dynamics.

Phytoplankton images were collected using an Imaging Flow Cytobot moored in the mesotrophic lake Lac Montjoie (Quebec, Canada). Cystodinium-an unusual dinoflagellate genus-was found during manual classification of the images into taxonomic groups while building an automated classifier. Cystodinium's particularity is that while it can take a typical motile dinoflagellate form, it is thought to exist primarily as an immotile photosynthetically competent parasitic cyst in the shape of a crescent moon. Observations presented here are of this immotile lunate cyst. Manually classified images revealed that the majority of the Cystodinium found (86%) were attached to other microalgae or detrital material while the rest were unattached. The established auto-classifier was only able to correctly identify unattached Cystodinium images and thus was used to generate time series as cells per 100 mL for the unattached cell subset. Our observations, coupled with a literature review, lead us to question the parasitic nature of this taxonomic group.