The Penguin Test

916 tracked penguins on the globe and three flat maps

Data. 916 penguins of 5 species tracked from 37 colonies around Antarctica and the sub-Antarctic islands, 1991–2013: 294,917 positions from satellite tags and GPS loggers, compiled by SCAR's Retrospective Analysis of Antarctic Tracking Data. The first selection is the emperor colony at Pointe Géologie, Terre Adélie, where March of the Penguins was filmed, tracked in 2005.

Test. Each track is drawn to scale on every map and measured there. In 3 days a penguin can travel no farther than 2.3 m/s of swimming, the fastest mean speed measured in any penguin species, plus 1.0 m/s of current, about twice the strongest jets of the Antarctic Circumpolar Current, with no time spent diving or resting, plus 50 km for position error: 11.9 km/h. A track turns red on a map where some 3-day stretch needs more.

Positions. The positions are those computed by the tags' location systems, Argos and GPS, which work on a globe. The test asks whether the same positions, drawn to scale on each map, describe journeys a penguin can make.

——
View

Globe

Sphere, radius 6,371 km

Track length, median
—
Beyond the limit
—
Fastest 3-day stretch
—

North-pole AE

Centered on the North Pole

Track length, median
—
Beyond the limit
—
Fastest 3-day stretch
—

Dual-pole AE

Centered on 0°N 0°E

Track length, median
—
Beyond the limit
—
Fastest 3-day stretch
—

Mercator

True on the equator, edges joined

Track length, median
—
Beyond the limit
—
Fastest 3-day stretch
—

Selected penguins

Result

Of the 784 tracks long enough to contain a 3-day stretch, none needs more than 11.9 km/h on the globe. On the North-pole AE map 213 (27%) do, on Mercator 16 and on the dual-pole AE map 1. At Pointe Géologie, 6 of the 11 emperors tagged in 2005 need more than the limit on the North-pole map and 0 on the globe.

The North-pole map keeps north–south distances true and stretches east–west ones more the farther they lie from its center: by a factor of about 7 at the latitude of the Antarctic coast. A median track is 3.2 times as long there as on the globe. Mercator stretches distances by the secant of the latitude, 2.5 at 66°S. The dual-pole map, centered on the equator, stretches the far south by up to about 2 in one direction, and these tracks alone do not rule it out. The flight times in The Best Flat Map rule out every flat map, this one included.

Method

Data
SCAR Retrospective Analysis of Antarctic Tracking Data, standardized tracking data version 1.5: penguin species only (916 tracks: 691 PTT, 194 GPS, 31 GLS). Each track is one tagged bird, from the capture record to its last position.
Positions
As published. Argos positions are computed from the Doppler shift of the tag's transmissions as recorded by polar-orbiting satellites; GPS positions from satellite ranging; both are expressed as latitude and longitude on the WGS 84 ellipsoid. No position is filtered on speed or on any map.
Daily positions
One per UTC day: the median time, latitude and longitude of that day's fixes (26,752 daily positions in all), to damp the error of single Argos fixes. Longitudes are unwrapped around the day's first fix before the median.
Lengths
On each map, the sum of straight steps between consecutive daily positions: great-circle steps on the globe, straight lines across the plane on the flat maps.
Stretches
Every pair of daily positions 3 days apart, with the elapsed time taken from their median times (at least 48 h). The speed a stretch needs is its distance on the map, less 50 km for position error at its two ends, divided by the elapsed time. A track is beyond the limit on a map if any of its stretches needs more than the limit there; tracks shorter than 3 days are drawn but not tested.
Limit
11.9 km/h: 2.3 m/s, the highest mean swimming speed measured with speed loggers in any penguin species (gentoo and chinstrap penguins; emperor 1.7, Adélie 2.0, king 2.1 m/s), plus 1.0 m/s of current flowing with the bird for the whole stretch. Surface currents in the Antarctic Circumpolar Current are typically about 0.1 m/s and reach 0.5–0.6 m/s in its frontal jets. Each term favors the flat maps; the bird is also assumed to swim without diving, resting or walking. The slider above changes the limit.
Maps
As in the Cable Latency Survey and The Sync Test: the North-pole and dual-pole maps are azimuthal equidistant, true in distance from their centers; Mercator is true on the equator, with its east and west edges joined so that a track can cross 180°. Lengths are map kilometers at that scale. Tracks fits each map to the selected tracks; the scale bar gives map kilometers.

Sources

  1. Ropert-Coudert Y, Van de Putte AP, Reisinger RR, et al. (2020). The retrospective analysis of Antarctic tracking data project. Scientific Data 7, 94. doi:10.1038/s41597-020-0406-x. Data: The Retrospective Analysis of Antarctic Tracking (Standardised) Data from the Scientific Committee on Antarctic Research, v1.5, SCAR AntOBIS, doi:10.15468/uzmdpm, CC BY 4.0.
  2. Sato K, Watanuki Y, Takahashi A, et al. (2007). Stroke frequency, but not swimming speed, is related to body size in free-ranging seabirds, pinnipeds and cetaceans. Proceedings of the Royal Society B 274, 471–477. doi:10.1098/rspb.2006.0005.
  3. Antarctic Circumpolar Current speeds: AMS Glossary of Meteorology; Stewart RH, Introduction to Physical Oceanography, §13.5.
  4. Natural Earth land, 1:50m south of 35°S and 1:110m elsewhere · D3.