Schools of fish have shown engineers how to squeeze much more power from wind farms.
A new wind farm design mimics a school of fish to exploit wind turbulence, and could dramatically improve power output.
Familiar propeller-style wind turbines with large sweeping blades have almost reached their limit of efficiency.
But in a wind farm, they must be spaced widely apart to avoid turbulence from the other turbines.
This has limited wind farm power output to around two watts per square metre of land at favourable sites.
But redesigned wind farms could perhaps get up to 10 times more power from the same land.
A test array in the California desert takes a whole new approach to the problem, according to a study published in the Journal of Renewable and Sustainable Energy.
This new study uses "vertical axis" wind turbines that resemble upright, spinning egg whisks. Although they are less efficient individually than the propeller-style turbines, they are able to use turbulent winds from many directions.
But the big step forward comes from the layout of the array which is based on fluid dynamics around schooling fish.
"Organising the arrangement of wind turbines based upon the vortices shed by schooling fish is definitely a new approach," said aeronautical engineer Robert Whittlesey of the California Institute of Technology (Caltech).
"The fish aim to align themselves to optimise their forward propulsion," he writes, and this can be adapted in a turbine array to maximise energy extraction.
The new design uses closely-spaced pairs of counter-rotating turbines that funnel air to their neighbours, with little energy lost to turbulence.
Not only do the neighbours benefit, but the funnelling effect is also important. In fact power generated by the paired turbines can actually be greater than that from the turbines working independently. In tests, a turbine five rows back still generated 95% of the power of the one on the front row.
A wind farm of this closely-packed design could produce 20 to 30 watts per square metre of land, around 10 times that of current wind farms.
Author of the study, Professor John Dabiri of Caltech, said: "While the connection between fish schooling and wind farms might seem non-intuitive at first, it is in fact a logical inference from the underlying flow physics."
The advantages don't stop there. At 10m high, the turbines used in this study were only around one tenth of the height of typical propeller-style turbines.
This means that they are less intrusive in the landscape, less visible to air-traffic control radar and could be less harmful to birds and bats.
The vertical-axis turbines are also "significantly more robust and probably less expensive. There are still some problems to be solved but they really deserve a second look" added Professor Charles Meneveau of Johns Hopkins University who was not involved in the study.
The big question now is whether this design works as a full sized wind farm. To work on this scale, energy from wind passing above the farm must be transferred to the turbines below by turbulence.
"It's a very interesting idea but this hasn't yet been shown," said Professor Loughhead of the UK Energy Research Centre. "Also, vertical-axis turbines face a lot of stress. It's difficult to make a tall turbine light enough to spin but rigid enough to stand up to the forces and vibrations that they're exposed to," he added.
"In this research field, the work seems to be met with great interest and a bit of healthy scepticism," observed Mr Whittlesey.
Further tests look promising though. "We have collected additional wind measurements this summer on an array of 18 turbines...The results suggest that the wind flow rates required for enhanced performance relative to horizontal-axis [propeller-style] wind turbines are regularly attained," said Professor Dabiri.