Cross-check 交叉核对

This page sets the site’s own computations beside the corresponding published values of the Hong Kong Observatory for the calendar year 2026. It records facts; it does not rank the two sources.

Reference: Hong Kong Observatory, Gregorian–lunar date concordance for 2026, Chinese-language table, file T2026c.txt, retrieved 2026-10-06: hko.gov.hk.

Lunar month beginnings

Twelve months begin in 2026; there is no intercalary month this year. Each row gives the Gregorian date on which the lunar month begins, as published and as computed here.

The twenty-four solar terms

The Observatory’s table gives each term to the day; this site computes it to the second. The comparison is therefore made on days. The instant shown is Beijing time, which is the same clock as Hong Kong time.

Notes on this comparison

• Both sets of values are expressed in UTC+8.
• Twelve lunar month beginnings and twenty-four solar terms were placed side by side; thirty-six of thirty-six agree on the day.
• The two sets are produced independently. Agreement on the day is reported as such and is not offered as proof that the two sides run the same algorithm.
• Close to midnight a term can fall on one day in one almanac and the next day in another. That is a difference of convention, not of correctness. No such case occurs in the year shown.
• The Observatory’s file was retrieved on 2026-10-06; later revisions upstream would not change this page by themselves.
• This site’s own values are computed from the JPL DE421 ephemeris; the method is described on the Data & Methods page.

Worked example · one record that carries a sexagenary day

The Chunqiu (Spring and Autumn Annals), Duke Yin year 3, records: “In the third year, spring, royal second month, on the ji-si day, there was an eclipse of the sun.” The conventional date for this eclipse is 22 February 720 BC (Julian calendar).

This site recomputes that day from the sexagenary day anchor (1949-10-01 = jia-zi; index = (JDN + 49) mod 60): JDN 1458496, sexagenary day ji-si (己巳). The day recorded in the Chunqiu is also 己巳. The two are set side by side here; this site reports only that the sexagenary names agree, and does not infer from this that the two calendars share a source.

• Year reckoning: 720 BC is written here as astronomical year −719. Taking −720 directly yields gui-hai (癸亥) for the same day, one day off, and raises no error.
• Delta-T: sexagenary days are counted as whole day numbers and do not pass through Delta-T. But a record that gives only “there was an eclipse” and no hour must be placed on a particular day through Delta-T, and before the Common Era Delta-T is measured in hours. This site uses Espenak & Meeus (2006), which computes to about 20,609 seconds (about 5 h 43 m) at astronomical year −719.
• Month reckoning: the Zhou calendar’s “royal second month” is not February in the Gregorian sense.

Related pages

• Data & Methods — how the values are computed and checked.
• Changelog — version history of the dataset.
• Dataset — the semantic endpoint and machine-readable description.