Would you rather have the longest totality of this eclipse, or a sky clear enough to watch it through?
Welcome to FreeAstroScience. We put NASA’s path data next to Jay Anderson’s cloud climatology for 12 August and found that the two answers point at different countries, so we ran the airmass arithmetic to see what the low Sun actually costs you in Spain.
The greatest duration of this eclipse is 2 minutes 18.2 seconds, and NASA’s prediction page marks the point where that happens as a separate spot from the point of greatest eclipse. Jamie Carter, writing in Sky and Telescope, places it off the coast of Iceland. Nobody is standing there. What NASA Science tells the rest of us is blunter: most people inside the path will see the Sun fully covered for less than two minutes, and even close to the center line it stays under two and a half.
Iceland owns the seconds and the Ebro owns the sky
Jay Anderson has been publishing eclipse-track climatology at Eclipsophile for years, and his numbers for 12 August cut the path clean in half.
Iceland has the geometry. It also has the North Atlantic storm track running straight through it, which is why Anderson puts island-wide cloud cover at 70 to 80 percent and records that Reykjavik sees sunshine in only 32 percent of its August daylight hours. Those are not comfortable odds for a two-minute event.
Spain then splits again inside itself, and the split runs north to south. Anderson calls the Bay of Biscay coast an uninviting destination for eclipse viewing, with cloud amounts universally close to 60 percent. Move inland and the picture inverts. Leon, Burgos and Valladolid run 68 to 78 percent of possible sunshine and would have shown the eclipse on 17 days out of 21. Best of all is the floodplain of the Ebro around Huesca and Zaragoza, under 30 percent cloud and clear on 18 years in 21.
Table 1 — Cloud climatology along the 12 August 2026 track (Anderson, Eclipsophile, 2026)
| Location | Mean cloud cover | Clear on the date |
|---|---|---|
| Reykjavik and west Iceland | 70 to 80 percent | not given as a count |
| Bay of Biscay coast | close to 60 percent | 10 of the August 12s |
| Leon, Burgos, Valladolid | not stated | 17 of 21 years |
| Huesca and Zaragoza | below 30 percent | 18 of 21 years |
One row of that table is deliberately ragged. Anderson writes that the Biscay coast would have delivered on 10 of the August 12s he reviewed, and he does not say out of how many, while every other figure on his page is stated as a count out of 21. Assuming the denominator would give us a tidy 48 percent and a false one.
As of today, 1 August 2026, the eclipse is eleven days out and climatology is still the strongest thing anyone can hand you.
A Sun at four degrees sits behind fourteen atmospheres
Anderson gives the Sun 10 degrees of elevation on the Spanish north coast at totality, and 4 degrees over the Balearic Sea. Every page we read skates past that, so here is the arithmetic. Airmass is the standard measure.
X ≈ 1 / sin h
Here X is the airmass, meaning the depth of atmosphere along your line of sight expressed in units of the depth straight overhead, and h is the Sun’s altitude above the horizon. Feed in Anderson’s 10 degrees and you get 5.8. Put in his 4 degrees and the same expression returns 14.3. Said plainly: from the Cantabrian coast you would view the corona through nearly six times the air a midday Sun sits behind, and from the Balearic shoreline the formula says fourteen.
That expression assumes a flat slab of atmosphere, which stops being true anywhere near the horizon, so 14.3 is a ceiling and not a measurement. Nobody disputes the direction of the effect. Light arriving at that angle is reddened and dimmed, the corona loses contrast against a sky still holding sunset color, and a single band of haze sitting on the horizon ruins everything while a clear zenith saves nothing. Anderson puts it in one line: the view to the Sun must be clear at the horizon whatever may be present in the rest of the sky.
He also names what tends to be sitting there. Saharan dust reaches the peninsula as a calima, tinting the sky orange and cutting visibility to very low values, and a second kind of calima is now common enough to plan around: haze from forest fires. Anyone who read our explainer on the Saharan Air Layer and Europe’s 2026 heat dome already knows how routinely that air makes the trip north.
None of this is bad luck, and none of it is a Spanish problem. NASA Science has totality falling at mid-day for a remote corner of northern Russia, in the late afternoon or early evening for Greenland and Iceland, and a little after half past eight in the evening for Zaragoza. A track that runs from Arctic noon to Iberian sunset was laid down by a shadow grazing the top of the planet instead of driving through its middle, and a grazing shadow buys length at the price of altitude. The 2 minutes 18.2 seconds and the 4-degree Sun are not two separate facts. They are one fact, read twice.
