How the answer is built
No agency publishes a yes or no for the aurora over your town, and this site does not pretend to speak for one. What exists are forecasts and models: a planetary Kp index forecast by the NOAA Space Weather Prediction Center, the OVATION model of where the aurora is right now, weather forecasts of cloud cover, and the astronomy of the Sun and the Moon. The site joins them for one place with the rules below. Each rule either comes from a named source or is marked as our own choice.
1. The magnetic latitude of the place
The aurora forms an oval around the magnetic pole, so distance to the aurora is measured in magnetic latitude. We use AACGM-v2, the altitude-adjusted corrected geomagnetic coordinates of Shepherd (2014), built on the International Geomagnetic Reference Field: AACGM-v2 at 110 km, epoch 2026-10-01, aacgmv2 2.7.1. For the 147 towns of the site the value is computed exactly; for a ZIP code or coordinates it is interpolated in a grid of one degree from 20° to 80° north, within 0.3° of the exact value in our tests. Town positions are the internal points of the Census Bureau 2024 Gazetteer, except a few large municipalities where the downtown is used; Canadian and British cities use city centre coordinates.
2. The Kp a place needs
Two published relationships give the magnetic latitude from which aurora can be seen at each Kp. Up to Kp 5, SWPC's Tips on Viewing the Aurora: the equatorward edge of the oval sits near 66° at Kp 0 and moves about 2° per step. From Kp 6 to 9, the NOAA Space Weather Scales, which give the typical magnetic latitude down to which aurora has been seen at G2 to G5. Between the points the line is linear, and at each Kp it never sits north of the oval edge itself: at Kp 6 the scale's 55° lies north of the 54° edge, so the edge is used until the two lines cross near Kp 6.3. The Kp a place needs is where the line crosses its magnetic latitude.
| Kp | Seen down to (magnetic latitude) | Source |
|---|---|---|
| Kp 0 | 66° | SWPC oval rule |
| Kp 1 | 64° | SWPC oval rule |
| Kp 2 | 62° | SWPC oval rule |
| Kp 3 | 60° | SWPC oval rule |
| Kp 4 | 58° | SWPC oval rule |
| Kp 5 | 56° | SWPC oval rule |
| Kp 6 | 55° | NOAA scale, G2 |
| Kp 7 | 50° | NOAA scale, G3 |
| Kp 8 | 45° | NOAA scale, G4 |
| Kp 9 | 40° | NOAA scale, G5 |
The Kp for aurora overhead uses the oval rule alone: (66 − magnetic latitude) ÷ 2. Below Kp 5 the view line is the edge of the oval itself, which is cautious: SWPC also says a bright aurora can be seen from up to 1000 km away from a hilltop with an open view north.
3. Darkness and the Moon
The Sun's altitude comes from the equations of the NOAA Global Monitoring Laboratory solar calculator; the Moon's from the low-precision formulas of the Astronomical Almanac. Our tests compare them with the JPL DE421 ephemeris (through the Skyfield library) for five places on five dates: the Sun within 0.05°, the start and end of astronomical night within two minutes, the Moon within one degree and its lit fraction within two percentage points.
4. The live feeds
Your browser reads, directly and without going through us: the SWPC Kp forecast (observed, estimated and predicted three-hour blocks), the latest OVATION grid, and the cloud forecast. In the United States the cloud forecast is the hourly sky cover of the National Weather Service API for the forecast grid cell of the place; elsewhere, or if it fails, the cloud area fraction of the Norwegian Meteorological Institute. SWPC and NWS data are in the public domain; MET Norway data are under CC BY 4.0. If a feed does not answer, the tool says so and lets you type the value.
5. The rules of the answer
Each hour between the end and the start of nautical twilight is rated twice. Activity: the Kp of the hour's three-hour block against the Kp the place needs; for the hour in progress, OVATION can raise it. Sky: the cloud cover of the hour. The hour's answer is the weaker of the two, capped by darkness and the Moon. The night takes its best hourly answer, and the best time is the longest run of hours that reach it. The thresholds are our own choices:
| Rule | Value | Why |
|---|---|---|
| Sky reach | 500 km | Half the 1000 km SWPC gives as the best case; a 100 km high border at that distance stands about 9° up. |
| OVATION strong / possible | 10% / 5% | Highest probability inside the sky reach, current hour only. |
| Clear / broken sky | ≤ 30% / ≤ 70% cloud | Above that, overcast. |
| Edge of activity | 1 Kp below the need | Gives a maybe, never a yes. |
| Bright Moon | 75% lit and 10° up | Turns a yes with less than 1 Kp of margin into a maybe. |
| Darkness | Sun below -18°; twilight to -12° | Astronomical night; twilight hours cap at maybe. |
Worked example: Chicago
Chicago sits at 51.3° magnetic latitude. That is between the Kp 6 point (55°) and the Kp 7 point (50°) of the view line, so the Kp needed is 6 + (55 − 51.3) ÷ 5 = 6.7, a G3 (Strong) storm. On a December night the Sun is 18° down for 11.5 hours. If SWPC forecasts Kp 7.00 for the block 03 to 06 UT, 9 pm to midnight in Chicago in winter, and the National Weather Service gives 20% sky cover for those hours, each hour is strong and clear: the answer is yes, from 9 pm to midnight. With 50% cover it would be maybe; with Kp 6.00 it would be maybe too, less than one Kp short; with Kp 5, no.
6. Monthly figures on place pages
The darkness table of each place uses the 15th of each month of 2027. The cloud amount is the monthly mean of the NASA POWER Climatology API v2, CLOUD_AMT, January 2001 - December 2025 (CERES SYN1deg) at the place, and clear dark hours are the dark hours times the share of sky not covered: a long-run average, not a forecast.
7. What this method cannot do
- It cannot predict substorms, the bursts that can light up a sky for twenty minutes on a quiet Kp night, nor the exact arrival time of a coronal mass ejection.
- Kp is planetary and averaged over three hours; the view line is an average relationship that SWPC itself calls approximate.
- It does not know your horizon, local light pollution or fog in your valley.
- Cloud forecasts are forecasts: a front can clear an hour early or late.
- AACGM coordinates drift as the magnetic field changes; the values are computed for October 2026 and are reviewed each year.
8. Tests and updates
Automated tests check the astronomy against DE421, the Kp rules against the SWPC statements quoted above, the parsers against saved payloads of each feed and the verdict against hand-built nights, before every release. The data files are rebuilt by scripts from their downloads. Report a night where the answer was wrong through the contact page, with the share link of the tool.