Earth System Science Data (Mar 2024)

Metazoan zooplankton in the Bay of Biscay: a 16-year record of individual sizes and abundances obtained using the ZooScan and ZooCAM imaging systems

  • N. Grandremy,
  • P. Bourriau,
  • E. Daché,
  • M.-M. Danielou,
  • M. Doray,
  • C. Dupuy,
  • B. Forest,
  • L. Jalabert,
  • M. Huret,
  • S. Le Mestre,
  • A. Nowaczyk,
  • P. Petitgas,
  • P. Pineau,
  • J. Rouxel,
  • M. Tardivel,
  • J.-B. Romagnan

DOI
https://doi.org/10.5194/essd-16-1265-2024
Journal volume & issue
Vol. 16
pp. 1265 – 1282

Abstract

Read online

This paper presents two metazoan zooplankton datasets obtained by imaging samples collected on the Bay of Biscay continental shelf in spring during the PELGAS (PELagique GAScogne) integrated surveys over the 2004–2019 period. The samples were collected at night with a 200 µm mesh-size WP2 net fitted with a Hydrobios (back-run stop) mechanical flowmeter and hauled vertically from the sea floor to the surface, with the maximum depth set at 100 m when the bathymetry was deeper than this. The first dataset originates from samples collected from 2004 to 2016 and imaged on land with the ZooScan and is composed of 1 153 507 imaged and measured objects. The second dataset originates from samples collected from 2016 to 2019 and imaged onboard the R/V Thalassa with the ZooCAM and is composed of 702 111 imaged and measured objects. The imaged objects are composed of zooplankton individuals, zooplankton pieces, non-living particles and imaging artefacts ranging from 300 µm to 3.39 mm in equivalent spherical diameter which were individually imaged, measured and identified. Each imaged object is geolocated and associated with a station, a survey, a year and other metadata. Each object is described by a set of morphological and grey-level-based features (8 bit encoding, 0 = black, 255 = white), including size, that were automatically extracted from each individual image. Each object was taxonomically identified using the web-based application Ecotaxa with built-in random-forest and CNN-based semi-automatic sorting tools, which was followed by expert validation or correction. The objects were sorted into 172 taxonomic and morphological groups. Each dataset features a table combining metadata and data at individual-object granularity from which one can easily derive quantitative population and community descriptors such as abundances, mean sizes, biovolumes, biomasses and size structure. Each object's individual image is provided along with the data. These two datasets can be used in combination for ecological studies, as the two instruments are interoperable, or they can be used as training sets for ZooScan and ZooCAM users. The data presented here are available at the SEANOE dataportal: https://doi.org/10.17882/94052 (ZooScan dataset, Grandremy et al., 2023c) and https://doi.org/10.17882/94040 (ZooCAM dataset, Grandremy et al., 2023d).