How to Read an Aurora Forecast

An aurora forecast is not a weather forecast. Nobody can tell you "aurora at 11:14 PM over your house." What forecasts give you instead is a stack of measurements — some predicting days ahead, some minutes ahead — and the skill of aurora chasing is knowing which number to trust at which moment. This guide teaches you that skill: how to read each layer of an aurora forecast, in the order a real chaser reads them, and how to turn the numbers into a decision.

The three time horizons of aurora forecasting

Every aurora forecast operates on one of three clocks, and confusing them is the number one beginner mistake.

Days ahead: watching the Sun

The longest warnings come from watching the Sun directly. When a coronal mass ejection erupts toward Earth, forecasters can estimate its arrival one to three days out — that is why you see "G2 storm watch for Thursday" headlines on Tuesday. Similarly, coronal holes rotate across the Sun's face on a roughly 27-day cycle, so their fast wind streams can be penciled in almost a month in advance.

Accuracy at this range is modest. Arrival times routinely miss by six to twelve hours, and a CME's most important property — its magnetic orientation — cannot be known until it arrives. Treat multi-day forecasts as "keep the evening free," never as promises.

Hours ahead: the upstream early-warning line

The best data in the sport comes from satellites parked at the L1 point, about a million miles sunward of Earth. Everything that flows past them hits Earth roughly 30–60 minutes later. This is real measurement, not prediction: when L1 reports fast wind and a southward magnetic field, that exact parcel of solar wind is about to arrive.

Our telemetry dashboard streams this feed live. When chasers say "the data looks great right now," this is the data they mean.

Right now: the sky and the ground

The shortest horizon is nowcasting: magnetometer needles twitching, all-sky cameras lighting up, and other observers reporting. Our live sightings map belongs to this layer — when reports bloom a few hundred miles north of you, your own display is often minutes away.

The numbers, one by one

Kp: the planning number

The Kp index summarizes global geomagnetic disturbance on a 0–9 scale in three-hour blocks. Its superpower is geography: each Kp level maps roughly to how far equatorward the auroral oval reaches. Use predicted Kp to answer one question — is tonight worth considering at my latitude? — and then stop leaning on it. It is a coarse, planet-averaged, three-hour number; it cannot tell you what the next thirty minutes hold.

Bz: the decision number

If you learn to read only one real-time value, make it Bz — the north–south direction of the solar wind's magnetic field. Southward (negative) Bz opens the door for solar wind energy to enter Earth's magnetosphere through magnetic reconnection; northward (positive) Bz keeps it mostly shut.

Reading it in practice:

• Bz around 0 to −5 nT: mildly favorable; good nights possible at high latitudes.
• Bz below −10 nT sustained for an hour or more: strongly favorable; storm-level activity likely.
• Bz strongly positive: unfavorable, even if every other number looks impressive. Many "failed" high-Kp forecasts died on a stubbornly northward Bz.

Bz flips constantly, so judge the trend over 30–60 minutes rather than a single reading.

Speed and density: the fuel gauge

Solar wind speed below ~400 km/s is sleepy; 500–800 km/s is the energetic range that keeps ovals active. Density spikes matter mostly as arrival flags — the compressed front of a CME or stream announces itself with a sudden density jump, often followed within hours by the magnetic-field swings that decide the night.

The combinations to know:

• High speed + strongly negative Bz = the jackpot pairing. Drop your plans.
• High speed + positive Bz = potential without payoff — stay alert for a Bz flip.
• Low speed + negative Bz = modest but real activity, worthwhile at high latitudes.

Hemispheric power: the brightness gauge

Hemispheric power estimates, in gigawatts, the total auroral energy raining into one hemisphere. Quiet nights run 5–20 GW; above ~50 GW the oval is genuinely bright; 100+ GW accompanies major storms. Because it refreshes frequently, rising HPI is a nice confirmation that the oval is intensifying right now.

