Monday, July 5, 2010

How to predict the lunar phase

So, you're an astronomy geek like me, and you want to know what the moon will be like during the Perseid meteor shower. Or you're planning a camping weekend and want to have a new moon to better see the stars. Or you're tired of how-many-cool-ways-to-break-a-bottle and you want a new party trick.

You want to calculate the phase of the moon on a given day? I can do that. It's actually pretty easy.

For 2000-2019:
Add 2 to the last 2 digits of the year.
Multiply by 11.
Add the month and day.
Cast out 30s.
The remainder is the number of days since the last new moon.

0 is a new moon; 7 is first quarter, 15 is full moon, 22 is last quarter.

For example: 5 July 2010

The moon today is thus 2 days past last quarter. It's a waning crescent and rises between midnight and dawn.

Here's the kicker: Thanks to the Metonic cycle, you can go forward or back 19 years, and the phases are the same. 1981, 2000, 2019, 2038, etc are all the same. This stays accurate for a few centuries either way; it gets off by a day every 219 years. For future dates, subtract one day for every 219 years from the calculated lunar age; for the past, add.

Just don't forget about the Julian/Gregorian switch, a 10-day gap in 1583 for Catholic countries and 13 days in 1752 for Britain et al.

Just an example of how accurate this method is: take November 12, 1594. 1594 is 418 (=19*22) years from 2012.
418 is about twice 219 years in the past, so add two days and it's 29. One day from new moon.

In fact, November 12, 1594 was a new moon and an annular solar eclipse. Over 4 centuries, this little arithmetic trick is accurate to one day.

Originally published in Sky and Telescope magazine, November 2005 issue, p.130. If you want the full text, let me know.

No comments: