A 21st century data and machine learning approach to address the risk from harmful algal blooms

Phytoplankton form the base of the marine food web. Through their primary production they are crucial to carbon cycling. However, some species form “harmful algal blooms” (HABs) which have had a severe impact on the finfish and shellfish aquaculture industries in Scotland over the last decades.

The aquaculture industry and associated policy makers require a rapid early warning of the development of HABs and a better understanding of their response to environmental forcing in a changing climate. This will help make informed management decisions such as deploying protective tarpaulins, reducing feeding, moving cages (fish farming), and early or delayed harvest or increased end product testing (shellfish farming) and could help protect human health (typically from HAB generated toxins vectored to humans by shellfish) and minimise mortalities of farmed fish as a consequence of other HAB genera.

A solution to the problem is the Imaging FlowCytobot (IFCB) https://mclanelabs.com/imaging-flowcytobot/. This is an in-situ automated submersible imaging flow cytometer, that also generates images of phytoplankton in-flow. The IFCB can track the progression of phytoplankton cycles or HAB events. Data collection is achieved through a novel combination of flow cytometric laser based and video technology to capture high resolution images of suspended particles. Collected images are processed using automated image classification software.

SAMS and Shetland UHI have been successful in obtaining funding for the 1st two IFCBs in the UK. These instruments will therefore provide a step change in the capability of UK environmental science to generate data to monitor and understand phytoplankton dynamics. This project will provide the capability to classify, display and analyse IFCB data.

The aim of the studentship is to achieve the 1st UK IFCB installation producing high resolution phytoplankton data for the early warning of HABs and interpretation of influence of environmental drivers on HAB developments.

A collaboration between SAMS, the Shetland-UHI, Marine Scotland Science and the University of Edinburgh with funding from Scottish Government and the Data Lab, this project will provide the capability to classify, display and analyse IFCB data. The studentship will address four challenges:

1) Achieving the 1st routine UK IFCB operation
2) IFCB image classification to discriminate and enumerate different phytoplankton species
3) Development of our dashboard https://ifcb-data.sams.ac.uk/ to produce new methods to display HAB data for easy interpretation by stakeholders
4) Application of data analysis approaches to improve risk assessment and ecological understanding

To study the relationship between environmental factors and phytoplankton abundance, we will utilise various modelling approaches including traditional ones like Random Forest to deep learning based approaches such as Long short-term memory (LTSM) that may be more capable of representing the complex factors that promote HAB formation than deterministic approaches.

The studentship requires a candidate that is interested primarily in methods to autonomously collect and process in data aspects of environmental science. Previous experience in high-level programming languages such as Matlab, R or Python is recommended. The analysis also requires the continuous update of an existing image classifier, therefore some knowledge in machine learning and working with Linux systems is desirable. An interest in phytoplankton taxonomy a willingness to get involved in practical work and a solutions mentality to instrument operation are would also be beneficial.

In year 1 the student will receive initial training at SAMS before visiting Shetland to participate in the deployment of the 2nd IFCB. We expect this visit to coincide with a week-long (virtual) IFCB training course from the instrument’s developers McLane Laboratories, this being best undertaken with an instrument to hand. Subsequent year 1 activities will relate to 1) collection of data during the summer phytoplankton growth season, 2) progressing the development of phytoplankton classifier and 3) progressing the development of the on-line data dissemination portal.

Interaction with the supervisory teams at SAMS, Shetland UHI, Marine Scotland Science and the University of Edinburgh will be through regular video-conferencing based meetings and, covid restrictions willing, 6 monthly face to face meetings. IFCBs are being deployed in a number of countries worldwide, the student will also collaborate with the wider IFCB user community on a dedicated GitHub hosted by Woods Hole Oceanographic Institute and through a European IFCB user group.

In years 2 and 3 the student will combing high resolution IFCB phytoplankton data with environmental data and novel deep learning modelling approaches to improve our understanding of the factors driving HABs and hence improve methods to provide early warning forecasts of these events for industry.


For further information

Please contact Professor Keith Davidson for further information (Keith.Davidson@sams.ac.uk)
Closing date for applications: Wednesday 12th January 2022 (see ‘How to Apply’ tab for details)
Interview date: Monday 31st January 2022
Studentship starts: Spring 2022

Research Facilities and Environment

SAMS has excellent facilities for marine research. The student will have access to computing, library and laboratory facilities as required. SAMS hosts approximately 40 postgraduate students who are supported by both SAMS and UHI graduate schools.


To comply with funder requirements the studentship is only available to UK or Irish citizens or applicants with UK settled status.

The 3.5 year studentship covers:

• Tuition fees each year (for 2020/21 this is currently £4,407 for full-time study)
• A maintenance grant each of around £15,000 per annum (for full-time study)
• Funding for research training

Applicants should normally have, or be studying for:

• A postgraduate Master’s degree from a degree-awarding body recognised by the UK government, or equivalent, or
• A first or upper second class honours degree from a degree awarding body recognised by the UK government, or equivalent.

How to Apply for this Project:

The documents you need to apply for the PhD studentship: A 21st century data and machine learning approach to address the risk from harmful algal blooms:

IFCBii_Reference form_21
IFCBii_PhD Studentship Application PART 1 1.0 2021-22

Please ensure you submit the application form – with the following supporting documentation – as one single pdf file – to SAMS Graduate School by email to: phd@sams.ac.uk

A current CV