Airplane tracking documents the fastest flight speeds recorded for bats

Abstract

The performance capabilities of flying animals reflect the interplay of biomechanical and physiological constraints and evolutionary innovation. Of the two extant groups of vertebrates that are capable of powered flight, birds are thought to fly more efficiently and faster than bats. However, fast-flying bat species that are adapted for flight in open airspace are similar in wing shape and appear to be similar in flight dynamics to fast-flying birds that exploit the same aerial niche. Here, we investigate flight behaviour in seven free-flying Brazilian free-tailed bats (Tadarida brasiliensis) and report that the maximum ground speeds achieved exceed speeds previously documented for any bat. Regional wind modelling indicates that bats adjusted flight speeds in response to winds by flying more slowly as wind support increased and flying faster when confronted with crosswinds, as demonstrated for insects, birds and other bats. Increased frequency of pauses in wing beats at faster speeds suggests that flap-gliding assists the bats' rapid flight. Our results suggest that flight performance in bats has been underappreciated and that functional differences in the flight abilities of birds and bats require re-evaluation.

Keywords: airplane tracking; bats; flight performance; ground speed; wind modelling.

Figure 1.

Flight trajectories and ground speeds. (a) Consecutive radio-tracking fix locations are indicated by circles. Detailed metrics were calculated for the trajectories based on the distances and times between fixes (table 1). Colours and individual identifications are consistent across all figures. Background colours and contour lines indicate topography in metres ASL. (b) Colour-coded ground speeds with arrows plotted on the trajectories to illustrate direction and speed measured for each segment. High speeds were distributed throughout the bats' flight trajectories.

Figure 2.

Flight speed distributions for bats. (a) Frequency distribution for the measured ground speeds of all bats and their kernel densities for each individual in metres per second. (b) Frequency distribution of air speeds for all bats and their kernel densities for each individual in metres per second.

Figure 3.

Response of bats to winds. (a) Modelled air speed (metres per second) isoclines in reference to wind support and cross winds based on wind data at 30 m AGL. Positive wind support values are tail wind and negative values are head wind. Cross winds are the length of the wind vector perpendicular to the bats' direction of travel. (see electronic supplementary material, table S1, for additional information). (b) Ground speeds and the azimuth of flight directions for seven bats. Prevailing winds on all nights were E to SE (90° to 135°). Faster ground speeds were not associated with prevailing winds.

Figure 4.

Ground speed versus wingbeat pauses. Wingbeat pauses increase with increasing ground speeds (both measurements shown on logarithmic scales), suggesting that bats are more likely to skip wingbeats and glide at very fast speeds. The linear relationship is based on a sample size of 25 observations (five sections of the flight of five individuals). The data were transformed to obtain normal Gaussian error distribution as follows: log(wingbeat pauses + 1) and log(ground speed). The model (F_1,23 = 23.1, p < 0.0001, adj R² = 0.5) showed a highly significant positive linear relationship between wingbeat pauses and ground speed as indicated in the figure (log(ground speed) effect size ± s.e. = 0.87 ± 0.18, t = 4.8, p < 0.0001). The model intercept was non-significant.