NOAA has a G2 (moderate) geomagnetic storm watch up for Friday, August 28, with G1 (minor) conditions expected the day before. A cloud of solar material (also called a coronal mass ejection or "CME") left the Sun on Tuesday morning and looks likely to clip Earth on its way past.
The models have this one clipping us on the flank (glancing blow), so there is a real chance it underperforms or misses us entirely. And the predicted arrival time lands close to sunrise for much of North America, which is about the least convenient timing we could have asked for.
The timing uncertainty on this one is huge, though, and that cuts both ways. An early arrival puts Thursday night in play for North America, and a late one puts Friday night in play. There is also a nearly total lunar eclipse on Thursday night regardless of what the CME does. The CME could also be strong enough, if it impacts according to schedule, for the effects to last until Friday night. We'll just have to see what happens!
What you're reading here went out to my email list first, a couple of days ahead of the predicted arrival, and I am posting it here so it is easier to find and share. If you would rather get these in your inbox as soon as I write them, you can sign up for my aurora alerts. A lot could change between now and Friday, so follow along on my socials for the latest.
The short version
On Tuesday morning, sunspot region AR 4513 fired off an M6.98 flare (call it an M7) and launched a CME, which is a giant cloud of solar plasma and magnetic energy. Actually it launched two of them. A smaller filament eruption went up alongside the main M7 CME, and the two are traveling together.
NOAA and NASA both put the arrival at around 11 UT on Friday, August 28, which is 7 AM Eastern, 6 AM Central, 5 AM Mountain, 4 AM Pacific, and 3 AM Alaska time. The uncertainty on that is roughly 10 hours in either direction. The HUXt model disagrees and lands about 14 hours later, on Friday evening.
On top of the CME, a coronal hole is sending faster solar wind our way starting Thursday, so the two could combine.
If it all lines up, that is a G2 storm with aurora possible across the northern U.S. and Canada, and a better shot for chasers in the southern hemisphere, where Friday's arrival time falls at night. If the CME misses us, Friday stays quiet. Either of those is still possible at this point.
What the Sun did
A CME (coronal mass ejection) is a giant cloud of solar plasma with a magnetic field tangled up inside it, blasted out into space. The magnetic field is the part that matters. When it reaches Earth and happens to point south, it links up with our own magnetic field and pours energy into the system, and that energy is what lights up the aurora.
This one came with an M6.98 flare from sunspot region AR 4513, peaking on Tuesday, August 25 at 10:02 UT. M-class sits in the middle of the flare scale, one step below the X-class monsters. An M7 is a solid flare, and it came with the signatures forecasters look for when they want to know whether a CME went with it: a Type II radio burst, and a wave rippling out across the solar surface.
Reading the coronagraphs
A coronagraph is a camera that blocks out the Sun's bright disk with a physical disk so we can see the faint material moving around it. It is the closest thing we have to watching a CME leave. Below are three frames from GOES-19's CCOR-1 coronagraph, taken over about two hours on Tuesday morning.
My read on them? There are two pieces to this eruption: a fainter, faster shock heading west, and the bulk of the cloud heading east. The bulk is the bright, obvious part, and it is mostly heading away from us. The western piece is the part that could clip Earth, and it is much harder to see. That ambiguity is most of why this forecast is such a tough call.
These are not impressive coronagraph images yet we still have a geomagnetic storm watch, so we will have to see what the solar wind actually brings in on Friday.
The coronal hole, and why it matters
There is a coronal hole on the Sun right now too, and it matters for two main reasons with respect to the incoming CME and chances for more widespread auroral visibility this week.
A coronal hole is a region where the Sun's magnetic field opens up and lets solar wind escape at high speed, typically 500-800 km/s instead of the usual 300-400 km/s. NOAA expects speeds in the 500-600 km/s range starting mid-to-late Thursday, and this rotation's hole has a more developed northerly extension than the last one, so the stream should be stronger than it was 27 days ago.
