Why Aircraft Runways Point the Way They Do

Apparently the huge number at the end of a runway isn't for inventory. It actually tells a pilot approximately which way the aircraft will face.

Runway 27 points roughly west, toward a compass heading of 270 degrees. Approach the same pavement from the other end and it becomes runway 09, pointing roughly east at 90 degrees. The numbers differ by 18 because the two approaches are 180 degrees apart. They are rounded to the nearest ten degrees, not exact headings. Where runways are parallel, 27L is the left runway and 27R the right from the pilot's approach.¹

So that explains the number, but not why someone pointed an enormous strip of pavement in that direction in the first place. For that, you need to look at the wind.

An aircraft needs air moving over its wings. A headwind supplies some of that movement while the aircraft is traveling more slowly across the ground, which helps during takeoff and landing. A crosswind is the part blowing sideways across the runway instead. It can push the aircraft off its intended path, even when the same wind is also helping from ahead.

A 20-knot wind arriving 30 degrees off the runway produces a 10-knot crosswind and about 17 knots of headwind. The wind produces effects in sideways and forward-facing directions.²

A runway cannot turn when the weather changes, so airport planners have to account for the range of winds a site actually gets. In the United States, the FAA normally orients the primary runway toward prevailing winds. Its design guidance considers a crosswind runway when wind coverage falls below 95 percent and the relevant aircraft use the airport regularly. Coverage here means winds remaining within an applicable crosswind limit, not winds pointing neatly down the runway 95 percent of the time. It is also a U.S. planning criterion, not a worldwide rule.²

I mapped 14,564 open, heading-known runways at 11,122 airports across 235 country and territory codes. The small accent-colored lines are sampled major-airport runways, each positioned at its airport's reference coordinates and pointed in the direction recorded by the dataset. Highlighted lines also mark Heathrow and two northern Canadian case studies. The circular chart includes all 14,564 runways, counting each strip once and mirroring its axis to show the two possible approaches. Both visuals use true geographic headings, not the magnetic numbers painted on most runways.³

The fullest ten-degree slice contains 1,423 runways, only about 9.8 percent of the total. There is no one direction airports everywhere prefer. The lines show physical orientations, not local wind observations, and they cannot tell us why any particular runway was built that way. Wind, available land, terrain, approach paths, noise and airport capacity can all shape the final arrangement.

The Heathrow example makes the difference between a fixed runway and a changing approach easier to see. Its two parallel strips are 09L/27R and 09R/27L. When aircraft face west, they use the 27 ends. When they face east, the same pavement becomes the 09 ends; left and right also swap because the pilot is looking from the opposite direction.

Heathrow's annual share of westerly runway operation ranged from 63 to 81 percent between 2000 and 2025. In 2025, it was 63 percent westerly and 37 percent easterly. Individual months moved much further: April 2025 was 29 percent westerly, meaning 71 percent easterly, while July was 88 percent westerly.⁴

Those percentages describe which way Heathrow operated its runways, not which way the wind was blowing. The airport attributes its usual westerly pattern partly to prevailing southwesterly winds, but also describes a daytime westerly preference and noise-related operating rules. Without actual weather-station observations, it's hard to separate those effects.⁵

There is one more complication. Most runway identifiers use magnetic north, the direction a magnetic compass points, while maps and the runway dataset use true north, the direction toward the geographic North Pole. The difference varies with location and changes over time. NOAA documents what that can mean in Fairbanks: in 2009, runway 1L/19R became 2L/20R because the magnetic reference had shifted. The runway itself did not turn.⁶

The exception I found most interesting is in Canada's designated Northern Domestic Airspace, where runway designators use true north instead.⁷ At Resolute Bay, 17T/35T follows a documented physical heading of about 167/347 degrees true. The 17T marking therefore refers to the same north reference as the map.⁸

Iqaluit is also far north, but its 16/34 runway uses magnetic designators even though the strip points approximately 137 degrees true. Its 16 marking represents roughly 160 degrees magnetic; that does not contradict the mapped direction because the two values use different references. Latitude alone does not tell you which rule applies.

So the painted number identifies an approach, the opposite end identifies its reciprocal, and the airport decides which end to use as conditions change. The strip itself stays where wind, geography, and the practical limits of building an airport put it. Sometimes even north depends on where you are.