Chinese Zodiac Year Start: Li Chun vs Lunar New Year

Two opening lines, and the table that carries both

A zodiac year does not have one starting date. It has two, and they seldom agree.

The first is the lunar new year, the first day of the first lunar month. The second is Li Chun, the solar term that opens spring. Both are published dates, both are used, and for a birthday that falls between them they name two different animals.

This page carries both lines for every year from 1940–2030: the Li Chun date, the festival date, the gap between them in days, and the two animals a birthday inside that gap can carry. Every date is read from the calendar tables published by the Hong Kong Observatory, one file per year, ninety-one files, and checked against a second file in the other language. Nothing here is computed and nothing is copied from another site.

Find your year, read the row, and check your birthday against the two dates.

What each line is, and what it governs

The lunar new year is the festival date. It opens the lunisolar year, and it is the line the Chinese national standard uses: GB/T 33661-2017, drafted by the Purple Mountain Observatory and in force since 1 September 2017, fixes the zodiac year as running from the first day of the first lunar month to the eve of that date in the following year.

Li Chun is the first of the twenty-four solar terms, and it is a moment rather than a day. In the branch record, the year pillar changes at that moment, so a birth chart reads its year position from Li Chun and not from the festival.

The two are used in different settings, and neither is the other's mistake. A standard names a year; a chart reads a moment. When a page asks only for your year and returns one animal, it has picked one of the two lines on your behalf without saying which. This site's own chart uses the Li Chun line for the year pillar; the national standard's line is the one behind the answer most people expect.

How to read the five columns

The first column is the Gregorian year. The second and third are the two dates, both in that Gregorian year, so the row is self-contained: 1985 means Li Chun on 4 February 1985 and the festival on 20 February 1985.

The fourth column is the gap in days, and its sign carries the order. A positive number means Li Chun came first and the festival followed; a negative number means the festival opened the year first; zero means the two fell on the same day.

The fifth column names the two animals in play for a birthday that sits between the lines. The pair is always this year's animal and the one before it. Which of the two you get depends on which line your birthday falls after — and the sign in the fourth column tells you which line that is.

The table, 1940–2030

Ninety-one rows, 1940–2030. The gap column is signed, so a row can be read without checking the two dates against each other first.

What ninety-one years show

Li Chun never falls outside 3–5 February in this window. It lands on 4 February in 75 of the 91 years, on 5 February in 12, and on 3 February in 4. Those three days are not evenly spread: the last 5 February Li Chun in the window is 1980, and the first 3 February one is 2017.

The festival travels much further. Its earliest date here is Jan 21 in 1966, and its latest is Feb 20 in 1985 — a spread of a full month, against the solar term's three days.

That spread is what makes the gap move. The widest gap is sixteen days in 1985, with Li Chun first; the widest the other way is fourteen days in 1966, with the festival first. Li Chun comes first in 48 of the 91 years and the festival first in 41. Only 2 years give no gap at all.

The animals divide almost evenly: 7 of the twelve carry eight years each in this window and 5 carry seven.

The two years the lines met

Twice in ninety-one years the two boundaries fall on the same day: 1943 and 1992. In 1943 both land on 5 February; in 1992 both land on 4 February.

Those two rows have no window between the lines, which makes them the exception rather than the rule. It also makes them the one case where the answer does not depend on which convention is being used — the two lines move at the same time, so both name the same animal from the same day onward.

The coincidence is worth keeping in proportion. Two years out of ninety-one does not mean the question is usually simple; it means the two calendars line up about once in a human lifetime.

One of those two rows has a page of its own: 1992 in the Chinese zodiac works through what a zero-day gap settles and the single case it leaves open.

Where the dates come from

Each row is read from two files that the Hong Kong Observatory publishes for the year: the English conversion table and the Chinese one. Both carry the solar terms in a column and the lunar date alongside, so the Li Chun date and the first day of the first lunar month are both printed in the source rather than derived here.

Three checks run over every year the file publishes, all two hundred of them. The Li Chun date must be identical in the English and Chinese files. The festival date must be identical in both. And the stem-and-branch pair named in each file's own header - the pinyin in the English one, the characters in the Chinese one - has to agree with the other header and with the sixty-year cycle, which is what ties the year label to the row. All two hundred years pass all three, with no exceptions filed. Accessed 2026-10-21.

The same rows are published as a file as well, extended to the full 1901 to 2100 range the Observatory publishes, so anyone who wants to reuse them gets every year rather than only the ones tabled here: the table as JSON and the same table as CSV. Both carry the two dates, the signed gap and the two animals; the source, the access date and the licence sit in the JSON, which a CSV has nowhere to put. The dataset page states the coverage, the method and how to cite it.

The hour, if your birthday is the boundary day itself

Li Chun is a moment, not a day, so a birth on that date needs the hour before the line can be applied. In 2026 it falls on 4 February at 04:02 Beijing time, as reported by Xinhua from the Purple Mountain Observatory. A birth on 4 February 2026 at 01:00 is before that moment; at 09:00 it is after.

This table gives days rather than moments, and that is a deliberate limit rather than an oversight. The observatory's own tables of solar-term times cover only a recent stretch of years, so a column of times running from 1940 to 2030 could not be filled from one source throughout. We left the column out instead of mixing a sourced decade into four unsourced ones.

If your birthday is a Li Chun date, treat the day as the boundary and take the hour into account, or keep both animals until you have it. Why solar time matters covers how the moment is converted for a birth place.

What a year boundary does not decide

A boundary decides which animal a date carries. It does not decide what that animal means, and it does not score anything. This table has no ratings and no ranking of the twelve animals; it is a lookup with two columns of dates and one of arithmetic.

It also settles only one of the eight positions in a birth chart. The year branch is the slowest of the four pillars and the least specific to a person; the month, day and hour pillars move faster and carry more of what a reading actually uses. What the year animal does and does not cover sets that boundary out in full.

So the honest use of this page is narrow: it tells you which date the two lines fall on in your year, and therefore whether your birthday needs a second look. The free chart works out the other seven positions from a date, a time and a place.

Where this page can be wrong

The dates here are calendar facts, printed by an official observatory and reproduced from two files that agree. The realistic failure is transcription, which is the reason every year is read twice rather than once.

Three limits are worth stating plainly. First, this page does not say which convention you should use; both are current, and the choice belongs to whoever is asking. Second, the two animals in the fifth column are always this year's and the previous year's, so we give both rather than pick one. Third, the minute of Li Chun is absent for most years, as explained above.

One more caveat belongs to the source rather than to this page. The observatory's tables are one authority among several, and where another authority prints a different boundary date for the same year the difference sits at the source. We read one authority consistently instead of blending several, so that any single row here can be checked against the same publication it came from.