The eastern edge of this coronal hole sits right next to AR 4513, and it appears to have acted like a wall, blocking the CME from spreading westward toward the Sun-Earth line. That is part of why so much of this eruption went east, away from us. Also, because the coronal hole was facing Earth directly a few days ago, the fast wind should reach us later today (August 26) or tomorrow (August 27) likely ahead of the CME. Even without any CME impacts, minor storming is a real possibility from the high-speed stream.
It could also make the arrival more effective. This is a positive polarity coronal hole, which is favorable heading into the equinox. There is an effect called the Russell-McPherron effect, where the tilt of Earth's magnetic field relative to the solar wind around the equinoxes makes it easier for the two to couple together. With a stream coming almost straight out of a positive-polarity hole, we may see more of the solar wind's magnetic field pointing south. That north-south direction is called Bz, and southward Bz is what opens the door and drives the aurora.
So the effects from the coronal hole stream arriving Thursday could combine with the CME arriving Friday. That is a complicated situation to forecast, and it nudges the overall disturbance level up a little. G1 to G2 is still the realistic ceiling here.
When would it arrive?
NOAA's Space Weather Prediction Center, the official U.S. forecaster, has run this CME through its ENLIL model and puts the arrival at about 11:00 UT on Friday, August 28. NASA's modeling agrees. In local time that is:
- 7 AM Eastern on Friday
- 6 AM Central on Friday
- 5 AM Mountain on Friday
- 4 AM Pacific on Friday
- 3 AM Alaska time on Friday
That is the awkward part of this event for North America. The central estimate lands right around sunrise for the eastern half of the continent.
Most CME forecasts have uncertainty at plus or minus 7 hours, but because this is a glancing blow, I am upping this to plus or minus 10 hours for NOAA's forecast, and that is important to realize. Ten hours early is 1 UT on August 28, which is 9 PM Eastern on Thursday night. Ten hours late runs into Friday afternoon and evening.
Even if the CME impacts at the predicted time, Friday night could still happen for North America. A strong impact could mean effects don't calm down until later on Friday with some residual activity possible Friday night into Saturday for North American chasers.
The HUXt model, run out of the University of Reading, takes the later view. It runs 500 slightly different versions of the same forecast (also called an ensemble model) so you can see the spread, and it lands the median arrival at 01:31 UT on August 29, which is Friday evening in North America (9:31 PM Eastern, 8:31 PM Central, 6:31 PM Pacific). It gives the CME a 79% chance of arriving at all, with a timing spread running from about 12 hours early to 22 hours late.
So Thursday night, Friday morning, and Friday night are all in play.
How strong could this get?
NOAA's watch, issued Tuesday evening, reads:
- Wednesday, August 26: no storming expected
- Thursday, August 27: G1 (minor), from the coronal hole stream
- Friday, August 28: G2 (moderate), from the CMEs
The G-scale runs from G1 (minor) up to G5 (extreme), so G2 sits near the low end. NOAA's own bulletin says the aurora may become visible over some northern and upper Midwest states, from New York to Idaho, and a good substorm can push it well south of this perhaps into Pennsylvania, Ohio, Missouri, Nebraska, Colorado, and even Northern California.
I have modest expectations around the chances for widespread auroral visibility extending into the central U.S. This is only forecasted to be a glancing blow, the coronagraph imagery is not impressive, and a good chunk of the eruption went the wrong way. What I would plan around is minor storming with a chance of moderate, and I would treat anything stronger as a bonus. Combination eruptions like this one can also get deflected off the predicted path, which is another reason the arrival time could still move.
One thing worth understanding about storm watches is that NOAA issues them so power grid operators and satellite companies can prepare for the worst reasonable case. For us, a geomagnetic storm watch means there is a chance at aurora if it pans out.
What else the Sun might do this week
AR 4513 is rotating through the Earth-strike zone right now, which is the band of solar longitude where an eruption has a straight shot at us. It still has a complex, mixed-polarity magnetic configuration, which is the setup that produces big flares. As of Tuesday, NOAA gave it a 50% chance of another M-class flare and a 15% chance of an X-class flare.
There is also AR 4517, which is growing as it comes around, sitting at 10% for an M and 1% for an X.
So this may not be the last word this week. If AR 4513 fires again while it is pointed at us, we would be talking about a much better-aimed CME than this one. I will post about it if that happens.
