Decoding Butterfly Flight with High-Speed Cameras and a Wind Tunnel
Fluid Dynamics Help Show Why Butterflies Flutter
On a hot summer day, Lund University biologists Christoffer Johansson and Per Henningsson waded through the grass and wildflowers of a meadow near the school’s field station in Sweden. They moved quietly and deliberately toward the brief flutter of a cantaloupe-colored insect, ready to collect it using a soft fabric net.
The capture of silver-washed fritillary butterflies is far more than just a summer hobby for Johansson and Henningsson. They use slow-motion cameras and high-speed flow measurements to determine what gives these butterflies their distinct flight pattern. This discovery not only helps scientists better understand the life of this insect but could also help inform the design of next-generation drones.
Determining what gives these butterflies their distinct flight pattern could also help inform the design of next-generation drones.
It’s a job that isn’t always easy, says Johansson, recalling the brutal heat during meadow hunting or the frantic search to recapture specimens within their experimental wind tunnel, but it’s always interesting.
“We normally spend most of the time in the lab,” says Johansson. “In this case, we actually spent a few days out in the meadows, capturing butterflies… But I would say the most fun and frustrating part is during the actual experiments in the wind tunnel. Even though we place the butterfly in a specific place, once it takes off, it can quickly move around the tunnel and disappear, only to reappear atop of the head of the researchers.”
A Monarch Mystery
Butterflies can fly not only in short bursts from flower to flower, but also in sustained migratory flights over long distances. The monarch’s migration, for example, from the U.S. to Mexico is more than 4,800 km (3,000 miles) one way.
And that’s not the only thing that makes butterfly flight unusual, says Johansson. They also have a body size–to-wing ratio unlike that of other insects, with unusually large wings for their small body size.
Johansson and his colleagues hope to provide new data to help answer the question of butterfly flight mechanics, which researchers have struggled to understand for nearly 50 years.
“Butterflies are extreme compared with other flying animals,” says Johansson. “They have a very low wing loading and a low aspect ratio of the wings, which means that the wings are essentially large, short, and broad compared with other flying animals.”
In the 1970s, scientists postulated that butterflies might be achieving this multipurpose flight by creating a jet of air when clapping their wings together at the top of the upstroke to propel them forward. However, researchers have struggled to confirm this mechanism for nearly 50 years because quantifying free-flying activity, as opposed to tethered flight under laboratory conditions, is easier said than done. In their research, Johansson and Henningsson provide new data to help answer the question of butterfly flight mechanics and understand how it could be used in other applications.
“It’s often difficult to imagine what basic research is going to be used for in the end, but in this case, there are direct applications in drones,” said Johansson. “There are drones flying today that use a clap mechanism to generate forces. It could be interesting for them to look into this proposed butterfly mechanism to improve drone wing flexibility and maximize the efficiency and force of the clapping.”
Capturing Butterfly Flight
The research team introduced six silver-washed fritillary butterflies into a recirculating wind tunnel at Lund University, using honey water feeders to entice the insects to take off. This wind tunnel is unique. It was initially constructed to study bird flight. A large fan in the tunnel circulates air at two meters per second (4.5 mph) to keep the butterflies from too easily flying away from the team’s measurement setup.
The team uses four high-speed cameras to record the air motion caused by the butterflies and another two cameras to capture the butterflies’ motion. The flow measurement technique, called tomographic particle image velocimetry, enables the researchers to create a 3D model of fluid flow; the team can then study the aerodynamic forces to understand how butterfly wings propel their flight. In this research, the focus was the jets of wind created when the butterflies clap their wings.
Tiny aerosol particles, only about 1 micrometer in size, are suspended in the wind tunnel. The high-speed cameras capture the motion of these particles as the butterflies fly through them in front of a sheet illuminated by laser light. In total, the team captured 25 sequences with one to three wingbeats each.
Johansson says the team used MATLAB® for data analysis, including using a user interface he designed for vector analysis to study the fluid dynamics of butterfly flight.
“Our field of research is technically and numerically demanding, and there is no single software that can do all the analysis in the way we want,” said Johansson. “Hence, we need to create most of the code ourselves.”
“MATLAB was used to visualize the flow and plot the results from the experiments,” said Sagar Zade, an education customer success engineer at MathWorks.
Johansson and Henningsson also used MATLAB to calculate aerodynamic forces, determine the aerodynamic power of the clapper, and estimate the background power using Monte Carlo simulations.
“The results from these experimental findings are extremely valuable for future researchers who would like to use MATLAB to numerically model fluid dynamics using complex Navier Stokes equations,” Zade added.
In addition to measuring the jet streams created by the butterflies’ flight, the team analyzed how the morphology of their wings worked to create a cupped shape on their upward strokes to improve these jets. They used this data to create mechanical butterfly wings to isolate the effect of the flexible wings and the cupped shape on the clapping performance.
Johansson says these mechanical wings are simple, right-angled triangles made of balsa wood and a latex membrane controlled by a servo engine and an Arduino® board. A hinge along the side of the wings enables them to rotate and clap together. The flexible membrane forms a cupped shape during the clap while the stiff balsa wood does not, allowing for comparisons of performance and investigation of the effect of the cupped shape alone. Between the biologists’ work in the tunnel and the work in the lab with the mechanical wings, Johansson says they gained some key insights into the dynamics of butterfly flight and how it could be recreated. One discovery, he says, was that a flexible membrane wing outperformed the balsa wing in impulse and efficiency by 25% during the clap. They also determined that a butterfly’s upstroke and downstroke serve two unique purposes in flight.
“The downstroke produces vertical force, while the upstroke and clap produce thrust,” says Johansson. “As with most flight, the vertical force dominates. In this particular case, vertical force is 9.4 times the thrust.”
The team’s full research was published in the Journal of the Royal Society Interface.
Because Johansson and Henningsson do not have formal training in MATLAB, Johansson says the accessibility of the software, its flexibility, and its ease of use were huge assets for this work.
“MATLAB was, by far, our most useful tool on this project,” says Johansson.
After the success of using this approach to capture butterfly flight, Johansson is interested in seeing how it applies more generally to model flight in other creatures, ranging from birds and bats down to microscopic organisms.
Nature has been perfecting flight much longer than any human. By studying the myriad of ways these creatures achieve this, engineers can build more efficient and dynamic flying—or even swimming—drones. These drones may one day deliver groceries to your front door or plumb the depths to study marine life.
“It’s been suggested that butterflies use every trick in the book to fly,” said Johansson. “There’s a lot of research to be done to figure out if that’s true and how different conditions affect those mechanisms.”