Underwater cameras to assess the global distribution of zooplankton

As part of an international collaboration, a research team from the Oceanographic Laboratory of Villefranche sur Mer (LOV, Sorbonne University/CNRS) has for the first time collected a large dataset of zooplankton data obtained by cameras on fleets globally. His analysis (1) allowed to model the oceanic composition and biomass of zooplankton. Calculations show maximum biomass values ​​in equatorial, temperate, and polar fertile zones, as well as minimum values ​​in oceanic deserts. The results of this study have been published Frontiers of Marine Science August 9, 2022.

Plankton is all the living organisms that drift in the water and sea compartments of the earth. It is extremely diverse and plays a central role in marine food chains and geochemical cycles. These organisms fill the entire water column and are fundamental to our societies as they provide important ecosystem services such as fisheries and carbon sequestration. It is important to be able to quickly measure their diversity and biomass in the oceans.

Although satellites can observe autotrophic plankton (2) on a global scale, they can only do so in the first few meters of the surface. As for the expedition ships, they can only get information locally. In addition, the processing of samples brought on board is often long and tedious. The total biomass of zooplankton is thus poorly known.

Recently, LOV researchers developed and commercialized an underwater digital camera that allows dozens of international colleagues (3) to obtain global observations of zooplankton between 1 and 50 mm in the world’s oceans at more than 3,500 sites. Recognition algorithms using artificial intelligence were able to identify millions of images, taking into account the biomass of 30 categories of macroplankton.

using algorithms machine learning, the research team correlated environmental conditions (temperature, salinity, nutrients and oxygen, etc.) with the organisms’ estimated carbon biomass to predict the biomasses of these categories of plankton on a global scale. Model results estimate biomass at least 0.403 gigatons of carbon between 0 and 500 m depth, i.e. on the same scale as all fish in the ocean (Bar-on et al., 2018).

Plankton biomass estimation maps show the spatial distribution of organisms according to environmental conditions (temperature, salinity, etc.) and can now be used for climate modeling or quantifying stocks for fish. These maps also allow for the first estimate of plankton biomass from millions of images collected in the oceans.

LOV is now equipping its autonomous floats with this imaging technology to obtain information in regions of the ocean where boats rarely sail. International deployment of these cameras on fleets of autonomous underwater robots is expected as part of United Nations-supported surveillance (4).

Legend: Summary diagram of methodology – 1) Acquisition of plankton and particle images using Underwater Vision Profiler 5 (UVP5, Picheral et al., 2010). 2) Classification of these images into different taxonomic groups and selection of the group of interest (here copepods). 3) Calculation of biomass for this planktonic group for each vertical profile produced by UVP5. 4) Restoration of environmental variables in international scientific databases (WOA, Copernicus, NOAA). 5) Implementation and use of habitat models, machine learning tools to estimate taxonomic group abundance in an unobserved location. To do this, we use the relationship between the group of interest and environmental conditions to estimate its global distribution. Thus, we obtain two model outputs: biomass change as a function of environmental variables and a global map of predicted biomass.

Press release from Sorbonne University and CNRS, 08/09/22

Reference:

Global Distribution of Zooplankton Biomass In Situ Imaging and Machine Learning, Laetitia Drago, Thelma Panaiotis, Jean-Olivier Irisson, Marcel Babin, Tristan Biard, François Carlotti, Laurent Coppola, Lionel Guidi, Helena Hauss, Leeard Faens, Leeard Karbi- , Andrew M. P McDonnell, Marc Picheral, Andreas Rogge, Anya M. Waite, Lars Stemmann and Rainer Kiko, Front. Tuesday Science, 09 August 2022 Sec. Ocean observation

(1) Prepared as part of Laetitia Dragon’s thesis at LOV.

(2) Autotrophic plankton refers to planktonic organisms that can use light and nutrients from water to convert CO2 into biomass through the process of photosynthesis.

(3) In France, scientists from Takuvik laboratory (CNRS/Laval University), Oceanology and Earth Sciences Laboratory (CNRS/Lille University/Littoral-Côte-d’Opale University), Mediterranean Oceanology Institute (Institute) were involved in this work. CNRS/IRD/University of Aix-Marseille/University of Toulon), Naval Stations Observatory (CNRS/University of Sorbonne)

(4) The Biogeochemical-Argo (BGC-Argo) project scientific and implementation plan published in 2016 supports the French contribution to the creation of a global network of biogeochemical profiling swims of the international BGC-Argo program.