All data were provided as part of the Flemish contribution to the LifeWatch ESFRI by Flanders Marine Institute (VLIZ)
The host institution
Flanders Marine Institute
Vlaams Instituut voor de Zee (VLIZ)
Flanders Marine Institute is an autonomous, non-profit organization witch mission is to create a platform for marine research. The institute offers an open library of data. One of the core activities is: supporting networks of marine scientists, detecting opportunities for marine research, providing equipment and infrastructure, developing data center, and managing library with extensive marine literature.
Data analyzed during professional practice, are going to be used in the monitoring program, performed under the framework of The Life Watch project. Life Watch is a part of the European Strategy Forum on Research Infrastructure and can be seen as a virtual laboratory for biodiversity research. Zooplankton monitoring data are part of open data library, accessible online, and free to use in scientific research.
Welcome to the world of marine drifters
The Secret Life of Plankton
Zooplankton is a group of living animals carried along in large water masses, unable to swim against currents. Some zooplankton organisms like Copepods or Appendicularias spend their whole lifetime as plankton. However, for many the planktonic larval stage is just the beginning of their life journey in the ocean.
The importance of studying zooplankton
Zooplankton organisms are an important part of ocean ecosystems. Many animals like fish and echinoderms have a planktonic larval stage in their life-cycle. What is more zooplankton species are an important part of the food web, many of them are a key element in the diet of fish and other animals. Species composition of zooplankton may tell us about water quality, as they are responding quickly to nutrient pollution. What is more, certain species ( e.g. Calanus finmarchicus ) are indicators of water masses and can be used in research about the effects of climate change. For this, and many other reasons, zooplankton is gaining popularity as a research topic. One of a many challenges that science is facing nowadays is developing better, less destructive methods of data collecting.
Meet Belgian Zooplankton
© LifeWatch - Flanders Marine Institute (VLIZ) 
Calanoida;
an order of copepods includes both marine and freshwater copepods. An important element in many food webs. They feeding on phytoplankton and algae. Either larval or adult forms of Calanoids are an important food source for many commercial fish species. They can be distinguished from other planktonic copepods by having the first antennae at least half the length of the body.
Appendicularia;
a class of free-swimming tunicates. They can be found throughout the world's oceans. Like most tunicates, they are filter feeders. Appendicularians are present in the pelagic zone, specifically in the upper photic zone (uppermost layer of the ocean that receives sunlight) or sometimes deeper. Until now, only Oikopleura dioica is known from the Belgian waters.
Echinoderms;
planktonic larvae of organisms from phylum Echinodermata: a widespread group of benthic (leaving on the seafloor) invertebrates, including starfish, brittle stars, sea cucumbers, and sea urchins. They are characterized by radial symmetry, and arms radiating from a central body.
Harpacticoida;
an order of copepods, characteristic in shape. It can be distinguished from other copepods by the presence of only a very short pair of first antennae. They typically have a wide abdomen, and often have a worm-shape body. In Belgium, the only known species of Harpacticoida is Euterpina acutifrons.
Chaetognatha;
phylum of predatory marine worms, a major component of plankton all over the world. Most chaetognaths are transparent and are torpedo-shaped, but some deep-sea species are orange. In Belgian waters, this taxon is rare and for now, just two species are known Parasagitta elegans and P. setosa.
Noctiluca;
commonly known as the sea sparkle, is represent of single-celled eukaryotes from phylum Dinoflagellate. They can create "blooms": seasonal high concentrations, that can be easily detected thanks to their bio-luminescence abilities.
Mysidae;
planktonic larvae of small, shrimp-like crustaceans. They are omnivores, capable of both filter-feeding and capturing prey. A feeding current is created by the thoracic legs, which also provide for regular swimming motion.
Obelia;
medusa stage of genus Obelia, class of Hydrozoans. Commonly known as sea fur. Most of the hydrozoans include two stages in their life-cycle: attached to the bottom polyp and freely swimming medusa. In zooplankton samples mainly medusa forms are present.How the zooplankton research is performed?
Video plankton recorder (VPR)
Video Plankton Recorder is a towed underwater microscope system. The metal frame is equipped in camera and strobe light with additional conductivity, temperature, pressure, and turbidity sensors. Measurements are obtained by the towing device behind RV Simon Stevin. The real-time images are transferred from device to deck unit, where the regions of interest (ROIs) are extracted from original photos.
Plankton net
The WP2 plankton net is a standard, recommended by the UNESCO tool used in plankton research. It consists of 57 cm diameter steel ring, mechanical flow meter, 2.6m long net (200 µm mesh size), and a plastic bucket at the end of the net.
ZooScan
ZooScan samples are collected with the use of plankton net, during monthly Life Watch campaigns, on the research vessel “Simon Stevin”. After splitting, in the laboratory, liquid samples are scanned in the waterproof ZooScan (HYDROPTIC) device. As a result, ZooScan generates high-resolution digital images of preserved zooplankton.
Sampling area
Sampling stations are presented on the map below.
Remote photo-identification
The initial plan of the practicum consisted of data collection on the research vessel and laboratory proceeding of the zooplankton samples. However, due to the COVID-19 pandemic and fallowing traveling restrictions, the plan has changed. Work was limited to tasks that can be done remotely, via VPN connection to institution servers. It was still mainly focused on zooplankton research, but based on photo-identification instead of sampling and laboratory analysis.
ZooScan
Automatic identification is possible in the Plankton Identifier software. In the first step, users build and evaluate the learning set, by manual assigning taxonomic names to pictures. Based on the learning set, the software automatically assigns sample images to the taxonomic groups and creates a prediction file. After prediction, the user is checking the correctness and moving the mislabeled element to the right folders.
VPR
Unlike in ZooScan analysis, there is no dedicated software to automatize this process. Pictures need to be manually dropped one by one or in groups from one folder to another.
Results
Comparison of two samples
A comparison of results from these two different techniques will help to evaluate which method is better. But how to do it? During sampling Video Plankton Recorder and WP2 net penetrate different water volumes, luckily we know their values. The next step is just to calculate densities of the samples and particular animal groups.
Result of preliminary comparison of ZooScan and VPR sample are presented below.
VPR data collected 27/11/2019, deployed for 3 hours, starting at station ZG02, ending at station 215, total water volume of sample: 0.3045m3
ZooScan data collected 24/11/2019, from station 215, total water volume of sample: 5.3559765m3
Comparison of Zooscan and VPR densities in different organism groups
VPR sample is characterized by a higher density of each group of organisms. In both samples, Calanoida is the densest. The results of two different methods are the most similar in Harpacticoida, VPR sample is less than twice more dens that ZooScan. The most different group is Annelida where density is more than fifteen times higher in the VPR sample.
It is not the end!
Based on the presented data, VPR seems to be a more accurate method, however high differences in densities of samples may be a result of an error. Video recording may be advantageous in research on fragile organisms as Cnidaria or Ctenophora. Those two groups occurred in the VPR but been absent in the ZooScan sample. They are also more likely to disintegrate during standard plankton sampling. Obviously we cannot conclude much based one competition of just single samples.
After replicating the sampling in different stations over time, statistically significant differences can be detected.
Which organisms are more or less detected with different techniques? Are plankton abundances comparable?
These questions still need to be answered, hopefully, results will give a reply.
Acknowledgments
I would like to thank Anouk Ollevier for the perfect organization of this practicum, regardless of difficult circumstances.