Moon Phase Finder

A lunar phase is a geometric quantity, not a physical change in the Moon. Half the sphere is always sunlit; what varies is how much of that lit hemisphere faces Earth, which is set by the Sun-Earth-Moon elongation angle. This tool tracks that angle through the moon age: the number of days elapsed since the most recent new moon, obtained by taking the days since a reference new moon modulo the mean synodic month of 29.53059 days. Age maps to phase in eighths of about 3.69 days each, so the quarter points fall near 7.38, 14.77 and 22.15 days. Illumination is then computed as (1 - cos(phase angle)) / 2, giving 0 percent at new, 50 percent at the quarters and 100 percent at full. That model gives the illuminated fraction of the visible disc, which is why a crescent looks fainter than its percentage suggests: brightness falls faster than area because of limb darkening. The main limitation is that a fixed 29.53059-day period is only an average. Because the lunar orbit has an eccentricity near 0.055 and is perturbed by the Sun, the actual interval between successive new moons ranges from roughly 29.27 to 29.83 days, so a uniform-rate model can drift by up to about half a day from the true instant of a phase. Professional ephemerides such as the JPL DE series and the routines in the IAU SOFA library evaluate hundreds of periodic terms instead. For observation planning, gardening, tide expectations or photography this approximation is adequate; for eclipse prediction or occultation timing it is not, since those need accuracy measured in seconds.

Systematic lunar phase prediction is one of the oldest quantitative sciences. Babylonian scribes recorded new-moon sightings in the astronomical diaries from at least the eighth century BC, and by the Seleucid period their arithmetic lunar theories, preserved in the tablet systems now labelled A and B, predicted lunation lengths well enough to anticipate eclipse possibilities. The 19-year cycle in which 235 synodic months almost exactly equal 19 solar years was known in Babylon and is named in the West for Meton of Athens, who introduced it around 432 BC. Hipparchus of Rhodes, working in the second century BC, derived a mean synodic month accurate to within about a second of the modern value. The modern figure of 29.530588 days comes from lunar theories developed in the eighteenth and nineteenth centuries, notably by Tobias Mayer, Peter Andreas Hansen and Ernest W. Brown, whose 1919 tables governed published almanacs until computers replaced them. Simple modulo methods of the kind used here descend from the ecclesiastical epact rules medieval churchmen used to date Easter, and were popularised in computing form by John Horton Conway and by Jean Meeus, whose 1991 book Astronomical Algorithms remains the standard reference for compact phase and illumination formulas. Authoritative tables today come from the US Naval Observatory and from NASA, working from numerically integrated ephemerides rather than a fixed mean period.