The lunar eclipse on Thursday night
The thing I am most excited about this week is the lunar eclipse.
On Thursday night into Friday morning there is a partial lunar eclipse that comes very close to total. At greatest eclipse about 96% of the Moon slides into Earth's shadow, so you get almost the full deep red color without it technically being a total eclipse.
Greatest eclipse is at 04:13 UT on August 28, which is:
- 12:13 AM Eastern, early Friday morning
- 11:13 PM Central, Thursday night
- 10:13 PM Mountain, Thursday night
- 9:13 PM Pacific, Thursday night
- 8:13 PM Alaska time, Thursday night
Eastern and central North America get the whole eclipse from start to finish. Farther west the Moon rises with the eclipse already underway, and the farther west you go the less of it you get. Up here in Fairbanks we miss almost all of it, because the Moon does not clear the horizon until about 9:13 PM on Thursday, roughly an hour after greatest eclipse, and it comes up low and into twilight.
Now put the two together. If the CME runs on the early end of its window it arrives during Thursday night, and there would be a chance to photograph a reddish Moon with the aurora in the same frame, somewhere well east of me. I would love to see somebody get that shot. I do not have high confidence in it happening, and I want to say that plainly, but it is why I am watching this event a lot more closely than I normally would for a forecast like this.
The Moon is full for this eclipse, but that is less of a problem than people expect. The aurora is still fully visible under a bright Moon (most true at high-latitudes), and moonlight lights up your foreground, keeps your eyes in color vision, and makes the whole night easier if you are shooting with a phone or learning a camera. At mid-latitudes, a full moon can definitely wash out faint aurora on the horizon, though. But for the hour or so around greatest eclipse, Earth's shadow will take most of that moonlight away allowing fainter emissions to be visible.
Watch for substorms
A substorm is the cycle Earth's magnetotail goes through as it loads up with energy and then lets it go, and it is what turns a faint green band into a sky full of moving color. During the growth phase the aurora sits low and quiet on the horizon. Then the expansion phase arrives and it brightens and spreads within a couple of minutes, sometimes pushing much farther south than the forecast map suggested. Then it settles back down in the recovery phase, and the whole cycle can start over.
On a night like this one, substorms matter more than the headline number. If you step outside and see nothing, give it some time before you pack up.
Where and when to look
An early arrival puts Thursday night in play for North America, with the best odds along the U.S.-Canada border: Washington, Idaho, Montana, the Dakotas, Minnesota, Wisconsin, Michigan, upstate New York, northern New England, and most of Canada. A substorm could push the aurora further south than this.
On-time, around 11 UT Friday, the timing works out much better for the southern hemisphere, where that lands on Friday evening. Tasmania, southern Victoria, and the South Island of New Zealand are the places to be for that one, and they have got long dark nights right now.
A late arrival or strong CME brings Friday night back into play for North America, which is what HUXt is arguing for.
Wherever you are, face north (or south, if you are in the southern hemisphere) and give it time. Across the northern U.S. this time of year you have real darkness from late evening until a few hours before sunrise, so look up the times for your own spot.
Up here in Fairbanks we have got our dark hours back, but I am still in Los Angeles to finish out my internship, so I will be watching whatever Maia is able to capture on her tours up in Fairbanks!
Don't lean on the Kp forecast or aurora apps
The 3-day Kp forecast is a planning tool rather than a schedule. It is built on a modeled arrival time that carries at least 10 hours of uncertainty on this event, so those tidy 3-hour boxes are not going to line up with what the aurora is actually doing. Most aurora apps run on the same numbers, which is why they will tell you it is happening when it is not, and stay quiet when it is.
Webcams and reports from people on the ground are the ground truth here. Solar wind data and the GOES magnetometers are useful real-time proxies on top of that, and they will tell you the CME arrived and the "charge state" of a substorm.
Check your clouds
The aurora happens roughly 60 to 200 miles up, far above all weather, which means any cloud at all sits between you and it. Even thin high cloud will mute displays.
