Protecting the UK's marine ecosystem

To effectively manage marine ecosystems, its essential to understand how fishing and other activities may impact interconnected fish populations. However, in a vast, dynamic system like the ocean, uncertainty is inherent, which is why we must manage it cautiously in order to achieve a sustainable balance between human needs and environmental health.

Mechanistic computer models, also known as simulators, are built to imitate a real-world system, like a fish population, by including the actual rules of nature. Their main purpose is to explain why a system acts the way it does, not just to predict what it will do. Scientists widely use these models to simulate marine ecosystems, including their physical, chemical, and biological processes. This helps them to better understand the state of the marine environment, including where marine mammals will travel, how many fish will be in the sea, and the specific impact of climate change.

While helpful, the predictions of these single marine ecosystem simulators aren’t completely accurate because they are simplified representations of an incredibly complex system. The ocean is influenced by countless variables and interactions, which a single model can't fully capture. 

That’s why Professor Paul Blackwell from the University of Sheffield’s School of Mathematical and Physical Sciences and Dr Mike Spence, former Postdoctoral researcher at the University of Sheffield, have developed statistical methodology for combining many previous ecosystem models. This multi-model approach is important because it improves the accuracy and reliability of predictions about marine ecosystems, which helps to reduce the uncertainty that's in any single model. It works by integrating the predictions of various individual models to produce a single, more accurate and robust outcome for different management or conservation scenarios, such as changing fish species catch quotas. 

Making fish population predictions more accurate

Marine scientists currently use various computer models to predict the future of fish populations, including the fish we catch as well as their predators and prey. Some of these models are limited to a single species, while others attempt to simulate the interactions of many species. There are also research groups who attempt to predict the interactions of many types of species, including seabirds or marine mammals.

The core challenge with this approach is that the different models, developed by different research teams, often produce contrasting answers and predictions. Funded by a joint grant from NERC and DERA, Professor Blackwell and Dr Spence, however, focused on developing a statistical method to combine these different models, creating a single, more reliable prediction.

“One of the main aims of the research project was to determine how long it would take for commercially fished populations in the North Sea to recover if fishing was significantly reduced. So, that work was mainly done by myself and by Dr Spence, who was a Postdoctoral Researcher at the University of Sheffield on that project. The other big contributor was Professor Julia Blanchard, a former lecturer in Biosciences at the University of Sheffield, who now works at the University of Tasmania” explains Professor Blackwell.

“We developed and published a methodology for combining ecosystems, which is also implemented with CEFAS as the software tool, EcoEnsemble. The main result was that the real uncertainty doesn't come from the details of the individual models or from how well you can estimate the parameters of a particular model, but it comes from the differences between models” adds Professor Blackwell.

“So, the idea was to provide a method for combining models and predictions. But, it also serves as a feedback mechanism for the scientists who created the individual models, allowing them to see where their model's predictions differ from the collective consensus. This doesn't reveal which model is right or wrong, but it highlights areas that require more attention and further investigation” explains Professor Blackwell.

“Aside from purely an academic point of view, a big motivation for the project was the hope that the research continues to lead to informed decision-making about things related to conservation and sustainability” adds Professor Blackwell.

Protecting the UK's marine ecosystem

Since 2018, the Centre for Environment, Fisheries and Aquaculture Science (CEFAS, an executive agency of the United Kingdom government Department for Environment, Food and Rural Affairs) have used the multi-model ensemble methods, developed by the Sheffield researchers, to estimate populations of various fish species in UK waters. Their use is led by Dr Spence, formerly of the University of Sheffield and now employed by CEFAS.

“Once we’d developed and published our modelling approach, Dr Spence went to a job at CEFAS which is the UK fisheries body and they're responsible for the scientific side of predicting fish stocks, setting fish quotas and deciding how commercial fishing is going to be managed for the UK. It’s worked out really well because he and I are still working together, but he's very much embedded in the practical side of things. Our role is all about trying to make sure that their modelling approaches are as up-to-date as possible” explains Professor Blackwell.

In 2024, a policy to ban the fishing of sandeels was implemented to protect the UK's marine ecosystem, particularly its seabird populations. It specifically targets the industrial fishing of sandeels in the English waters of the North Sea and all Scottish waters.

Predictions from the multi-model approach developed by Professor Blackwell and Dr Spence were a key part of the evidence used to ban sandeel fishing. The models forecasted the positive effects of a fishing ban on both sandeel stocks and seabird populations, providing a scientific basis for the policy.

The UK recently banned the fishing of sandeels. The reason for that was because they're a really important food source for puffins, kittiwakes and other seabirds. So, the idea was to ban those in order to protect seabird populations. From a conservation point of view, the policy has been really well received.

Professor Paul Blackwell

 School of Mathematical and Physical Sciences

Expanding the research

Professor Blackwell and Dr Spence are now working on a paper to make their modelling approach more accessible to a wider audience. They are also looking to extend their methods to deal with a wider range of underlying models.

“We’re looking to develop models with greater spatial detail, moving beyond just looking at particular populations in, say, the North Sea, to creating a map showing the abundance of marine life across an entire region. While this presents an interesting mathematical challenge, the project is primarily driven by policy questions from CEFAS, who want to make predictions at this more detailed level” adds Professor Blackwell.