How to Photograph the Aurora: Kp Index, Camera Settings, and Reading the Space-Weather Forecast

You have an aurora app on your phone, and at some point it has pushed you a notification with a number on it. Whether that number means you should drive ninety minutes north of the city or go back to sleep is the entire question, and the app almost never answers it.

That gap is the reason this piece exists. Aurora photography is a forecasting problem first and a camera problem second, and most of the people who miss a display miss it because they read the right number at the wrong hour.

This solar cycle has raised the stakes considerably. Solar Cycle 25 moved through its maximum period in the mid-2020s, and the declining phase that follows a peak has a long history of producing a cycle's most dramatic storms — the Halloween storms of October 2003 arrived years after that cycle's sunspot peak.

What Does The Kp Index Actually Measure?

Kp is derived from a network of mid-latitude magnetometer stations that record how much the Earth's magnetic field is being pushed around. The scale is quasi-logarithmic, so the jump from Kp 6 to Kp 7 is far larger than the jump from Kp 2 to Kp 3.

You will also see it written in thirds — 5-, 5o, and 5+ are all meaningfully different conditions inside the same integer. Keep in mind that a forecast of Kp 5 and an observed Kp 5+ are not the same night.

Two consequences follow from that definition, and both matter for planning. First, Kp is planetary rather than local, so it describes the whole magnetosphere and not the sky above your driveway.

Second, it is computed across three-hour blocks, which means the published value is always describing a window that has partly or entirely already happened. That lag is exactly why the number on the notification is so often useless at the moment it arrives.

NOAA's Space Weather Prediction Center maps Kp onto the G-scale used in public alerts. Here is how the two line up:

Kp value	NOAA G-scale	What it means for a photographer
Kp 5	G1 — Minor	Northern-tier states and southern Canada. Worth driving for if your skies are dark.
Kp 6	G2 — Moderate	Upper Midwest and northern New England become realistic. Camera will see more than the eye.
Kp 7	G3 — Strong	Mid-latitudes. This is the threshold where cities well south of the border start reporting sightings.
Kp 8	G4 — Severe	Rare. Visible overhead across much of the northern half of the country.
Kp 9	G5 — Extreme	Once-in-a-decade. The May 2024 storm reached this level and was reported into Florida and Mexico.

All of these thresholds assume clear skies, real darkness, and a view toward the pole. The Kp value sets the ceiling on what is possible; everything else in this post determines whether you actually get it.

The Live Number To Watch Is Bz

The solar wind carries a magnetic field with it, and the north-south component of that field is called Bz. When Bz points south, it is antiparallel to Earth's own field at the dayside boundary, and the two reconnect.

That reconnection is the valve. Energy pours into the magnetotail, loads up, and then unloads into the auroral oval as a substorm.

This is why two nights with identical Kp forecasts produce completely different results. A night with Bz hovering at negative fifteen for four straight hours will outperform a night that spikes to negative twenty for ten minutes and then flips north.

The data comes from spacecraft parked at the L1 Lagrange point, roughly 1.5 million kilometres sunward of Earth. That distance buys you somewhere between fifteen and sixty minutes of warning depending on solar wind speed, which is enough time to get out the door but not enough to plan an evening.

Alongside Bz, two other real-time values are worth a glance. Solar wind speed above roughly 500 kilometres per second and elevated particle density both make a given Bz more productive.

How Far South Will It Actually Reach?

The auroral oval is centred on the geomagnetic pole, which does not sit at the geographic pole. It has been drifting for as long as we have measured it, and the offset currently favours North America substantially.

The practical effect is that a photographer in the American Midwest gets aurora opportunities that a photographer at the equivalent line of latitude in Europe simply does not. Anyone comparing sighting reports across the Atlantic without accounting for this will conclude the forecast is broken when it is working exactly as described.

NOAA publishes a viewline on its aurora forecast maps — a line indicating how far equatorward the aurora may be visible low on the horizon. Note that the viewline is a horizon-glow boundary, not an overhead boundary, so being just inside it means looking north at something faint rather than standing underneath a corona.

