Guide · 8 min read
Umm al-Qura, tabular and moon-sighting calendars compared
Ask three sources for today's Hijri date and you can get three answers within a day of each other. That is not a bug in any of them. There is more than one Islamic calendar in use, they answer slightly different questions, and the day a Hijri month begins depends on which one you asked.
This guide describes the three you are most likely to meet - the Umm al-Qura tables, the arithmetic tabular calendar, and local crescent sighting - and what each is actually for. HijriConv implements the first two; the third is not a calculation at all, and no software can produce it. Nothing here takes a position on which method a community should follow.
The question every Islamic calendar has to answer
A lunar month is roughly 29.53 days, so a Hijri month must be either 29 or 30 days long. That single choice, repeated twelve times a year, is the entire design space. Everything that separates one Islamic calendar from another comes down to how it decides when the current month ends and the next one begins.
There are three broad answers. You can look at the sky and decide when the new crescent has actually been seen. You can compute the answer in advance from astronomical data and publish it as an official table. Or you can drop astronomy entirely and use a fixed arithmetic rule that anyone can reproduce.
The three methods usually agree. When they disagree, it is almost always by exactly one day.
Umm al-Qura - the official Saudi civil calendar
Umm al-Qura is the official civil calendar of Saudi Arabia. Its month lengths are set in advance and published month by month in official tables, based on astronomical criteria for the new moon as observed from a defined reference location. It is used across the Gulf for civil and administrative purposes and is the calendar most international software means when it says "Islamic calendar".
Because it is a published table rather than a formula, there is no rule you can apply to extend it. HijriConv embeds the official month lengths for 1318-1499 AH - from 30 April 1900 to 26 November 2076 in Gregorian terms - as a compact bit mask generated at build time, so the conversion is exact for every date inside that window and returns nothing outside it.
Umm al-Qura is the default engine on this site. It is what most people expect for a civil date, it is what appears on documents and calendars across the Gulf, and inside its range it is exact rather than approximate. It is still a calculation: it is designed as a civil calendar, and Saudi Arabia announces the start of Ramadan and the Eids by sighting, which can land a day either side of the tabulated date.
The tabular calendar - arithmetic with no astronomy
The tabular or arithmetic Islamic calendar throws out observation altogether. Months alternate 30 and 29 days, odd months getting the longer count, and the twelfth month gains a day in a leap year. Leap years follow a fixed 30-year cycle containing 11 of them: a year is a leap year when (11 × year + 14) mod 30 is less than 11.
That is the whole calendar. It needs no table, no ephemeris and no data of any kind, so it is exact and reproducible for any year in either direction - which is why HijriConv uses it for years outside the Umm al-Qura range and why it is the engine behind the year pages that reach to 1524 AH.
There are two variants, and the only difference between them is the epoch - the Gregorian day that 1 Muharram of year 1 AH is anchored to.
The astronomical variant therefore always runs exactly one day ahead of the civil variant. Both are offered on the converter, because different software and different reference works assume different ones, and a one-day discrepancy is usually this and nothing more.
- Tabular civil - the Friday epoch, Julian Day Number 1948440, which is 19 July 622 CE in the proleptic Gregorian calendar. This is the rule the Unicode and ISO "islamic-civil" calendars implement, and it is HijriConv's tabular default.
- Tabular astronomical - the Thursday epoch, one day earlier at Julian Day Number 1948439. This is the "islamic-tbla" rule.
Moon sighting - not a calculation
The third method is direct observation of the new crescent, with the start of the month declared by a religious authority once the sighting criteria that authority uses have been met. This is the basis on which religious observances are actually determined in most of the world.
It cannot be reduced to a formula, and HijriConv does not try. Whether the crescent is visible on a given evening depends on where the observer is standing, the altitude of the Moon above the horizon at sunset, the elongation from the Sun, atmospheric conditions and whether optical aid is accepted. Communities and countries apply different criteria, and some follow the announcement of a particular authority regardless of local visibility.
This is why a computed calendar and a local announcement can disagree by a day, and why two countries can begin Ramadan on different days without either being wrong on its own terms. Choosing between these approaches is a matter for religious authorities and not something a date converter can settle.
How far apart do they actually land?
In practice the calculated calendars are close. Taking every month start across the eleven Hijri years from 1440 to 1450 - 132 month beginnings - the Umm al-Qura table and the tabular civil calendar give the identical Gregorian day 82 times and differ by exactly one day the other 50 times. They never differ by two.
A worked case shows what that looks like. The Hijri year 1447 begins on 26 June 2025 under Umm al-Qura and on 27 June 2025 under the tabular civil rule, so the two calendars are a day apart on New Year's Day. By the ninth month they have converged again: 1 Ramadan 1447 falls on 18 February 2026 in both.
Read the other way round, 26 June 2025 is 1 Muharram 1447 in Umm al-Qura, 1 Muharram 1447 in the tabular astronomical variant, and still 30 Dhu al-Hijjah 1446 in the tabular civil one. Same day, three engines, two different Hijri dates. You can reproduce all of these on the converter by switching the calendar selector.
So the practical rule is simple: within a day or so, all of these agree; beyond a day, check which calendar you are reading.
Which one should you use?
It depends entirely on what the date is for, and the honest answer is that only the last case is not HijriConv's to give.
Whichever you pick, state it. A Hijri date quoted without its calendar is ambiguous by about a day, which is exactly the margin that matters.
- For a civil or administrative date, especially anything Gulf-facing, use Umm al-Qura. It is the default here and the one most documents and calendars follow.
- For a historical date, a date outside 1318-1499 AH, or anything that must be reproducible from a published rule, use the tabular calendar. Note which variant you used, because civil and astronomical differ by a day.
- For matching another piece of software, check which calendar it implements before assuming one of you is wrong. "islamic-civil" and "islamic-tbla" are the tabular variants; "islamic-umalqura" is the Saudi table.
- For fasting, Eid, Hajj or any other religious observance, follow the announcement of your local religious authority. That is a sighting decision, and no calculated calendar - including this one - is a substitute for it.
Frequently asked questions
What is the difference between Umm al-Qura and the tabular Islamic calendar?
Umm al-Qura is a published table: the length of each month is fixed in advance from astronomical criteria and issued officially by Saudi Arabia, so there is no formula behind it. The tabular calendar is pure arithmetic - alternating 30 and 29 day months with a fixed 30-year leap cycle - so it can be computed for any year but is not tied to any observation.
Why does my phone show a different Hijri date from this site?
Most likely because it is using a different calendar. Operating systems ship several: islamic-umalqura is the Saudi table, islamic-civil is the tabular calendar with the Friday epoch, and islamic-tbla is the same arithmetic with the Thursday epoch, which always reads one day ahead. Some devices also let the user apply a manual day offset.
Can a computer predict the moon sighting?
It can predict where and when the crescent is likely to be visible, but visibility depends on the observer's location, the Moon's altitude and elongation at sunset, atmospheric conditions and the criteria a given authority applies. The decision to begin a month is made by a religious authority, not by a calculation, so HijriConv reports calculated calendars only.
Which calendar does HijriConv use by default?
Umm al-Qura, because it is what most people expect for a civil date and it is exact inside the 1318-1499 AH range the official tables cover. Outside that range the site falls back to the tabular calendar and says so on the page. The converter lets you switch between Umm al-Qura, tabular civil and tabular astronomical on any date.
How far apart can two Islamic calendars be?
The calculated ones stay very close. Across the 132 month starts in 1440-1450 AH, Umm al-Qura and the tabular civil calendar match on 82 and differ by exactly one day on the other 50, never by two. A local sighting announcement can also fall about a day either side of a calculated date.