Data vintage Forecasts last recomputed 2026-06-13 17:38 UTC · latest observed planetary Kp 2.67.
Where the data comes from
- NOAA SWPC planetary Kp forecast — the 3-day, 3-hourly forward forecast of global geomagnetic activity. This is the backbone of the "tonight" and 3-day scores. (services.swpc.noaa.gov, planetary-k-index-forecast)
- NOAA SWPC observed planetary Kp — the most recent measured Kp, shown as the "right now" reading. (services.swpc.noaa.gov, planetary-k-index)
- NOAA SWPC OVATION model — the auroral-oval probability map; we read the overhead viewing probability for each destination as a now-cast enhancement. (services.swpc.noaa.gov/json, ovation_aurora_latest)
- Open-Meteo — free hourly cloud-cover forecasts plus sunrise/sunset and daylight duration for each destination's exact coordinates. (api.open-meteo.com)
We are not affiliated with NOAA, the University of Alaska, or Open-Meteo; we reproduce and combine their public data and cite it here.
How the score is computed
For each destination and each 3-hour forecast window we combine three factors into a 0–100 score:
- Solar activity vs. the local threshold. Each destination has a minimum Kp baked into its profile, derived from its geomagnetic latitude (the latitude that actually governs where the auroral oval sits — not the map latitude). At or above the threshold, the aurora reaches roughly overhead; below it, the credit falls off as the lights drop toward the horizon.
- Cloud cover (the dealbreaker). The Kp credit is multiplied by a sky factor from the Open-Meteo cloud forecast over the dark-hours window. Overcast skies crush the score even when activity is high — because you genuinely won't see it. We keep a small floor (15%) because clouds can break; we never claim a sighting is impossible.
- Darkness. A night with less than 5 hours of real darkness is treated as out of season (midnight sun / white nights) — no usable window regardless of Kp. We use sun-below-horizon hours from Open-Meteo and err toward the honest "out of season" call near the season edges, where twilight, not true darkness, dominates.
The score maps to a plain band: Low (0+), Fair (25+), Good (50+), Excellent (75+).
How fresh it is
An automated job re-pulls all four feeds every hour, recomputes every destination's score, and rebuilds these pages. We track 35 destinations. If a feed is stale or failing, we omit the affected reading and flag it rather than show a number we can't stand behind.
This is a probabilistic forecast, not a guarantee. Source feeds can be late, revised, or briefly unavailable; solar activity and cloud cover are inherently uncertain. The absence of good odds is not proof the aurora won't appear, and great odds are not proof it will. Full terms.