CHROME Project: Cytosense and Ferrybox data collected on the Tunis-Genova route of the CTN ferry le Carthage

The Ferrybox has been installed on the CTN ferry Le Carthage in the frame of the DYMEDTUN Tunisian National project funded by the Marine Environment Laboratory LR16INSTM04 and was combined with a cytosense flow cytometer in the frame of the A*MIDEX CHROME project (n° ANR-11-IDEX-0001-02) funded by the « Investissements d’Avenir » French Government program, managed by the French National Research Agency (ANR). 
The Ferrybox system (4HJena Engineering) is a combination of different automated sensors along a controlled seawater flow-through installed below the waterline. The sea surface temperature (SST, °C), salinity (SSS), turbidity, oxygen, and chlorophyll a (chla) fluorescence (r.f.u) were recorded along the route. The temperature was directly recorded at the inlet of the pumped seawater by an SBE38 (Seabird). Salinity, oxygen, turbidity and chla fluorescence were recorded within the Ferrybox by a Seabird thermosalinometer SBE45, an Aandera optode, and Seapoint turbidimeter and fluorometer, respectively. A regular cleaning using a 24 % sulfuric acid limits the biofouling in the Ferrybox tubing, and the system starts its recording 1 to 3 nautical miles (~2-5 km) from the coast, controlled by an external GPS. The FerryBox was installed shortly before the flow cytometer, in February 2016, with recently calibrated sensors. This calibration was used for the data sets presented in this paper. While the fluorometer was manufacturer calibrated, in the absence of discrete chla measurement we prefer to report these values as r.f.u. (Marrec et al., 2025). The water inlet was implemented at the front of the ship at -5 m.
The CytoSense
The CytoSense (Cytobuoy, NL) is an automated pulse-shape recording flow cytometer (AFCM) dedicated to being deployed on vessels and remote access sampling systems. The seawater passing through the Ferrybox was collected in a subsampler system made of an enclosed volume of 200 cm-3 continuously flushed at a flow rate of 5 dm-3.min-1, combined with a controlled pinch valve isolating the sample before analysis. This strategy considerably reduces the spatial coverage of the sampling as the flow cytometer requires several minutes to finalise its acquisition. The CytoSense records optical pulse shapes per particle contained in the seawater, separating phytoplankton cells based on their autofluorescence by photosynthetic pigments. The sample is sent to an optical chamber, thanks to a weight-calibrated sample peristaltic pump, and is then surrounded by an isotonic sheath fluid, generating a laminar flow, making each particle aligned one by one, ready to be intercepted by a 488 nm laser beam (Coherent sapphire). A set of optical curves, called pulse shapes, is thus generated for each particle. Settings have been optimised and used as a reference to build the best practices (Gallot et al., 2025). The pulse shapes of sideward scatter (SWS, 488 nm) and fluorescence emissions were separated by a set of optical filters – orange fluorescence (FLO, 552–652 nm) and red fluorescence (FLR, > 652 nm) – and collected on photomultiplier tubes. The pulse shapes of forward scatter (FWS) were collected on left- and right-angle photodiodes and used to validate the laser alignment. The samples were acquired every 30 min. during the functioning of the Ferrybox, through two distinct protocols, one dedicated to the smallest autofluorescing cells, and a second one, with a larger analysed volume, dedicated to the large and less concentrated cells. The instrument and the acquisition protocol are described in Marrec et al. (2018) and Barrillon et al. (2023).
To ensure a reliable clean sheath fluid and the analysis of beads regularly to monitor the stability of the optics and the fluidics, a prototype was fixed to the CytoSense frame, including a syringe filled with a solution containing polystyrene 2 µm beads (Polyscience ®), programmed to be analysed every day at a dedicated time, and a set of coal filters (Pall filters ®) to clean the sheath from chromophoric material, and a bag with biocide (Proclin 150 ®), injected at a concentration of 150 ppm into the sheath to reduce the fouling in the tubing.
The phytoplankton functional groups, biomass and chlorophyll estimation
Phytoplankton functional groups (PFG) were manually classified using the Cytoclus3 software following the description of Thyssen et al., 2022. The mean of the area under the curve of FWS and FLR (hereafter identified as FWS and FLR) was retrieved for each PFG and converted into estimated size diameter (ESD, µm), carbon content per cm-3 and chlorophyll a (chla) per cm-3. The FWS and the FLR have been normalised toward 2 µm red fluorescing beads (Polysciences ®) analysed along the experiment. Flow cytometry normalised FWS was converted into ESD thanks to a regression between silica beads of different sizes (1.0, 2.01, 3.13, 5.02, and 7.27 μm, non-functionalised silica microspheres, Bangs Laboratories, Inc.) as described in (Marrec et al., 2018 and Barrillon et al., 2023). The conversion of ESD into biovolume (BV, μm3) was estimated from the regression computed during the Fumseck Cruise (May 2018, Barrillon et al., 2023), which used the same instrument and the same settings. The normalised FWS during this experience was then multiplied by the 2 µm red beads FWS value as the size regression was estimated before the Fumseck cruise (May 2018). The following regression was used to convert FWS into biovolume: log(biovolume)= log(FWS) x 0.912 - 5.5407 (n=5, r=0.95, p=0.015). The phytoplankton biomass per group was computed (in ngC.cm-3) from the power law of the form aBVb, to get a mean carbon cellular quota (C, pgC per cell), with a and b conversion factors reported by (Menden-Deuer and Lessard, 2000) applied to pico-sized cells (a=0.26, b=0.86), and (Verity et al., 1992), applied to nano-sized cells (a=0.433, b=0.863).
To convert the FLR recorded by the CytoSense into chlorophyll a concentration, the instrument was calibrated with discrete extracted chlorophyll a in November 2015, February 2017 and February 2018 as described in Barrillon et al., 2023. The phytoplankton FLR collected during this study was normalised with the regularly analysed 2 µm polystyrene beads FLR. The linear regression was chla (ng.cm-3)= 0.11 x normalised FLR (a.u. cm-3) (n=41, r=0.86, p<0.001).