January 1 has no astronomical or agricultural meaning. Julius Caesar set it in 46 BCE to match the day incoming Roman consuls took office — ten days after the solstice, for administrative reasons that had nothing to do with the sky. Pope Gregory XIII's 1582 reform fixed the calendar's drift but never moved the date.
Most of the world's other calendars put the new year somewhere else on the circle, and most tie it to something the sky or the harvest actually does. A year has no natural starting line — only a natural shape, since it repeats. Solar calendars fix the date to a solstice or equinox. Lunisolar calendars run on the moon but insert a leap month so the date never drifts far from a season. Purely lunar calendars use the moon and never correct, so the new year visits every season once every 33 years. Below are seven of them, placed on the year they actually fall in for 2026, next to the temperature record of the place each one comes from.
Flattened onto a line starting January 1, these dates look scattered. Wrapped around the year they actually describe, six of the seven cluster near a solstice, an equinox, or a harvest — and one, by design, sits nowhere in particular.
Angle = day of year, clockwise from January 1 at the top. Short inner ticks mark the equinoxes and solstices (approximate). Rosh Hashanah and Enkutatash fall a day apart in 2026 — two unrelated calendars, both keyed to the end of a local wet or harvest season — and their dots are nudged apart here only so both are readable. The thick colored bands are each calendar's true 20-year range, to scale, drawn only where there's a real range to show: Gregorian has none at all (always Jan 1), and the sidereal and fixed-date calendars move by at most a day, too little to plot. Chinese New Year and Rosh Hashanah get the full month-long window their leap-month rule allows — a 20-year span is long enough to have sampled nearly all of it, since the correction cycle repeats roughly every 19 years. Look closely at the bottom left: Enkutatash's one-day range (Sep 11–12) sits entirely inside Rosh Hashanah's month-long one, drawn as a thinner arc just inside it — two unrelated calendars whose possible dates overlap. The Islamic New Year gets no band here; its number is in its caption below, since a 229-day range would swallow the rest of the chart.
Each panel below pairs two views of the same data: the mean daily temperature range for the city closest to that calendar's origin, and directly beneath it, the month-over-month rate of change — the temperature curve's own first derivative. The derivative's peak and trough are the curve's two inflection points: the fastest-warming day of the year and the fastest-cooling day. A solstice sits near a temperature extreme, where the derivative crosses zero. An equinox sits near an inflection point almost by definition — it's where the sun's declination curve is steepest — and if temperature tracked sunlight with no lag, an equinox-based new year would land exactly on the derivative's peak. It doesn't, because land and air take a few weeks to catch up to the sun; that lag is the gap you'll see below between Nowruz and Tehran's actual fastest-warming day. Hover either chart for exact numbers.
| Calendar | System | New year rule | 2026 date | Local season | Agricultural marker |
|---|---|---|---|---|---|
| Gregorian | Solar, fixed | Fixed civil date, no astronomical anchor | Jan 1 | Deep winter (Rome) | None — a bookkeeping date |
| Chinese New Year | Lunisolar | 2nd new moon after the winter solstice | Feb 17 | Late winter (Xi'an) | End of dormancy, start of field prep |
| