Home PageMay 15-16, 2003 Total Lunar EclipseArticle Page

The Saros Eclipse Cycle

For a lunar eclipse to take place, the Moon has to be full. Secondly, the Moon has to be appreciably close to one of its two nodes -- points where the Moon's orbit intersects the ecliptic (Earth's orbital plane). At the descending node the Moon moves from north to south, and at the ascending node it moves from south to north. (See May eclipse article.)

The discontinuity between successive full Moons and the Moon's return to its node prevents lunar eclipses from happening monthly. Successive full Moons (the synodic month) average 29.530589 days, whereas the Moon's return to the same node averages 27.21222 days (the draconic month). Eclipses depend upon the synchronization of synodic and draconic periods. For example, a particular eclipse cycle has 47 synodic months (1387.938 days) corresponding with 51 draconic months (1387.822 days), featuring an eclipse every 3.8 years. This, however, represents but one of many eclipse cycles; and almost always, when one speaks of an eclipse cycle, it's in reference to the Saros.

Reportedly, the Chaldeans knew of the Saros a good many centuries before the birth of Christ. 223 synodic months (6585.321 days) almost equal 242 draconic months (6585.357 days), a period of 18 years 10 and 1/3 (or 11 and 1/3) days -- depending on the number of intervening leap years. When you observe a lunar eclipse, you can figure one Saros period (223 full Moons) later that the eclipse will take place about 8 hours later during the day, and about 120 degrees west of the previous eclipse. The second Saros (446 full Moons) afterwards, the eclipse comes some 16 hours later and 240 degrees westward. The third Saros (669 full Moons) finds the eclipse coming full circle, happening at about the same time and place. This Triple Saros is called an Exeligmos, a period representing some 54 years and 34 days.

Regarding the May 15-16, 2003 total lunar eclipse, the Triple Saros is projected into the year 2057. Click on the date to see an eclipse map, all maps courtesy of Fred Espenak, NASA/GSFC.

DateTime of Mid Eclipse
May 15, 200323:40 Eastern Daylight Time
May 26, 20217:18 Eastern Daylight Time
June 6, 203914:53 Eastern Daylight Time
June 16, 205722:24 Eastern Daylight Time

As is evident in the table above, the approximation fits quite nicely. You can project the Saros of any eclipse (whether it be lunar or solar) into the future with reasonable assurance -- though there are around 40 different Saros in operation at any one given time. It's thought that ancient sky watchers were more adept at keeping track of Lunar Saros Cycles than Solar Saros Cycles. Yet, tales abound about the philosopher Thales using the Saros to predict the solar eclipse of 585 B.C.

Saros cycles don't last forever or indefinitely, because realignments of full Moon with its node are not perfect. A typical Saros series consists of 69 to 86 eclipses spanning well over a thousand years. The May 15-16, 2003, total lunar eclipse is the 55th of 84 eclipses in Saros Series #121. Beginning in the year 1029 and ending in 2526, Series #121 in the table below shows the following sequence: penumbral, partial, total, partial, penumbral.

Lunar Saros Series #121

Type of Lunar EclipseDate of Eclipse
First PenumbralSeptember 25, 1029
First Partial UmbralApril 29, 1390
First TotalJuly 13, 1516
Last TotalMay 26, 2021
Last Partial UmbralAugust 11, 2147
Last PenumbralMarch 29, 2526

All Saros Series follow this progression. If the Saros Series has the Moon at the descending node (moving north to south), which #121 does, the first eclipse starts as penumbral at the southjern limb of the lunar disk, and the last eclipse of the series ends as penumbral at the Moon's northern limb. A Saros Series with the Moon at the ascending node has this same progression, except that it starts at the Moon's northern limb and finishes at the southern one.

I am greatly indebted to Fred Espenak for the information in this article, and highly encourage you to visit his web site .

by Bruce McClure

Five Thousand Years of Lunar Eclipses