This is an early forecast and it will change, but as it stands right now the East Coast looks poor for Thursday night and the northern Plains look better. Check your own local forecast before you commit to a drive, and then check it again on Thursday.
Want to track this yourself?
A few free guides on my site that cover the tools I use on a night like this: how to spot substorms, how to use the GOES magnetometers, my list of 100+ aurora webcams, and a walkthrough of the HUXt model. If you are starting from scratch, my free beginner's guide to aurora chasing covers the basics in one place.
A few quick tips for the night
Get away from city lights, find a clear view to the north, and give your eyes 15-20 minutes to adjust. Your phone's night mode or a camera on a tripod will pick up color your eyes miss, especially when the aurora is faint. If you want to go deeper on settings and technique, that is what my complete guide to aurora photography is for.
In a nutshell
An M7 flare from AR 4513 on Tuesday launched a double CME that may clip Earth on Friday, August 28. NOAA has a G1 watch for Thursday from coronal hole effects and a G2 watch for Friday from the CMEs. The predicted arrival is 11 UT Friday, give or take 10 hours, and HUXt argues for Friday evening instead. This is a glancing blow, so it may underperform or miss us entirely.
Thursday night is the eclipse night, and an early arrival would put aurora and a nearly total lunar eclipse in the sky together. Friday's predicted arrival time works better for the southern hemisphere than for North America.
Check your clouds, look north, and keep your hopes up and your expectations realistic. I wrote this a couple of days ahead of the arrival, so follow me on socials for updates as the models firm up, and get on my alert list if you want the next one of these before it hits the blog. Good luck!
A few other things from Fairbanks
The rest of this post is the news I sent along with the alert. If you came here only for the storm, that is all of it above.
I passed my written comprehensive exams!
Some good news on my end: I passed my written comprehensive exams! That was the big hurdle.
For anyone new here, alongside all of this I am a Space Physics Ph.D. student at the University of Alaska Fairbanks, studying the aurora. Next up are my oral comprehensive exam and my qualifying exam, and both are comparatively easy after the written comps. Once those are done I officially become a Ph.D. candidate, which puts me on track to graduate in spring 2028.
My thesis topic is still being decided, but it will definitely feature auroral beads and auroral substorms. Auroral beads are the small bright spots that appear along an auroral arc and line up like a picket fence when you view them edge-on. They show up in the minutes before the expansion phase begins, which makes them one of the few visible clues that a substorm is about to break out.
The best part of having this behind me is the time it frees up. I have a lot more of it this fall to put into making content and getting you timely space weather updates, and I am excited about that. If you ever want the actual papers rather than the short version, they are all on my Google Scholar page.
Maia ran her first tour of the season
Maia ran her first tour of the 2026-2027 season on Monday night in Fairbanks, and the aurora showed up for it. Aurora season is back!
Every guest on a tour with Maia or Levi gets aurora portraits, along with group shots and landscape shots, and they are all sent to you afterward. Both of them also shoot real-time video with specialized low-light cameras when the aurora is bright enough for it, and you keep those clips. That is not something every tour includes, and it is one of the reasons I send people to them.
Maia's season runs August 10 through April 20. Levi starts on November 26 and runs through April 20 as well.
If you want a private tour, book it now
I want to explain how private tours actually work on our calendars, because the way they get booked surprises people.
A private tour means you reserve the whole vehicle. There is no minimum guest count on our standard tours, so the moment one person books a standard tour on a given night, that night is off the board for a private tour. It only takes the one booking, which is why private dates go so much earlier than standard ones do.
Look at what that has already done to Maia's March 2027. March fills first every year, because the skies are usually the clearest of the season and geomagnetic activity gets a boost from the equinox (the Russell-McPherron effect I mentioned up top). Green means the night is still open for a private tour, and gray means it is gone.
So if a private tour in March is what you want, Levi is the one with room right now, and I would not wait on that either.
The same thing is already happening for this fall. This is Maia's September:
On price, a private tour is a better deal per person than you might expect. A standard tour is $350 per person. A private tour is a flat $1,500 for a group of five, which works out to $300 each, and $1,600 for six, which is about $267 each. Five or six people traveling together pay less per head for the whole vehicle than they would for individual seats.
