Microalgae and artificial intelligence come to the rescue of the Baltic Sea

Dead fish, slimy seaweed, and a seabed without oxygen, where nothing can survive. That’s the reality in large parts of the Baltic Sea today.

Oxygen depletion began seriously affecting Danish waters in the 1980s, primarily due to nutrient runoff from agriculture and wastewater. The situation has worsened dramatically over the past 20 years, and today it’s more severe than ever since the early 2000s.

The answer may lie in the microalgae. Floating in their billions and invisible to the naked eye, these tiny organisms know more about the state of the Baltic Sea than any sensor probe.

They respond to even the most minor changes in nutrients, temperature, and currents. They grow explosively when the ecosystem is out of balance — and when they die, they leave oxygen-deprived dead zones behind.

Now, researchers want to use these algae as biological alarm bells in a system that can predict — and help preventing — the Baltic Sea’s collapse.

What if we could listen to what the algae tell us in real time? What if farmers, policymakers, and environmental authorities could see the effects of different interventions on the sea — before spending a single euro or closing a single fish farm?

That’s the vision behind RECOVER, a Danish-German project that will work towards building a digital twin of the southwestern Baltic Sea over the next three years. A virtual ocean powered by data from the sea’s most sensitive early warning system: the algae themselves.

“Microalgae are at the heart of the project because they are indicators of the ecosystem’s health,” says Professor Anja Engel from the GEOMAR Helmholtz Centre for Ocean Research in Kiel, who leads the cross-border project.

“Depending on the nutrient load, the biodiversity of the algal community changes. They tell us exactly what’s wrong — if we learn to listen,” Engel explains.

The ocean’s detectives

Microalgae act as the ocean’s sensors — living instruments that reveal everything from eutrophication to climate change. When there’s too much phosphorus relative to nitrogen, as is the case in much of the Baltic Sea, the composition of the algal community shifts.

“In a system like the Baltic, which is partly anoxic and loses a lot of nitrogen, we end up with a relatively high phosphorus-to-nitrogen ratio,” Engel explains.

“That favours algae that can fix atmospheric nitrogen — cyanobacteria.”

These massive blooms of filamentous cyanobacteria are more than just an eyesore when they wash up on beaches during the summer months. They often arre toxic.

“Sensitive individuals, children, and even pets like dogs can become seriously ill if they touch or drink the water,” says Engel.

That’s why beaches have to close. But before that happens, the algae have already warned us — we need to understand their language better.

That’s precisely what RECOVER aims to do. By mapping algal species, abundance, and composition in real time, researchers can assess the current state of the ecosystem and predict what will happen if conditions change.

“It’s the algal biomass and the composition of the algal community that truly inform us about the ecosystem’s health,” Engel says.

“With RECOVER, we’ll use the information the algae provide as indicators of ecosystem status.”

From biological alarm to digital intelligence

This is where the University of Southern Denmark (SDU) comes in. Researchers at the Mads Clausen Institute in Sønderborg are developing sensors and AI tools to translate the algae’s biological signals into usable data.

“We combine advanced sensor technology with artificial intelligence to create a system that can monitor the sea far more comprehensively than traditional methods,” says Associate Professor Jacek Fiutowski from MCI.

“Instead of costly, time-consuming analyses, we’ll get real-time data that can be directly integrated into the digital twin. It opens up entirely new ways of understanding how the Baltic reacts to changes — and it all starts with listening to the algae.”

At SDU’s Centre for Industrial Software, researchers are investigating tools for building the system’s brain — the AI models and visualisation tools that will make complex oceanographic data accessible to the public.

The project uses high-resolution camera systems that automatically identify and count algal cells and cutting-edge DNA analyses that reveal the complete algal community composition from a single water sample. These data are then integrated with oceanographic models and artificial intelligence.

A pot without a plug

The challenge is not only technical — it’s also political and geographical. As Anja Engel explains, the Baltic Sea behaves like a pot:

“With respect to the key nutrient phosphorus, the Baltic Sea largely retains what’s put in.”

That also means that decisions on nutrient input made in Germany affect conditions in Denmark — and vice versa. The water in Sønderborg today was in Flensburg Fjord yesterday, and in Kiel Fjord the day before. Currents and winds connect the ecosystem across borders.

“That’s why the cross-border aspect is so important,” says Engel.

“For example, when we think about marine protected areas, we can’t just focus on the seabed. We have to understand the spatial and temporal dynamics of the water column — and that requires a broader perspective.”

RECOVER therefore also compares German and Danish approaches to marine management. How do regulatory systems differ? What are farmers and fishers in each country willing to accept? And which strategies actually work?

From data to democracy

But the digital twin the project is aiming for isn’t just a tool for scientists. The project’s most ambitious goal is to make the algae’s warnings — and the complex oceanography — understandable to everyone: from local politicians to fishermen to ordinary citizens swimming at the coast.

“We can simulate the effects of different climate scenarios and test management strategies before they’re implemented,” explains Fiutowski.

“Imagine a municipality considering a new protected area. With our tool, decision-makers can first explore the consequences in a virtual model — how will the ecosystem respond? Will water quality improve? What does it mean for fisheries? It democratises the decision-making process.”

Through workshops, public engagement, and citizen science initiatives — where local fishers, for example, are equipped with sensors to collect algal data — the project seeks to bridge the gap between science and society.

When science shows the way

RECOVER alone can’t save the Baltic Sea. But it can deliver the knowledge and tools that policymakers urgently need.

“RECOVER can’t by itself bring the Baltic back to good ecological status,” says Engel, “but it can generate knowledge and tools  for policymakers and decision-makers to support ecosystem restoration.”

And perhaps most importantly, the project can help finding measures that actually work.

“Politicians implement decisions and must subsequently monitor the system to assess their effectiveness.,” Engel says.

“Current monitoring is often carried out using outdated and time-consuming methods. We now have the technology to move to real-time assessment.”

When RECOVER concludes in 2028, the researchers will deliver concrete tools: affordable sensors for large-scale monitoring. AI models that can visualize algal blooms, and an interactive digital platform where anyone can explore the Baltic Sea’s future.

A digital twin that makes the invisible visible — and the algae’s alarm signals understandable.

And perhaps, the researchers hope, it will mark the beginning of a new era for the Baltic Sea: one where decisions are made not in the dark, but based on simulation, data, and knowledge — and where the algae finally get a voice in the political debate, before it’s too late.