Project
APICS
Completed project
Project start: June 2022 | Project end: March 2024Funder: Natural Environment Research Council (NERC)
Principal Investigator: Dr James Clark
Other participants from PML: Dr James Fishwick, Claire Widdicombe, Elaine Fileman
The Background
Marine plankton are defined as organisms with zero or small swimming velocities, which move passively with ocean currents. The definition covers a taxonomically and morphologically diverse group of organisms that span many phyla and tens of thousands of species. In size alone, plankton span many orders of magnitude, ranging from sub-micron scale unicellular life forms up to large jelly fish that can measure up to a meter in diameter.
Plankton sustain all other forms of multicellular life in the ocean. Photosynthetic members account for approximately 50 % of global primary production. The energy captured by photosynthetic plankton is passed up the food chain through predatory interactions, where it supports the growth of commercially important species of fish and shellfish. Plankton also regulate Earth’s climate, through their role in the global carbon cycle.
Phytoplankton microscope image by Claire Widdicombe (taken in September 2022 from an L4 sample)
The primary requirement for high frequency observations of plankton arises from the need to align observing frequencies with time scales relevant to plankton population dynamics. Small unicellular phytoplankton have short generation times and can double in number over time-scales of hours to days. This ability often results in spectacular blooms which are readily visible from space. Blooms are characterised by successional phases, where a given species rises in dominance only to be replaced later by another. With infrequent ship-based observations, the successional phases that characterise bloom dynamics cannot be resolved at fine temporal scales. Furthermore, the arrival of invasive species, or the emergence of Harmful Algal Blooms–caused by some algal species which produce deadly toxins as a by-product of their metabolism–may be missed. From infrequent ship-based observations, it is also impossible to observe many short-term phenomena, such as the vertical migration of larger mesozooplankton to find food or avoid being eaten; and the movement of meroplankton–organisms that spend just part of their life cycle as plankton (e.g., shellfish larvae), and are known to change their vertical position in the water column on sub-daily time scales with implications for larval transport, dispersal and recruitment.
Long-term, high frequency observations of the marine environment require automated sampling procedures. However, while the technology for autonomously measuring many physical and chemical variables in situ has existed for decades, in the case of plankton, it has taken time for suitable, commercially available instruments to become available. By working closely with equipment manufacturers, and leveraging the facilities within the WCO, APICS will enable a step change in our ability to make autonomous, long-term, broad size-spectrum measurements of plankton to be made at frequencies comparable to interdependent, co-located physical and chemical variables from a remotely operated offshore platform.
The APICS Platform
The project will configure and deploy an Automated, in situ Plankton Imaging and Classification System (APICS). The asset will be used in long-term, autonomous deployments at sea within the Western Channel Observatory (WCO) off the south coast of the UK and forms part of the National Centre for Coastal Autonomy. APICS consists of an Imaging FlowCytobot (IFCB; McLane Research Laboratories, Inc.) for automatically imaging and classifying plankton that range in size from <10 μm to 150 μm; and a Plankton Imager (Pi-10; Plankton Analytics Ltd.), for automatically imaging and classifying plankton from 180 μm to 20 mm in size. Ancillary equipment, including a buoy and mooring system, are being used to facilitate remote deployment. Images from the two devices will be automatically classified using associated machine learning software. APICS will be deployed at the L4 monitoring site, where it will join a suite of complementary state-of-the-art scientific instruments. The asset will be managed by Plymouth Marine Laboratory (PML). PML operate the WCO, in collaboration with the Marine Biological Association(MBA), using NERC National Capability funding.
Plankton Classification
Zoo Plankton microscope image by Claire Widdicombe
The two camera systems can collect far more data than it is possible to process by hand. To fully exploit their capabilities, it is necessary to automate the classification process. This can be achieved using Machine Learning software. During classification, taxonomic labels are associated with organisms within the images, and organism attributes (e.g., their size and shape) are computed and saved. Within the project, we will use bespoke classification software that ships with the two camera systems. However, we are also actively working on developing our own classification software, with a view to potentially using this with APICS. For an example of our work in this area, see Kerr et al (2020).
Kerr, T., Clark, J. R., Fileman, E. S., Widdicombe, C. E. and Pugeault, N. (2020) Collaborative deep learning models to handle class imbalance in FlowCam plankton imagery. IEEE Access, 8, pp. 170013-170032.
Impact
As part of the WCO, APICS will be added to the set of assets which make up Smart Sound Plymouth. Smart Sound Plymouth is the UK’s marine robotics proving area for developing cutting edge products and services for the commercial marine sector and is led by PML. The Smart Sound Plymouth group operates a network of autonomous assets including buoys, unmanned surface vessels, a fleet of underwater robots and an advanced communications network (Smart Sound Connect). APICS will allow critical plankton data to be autonomously collected at the same temporal resolution as physical and chemical variables, yielding a consistent dataset and world first net-zero autonomous package, which service and product designers can utilise when testing new technologies (e.g.,sensor packages).
APICS itself will be a demonstrator project for the autonomous collection of high frequency biological data in the global coastal ocean, which will be critical for achieving the UN Decade of Ocean Science for Sustainability objective to have an ‘Observed Ocean’ in support of effective marine management.
Timeline
June 2022 - March 2024.
The main aim of the project is to fully deploy APICS within the WCO ready for use in scientific applications.
Project Updates
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Who Is Involved
| James Clark Marine Systems Modeller Plymouth Marine Laboratory |
Claire Widdicombe Plankton Ecologist Plymouth Marine Laboratory |
| Elaine Fileman Plankton Ecologist Plymouth Marine Laboratory |
Dr James Fishwick |
