Handling Weather and Recovering From a Late Aircraft

Summary

Across 52 airports the correlation between points of on-time rate lost in bad weather and minutes short of expected after a late landing is -0.04, too small to call a relationship. 24 airports sit on the same side of the median on both measures, where 26.0 would be even if the lists were unrelated. DEN lost the most in bad weather (27.3 points) and ranked 4th on recovery.

An airport can hold up when the weather turns, and it can take a late aircraft and send it on with less delay than it arrived with. Each is read as a sign of a well-run airport, and this report asks whether the same airports have both. For 52 covered airports it sets the points of on-time rate lost going from normal weather to bad weather against the minutes given back on the ground after an aircraft landed 30 or more minutes late, beyond what the airport’s own turn lengths predict. The typical airport lost 12.6 points in bad weather and gave back 1.3 minutes less than expected. The correlation between points lost and minutes short of expected is -0.04, too small to call a relationship. Ranked instead of measured the correlation is -0.06, and against the bad-weather on-time rate alone it is -0.02. Knowing how an airport does in bad weather tells almost nothing about how it does with a late aircraft.

Airports are split at the median on each measure: 12.6 points lost in bad weather and -1.3 min beyond the expected. If the two were unrelated each group would hold about 13.0 airports. Airports with the same figure as the median fall in the second group of each pair. Points lost is the on-time rate in normal weather minus the rate in bad weather.

Of the 52 airports with both measures. The first rank on weather is the smallest loss of on-time rate in bad weather and the first rank on recovery is the most minutes given back beyond expected. Recovery is for aircraft that landed 30 or more minutes late, with the minutes expected from each airport’s own mix of scheduled turns.

24 of the 52 airports are on the same side of the median on both measures, 12 on the good side and 12 on the poor side. Taking the best quarter of each list (13 airports), 3 airports are in it on both (OAK, SFO, and SMF), against about 3.3 if the lists were unrelated. In the worst quarter of each there are 3 (FLL, MCO, and MIA). DEN lost the most in bad weather, 27.3 points, and ranked 4th of 52 on recovery. ORD gave back the most beyond the expected, +6.4 min, and ranked 48th on weather.

Both measures carry what the airports bring with them. The loss in bad weather depends on how severe that weather was and which kinds of it the airport gets, and a thunderstorm and a snowfall are counted together. Recovery depends on the airlines that fly there and on how they schedule their turns; the expected figure takes out the turn lengths and nothing else. Each list has its own floor (5,000 completed flights in each kind of weather, 5,000 departures after a late landing), and 7 airports with a recovery figure and no weather figure are left out. A correlation over 52 airports is itself uncertain, and one that is too small to call a relationship does not rule out a small one. Some of the aircraft that landed late were late because of weather, so the two measures overlap in the flights they are drawn from.