Private tours are also the right call if you are traveling with small children, if anyone in your group has limited mobility, or if you are planning a proposal. Maia and Levi's tours do not require walking.
If your dates are already booked
Maia and Levi both cap their groups on purpose, so there are only so many nights to go around. If the nights you want with Maia are already taken, and Levi does not start until November 26, here is what I would do.
First, get on the waitlist. It is new this season for both of them, and it means I can email you if a cancelation opens a spot back up.
Second, book either Greatland Adventures or Face the Outdoors. Both run slightly larger groups, so they tend to have availability when the small-group calendars do not.
Greatland Adventures is where I would look first. They run the highest-rated aurora tours in the area, in both Fairbanks and Anchorage, so they are also the answer if your trip routes through Anchorage or you are traveling with a bigger group. $325 per person, ages 13+, about 8 hours, transportation included. Private tours are a flat $1,995 for up to eight people, plus $100 each up to twelve.
Face the Outdoors has limited availability in September, but their calendar opens wide in October, which makes them a good fit for a fall trip. Theirs is a lodge tour. Pickup is still from Fairbanks, then about 90 minutes south to their private Aurora Viewing Lodge, where you wait for the lights in the warm with snacks, coffee, and a bathroom, under dark skies with wide-open views all around. They are the only lodge tour I know of that loads the van and chases when the sky starts to cloud up, so you are not stuck at the lodge. $325 per person, ages 8+ and all ages on a private tour, about 9 hours, pickup around 8 PM and back around 5 AM. They are also a good option if anyone in your group has limited mobility.
A new partner: Midnight Sky Adventures
I have got a new partner to tell you about: Midnight Sky Adventures, run by Ron Brown out of North Pole, Alaska.
This one fills a gap I have been getting asked about for a long time. Ron runs an Arctic Circle and Northern Lights tour up the Dalton Highway, and until now I had nobody to send you to for it.
It is a 14-hour day. Pickup is late morning from hotels and Airbnbs in Fairbanks and North Pole, and you are back in town by about 2 AM. You cover close to 200 miles of the Dalton, stop at the trans-Alaska pipeline, break for dinner at the Yukon River Camp, and stand at the Arctic Circle wayside at 66°33' north. Snacks, water, and hot drinks are included, and there is a grocery stop in Fairbanks before you leave. The van is all-wheel drive with a guaranteed window seat for everyone, capped at six guests in winter and seven in summer. $369 per person, minimum age 15.
Ron watches real-time solar wind data on the drive back down the Dalton and picks his pullouts accordingly. That is exactly how I would run it, and it is the reason I wanted to work with him.
You can read the full writeup on my Fairbanks Visitor's Guide, or go straight to dates and booking.
The 2027 Awesome Auroras calendar is coming
My 2027 Awesome Auroras calendar launches this October, and the design is coming together now.
The first 100 people on my email list to buy one will get it for 20% off. That is the same deal I ran on the 2026 calendar, and it went quickly.
Two pages of that calendar are yours
There is one more thing about the 2027 calendar. I want two of your photos in it!
The June and July pages are going to aurora chasers on my email list. The Sun barely sets in Fairbanks in June and July, so there is not enough darkness for aurora viewing either month, which is why those two pages have always been the odd ones out here. This year I am giving them away.
June is decided by a public vote. July is the one I pick by hand. You can enter either or both. Each winner gets $100 and a free calendar, with their name and their own description of the shot printed next to the photo.
Your photo has to have been taken during the month you are entering, but it can be from any year, so there is a good chance the shot is already sitting on your hard drive.
Submissions open Monday, August 31, and this one is for my email list. I will send out the full rules and the entry form before then, so make sure you are on the list.
Final thoughts
This one could go either way. Watch the solar wind, check your clouds, and if you are far enough east, go outside on Thursday night for the eclipse whether or not the CME turns up.
Happy chasing!
— Vincent (The Aurora Guy)
One more thing: my Aurora Borealis Magic Mug. Pour in a hot drink and the northern lights appear on the cup. It is my most popular product, and it helps keep these alerts free.
