Flight map
Every nonstop to Hawaii, and what the wind costs you
68 nonstop city pairs. 88,387 flights. Each airport is placed at its real distance and compass direction from Honolulu.
Flying west to Hawaii takes longer than flying home from almost everywhere. Every figure below is that difference: the same route flown both ways, and how many more minutes it takes in the direction of Hawaii. The interesting part is what happens as the routes get longer.
Slower flying out Faster flying out Line weight is flights per day. Crowded dots stay unlabeled until you pick one.
The extra time grows faster than the distance
That flying west takes longer is not news. What surprised me is that the extra time does not stay proportional to the distance. The longer the route, the bigger that gap becomes as a share of the return flight.
One route per airport below, the Honolulu one, so the rows compare like with like. The ranked table further down averages every route an airport flies, which is why a few numbers differ between the two.
| To Honolulu from | Miles | Longer flying out | Share of the return |
|---|---|---|---|
| San Francisco SFO | 2,398 | 22 min | 8.0% |
| Los Angeles LAX | 2,556 | 30 min | 10.3% |
| Seattle SEA | 2,677 | 27 min | 8.7% |
| Denver DEN | 3,365 | 44 min | 12.1% |
| Chicago ORD | 4,243 | 66 min | 14.8% |
| Atlanta ATL | 4,502 | 81 min | 16.8% |
| Newark EWR | 4,962 | 93 min | 17.5% |
San Francisco gives up about 8 percent of its return time. Newark gives up 17.5 percent over roughly twice the distance. Across the 66 westbound pairs the extra time tracks distance at r = 0.94 with a negative intercept.
That is a strong trend and it is not a rule. Seattle is the row that makes that clear. It is 121 miles further out than Los Angeles and gives up three minutes less. Whatever is happening is not simply a function of distance.
I do not know why
The obvious guess is the jet stream, and that the longer routes spend more of the flight in the stronger part of it. That makes sense until you notice Seattle is the northernmost departure in the table and one of the smallest gaps, so the explanation does not survive its own example.
The DOT data gives me flight times and distances, not winds aloft. I cannot settle the question from this dataset and I am not going to pretend I can. If you have historical winds-aloft data that would let me test it, I would like to see it.
Why the two on the far side are not exceptions
The jet stream generally runs west to east across the Pacific. Fly west into it and it costs you time. Fly east with it and it gives some back.
For a while this page said 66 of 68 routes, with two exceptions. That is technically true and it makes the two routes sound mysterious. They are Guam and Pago Pago. Both are west and south of Hawaii, so flying from either one to Hawaii means flying east, and both are faster flying out, Guam by 37 minutes and Pago Pago by 16.
Once you account for direction it is 68 out of 68. Every route that flies west to reach Hawaii is slower doing it. Both routes that fly east to reach Hawaii are faster.
The map is what made this obvious. Each airport is placed by its initial great-circle bearing from Honolulu rather than on a conventional geographic map, so Guam and Pago Pago land on the opposite side of the circle from everything else and the colour flips there too.
Ranked by extra time flying out
| From | Miles | Bearing | Extra time flying out | Flights a day |
|---|---|---|---|---|
| Newark EWR | 4,955 | 283° | +93 min | 0.4 |
| Atlanta ATL | 4,495 | 280° | +90 min | 1.5 |
| Washington Dulles IAD | 4,809 | 281° | +89 min | 0.6 |
| New York JFK | 4,975 | 283° | +84 min | 1.9 |
| Detroit DTW | 4,468 | 277° | +75 min | 0.9 |
| Chicago ORD | 4,237 | 274° | +69 min | 2.2 |
| Dallas Fort Worth DFW | 3,779 | 274° | +65 min | 4.1 |
| Houston IAH | 3,898 | 277° | +63 min | 1 |
| Minneapolis MSP | 3,967 | 268° | +61 min | 1 |
| Salt Lake City SLC | 2,991 | 258° | +46 min | 4 |
| Denver DEN | 3,360 | 264° | +44 min | 5 |
| Phoenix PHX | 2,914 | 265° | +43 min | 8.3 |
| Las Vegas LAS | 2,759 | 260° | +38 min | 10.2 |
| Everett PAE | 2,694 | 240° | +37 min | 0.6 |
| Long Beach LGB | 2,566 | 261° | +34 min | 3 |
| San Diego SAN | 2,610 | 263° | +33 min | 9.7 |
| Ontario ONT | 2,600 | 261° | +30 min | 1.1 |
| Burbank BUR | 2,559 | 260° | +29 min | 0.3 |
| Portland PDX | 2,603 | 242° | +27 min | 6.5 |
| Los Angeles LAX | 2,553 | 260° | +27 min | 30.3 |
| San Jose SJC | 2,415 | 253° | +27 min | 4.2 |
| Seattle SEA | 2,677 | 240° | +26 min | 20 |
| Sacramento SMF | 2,460 | 251° | +25 min | 5.3 |
| Oakland OAK | 2,407 | 252° | +23 min | 5.8 |
| San Francisco SFO | 2,397 | 252° | +20 min | 17.6 |
| Anchorage ANC | 2,781 | 191° | +14 min | 1.5 |
| Pago Pago PPG | 2,611 | 20° | -16 min | 0.3 |
| Guam GUM | 3,796 | 73° | -37 min | 1 |
Bearing is the initial great-circle heading from that airport to Honolulu. Extra time is the mean across every route that airport flies, from median air times over 2026-01-01 to 2026-06-30, which is why an airport serving more than one island reads differently here than in the Honolulu-only table above. Miles here are measured from Honolulu rather than taken from the DOT's filed route distance, so they can differ by a few miles for the same reason.
Route-by-route figures are on the flights page. The whole dataset is downloadable as CSV, every row carrying its source and check date, CC0.
How the map works
The projection is azimuthal equidistant centred on Honolulu, so distance from the centre is true distance and the angle is true bearing. The geometry is the map, which is why there is no basemap and no map library: it is inline SVG. The 32 airport coordinates come from OurAirports and are public domain. Nothing in the map is fetched at runtime: no tiles, no map library, no third-party JavaScript. The page itself still pulls webfonts from Google like the rest of the site, so it is the map that makes no external requests rather than the document.
