Why Flight Paths Look Curved on Flat Maps

A traveller on a Los Angeles → Tokyo flight glances at the seat- back map and sees the aircraft routing far north — over Alaska, past the Aleutian Islands, then south across the Pacific to Japan. The path looks needlessly long. Why isn't the plane flying due west, along the Tropic of Cancer, where the route looks shorter on the map?

The answer is geometry. The seat-back map is a flat projection of a curved Earth. The route over Alaska is the shortest path — called a great-circle route — but the Mercator projection on the seat-back display stretches the distances at high latitudes, making the actual shortest path appear longer. The path that looks shortest on the flat map (due west along the Tropic of Cancer) is in fact the longer route, called a rhumb line (when shipped).

This article walks through the geometry, the worked examples for the most-cited routes, and the secondary factors (jet streams, ATC, ETOPS rules) that bend real flight paths slightly off the geometric optimum.

The geometry

Take a globe (a physical one if you have one). Stretch a string between two points on it — say, Los Angeles and Tokyo. The string lies along the unique great circle through those two points, and its length is the shortest possible distance over the surface.

Now imagine peeling that globe and flattening it into a Mercator projection. The high-latitude regions stretch horizontally (a ~30° band at 60° N becomes ~2× as wide as the same angular band at the equator), but the path of the string doesn't move on the underlying Earth — only its representation on the map shifts. The result: the string's path, which was straight on the globe, appears as a curve bowed toward the nearer pole on the flat map.

The visual rule of thumb on a north-up Mercator map:

• Two points on the same meridian → great-circle path is straight (the meridian itself).
• Two points on the equator → great-circle path is straight (the equator).
• Two points in the same hemisphere at moderate latitudes → great- circle path bows toward the nearer pole.
• Two points on opposite hemispheres → great-circle path crosses the equator at some intermediate longitude, with both hemispheres contributing curvature.

The curvature is the projection's fault, not the path's.

A worked example: LAX → HND (Tokyo)

Los Angeles (LAX) at approximately 33.94°N, 118.41°W. Tokyo Haneda (HND) at approximately 35.55°N, 139.78°E.

The straight-line longitude distance is roughly 102° (going westward) or 258° (going eastward). The westward direction is shorter on the map — and the rhumb-line distance is about 8,820 km along the 34th parallel. But the great-circle path goes north — way north. It crosses through about 52°N at its highest latitude, passing over the Aleutian Islands. The great-circle distance is about 8,790 km — only 30 km shorter than the rhumb line, but tracing the actual flight path that 747s and 787s fly.

Why such a small saving on a 9,000 km route? Because the geometry: LA and Tokyo are at similar latitudes, so the great- circle path doesn't deviate dramatically. The savings are larger for paths between very different latitudes.

A bigger example: JFK → HKG (New York → Hong Kong)

New York (JFK) at 40.64°N, 73.78°W. Hong Kong (HKG) at 22.31°N, 113.91°E.

Rhumb-line distance (constant bearing roughly 286° west of north): about 16,200 km. Great-circle distance: about 12,990 km — a saving of about 3,200 km (~20 % shorter). The great-circle path passes over Greenland, Siberia, Mongolia, and northern China, far from the rhumb-line track. Modern transpacific routes between the US east coast and East Asia all take this northern path because the saving is enormous.

The flight time saving at 900 km/h cruise speed is about 3 hours 30 minutes — and the fuel saving is proportional. This is why the geometric difference between great-circle and rhumb-line paths matters: it's not just academic. The 20 % shortest- path effect is what makes the route economically viable in the first place.

What real flight paths look like

A real great-circle path between two cities is the target. Actual flight paths differ slightly because of:

• Jet streams. The jet stream is a ~400 km/h tailwind flowing west-to-east at altitude. Eastbound flights ride it (saving fuel); westbound flights deviate to minimise headwind. Adjusting the route ±200 km off the geometric optimum can save enough fuel to justify the slightly longer path.
• Air Traffic Control airways. Airspace is organised into defined airways (like highways in the sky), and aircraft fly along them rather than arbitrary great-circle bearings. The airways approximate the great-circle path but include rectilinear deviations at navigation waypoints.
• ETOPS rules. Twin-engine aircraft (most modern long-haul jets) must remain within a certain flight time of an emergency diversion airport. ETOPS-180 means 180 minutes; ETOPS-330 means 330. Over remote stretches of ocean and ice, the rule constrains the route to keep one diversion airport always within reach. This historically forced flights to deviate from great-circle paths over the Pacific and Atlantic; modern ETOPS-330 has largely removed the constraint.
• Airspace restrictions. Military exclusion zones, political closures (e.g., flights have avoided Russian airspace since 2022), and special-use airspace force re-routing.
• Weather avoidance. Convective storms, turbulence, and volcanic ash plumes route flights around them tactically.

