The emergence of automated high-frequency flow cytometry: revealing temporal and spatial phytoplankton variability

Phytoplankton observation is the product of a number of trade-offs related to sampling processes, required level of diversity and size spectrum analysis capabilities of the techniques involved. Instruments combining the morphological and high-frequency analysis for phytoplankton cells are now available. This paper presents an application of the automated high-resolution flow cytometer Cytosub as a tool for analysing phytoplanktonic cells in their natural environment. High resolution data from a temporal study in the Bay of Marseille (analysis every 30 min over 1 month) and a spatial study in the Southern Indian Ocean (analysis every 5 min at 10 knots over 5 days) are presented to illustrate the capabilities and limitations of the instrument. Automated high-frequency flow cytometry revealed the spatial and temporal variability of phytoplankton in the size range 1−∼50 μm that could not be resolved otherwise. Due to some limitations (instrumental memory, volume analysed per sample), recorded counts could be statistically too low. By combining high-frequency consecutive samples, it is possible to decrease the counting error, following Poisson’s law, and to retain the main features of phytoplankton variability. With this technique, the analysis of phytoplankton variability combines adequate sampling frequency and effective monitoring of community changes.

Microbial spatial variability: An example from the Celtic Sea

In July 2004, dominant populations of microbial ultraplankton (<5 μm), in the surface of the Celtic Sea (between UK and Eire), were repeatedly mapped using flow cytometry, at 1.5 km resolution over a region of diameter 100 km. The numerically dominant representatives of all basic functional types were enumerated including one group of phototrophic bacteria (Syn), two groups of phytoplankton (PP, NP), three groups of heterotrophic bacterioplankton (HB) and the regionally dominant group of heterotrophic protists (HP). The distributions of all organisms showed strong spatial variability with little relation to variability in physical fields such as salinity and temperature. Furthermore, there was little agreement between distributions of different organisms. The only linear correlation consistently explaining more than 50% of the variance between any pairing of the organism groups enumerated is between two different groups of HB. Specifically, no linear, or non-linear, relationship is found between any pairings of SYB, PP or HB groups with their protist predators HP. Looking for multiple dependencies, factor analysis reveals three groupings: Syn, PP and low nucleic acid content HB (LNA); high nucleic acid content HB (HNA); HP and NP. Even the manner in which the spatial variability of Syn, PP and HB abundance varies as a function of lengthscale (represented by a semivariogram) differs significantly from that for HP. In summary, although all microbial planktonic groups enumerated are present and numerically dominant throughout the region studied, at face value the relationships between them seem weak. Nevertheless, the behaviour of a simple, illustrative ecological model, with strongly interacting phototrophs and heterotrophs, with stochastic forcing, is shown to be consistent with the observed poor correlations and differences in how spatial variability varies with lengthscale. Thus, our study suggests that a comparison of microbial abundances alone may not discern strong underlying trophic interactions. Specific knowledge of these processes, in particular grazing, will be required to explain the causes of the observed microbial spatial variability and its resulting consequences for the functioning of the ecosystem.

Nanoplankton and picoplankton in the Western English Channel: abundance and seasonality from 2007–2013

The nano- and picoplankton community at Station L4 in the Western English Channel was studied between 2007 and 2013 by flow cytometry to quantify abundance and investigate seasonal cycles within these communities. Nanoplankton included both photosynthetic and heterotrophic eukaryotic single-celled organisms while the picoplankton included picoeukaryote phytoplankton, Synechococcus sp. cyanobacteria and heterotrophic bacteria. A Box–Jenkins Transfer Function climatology analysis of surface data revealed that Synechococcus sp., cryptophytes, and heterotrophic flagellates had bimodal annual cycles. Nanoeukaryotes and both high and low nucleic acid-containing bacteria (HNA and LNA, respectively) groups exhibited unimodal annual cycles. Phaeocystis sp., whilst having clearly defined abundance maxima in spring was not detectable the rest of the year. Coccolithophores exhibited a weak seasonal cycle, with abundance peaks in spring and autumn. Picoeukaryotes did not exhibit a discernable seasonal cycle at the surface. Timings of maximum group abundance varied through the year. Phaeocystis sp. and heterotrophic flagellates peaked in April/May. Nanoeukaryotes and HNA bacteria peaked in June/July and had relatively high abundance throughout the summer. Synechococcus sp., cryptophytes and LNA bacteria all peaked from mid to late September. The transfer function model techniques used represent a useful means of identifying repeating annual cycles in time series data with the added ability to detect trends and harmonic terms at different time scales from months to decades.