Why Two Moon Phase Calendars Disagree About the Date

Reason one: the 29.53-day shortcut

The commonest way to compute a phase is to count the days since a known new moon and divide by 29.530588. That number is the mean synodic month — the long-run average interval between new moons — and the actual interval is never exactly it. The Moon's orbit is an ellipse, so it moves faster when it is closer to Earth and slower when it is further away, and consecutive lunations run anywhere from about 29.27 to 29.83 days.

The error does not cancel out within a lunation; it accumulates and then unwinds. At its worst the remainder method is roughly fourteen hours away from the truth, which is comfortably more than enough to put a full moon on the wrong calendar day, several times a year. It gives no indication when it has done so, which is what makes it worse than merely imprecise. This one is a mistake, and it is the reason the moon phase page computes positions instead.

Reason two: time zones

A phase is a single instant. The Moon reaches exactly new at one moment, and that moment happens simultaneously everywhere — what differs is the local date it lands on. A new moon at 01:30 UTC on the 12th falls on the 12th in London and Tokyo, and on the 11th everywhere in the Americas. Both calendars are right about the same event.

This is not a small effect. Roughly one phase in three falls close enough to a date boundary that some part of the world disagrees with another about which day it was, and a calendar that converts to local time without saying so gives its readers no way to tell whether they are looking at a real disagreement or a timezone one. Times on this site are labelled UTC for exactly that reason.

Reason three: what counts as "full"

The four named phases are instants, not days. But many calendars assign the names by illumination band instead — anything above about 98 per cent gets called full — which spreads a single instant across two or three consecutive days. That is friendly, and it is why a calendar can show "full moon" on both the 12th and the 13th while another shows it only on one.

A related surprise: a full moon is almost never 100 per cent lit. The Moon's orbit is tilted about five degrees to the plane of Earth's orbit, so at most full moons it sits slightly above or below the exact alignment and a sliver of the disc stays dark. The full moons that do reach a true 100 per cent are the ones where the alignment is exact — and those are lunar eclipses, where the Earth's shadow immediately takes the light away again.

How to tell which you are looking at

A calendar that prints an exact time and names its time zone is doing the real calculation; one that prints only a day has either done the real calculation and rounded, or used the shortcut, and you cannot tell which. A calendar that names the same phase on two consecutive days is using illumination bands. A calendar that disagrees with a well-sourced one by a day, near a date boundary, in a different time zone, is probably not wrong at all.

The moon phase page here prints the exact instant in UTC, names a day for a principal phase only when that instant falls inside it, and the method page describes the external table its output is checked against.