How we make our numbers

Last reviewed 24 October 2026.

Most pages here rest on two things we work out ourselves: how long totality lasts at each place, and whether a meteor shower is worth a trip. This page shows the method, the tests, and the limits.

Eclipse lengths for each place

We use the open-source Astronomy Engine library (MIT licence, version 2.1.19). For each place we give it the latitude, longitude and height above sea level, and it returns when the eclipse starts, peaks and ends there, how much of the Sun is covered, and how long totality lasts. We then convert the times to local time with the time-zone rules known today.

The correction we apply

At the point of greatest duration (Point of greatest duration, 26 deg 48.8' N, 31 deg 08.0' E), EclipseWise publishes 383.24 seconds. Our code gives 386.3 seconds there, so it runs about 3.1 seconds long. We subtract 3.1 seconds from every place. This is a single flat correction, chosen at one point and tested at others.

The test

After the correction, we compared three points that were not used to set it:

Two of the three published figures are rounded to the second or the minute by their authors, so differences of a second or two are inside the rounding. Three points is a small test. It is enough to tell us the method is sound, not enough to promise any single number to the second.

What the numbers mean for you

• Lengths are for one point per place, usually the town centre. Move a few kilometres and they change, a lot near the edge of the path.
• Totality can be shortened by hills, buildings and cloud on the horizon. We do not model any of these.
• Calculated on 2026-10-24. We will recalculate if the time-zone rules or the library change.

The 2027 eclipse in 215 cities

For the city pages we run the same calculation for each city centre. We compared twelve big cities with National Eclipse, 2027 partial solar eclipse cities (read 24 October 2026). Their figures were not used to set anything.

Our times run about eight to ten seconds earlier throughout, most likely a small difference in how the two calculations model the Earth’s rotation. City pages show times to the minute, so it does not change what you read there.

Every eclipse to 2030

For the list of eclipses we use the same library to find every solar and lunar eclipse to the end of 2030, and work out what 47 cities see. We do not apply the 2027 correction there, so lengths can run a few seconds long. We compared them with published figures that we did not use to set anything:

The Seville figure is published only to the nearest minute. Our city points are the city centre at sea level, so a few seconds either way is expected. The type and date of every eclipse on the list match NASA and National Eclipse.

Meteor-shower verdicts

Each shower gets one of three verdicts, from two numbers only: how many meteors you could see in the best case, and how bright the Moon is on the peak night.

• Go: 20 meteors an hour or more, and the Moon 35% lit or less.
• Skip: fewer than 10 an hour, or the Moon 80% lit or more.
• Maybe: everything in between. If the Moon is waxing, it sets in the evening or night and the dark hours come after it sets. If it is waning, it rises late and the dark hours come before it rises.

The rate is the ZHR (zenithal hourly rate) from the International Meteor Organization's calendar. It is the number one observer would count under a perfectly dark, clear sky with the shower's source point overhead. Real counts are lower, often much lower from a town.

Moon light is our own calculation: Astronomy Engine 2.1.19; Moon light at 00:00 UTC on the second date of each peak night. We checked it against the IMO's own Moon figure for all twelve showers on our list and it agrees within 1 percentage point.

The verdict does not look at your weather, your horizon or your light pollution. It tells you whether the shower and the Moon are on your side. Whether it is clear where you are is up to the forecast.

How the pages are written

Numbers come from the calculations above or from the sources listed on each page, never from memory. Text is drafted with the help of an AI assistant and then checked against those sources before it is published. If we find a mistake, we fix it and say so on the page. Our editorial policy has the details.

What we have not checked

• Cloud statistics for each eclipse location. Until we have a source we can verify, we do not rank places by weather.
• Hotels, tours, flights and prices.
• Visa and entry rules, and government travel warnings. Check them yourself for every country.

Found an error?

Write to us. Tell us the page and the number, and, if you can, where the correct one is published.