Novel methods for automated in situ observations of phytoplankton diversity

This is a summary of activities and results from months 1-24 of JERICO-NEXT Work Package 3 Innovations in Technology and Methodology, Task 3.1 Automated platform for the observation of phytoplankton diversity in relation to ecosystem services. The aim is to provide a synthesis report after developments dedicated to the observation of the phytoplankton diversity by applying novel techniques on automated platforms. The work has been carried out in close connection with task 4.1 Biodiversity of plankton, harmful algal blooms and eutrophication. The partners involved are CNRS, SYKE, SMHI, HZG, RWS, VLIZ CEFAS, Ifremer and AZTI. Subcontractors in WP4, task 4.1 are WHOI, Scanfjord AB, Tomas Rutten b.v., CytoBuoy b.v. and UGent - PAE.
The work has been carried out mainly in the field with activities in the Baltic Sea, the Kattegat-Skagerrak, the English Channel-North Sea Area, the Western Mediterranean, as well as in shared studies with other WP3.4 and WP4.4 in the Bay of Biscay and, out of Europe, in the Benguela Current. Instrument platforms include research vessels, Ferrybox systems on merchant vessels, instrumented oceanographic buoys/fixed platforms and land based systems. Ocean observatories, i.e. multi-sensor and multi-platform systems, have been used in some locations, allowing inter comparison of techniques and sensors between at least two partners. In addition, work on developing and testing new algorithms have been carried out in offices and laboratories. Some work on microalgae cultures has also been carried out. Two international workshops have been successfully arranged, one in Wimereux, France (June 2016 – organised by CNRS-LOG) and one in Gothenburg, Sweden (September 2016, organised by SMHI) in which partners presented, discussed and were also able to inter compare the sensors and techniques that were or will be implemented in the field. The work was divided into three sections but there is substantial overlap and cooperation. One example is that reference samples analysed in the microscope were used for completing and/or evaluating the quality of some of the automated methods.
Imaging in flow and in situ imaging of plankton (led by SMHI)

The work includes evaluating instruments and developing algorithms for automated identification of phytoplankton from automated image acquisition (in flow or in situ). Three different commercial instruments and one instrument prototype were used. On the Swedish west coast (Skagerrak coast) a study of harmful algae and other phytoplankton was carried out near a mussel farm. The Imaging Flow Cytobot was deployed in situ and collected samples at six different depths for approximately two months. In the English Channel the old generation of FlowCAM and a prototype system, the FastCam, were used to analyse samples on research vessels or in the laboratory. In addition, the CytoSense and CytoSub were used to collect images. The in situ video system UVP5 was implemented during a cruise in the Baltic Sea-Skagerrak-Kattegatarea, together with a new generation of FlowCAM of faster acquisition and providing colour images and CytoSense. A major task was to develop and evaluate plankton identification algorithms. This includes using a subset of images of organisms for training the systems. Existing software were improved (as the PhytoZooImage) and an image data system/platform named EcoTaxa was described and is currently available for storing and cooperative analysis/discrimination of plankton images.
Single-cell optical characterization (led by CEFAS)
Automated flow cytometers (FCM, CytoSense/CytoSub, Cytobuoy b.v.) were implemented on a Ferry line and on research vessels to investigate functional groups of phytoplankton. In the Western Mediterranean the main targets were the picoplankton and the nanoplankton while in the other areas pico-, nano- and microplankton were in focus. Several cruises were carried out in the Channel and North Sea to follow combined diatoms and Phaeocystis bloom development. A cruise covering the Baltic Sea and Skagerrak-Kattegat area had a main focus on cyanobacteria and dinoflagellates. Moreover, inter comparisons of machines and on clustering analysis methods were performed. Finally, a combination of FCM and multi-spectral fluorometer continuous recording was coupled with physical and hydrological continuous measurements in the southern Bay of Biscay.


Novel multi-wavelength fluorometers for detecting phycoerythrin indicative e.g. of certain cyanobacteria and of cryptophytes were evaluated in the Baltic Sea. Multi wavelength fluorometers were also used in the Benguela current, during the Gothenburg workshop, as well as on a variety of field cruises from the southern Bay of Biscay to the E. Channel and North Sea, in order to discriminate amongst main phytoplankton pigmentary groups. The manufacturers’ algorithms were found to be partly inaccurate for detecting algal groups based on photosynthetic pigment composition. New dedicated fingerprints were used in field work to improve discrimination amongst phytoplankton groups. A principle component analyses approach was also evaluated. Single wavelength fluorometers were evaluated in several sea areas. Sun induced photoquenching had a strong effect on fluorescence yield. In the North Sea and the Norwegian Sea multi spectral absorption was used to detect chlorophyll and phytoplankton groups based on pigment content.
Variable fluorometers were implemented on both samples, continuous recording and profiles in the E. Channel and North Sea, as well as in the Baltic and Skagerrak-Kattegat, for studying photosynthetic parameters and potential primary productivity. Recommendations are made on the strategy and type of measurements to carry out.
Future work in task 3.1
Most field work has ended but some will continue, e.g. at the Utö observatory in the Baltic Sea. The data collected during months 1-24 and the new data will be processed further and used for improving the discrimination of phytoplankton taxa or functional groups by inter-comparison of techniques and continued algorithm development, as well as for preparing JERICO-NEXT delivery 3.2. In addition, scientific publication of results is in progress or being planned. A special issue in the open access journal Diversity (MPI) is being discussed. Some results and strategy will be presented during a symposium in Hannover, Germany, in October 2017 and during the FerryBox workshop on board the ship Colour Fantasy later in October 2017. Results will also be presented during the third JERICO-NEXT plankton workshop to be arranged in Marseille in March 2018, during the International Conference on Harmful Algae in Nantes in October 2018, and in other meetings to be determined.

Main conclusions:
1. The methods used are reliable for automated observation of phytoplankton biodiversity (functional groups, size classes, taxa when possible) and biomass, complementing manual methods for sampling and microscope analyses.
2. Operating the equipment and interpreting the results still need a lot of knowledge and time. Even though some operational procedures can be established, the standardization of analytical and data processing as well as data management need more development. The degree of automation varies depending on the method considered.
3. Imaging in flow and in situ imaging provide means for identifying and counting phytoplankton at the genus or species level. Also, biomass based on cell volume of individual cells can be estimated. Development of classifiers for automated identification of organisms is time consuming and requires specific skills on signal analysis and on taxonomy.
4. Flow cytometry has proven to be a useful tool for counting phytoplankton and for describing the phytoplankton community as size based classes and functional groups. There was an agreement to report the phytoplankton count in four groups for inter comparison purposes: Synechococcus (pico-cyanobacteria), pico-eukaryotic organisms, nanoplankton and microplankton.
5. Single and multi-wavelength fluorometry makes it possible to estimate phytoplankton biomass (at a chlorophyll-a basis) and to differentiate phytoplankton based on photosynthetic pigments. Sunlight induced photoquenching is a problem for estimating chlorophyll a from fluorescence. For instruments mounted buoys or vessels, night time data can be used to minimize the problem.
6. Multi-wavelength absorption is a useful tool for estimating chlorophyll a-a and is also useful for discriminating between phytoplankton groups based on pigment content.
7. Variable fluorescence is available for addressing phytoplankton physiology, photosynthetic parameters and to estimate primary productivity on both continuous sub-surface recording and water column profiles, mediating careful coupling with other optical and also biogeochemical analysis.