Average cloud cover over Spain considering cloud shadows

https://view.eumetsat.int/ Jul 31 18:10 UTC Eumetview

Weather is always a critical factor when observing a total solar eclipse. For the August 12, 2026 eclipse in Spain, the challenge is even greater: the Sun will be very low on the horizon, meaning visibility depends not only on overhead conditions but on the entire line of sight, which spans hundreds of kilometers.

Meteorologist Jay Anderson, in his very detailed weather statistics analysis for upcoming eclipse (Total Solar Eclipse 2026 August 12 | Eclipsophile), in the section “What can go wrong?”, also clearly warns about this challenge:

Cloud-cover statistics do not make allowances for intervening clouds – they only reflect the overhead sky condition. Pay particular attention to the cloud shadows in the images—they will be much longer at totality.

Not all clouds obstruct in the same way

In our previous article Which clouds can ruin your eclipse view – Besselian Elements we analyzed which types of clouds can block the view of the eclipse depending on their altitude: low, medium and high. We’ve also added an indicator to our Eclipse Map 2026 – Besselian Elements map to help you understand how far the clouds on the horizon may be along the viewer’s line of sight.

A concrete example: Mallorca

In Mallorca, the Sun will be visible toward the northwest, at just 2.6° elevation. Under these conditions:

High clouds or distant storms may align directly with the Sun.
A high cloud even 200 km away can obstruct the view.

This makes “local” weather forecasts insufficient for evaluating eclipse visibility.

Are local forecasts really useful?

It is natural to ask: “Is the probability of clear skies at the observation point enough to predict the visibility of this eclipse?” When the Sun is at its zenith, yes. When the Sun is almost on the horizon, no.

When the Sun is high, local clear-sky probability is a good predictor. Near the horizon, it is not. The sunlight passes through multiple cloud layers spread across large regions, requiring a more detailed assessment of cloudiness along the entire viewing path, not just overhead.

Methodology: Using DWD Cloud Data

We used cloud-layer data (low, medium, high) from the German Weather Service (DWD) www.dwd.de

and analyzed the entire summer of 2025, limiting the dataset to the 30-day window centered on August 12 to reflect stable, representative summer conditions in Spain.

For each location, we calculated:
– The Sun’s elevation at totality (~18:30 UTC)
– The coordinates where the line of sight intersects each cloud layer
– The probability that no cloud layer obstructs the Sun along that path

The resulting map

By adding up the contributions of the various layers, we obtained, for each point in Spain, the probability of seeing the eclipse free of clouds along the horizon.

This approach:

Reflects the actual geometry of the event
Highlights increased cloud obstruction risk where the Sun is lowest

Provides visibility estimates based on line‑of‑sight conditions, not simply overhead cloud cover

Conclusion

For the 2026 eclipse, it is not enough to look at the local weather. We must consider all the atmosphere between us and the Sun, which in this case runs close to the horizon for hundreds of kilometers.

With our real-world data-based approach and line-of-sight modeling:

we estimated the probability of clear skies for each point in Spain;
we generated a map that reflects the true geometry of the event;
We have offered a more realistic way to assess the visibility of the eclipse.

Note: How to combine the clouds of the various layers?

For each location we reconstructed the geometry of the line of sight and verified:

which portions of the sky intercept the cloud layers;
what is the percentage of coverage in each of them.

To combine the cloud covers, we adopted the hypothesis of independent layers: we assumed that the percentage of cloud cover in one layer does not depend on the percentage of cloud cover in the other layers.

We can express the overall percentage of cloud cover (0 = no clouds, 1 = overcast) as:

CLOUD COVER PERCENTAGE = 1 – CLEAR SKY PERCENTAGE

The independent layers assumption allows expressing the overall clear sky percentage as:

CLEAR SKY PERCENTAGE = CLEAR SKY PERCENTAGE IN THE LOW LAYER x

CLEAR SKY PERCENTAGE IN THE MEDIUM LAYER x

CLEAR SKY PERCENTAGE IN THE HIGH LAYER

If we call TOTAL the overall cloud cover percentage, LOW the cloud cover percentage in the lower layer, MEDIUM the cloud cover percentage in the intermediate layer, and HIGH the cloud cover percentage in the higher layer, we can finally write:

TOTAL = 1 − (1 − LOW) (1 − MEDIUM) (1 − HIGH)

In reality, we know that, especially between low and middle layers, there are correlations: clouds often have a vertical development that occupies several levels. We tested the formula against real total coverage, finding that:

the approximation works surprisingly well;
the correlation exists, but it is slight and not significant enough to compromise the accuracy of the model.

In our case, however, the situation is even more favorable: the cloud layers intercepted by the line of sight are often geographically very distant, so the hypothesis of independence becomes even more reasonable.