Deep Space Tuesdays · the far-out beat

From The Decisional Pass

Tomorrow afternoon the Moon's shadow crosses Greenland, clips western Iceland, runs across the North Atlantic and comes ashore in northern Spain just before sunset. For about two minutes at any given spot, the Sun's corona will be visible to the naked eye.

Several teams have already published what they think it will look like. Tomorrow they find out if they were right.

That framing — an eclipse as a scheduled exam for a physics model — is the thing I want to hand you today, because it's not how most coverage treats it, and it's the reason ESA has three separate assets pointed at this.

Where and when

The path of totality crosses eastern Greenland, western Iceland, northern Russia, the North Atlantic, the northern half of Spain, and a small corner of northeastern Portugal.

• Reykjavík goes dark around 17:48 UTC, with roughly a minute of totality
• León, Spain gets it much later in its local evening — about 20:28 local, low and close to sunset
• Maximum totality anywhere on the path is about 2 minutes 18 seconds, per ESA. NASA's guidance is blunter: most people inside the path will get less than two minutes
• A partial eclipse is visible across much of Europe, the UK and Ireland, northwestern Africa, most of Canada, and the northern US from Alaska to North Carolina

It is the first total solar eclipse visible from mainland Europe since 1999, and the first from the Iberian Peninsula in more than a century.

The obligatory and genuinely important note: during the partial phases you need ISO 12312-2 solar viewers. Sunglasses are not eclipse glasses, and no amount of stacking makes them one. The only moment it's safe to look directly is totality itself. NASA streams from 1:15 p.m. EDT if your sky is the wrong one.

The exam

Here's the part that makes this a deep-space story rather than a travel story.

The corona is the Sun's outer atmosphere, and it is where essentially every space-weather problem we have begins — the solar wind that erodes satellite orbits, the storms that trip power grids, the particle events that make crewed flight beyond low Earth orbit a radiation-management problem. We model it. We would very much like to know how well we model it.

The trouble is that verifying a corona model is hard, because the corona is a million times fainter than the disc in front of it. Normally you can only see it by manufacturing an artificial eclipse inside a coronagraph, and every coronagraph has a blind zone right where the interesting physics is.

So: ESA has partnered with Predictive Science Inc. and KU Leuven — whose model is named Coconut — to generate simulations forecasting what tomorrow's corona should look like. Predictive Science's outputs include radial-filtered K-corona brightness and the Fe XIV green-line emission. These are on the record, before the event.

Jorge Amaya, ESA's space weather modelling coordinator, put the logic plainly: a total solar eclipse is how you verify the models are correct.

I love this because it inverts the usual relationship. Ordinarily nature is the thing you're trying to photograph. Here nature is the answer key, published once every few years, at a time and place fixed by celestial mechanics and available to anyone who can get under the shadow with a camera.

The machine that makes its own eclipses

The second asset is the reason I've been waiting to write about this.

Proba-3 is two spacecraft flying in formation about 150 metres apart, holding that separation to millimetre precision. One of them, the Occulter, is a flying disc whose entire job is to cast a shadow onto the other one, the Coronagraph, which carries an instrument called ASPIICS.

It is, functionally, a total solar eclipse that you own and can schedule.

The numbers are absurd in the best way. Since July 2025 the pair has produced 57 artificial eclipses and collected more than 250 hours of high-resolution coronal video — an amount of coronal observing time roughly equivalent to 5,000 ground-based eclipse expeditions. ASPIICS sees down to about 70,000 km from the solar surface, a tenth of a solar radius, closer in than any other space-based coronagraph manages.

That capability has already changed something. Proba-3's first science results found that the "slow" solar wind in the low corona is moving at 250–500 km/s — substantially faster than models had it. When the slow wind isn't as slow as your model thinks, your acceleration physics is wrong somewhere, and every downstream space-weather forecast inherits the error.

For tomorrow, ASPIICS is running continuous coronagraphy from 28 July through 19 August. Andrei Zhukov, the ASPIICS principal investigator at the Royal Observatory of Belgium, noted that eclipse day marks the 62nd eclipse of Proba-3's nominal operations.

Now, an operational detail I find delightful, and which nobody advertises: Proba-3's observing windows on August 12 don't cover the eclipse. Its imaging runs 03:38–09:23 UTC, then picks up again at 23:18 UTC and runs into the 13th. Totality happens in the gap. Proba-3 is on a highly elliptical orbit and can only do this trick near apogee, so its schedule is set by orbital mechanics, not by what would be convenient.

Which is why Zhukov's approach is to mirror observations taken two weeks earlier to predict the corona's appearance. Two weeks is roughly half a solar rotation — the face of the Sun that will be turned toward Spain tomorrow is, broadly, the face that was turned away a fortnight ago.

That's a real constraint handled with a real trick, and it's the kind of thing I'd never have appreciated before I spent time in operations. You don't get the observation you want. You get the one your orbit allows, and then you do arithmetic.

The wildcard on the far side

Third asset, and the reason tomorrow's prediction is harder than it looks.

The ESA-led Solar Orbiter is running a special observing campaign, and its Polarimetric and Helioseismic Imager has picked up a newly emerging active region — one not visible from Earth's vantage point. New active regions restructure the magnetic field that shapes the coronal streamers, which means the corona people photograph tomorrow may not be quite the corona the models were built to predict.

That's not a failure. That's the experiment getting interesting. A model that nails a quiet, well-behaved corona has told you less than a model that gets tested against a Sun doing something unannounced.

Solar Orbiter can see it because it isn't parked where we are. That's the whole case for flying spacecraft away from the Earth–Sun line: not novelty, but coverage of the hemisphere that's about to rotate into view and ruin your forecast.

What I'll be watching for

Not the photographs. There will be thousands of gorgeous photographs and you will see them everywhere by Thursday.

I'll be watching for the comparison images — the published prediction laid next to the observation, streamer by streamer. That's the actual scientific product of tomorrow, and it's the one thing you can only get on a day when the Moon does the occulting for free, out to solar radii no instrument can reach.

Two minutes and eighteen seconds. Bring the right glasses.

— Elio

References

• Total Solar Eclipse on August 12, 2026 — NASA Science
• How ESA mimics and models the 2026 total solar eclipse — ESA, August 10, 2026
• Proba-3 and the total solar eclipse of 2026 — ESA Cosmos
• First Proba-3 science: surprisingly speedy solar wind — ESA
• Proba-3's first artificial solar eclipse — ESA
• Total solar eclipse 12 August 2026 – Global map — ESA
• Total Solar Eclipse on August 12, 2026 — timeanddate.com

Forwarded this? Get The Decisional Pass in your inbox: