Table of Contents

• Coverage
• Dataset Description
  • Methods & Sampling
  • Data Processing Description
  • BCO-DMO Processing Description
  • Problem Description
• Data Files
• Supplemental Files
• Related Publications
• Related Datasets
• Parameters
• Instruments
• Deployments
• Project Information
• Funding

See the closely related datasets listed under the section heading "Related Datasets." In particular, BCO-DMO dataset (960033): "EcoPart particle output from UVP5 of particles and plankton collected from CTD casts during four US GO-SHIP cruises from 2018 - 2022" https://www.bco-dmo.org/dataset/960033

The related CTD data are available at CCHDO:

S04P: https://cchdo.ucsd.edu/cruise/320620180309 (doi: 10.7942/C2F08X)
I06s: https://cchdo.ucsd.edu/cruise/325020190403 (doi: 10.7942/C29660)
A22: https://cchdo.ucsd.edu/cruise/325020210420 (doi: N/A)
P2 Leg 1: https://cchdo.ucsd.edu/cruise/33RR20220430 (doi: N/A) - Identified at CCHDO as P02W
P2 Leg 2: https://cchdo.ucsd.edu/cruise/33RR20220613 (doi: N/A) - Identified at CCHDO as P02E

The Underwater Vision Profiler 5 (UVP5) was utilized to collect in situ images of particles and plankton across the four repeat hydrography transects. The UVP5 was integrated within the conductivity temperature depth (CTD) rosette, and several images per second were acquired during the downcasts spanning from the surface down to near the bottom or 6000 m. The UVP captured in situ images of particles and plankton in a mixed processing mode in which particles were sized, counted, and tabulated in real time during the profile. Subsequent processing bins these particle counts into discrete, pre-defined size bins ranging from 102 µm to >26 mm in equivalent spherical diameter. Particle biovolume was also computed for each size bin. All data contained here are inclusive of both living and non-living particles and plankton. 

Images larger than about 500 µm in equivalent spherical diameter were saved and imported into Morphocluster, an unsupervised clustering algorithm until 91% classification was achieved. Then the resulting classified and unclassified data was imported to the online Ecotaxa database and image sorting tool (https://ecotaxa.obs-vlfr.fr/). There, unclassified images were first sorted into predicted taxonomic categories using a machine learning algorithm, and then predicted zooplankton images from a subset of vertical profiles were manually validated into the most appropriate taxonomic categories. The limited resolution of the images often precludes definitive taxonomic assignment. As such, data users should explore the actual images and their associated classifications by accessing the project via the Ecotaxa website in order to determine whether or not the taxonomic data is correct and appropriate for their purposes. Data was exported from Ecotaxa on 2024-09-05.

The UVP5 software acquires and processes images in real time. The gain, shutter and LED pulses are controlled and the background image is removed. Images are acquired and processed to get size and grey level for each image. Size information on all detected particles is stored and images of individual particles and plankton larger than 500 µm in equivalent spherical diameter are segmented and saved for later identification. Image post processing and metadata creation is accomplished with the Zooprocess software (Version 7.39). Tabulated particle data are used to sum the number and volume of particles within predefined size and depth bins, allowing for the computation of the Datasets. Data and images have been uploaded to the Ecotaxa website (http://ecotaxa.obs-vlfr.fr/) which serves as a tool for particle and zooplankton identification with machine learning and human verification, as well as a repository for all globally collected UVP data. Data files for particle and zooplankton abundances are exported from the Ecotaxa particle module in detailed format. The size distribution data is reported in non-differential forms (simply the concentration of particles in each size bin) as well as on a numeric and particle volume basis assuming all particles are spherical. The size bin limits are defined in equivalent spherical diameter (ESD), where ESD = (4 Sm π-1)-0.5 where Sm is the projected area of each particle in mm² and the particle concentrations are reported for each size bin defined by the log-transformed center of each size bin in µm. The biovolume is computed as the sum of the individual spherical volumes of each particle issued from the calibrated ESD.

* Data tables within "*export_detailed_.txt" files were not able to be imported into the BCO-DMO data system directly due to incompatibilities in lat/lon and datetime formats in this and related BCO-DMO dataset "960033" "EcoPart particle output from UVP5..."
(see problems and issues section)

Data attached to this dataset include the original format submitted to BCO-DMO:

* Ecotaxa images provided as .zip files attached as "Data Files" to this dataset.
* Ecotaxa output data files (one per cruise) was zipped into Data File "EcoTaxa_Export_Files.zip".
* Example images attached as supplemental files

Supplemental Cruise metadata:

* Individual cruise metadata text files were concatenated into the supplementary file "cruise_metadata.csv". Additional columns were added to the combined table "Cruise_ID_R2R" to clarify which cruise id in the dataset corresponds to the cruise identifier at the Rolling Deck to Repository (R2R). See problems/issues section, which documents the format of latitude and longitude in this table.

NSF Award Abstract:
Microscopic plants and animals in the surface ocean remove the atmospheric carbon dioxide dissolved in seawater by using it to make biological materials. After organisms die, some portion of this carbon sinks into the deep sea where it dissolves back into the water or lands on the seafloor. The Biological Carbon Pump is the name for this carbon transfer out of the sunlit surface ocean, and it is a very important process controlling the earth's carbon cycle and climate. Surprisingly, the size and rate of this transfer remains unclear because data for carbon in particles are not available for many times and places in the ocean. The researcher for this project plans to collect a remarkable new data set and examine it to answer major questions about the Biological Carbon Pump. The data will come from joining existing cruises on research ships sailing through all of the world's oceans, creating a systematic global survey of particles in the ocean. These data will be the heart of a new public database for use by other international researchers in their continued study of the Biological Carbon Pump. For this young researcher, the project will also lay the foundation for a career of integrated research, education, and outreach in oceanography. An Alaskan native will go to sea and analyze data, using these results to complete a Ph.D. degree. In addition, the project will design and produce a new aquarium exhibit at the Alaska SeaLife Center (ASLC) to educate the public about the complex workings and global importance of the Biological Carbon Pump. These activities should increase diversity in oceanography and inspire a new generation of scientists.

Reliable and useful data on the Biological Carbon Pump is sparse, mostly due to the experimental, logistical, and technical challenges of studying a complex system over the great expanse of the global ocean. This project aims to tackle a fundamental and ongoing problem in studying the Biological Carbon Pump by providing a global and consistent dataset on particles and plankton in the ocean. In collaboration with the US Repeat Hydrography program, this project will use in situ imaging technology to determine the total abundance, size distribution, and functional groups of particles and mesozooplankton along seven global ocean transects. The design will test fundamental hypotheses related to the presence and nature of regional particle hotspots and examine the global patterns of zooplankton activity and vertical migration as they affect the biological pump. Links between satellite data, calculated flux patterns, attenuation of particles through the mesopelagic, and constraints on biogeochemical models will also be investigated.