Quantitative imaging datasets of micro to mesoplankton communities and surface microplastic across the Pacific Ocean from the Tara Pacific Expedition

Abstract. This paper presents the quantitative imaging datasets collected during the Tara Pacific Expedition (2016–2018) on the schooner Tara. The datasets cover a wide range of plankton sizes, from micro-phytoplankton > 20 μm to meso-zooplankton of a few cm, as well as non-living particles such as plastic and detrital particles. It consists of surface samples collected across the North and South Pacific Ocean from open ocean stations (a total of 357 samples) and from stations located in coastal waters, lagoons or reefs of 32 Pacific islands (a total of 228 samples). As this expedition involved long distances and long sailing times, we designed two sampling systems to collect plankton while sailing at speeds up to 9 knots. To sample microplankton, surface water was pumped onboard using a customised pumping system and filtered through a 20 µm mesh size plankton net (here after Deck-Net (DN). A High Speed Net (HSN; 330 μm mesh size) was developed to sample the mesoplankton. In addition, a Manta net (330 µm) was also used when possible, to collect mesoplankton and plastics simultaneously. We could not deploy these nets in reef and lagoon stations of islands. Instead, two Bongo nets (20 µm) attached to an underwater scooter were used to sample microplankton. Microplankton (20–200 μm) from the DN and Bongo nets was imaged directly on-board Tara using the FlowCam (Fluid imaging, Inc.) while the mesoplankton (> 200 μm) from the HSN and Manta nets was analyzed in the laboratory with the ZooScan system. Organisms and other particles were taxonomically and morphologically classified using the web application EcoTaxa automatic sorting tools, followed by taxonomic expert validation or correction. More than 300 different taxonomic and morphological groups were identified. The datasets include the metadata with the raw data from which morphological traits such as size (ESD) and biovolume have been calculated for each particle, as well as a number of quantitative descriptors of the surface plankton communities. These include abundance, biovolumes, Shannon diversity index and normalised biovolume size spectra, allowing the study of their structures (e.g. taxonomic, functional, size structure, trophic structure, etc.) according to a wide range of environmental parameters at the basin scale. In addition to describing and presenting the datasets, the complementary aim of this paper is to investigate and quantify the potential sampling biases associated with the two high speed sampling systems and the different net types, in order to improve further ecological interpretations.