Accurate prediction of Sun's position at low altitude -- total solar eclipse August 12, 2026

[Ramon Oliver]

Hi Stellarium!

From the Balearic Islands, Spain, the total phase of the August 12, 2026 solar eclipse will be observable at a very low height above the horizon, around 2.5º. I would like to know how accurate is the Sun's position provided by Stellarium for objects so close to the horizon. What is the typical error of the azimuth and altitude of the Sun, in particular when the altitude is small?

The reason for asking this is that it is very hard to predict the visibility of the totality for all locations in the islands: any small elevation can hide the totally eclipsed Sun. On April 28 and 29, 2026 the Sun follows a very similar path on the sky as on August 12. I want to use Stellarium to find the range of times on April 28 and 29, 2026 when the Sun is the nearest to the position of the totally eclipsed Sun on August 12, 2026. This will allow anybody to test well in advance whether the total phase will be observable from their favourite location.

Thanks in advance,

Ramon

[Georg Zotti]

Stellarium uses the refraction model from Jean Meeus' Astronomical Calculations. In that altitude the atmospheric effects are unpredictable (maybe even a few arcminutes? Rather less, but please do not quote me, search for refraction deviation from standard atmosphere models. And tell us your results! Error will be larger below 1°.). Azimuth should generally not be affected, but of course a "contact point of Sun&horizon", like a faraway monument that you want to frame with the eclipsed sun behind, may change when refraction is an arcminute different. Testing a site on the corresponding date in April is a great idea! (Under certain circumstances, refraction may still differ slightly between April and August, though. This is really an interesting experiment!)

You can experiment with refraction settings for air pressure and temperature for a general feeling of "how much error" to expect. This can influence "average" refraction, but we cannot model things like Fata Morgana.

[Ramon Oliver]

Hi Georg,

Thanks for the quick and informative reply! As you suggested, I played a little bit with the refraction parameters of Stellarium (wow, did not know you could do so!) and found out that at my location and with a solar altitude of 2.5º, the maximum variation of the Sun's height is 2 arcmin; and this was after using a very high temperature and very low pressure. This uncertainty is more than enough for our purposes because my estimate is that we do not need to achieve a precission higher than 10-15 arcmin.

Cheers,

Ramon