Hovering flight in hummingbird hawkmoths: kinematics, wake dynamics and aerodynamic power

Hovering insects are divided into two categories: 'normal' hoverers that move the wing symmetrically in a horizontal stroke plane, and those with an inclined stroke plane. Normal hoverers have been suggested to support their weight during both downstroke and upstroke, shedding vortex rings each half-stroke. Insects with an inclined stroke plane should, according to theory, produce flight forces only during downstroke, and only generate one set of vortices. The type of hovering is thus linked to the power required to hover. Previous efforts to characterize the wake of hovering insects have used low-resolution experimental techniques or simulated the flow using computational fluid dynamics, and so it remains to be determined whether insect wakes can be represented by any of the suggested models. Here, we used tomographic particle image velocimetry, with a horizontal measurement volume placed below the animals, to show that the wake shed by hovering hawkmoths is best described as a series of bilateral, stacked vortex 'rings'. While the upstroke is aerodynamically active, despite an inclined stroke plane, it produces weaker vortices than the downstroke. In addition, compared with the near wake, the far wake lacks structure and is less concentrated. Both near and far wakes are clearly affected by vortex interactions, suggesting caution is required when interpreting wake topologies. We also estimated induced power (P-ind) from downwash velocities in the wake. Standard models predicted a P-ind more than double that from our wake measurements. Our results thus question some model assumptions and we propose a reevaluation of the model parameters.

Automated summary

This study used tomographic particle image velocimetry to show that the wake shed by hovering hawkmoths is best described as a series of bilateral, stacked vortex 'rings'. Despite aerodynamically active upstroke with an inclined stroke plane, it produces weaker vortices than the downstroke, and the far wake lacks structure and is less concentrated compared with the near wake. The wake topologies of both near and far wakes are clearly affected by vortex interactions, suggesting caution is required when interpreting wake structures.