In total, real flight paths typically deviate 1–5 % from the geometric great-circle optimum. For most travellers the deviation is invisible; for fuel-planning engineers the few-percent adjustment is the difference between profit and loss on a route.

On non-Mercator projections

Web map tiles and seat-back displays use Web Mercator, which has the characteristic high-latitude stretching that exaggerates the visual curvature of polar-routing flights. On other projections:

• Robinson (used by National Geographic 1988–1998) — great- circle paths still curve, but less dramatically because the projection is a compromise between conformal and equal-area.
• Winkel Tripel (current National Geographic standard) — similar to Robinson; mild curvature for polar routes.
• Mollweide / Eckert IV (equal-area projections) — great- circle paths look curved but the shape of the curve is different because the projection doesn't preserve angles.
• Gnomonic projection — the unique projection where great- circle paths appear as straight lines. Rarely used in consumer mapping because everything else (shapes, areas, distances away from the centre) is severely distorted. Specialised aviation and marine charts sometimes use Gnomonic precisely because of the straight-great-circle property.

The choice of projection on a route map is a deliberate display choice. Most airlines use Web Mercator because that's what their tile servers serve; the visual “curving” artefact is incidental to the engineering choice.

The interactive demonstration

A simple demonstration: use any web mapping tool (Google Earth, the Coordinately /tools/distance-calculator) to draw a path between two distant city pairs. On the flat web-Mercator view, the path bows. Switch to a globe view (Google Earth's 3D mode is ideal) and the same path is a perfectly smooth arc — the visual “curve” is gone because the globe doesn't need to flatten anything.

The /tools/distance-calculator emits the great-circle path as a polyline overlay on the map — sampled at 64 points and antimeridian-aware so the curve renders correctly even for trans-Pacific routes that cross 180° longitude.

Common misconceptions

“Pilots take longer routes to save fuel.” Sometimes — for jet-stream optimisation, yes. But the baseline route is the great-circle (shortest geometric path); fuel-optimal routes are deviations from that baseline by a few percent. The path is not deliberately longer; the optimisation is over the great- circle baseline.

“The Earth's rotation makes flights faster going east.” Earth's rotation has no direct effect on flight time relative to the ground (the atmosphere rotates with the Earth). Eastbound flights are usually faster because the jet stream blows west-to-east at high altitude, not because the planet is spinning underneath.

“Polar routes are dangerous.” They're heavily regulated (ETOPS rules, dedicated diversion airports, cold- weather operating procedures), but routine. Hundreds of polar flights operate daily. The risk per flight is comparable to or lower than mid-latitude oceanic routes.

“Maps show distance accurately.” Most consumer flat maps don't. Mercator's distance is exact only along the equator and along meridians, and only locally (the scale changes with latitude). For accurate distance, use a geodesic formula on geographic coordinates — see /learn/great-circle-distance.

“Great-circle routes only matter for very long flights.” The effect is most visible on long flights but applies at every scale. A 500 km regional flight uses a great-circle path that's indistinguishable from a rhumb line to the eye; a 12,000 km transpacific flight uses a great-circle path that visibly arcs over the poles. The geometric principle is the same; the visible effect scales with distance.

“Why don't airlines just use globe-view maps?” Engineering cost. Slippy-map tile servers serve flat tiles because tile arithmetic is trivially fast in Web Mercator. A true globe view is computationally more expensive and harder to zoom and pan smoothly. Some seat-back displays have a toggle between flat and globe modes; the underlying route data is the same.

“Old paper maps had the same problem.” They did, but paper-era navigators were trained to read Mercator charts with the projection's distortion in mind. The classical solution: use a Gnomonic chart to plot the great-circle path (where it appears as a straight line), measure waypoints off it, then transfer the waypoints to a Mercator chart for actual steering. Two-chart navigation was standard practice in commercial aviation through the 1970s; GPS and FMS (Flight Management System) computers replaced it in the 1980s onwards. Today the great-circle path is computed by software and the projection question becomes purely cosmetic.

“Curved flight paths are an optical illusion.” They aren't illusions — they're mathematically real curves on the projected map. The path on the real Earth is straight (a great-circle arc). The projection from sphere to flat map is what makes the path appear curved. Both the “straight on the globe” and “curved on the flat map” descriptions are correct; they apply to different representations of the same physical path.