How dangerous are offshore wind turbines to birds? The answer has eluded researchers for decades. But new studies in the Netherlands and the UK use the latest technology—and the results surprise even the experts.

By Lars-Magnus Kihlström

Offshore wind power is one of the most promising ways to produce the fossil-free electricity needed for the energy transition, and in the North Sea area, there are major plans for expansion. At the same time, there are growing concerns about how offshore wind turbines affect bird populations. While the visual impact of towering turbines is obvious, the invisible toll on birds flying through offshore wind farms has remained largely speculative. Now, with AI and other advanced technologies, an answer could be within reach.

“We’ve been working with offshore wind for 20 years and doing bird studies for almost as long,” says Jesper Kyed Larsen, Bioscience Expert at Vattenfall. “But we haven’t really been able to approach the collision question with the kind of evidence we need. We’ve been left to speculate.”

No trace in the sea

The problem is as practical as it is scientific. Birds that collide with turbine blades at sea fall into the water and drift away, leaving no trace. This has made it nearly impossible to quantify the true impact of wind farms on avian life. As a result, environmental assessments have relied on educated guesses.

A great number of studies have been conducted, and a lot of knowledge has been collected, suggesting that the number of collisions might actually be lower than some fear. At Vattenfall’s Aberdeen Bay wind farm, a study using radar and cameras on the turbines detected no collisions during the two-year study period.

“There are quite a few studies about how birds fly in wind farms, and it’s quite obvious that they are very sensitive and, in general, really good at adapting their flight paths to fly around the turbines. The study in Aberdeen was promising, but not actually designed to detect collisions. We are now looking into refining our methods in that regard,” says Jesper Kyed Larsen.

Bird flight in three dimensions

A first step on this path was a tech test conducted in Aberdeen Bay together with Spoor, a Norwegian company that has developed ground-breaking AI technology to track a bird’s flight path in 3D in the immediate vicinity of turbine blades. This could allow wind farm operators to document ‘micro-avoidance behaviour’—the last-minute flight adjustments birds make to avoid colliding with the rotor blades.

The test to prove the technology and concept was concluded earlier this spring with the help of bird experts from the British Trust for Ornithology. In the next phase, Spoor’s experts will teach the AI system to analyse and identify actual collisions.

“It’s a painstaking process,” Larsen admits. “Training the algorithms to pick out deviating bird tracks from the many thousands recorded, and checking that it didn’t miss any of them, takes hundreds of hours of bird expert screen time. Human input and validation is key to making AI trustworthy.”

Experts from the British Trust for Ornithology helped testing the AI-camera concept at the Abrerdeen Bay windfarm

Millions of birds migrate 

The Aberdeen Bay studies focus on seagulls, gannets, and other seabirds that live and feed among the wind farms. In the Dutch wind farm Hollandse Kust Zuid, Vattenfall is conducting another kind of study on migrating birds such as starlings and finches. Twice a year, millions of birds migrate across the North Sea or along the coast, and there is widespread concern about what the cumulative growth of offshore wind farms along the North Sea coast will mean for these species.

“Seabirds are active in daytime, but bird migration, which happens during a few weeks in spring and autumn, takes place mainly at night. Therefore, we use AI together with thermal cameras in this case. Basically, we want to document how many birds collide and also understand the relationship between collisions and weather, migration intensity, flight altitude, etc., to decide on the best way to mitigate the risks. One way of reducing these risks is to simply stop the turbines during the migration period, but then we want to know that it has a real effect. Otherwise, we waste valuable fossil-free electricity for nothing.”

Surprising results

The thermal cameras are very sensitive and detect everything in the air—from insects to airplanes. For this study, 16 such cameras are placed on one of the turbines in the wind farm. Pointing in all horizontal directions, the cameras are mounted on the turbine tower below the rotor to detect anything that falls into the water.

This spring, the first migration period was studied. The preliminary results were surprising also to Jesper Kyed Larsen: during the peak period in March, only a couple of collisions were detected—and neither occurred at night, when birds migrate in the thousands.

“At this point in time, we need to be careful about drawing strong conclusions. Further analysis of the data will take place, and caution is needed given the study only covers one turbine. Still, if it holds up under closer scrutiny, this is very positive. And it wasn’t like the cameras didn’t pick up the migrating birds—they did, in large numbers. Just no one falling into the sea. Now we’re excited to see what we’ll find during the autumn migration, which is much bigger in terms of bird numbers.” 

Another study on seabirds is being done together with Oxford University. Here the researchers knowledge of bird vision is combined with realistic computer models of operating turbines to simulate how different combinations of colours and patterns on the blades might reduce collision risk. By directing birds’ attention to the moving blades, increasing their chances of taking aversive action in time to avoid a potentially fatal collision. The results are not yet ready, but blades with black and red striped patterns show good promise.

A culture of openness 

Jesper Kyed Larsen emphasizes the importance of transparency in this work.

“We want to be open and public about what we’re doing,” he says. “There’s a lot of concern out there, and rightly so. But now we have tools that can move us from speculation to science.”