For a horizon-glow night, your northern sightline matters more than your gear. A hill, a lake shore, or any position with an unobstructed low northern horizon is worth more than two stops of lens speed.

The Four Gates Every Aurora Night Has To Pass

Four conditions have to align on the same night, and failing any one of them ends the attempt regardless of how good the other three look:

• Geomagnetic activity. The Kp forecast has to clear your location's threshold, and Bz has to actually go south when the window opens. This is the only gate you cannot influence.
• Cloud cover. Aurora happens above roughly 100 kilometres, far above every weather layer, so any opaque deck between you and it is total. High cirrus is the treacherous one because it barely registers on a casual glance at a forecast app — our guide to reading cloud layers covers how to spot the layer that will ruin you.
• Moon phase. A bright gibbous or full moon raises sky brightness enough to flatten a faint display, though a strong storm will punch through it. If you have flexibility, check moon phase and moonrise timing before committing to a date.
• Timing and darkness. You need full astronomical darkness, which at high latitudes in summer simply does not arrive. This is the gate that quietly disqualifies entire months for northern shooters.

Run all four before you drive anywhere. Most disappointing aurora chases are not failures of forecasting skill so much as failures to check the boring gate — the one that was knowable at four in the afternoon.

When Do You Actually Go Outside?

The aurora does not run continuously through a storm. It loads and unloads, and the unloading events are the ones that produce the vertical rays and the fast-moving bands worth photographing.

Activity concentrates around magnetic midnight, which is not the same as clock midnight and drifts depending on your longitude relative to the magnetic pole. In practice, the hours between roughly 10 p.m. and 2 a.m. local time carry most of the good material.

Do not treat a quiet sky at 10:30 as a verdict. The most common mistake we see is a photographer packing up twenty minutes before a substorm breaks.

For exact darkness windows, use published sun ephemeris data from NOAA or the US Naval Observatory rather than a general-purpose weather app. If you already track twilight for other work, the same discipline described in our piece on blue hour and the twilight phases applies directly — astronomical twilight is the boundary that matters here, not civil.

What Camera Settings Work For Aurora?

Aperture is where the decision gets made. A fast wide lens between f/1.4 and f/2.8 is worth more than any amount of ISO performance, because the shutter speed it buys you is what preserves structure in a moving display.

Shutter speed is the setting most people get wrong, and they get it wrong in the same direction every time. A fifteen-second exposure of an active aurora produces a green smear where there were discrete rays.

What the sky is doing	Shutter	ISO
Faint static arc on the horizon	8–15 sec	1600–3200
Defined bands, slow drift	4–8 sec	3200
Fast-moving curtains, visible rays	1–4 sec	3200–6400
Substorm breakup or overhead corona	0.5–2 sec	6400+

Move down that table as the display intensifies, and be willing to change settings every few minutes. An aurora that goes from arc to corona in ninety seconds will leave you with a folder of unusable frames if you set once and forget.

Focus manually, always. Magnify live view onto the brightest star available, focus until it is a point rather than a disc, then tape the ring or disable autofocus so a stray hand does not undo it.

Shoot RAW without exception, and set white balance somewhere around 3400K to 4000K as a starting point. The file is where the colour decision actually gets made, and the same latitude that helps with Milky Way exposures is what lets you recover an aurora you underexposed by a stop.

Why Your Photos Show Colours You Could Not See

Aurora colour comes from specific atmospheric emissions at specific altitudes. The familiar green is atomic oxygen at 557.7 nanometres, produced across roughly 100 to 250 kilometres, and it dominates most displays.

The red above it is also atomic oxygen, at 630.0 nanometres, from higher in the thermosphere. Blues and violets come from ionised molecular nitrogen lower down, and they tend to show up only in genuinely strong events.

Your night vision runs on rod cells, which peak in sensitivity near 500 nanometres and fall off badly toward the deep red. The camera has no such bias, which is why the photograph of a mid-latitude display so often looks nothing like the memory of it.

