Comparing lunar-day concepts across systems
The comparison record
“Lunar day” can describe different things. In a Chinese calendar, it ordinarily means a numbered civil day within a lunar month. In Tibetan-derived calculation, it can mean a variable-length unit of lunar phase whose number is assigned to a calendar day. Similar-looking month and day numbers can therefore describe different kinds of date.
A comparison also needs a common starting point: either the same instant or a civil date under a stated convention. An instant can be expressed in several time zones. A civil date covers a whole day, whose boundaries depend on the calendar and reference time.
| Field | Detail to preserve | Why it matters |
|---|---|---|
| Comparison target | Exact instant with time zone, or a stated civil date | Distinguishes a point in time from a whole day |
| Civil calendar | Gregorian or Julian, with year, month and day | The same written numbers can identify different days |
| Traditional calendar | Chinese, Vietnamese, Korean, or a named Tibetan-derived version | Identifies the rules behind the date |
| Reference convention | Standard time, reference meridian and day boundary where relevant | Determines which day contains an event |
| Year | Full year or era designation, including cycle information when needed | Avoids ambiguity in recurring year names |
| Month | Number or original name, plus ordinary or leap status | Distinguishes two months with the same designation |
| Lunar day | Number and, where applicable, which occurrence of a repeated date | A bare number may identify more than one day |
| Calendar record | Publication or calculation version used for the conversion | Preserves the basis of the result, especially for historical dates |
Keep the original date and its civil equivalent together. A translated month name alone does not show which rules produced the conversion.
Different boundaries
In the Chinese calendar, lunar day 1 is the civil day containing the astronomical new moon. Part of day 1 may therefore occur before the new-moon instant. Subsequent day numbers advance with the civil dates, through a month of 29 or 30 days. Vietnamese and Korean calendars use the same general kind of numbered day, although their reference times can put an astronomical event on a different date.
Day 15 is a position in that sequence, not a precise full-moon timestamp. The interval between new moon and full moon varies, and day 1 begins at the civil-day boundary rather than at conjunction. Exact fullness can consequently occur before or after the fifteenth day.
In Tibetan-derived calculation, a lunar day represents a 12-degree increase in the Moon's calculated longitude relative to the Sun. Thirty such intervals make a complete phase cycle, but their durations vary. The traditional astronomical model supplies these intervals, and a conventional dawn boundary is used to assign their numbers to ordinary days.
Under the Phugpa assignment rule, an interval that passes entirely between two dawn boundaries receives no date in the ordinary-day sequence. An interval spanning two dawn boundaries supplies its number twice. The following schematic sequences show the difference.
| Printed lunar-date sequence | Interpretation |
|---|---|
| 5, 7, 8 | Date 6 is omitted |
| 8, 8, 9 | Date 8 is repeated on two successive ordinary days |
The weekdays continue normally in both cases. A leap month is separate from either pattern: it adds a month to the year rather than changing a day number within a month.
Worked examples: why lunar dates disagree
These examples compare the same written Gregorian date under each calendar’s reference clock, rather than one timestamp converted between time zones.
February 17, 2007: opposite sides of New Year
| Calendar | Lunar year | Lunar month | Lunar day | New Year in 2007 |
|---|---|---|---|---|
| Chinese, UTC+8 | 2006 | Ordinary month 12 | 30 | February 18 |
| Vietnamese, UTC+7 | 2007 | Ordinary month 1 | 1 | February 17 |
Here, check which calendar the tool uses and when month 1 begins. The same new moon falls on February 17 at UTC+7 but February 18 at UTC+8. February 17 is therefore New Year in Vietnam but the preceding year’s last day in China. This is a particular boundary difference, not a permanent one-day offset between the calendars.
January 21, 1985: the day agrees, but the month differs
| Calendar | Lunar year | Lunar month | Lunar day | New Year in 1985 |
|---|---|---|---|---|
| Chinese, UTC+8 | 1984 | Ordinary month 12 | 1 | February 20 |
| Vietnamese, UTC+7 | 1985 | Ordinary month 1 | 1 | January 21 |
Both calendars give day 1. Moving the input by one day will not explain this disagreement; trace the month numbering back to the winter solstice.
The month containing the winter solstice is numbered 11. In 1984, the solstice fell on December 21 at UTC+7, but December 22 at UTC+8. Vietnam therefore assigned month 11 to the month beginning November 23; China assigned it to the month beginning December 22. Counting forward makes January 21 month 1 in Vietnam but month 12 in China.
China’s sequence included a leap tenth month beginning November 23, 1984. Vietnam instead had a leap second month from March 21 through April 19, 1985, after which the month numbers aligned again.
These documented differences do not indicate a broken Chinese converter. However, two tools claiming the same Chinese calendar, UTC+8 reference and civil-date rules should agree. If they disagree, check their actual settings and compare the disputed date and nearby month boundary with an independent Chinese calendar table. A difference between calendars does not explain a genuine same-calendar error.
Month boundaries and conversion errors
A difference of one day can arise when a new moon near midnight falls on different civil dates under two reference times. Different leap-month placements can create a much larger difference in month-and-day labels. The location of the discrepancy helps identify what needs comparing.
Begin with the calendar version and reference convention, then compare the starts of the months concerned. Check whether each month is ordinary or leap before counting its days. In a Tibetan-derived calendar, inspect omitted and repeated numbers as well. Counting ordinary days from the month's start without those adjustments will not reproduce its lunar dates.
Where two results claim to use the same rules, a discrepancy may come from numerical accuracy, an outdated table, a transcription error or an implementation error. Events very close to a day boundary are particularly sensitive to small calculation differences. Compare the underlying event times or a published month table, rather than treating every disagreement as a regional variation.
A festival's scheduled celebration is a separate question. An observance moved to a weekend or assigned a special rule for an omitted date does not change the calendar's sequence of dates.