A worked example: reading a storm night

Here is how the layers stack on a classic CME night:

1. Two days out: headlines announce a G2 watch. You check the arrival window — late evening, workable — and keep the night free.
2. Afternoon of arrival: telemetry shows quiet wind, ~380 km/s. Nothing yet. Arrival estimates have hours of slack, so you don't conclude anything.
3. 8:40 PM: density spikes from 4 to 25 per cm³ and speed jumps to 550 km/s in minutes — the shock has arrived. Bz is positive. You get dressed anyway.
4. 9:30 PM: Bz swings to −12 nT and holds. This is the green light. You drive to your dark site.
5. 10:15 PM onward: substorms begin — the sky erupts in cycles for the next three hours, matching fresh reports stacking up on the sightings map.

Notice what made the call: not Kp (still showing the pre-storm value for its three-hour block), but the L1 sequence — density spike, then sustained negative Bz.

The checklist that prevents wasted nights

Numbers can be perfect while the night fails for ordinary reasons. Before driving:

• Darkness: is the Sun far enough below the horizon? High-latitude summer makes auroras invisible regardless of data. See Best Time to See Northern Lights.
• Clouds: any overcast layer defeats any Kp. Check a cloud forecast with the same seriousness as the space weather — our guide to aurora cloud cover covers strategy.
• Moon: a full moon washes out faint displays; strong ones punch through.
• Horizon: at mid-latitudes the show sits low in the north. Trees and town glow cost you more than a Kp point.

Common forecast-reading mistakes

• Treating predicted Kp as a promise. It is a probability statement about a three-hour window, frequently revised.
• Quitting because "nothing is happening." Substorms load for one to three hours between eruptions. Quiet arcs frequently precede breakups — see How Do the Northern Lights Work? for the loading cycle.
• Ignoring the 30–60 minute L1 lead. When Bz goes strongly south, you have roughly half an hour of warning before that parcel of wind arrives. Use it to get in position, not to start deciding.
• Watching one number. Every veteran reads speed, density, Bz, and HPI together, the way a sailor reads wind, swell, and sky.

Building your personal alert setup

Forecast reading pays off most when you stop checking manually and let thresholds watch for you. The setup takes ten minutes and lasts for years:

Calibrate to your magnetic latitude first. Look up your magnetic latitude once — it may differ meaningfully from your geographic intuition, especially in North America, where the geomagnetic pole's offset gives the Midwest a hidden advantage. Then set your Kp alert one notch below your visibility threshold: if your latitude typically needs Kp 5, alert at Kp 4. Forecasts under-call as often as they over-call, and the early alert buys you scouting time.

Layer the alerts. A robust setup watches three tiers: storm watches (days ahead — plan), predicted Kp crossing your threshold (evening — prepare), and a real-time trigger such as Bz below −10 nT for 30+ minutes or hemispheric power above 50 GW (now — go). Any one alone produces false alarms or missed nights; together they filter each other.

Budget for false alarms without resentment. Even a well-tuned setup will send you outside for perhaps one fizzle in every three or four calls — that is the honest hit rate of a chaotic system measured an hour upstream. Veterans reframe it: every response is a rehearsal that keeps your dark site, gear ritual, and drive time sharp for the night that delivers. The chasers with the best photo libraries are rarely the best forecasters — they are the ones who answered the most alerts.

Review after every event. Win or fizzle, glance back at the telemetry record afterward and ask what the data looked like an hour before the outcome. A dozen such reviews builds the pattern recognition no article can install directly — the ability to glance at speed, density, and Bz together and simply know the night's character.

Your reading order, condensed

1. Days ahead: any CME watches or coronal-hole streams inbound? (Plan.)
2. Evening of: predicted Kp vs your latitude. (Consider.)
3. Real time: speed, density, and above all sustained Bz on telemetry. (Decide.)
4. Confirmation: hemispheric power rising, reports on the sightings map. (Go.)

Our homepage forecast automates most of this stack for your location, but understanding the layers changes how you use it: you will know why tonight's odds are what they are — and you will recognize the golden combinations the moment they appear in the data.