Competitive angling as a scientific tool
Figure 2. Recreational anglers releasing a smoothhound back into the water.
It is a glorious June day on England’s south coast and the sun is shining onto the still waters of Portsmouth harbour. The marina is buzzing with activity as anglers prepare their hooks and lines, sort their bait, and make their plans for the next 2 days. Among them is a group of researchers, ensuring each boat is equipped with a fish-measuring board, GPS tracker, and GoPro (see Fig. 1). This collaboration between anglers and researchers began in 2021 when Ross Honey, owner of Angling Spirit and organizer of prestigious angling competitions across Europe, approached the University of Portsmouth with an offer to share the vast quantities of data being collected by the anglers. Ross could see the immense value in the hundreds of fish records being collected over the annual 2-day ‘Sea Angling Classic’ competition and wanted to see these records put to good use.
This collaboration between the University of Portsmouth, Angling Spirit, and Southern Inshore Fisheries Conservation Authority (IFCA) became the research project named Competitive Angling as a Scientific Tool, or CAST for short. Funded by DEFRA under the Fisheries Industry Science Partnerships scheme, CAST aims to increase our knowledge on the biology and ecology of data-deficient fish species in the Solent and use this to inform the development and improvement of sustainable fishery management plans. Collecting fisheries ecological data requires a significant investment of time and money, so utilizing recreational anglers as citizen scientists opens up a vital data source that would otherwise be unattainable.
The Sea Angling Classic, which provides the data for the CAST project, is a strictly catch-photograph-release competition. This ensures that all fish go back into the water (see Fig. 2). The photographing in each competition generates large numbers of images that we can use for our research. The competition has been running for 3 years, plus a launch event in 2021, and has generated data on over 2,500 fish so far. The competition targets five species/groups that are considered data deficient within the UK: European seabass (Dicentrarchus labrax), black sea bream (Spondyliosoma cantharus), skates/ rays (Raja spp.), smoothhound (Mustelus spp.) and tope (Galeorhinus galeus).
The competition attracts competitors from across the UK, with a prize of a £100,000 boat for the winning team. With such a large prize at stake, verifying anglers’ catches is crucial. The anglers must submit a geo-referenced photo of each fish on a competition measuring board to enable it to be counted. These photos are submitted through an app, enabling the competition HQ to immediately verify the species and length of fish as the images are submitted. From these photos, the CAST research team can collect data on the species, size, catch location, and the presence of disease or external parasites.
Aside from the biology and ecology research, another aspect of the CAST project is to develop machine-learning models (a type of Artificial Intelligence) that can be used for fisheries research. We are using a computer vision model to detect and remove the image background and obstructions in the images submitted by anglers (such as hands on the fish, see below), then identify and measure the fish. The goal is for the models to be able to accurately identify and estimate the length of fish without the need for a measuring board or scale bar in the image. Although the models haven’t quite reached the necessary level, the more images we can use, the better the accuracy.
If any readers have images of our target species on measuring tapes or boards that they would be happy to share, then please get in touch via the project website (see below).
Figure 3. Demonstration of the AI model’s capability to detect the object of interest (smoothhound) and potential obstructions (anglers’ hands and ID sticker).
In addition to the photo data, each competition boat also has a GPS tracker and GoPro. The trackers allow us to identify the areas where boats are fishing and estimate the length of time they spend fishing there. The advantage of GPS trackers is that they show areas where boats have been fishing but didn’t catch anything, whereas the georeferenced images only show locations where fish were caught. This summer we introduced GoPros to record the fishing activity on deck as a more accurate method of recording fishing effort, as we can review footage to record the exact length of time spent fishing and how many fishing rods were used. Together, these datasets will allow us to calculate catch per unit effort (CPUE): the number of fish caught per fishing rod, per hour at each fishing location.
Citizen science is a large part of the CAST project, but we are also collecting additional data to further expand our knowledge on the biology and ecology of the target species. For example, we are collecting DNA swabs of smoothhounds to determine whether both starry and common species are present in the Solent, and we will assess habitat associations and prey availability at a range of fishing sites to assess habitat preferences of our target species.
The overall aim of the CAST project is to create a standardized data collection workflow that will be repeated at the Sea Angling Classic in future years and could also be used at other angling competitions in the UK and abroad. This would include the collection of images submitted by anglers, analysis of GPS tracker data, and automatic identification, classification, and estimation of fish length from images by the AI model. Our data collection workflow provides a cost-effective way of collecting long-term data on local fish stocks, thereby overcoming fisheries research budget limitations. We have already collected fish data from the equivalent of 350 boat-days, something that would be economically and logistically unattainable with traditional research projects. The long-term datasets generated through our data collection workflow will be essential for the development and continuous improvement of sustainable fisheries management plans.
• Dr Christina Hunt (christina.hunt@port.ac.uk), Senior Research Associate, University of Portsmouth.
Georgina Banfield, Dr Obinna Umeh, Dr Ian Hendy, Professor Gordon Watson.
All co-authors at University of Portsmouth.
For more information, please visit our website: https://castproject.co.uk/