This is worth saying plainly to anyone you bring along. Manage the expectation before you drive two hours, because the gap between the sensor and the eye has ruined more aurora trips than cloud cover has.

What Ruins An Aurora Night

Assuming the forecast holds and the sky is clear, the remaining failure modes are almost all logistical:

• Lens condensation. A cold front element under a humid air mass will fog or frost, and it usually happens right as the display peaks. Check the spread between air temperature and dew point before you go, and bring a lens warmer if it is narrow.
• Battery failure. Lithium cells lose capacity fast in the cold, and a battery that shows two-thirds at the car will show empty an hour later. Carry spares against your body, not in the bag.
• Light pollution to the north. A city glow on your northern horizon competes directly with a horizon-glow display. Drive past it rather than around it.
• Forecast whiplash. Models disagree on cloud timing more than people expect, and a night that looked clear at noon can close by ten. Our piece on how forecast models disagree explains which one to trust in which situation.

None of these are exotic. All of them are the reason experienced aurora photographers pack the night before rather than the hour of.

How We Read It

We think most aurora coverage fails readers in the same specific way — it publishes the Kp number as though the number were the answer. The number is the beginning of a decision, and the decision is what a reader actually came for.

So the sequence we use is fixed. Kp and the three-day outlook decide whether the night stays open, cloud and moon decide whether it is worth driving, Bz decides when to step outside, and the display itself decides the shutter speed.

That is the same translation discipline behind Sunset Verify and behind every daily brief we publish — data becomes sensation becomes a decision, in that order. You can find the rest of our light and atmosphere work in the journal, and the reasoning behind the format in how we write a brief.

Set an alert threshold appropriate to your magnetic latitude, and then trust it enough to actually get in the car. The photographers who come home with the frames are not the ones with better forecasts — they are the ones who went outside on the marginal nights.

Common Questions

Can I photograph the aurora with a phone?

Yes, and modern computational night modes do it surprisingly well. Use a tripod or a stable surface, enable the longest available night exposure, and turn off any automatic flash. The limitation is control — a phone will pick its own shutter speed, which means an actively moving display often comes out smeared. A dedicated camera with a fast lens still wins on structure.

What lens should I use for the aurora?

A wide-angle prime between 14mm and 24mm at f/1.4 to f/2.8 is the standard choice. The wide field captures a full curtain and includes foreground for scale, while the fast aperture keeps your shutter short enough to preserve rays. If you only own a kit zoom at f/3.5, it will still work on a strong display — you will simply need higher ISO and more noise reduction in post.

Why did my aurora photo come out as a green blur?

Your shutter was open too long for how fast the display was moving. An active aurora can restructure itself in under a second, so a ten or fifteen second exposure averages all of that motion into a formless wash. Drop to two to four seconds and raise ISO to compensate. Sharper rays at higher noise beat a clean, structureless smear every time.

Does the aurora happen more often around the equinoxes?

Statistically yes. Geomagnetic activity shows a well-documented peak near the March and September equinoxes, driven by the seasonal geometry between Earth's magnetic axis and the interplanetary magnetic field. The practical bonus is that equinox nights also give northern photographers genuine astronomical darkness at a reasonable hour, which midsummer does not.

How much warning does a space-weather forecast actually give?

It depends on which signal you are watching. A coronal mass ejection can be forecast one to three days out from solar imagery, though arrival-time estimates carry hours of uncertainty. The real-time solar wind data from L1 spacecraft gives only fifteen to sixty minutes. Use the multi-day forecast to keep the night free and the L1 data to decide when to walk outside.

Do I need a dark-sky site, or will suburbs work?

For a strong overhead display, suburbs are workable — a G4 or G5 event is bright enough to compete with moderate light pollution. For anything at or below G2, where the aurora sits low on the northern horizon as a glow, city light in that direction will erase it. The rule is simple: the weaker the storm, the darker your northern horizon needs to be.