Two published altitudes for the same Sun
Sky and Telescope does not agree with Eclipsophile about how low that Sun gets. Carter, writing in April 2024, gives 12 degrees above the northwestern horizon in Spain sinking to 2 degrees above the Balearic Sea, and 5 degrees at the Castillo de los Calatravos. Anderson gives 10 and 4. Neither page says which point on the track its figures belong to, and on the expression above the difference between 2 and 4 degrees is the difference between an airmass of 29 and an airmass of 14, so this is not a rounding quibble.
We use Anderson’s numbers for the rest of this piece. His page is a climatology study that states its satellite record and its period; Carter’s is a travel-planning article whose recommendation is a hotel. That is a judgment about which page shows its work, not a claim that Carter is wrong.
One thing we are leaving out on purpose. The International Meteor Organization puts the Perseid maximum on 12 and 13 August at a zenithal hourly rate of 100, and notes that the Moon is new on 12 August, so conditions are optimum for the shower. That new Moon is this new Moon. Earth’s only natural satellite works two jobs that night: it covers the Sun over Zaragoza at sunset, and its absence a few hours later is exactly what leaves the sky dark enough for the shower. Knowing that is a pleasure. It is not a reason to bolt a meteor guide onto the end of an eclipse argument, so the Perseids will get their own page.
The Moon’s mountains move the edge by three kilometers
NASA’s prediction page carries a warning that almost nothing quoting it reproduces. The predictions do not include the effects of mountains and valleys along the edge of the Moon, and the page states the cost: duration can shift by one to three seconds, and the position of the path by one to three kilometers.
Against a greatest duration of 2 minutes 18.2 seconds, three seconds is roughly 2 percent, which is a shrug. Set beside a path Carter gives as 293 kilometers wide, three kilometers is nothing at all. Standing three kilometers inside what you believe is the edge of totality, it is the whole trip.
The same page prints the machinery it used, which is why we trust it: the VSOP87 and ELP2000-85 solar and lunar ephemerides, and a delta T value of 71.4 seconds. Set that against the times you can actually plan from. NASA Science gives Reykjavik totality as 5:48 to 5:49 p.m. and Zaragoza as 8:29 to 8:33 p.m., rounded to the minute, which is enough to organize a day around and not enough to compare two sites properly. Timeanddate lists maximum eclipse at 17:46:06 UT, quoted to the second, without saying what it means by maximum. We could not tell from either page whether that instant is the one NASA marks as greatest eclipse, and we are not going to guess.
Everything here is about where to stand rather than how to look. Filters, the ISO 12312-2 standard and the single moment when the glasses may come off are covered in our guide to safe eclipse viewing on 12 August, which is the companion to this piece and not a repeat of it.
For the other half of the geometry, our piece on how the Moon’s apparent size changes with its distance explains why some new Moons cover the Sun completely and others leave a ring of it showing.
Where we would stand on the twelfth
The Ebro. Under 30 percent cloud and 18 clear August 12s in 21 beat a greatest duration of 2 minutes 18.2 seconds that exists only over open water, and they beat Reykjavik’s 32 percent of sunny daylight hours by a margin no extra seconds repair. The cost is altitude, a Sun at 4 to 10 degrees seen through six to fourteen atmospheres, with a horizon that has to be clean even when the rest of the sky need not be. We think that is the right trade, and we would rather be argued out of it than agree with everybody early. Come back when 12 August enters the medium-range forecast and we run it against Anderson’s climatology in public. FreeAstroScience, Rimini. Gerd Dani.
Sources
- Espenak, F. Total Solar Eclipse of 2026 Aug 12. NASA Goddard Space Flight Center. https://eclipse.gsfc.nasa.gov/SEgoogle/SEgoogle2001/SE2026Aug12Tgoogle.html
- NASA Science. Total Solar Eclipse on August 12, 2026. National Aeronautics and Space Administration. https://science.nasa.gov/eclipses/future-eclipses/total-solar-eclipse-on-august-12-2026/
- Anderson, J. Total Solar Eclipse 2026 August 12. Eclipsophile. https://eclipsophile.com/tse2026/
- International Meteor Organization. Meteor Shower Calendar. https://www.imo.net/resources/calendar/
- Carter, J. (2024). Plan Now for the 2026 Total Solar Eclipse in Europe. Sky and Telescope. Published 10 April 2024. https://skyandtelescope.org/astronomy-news/plan-now-for-the-2026-total-solar-eclipse-in-europe/
- Total Solar Eclipse of 12 August 2026. Timeanddate. https://www.timeanddate.com/eclipse/solar/2026-august-12
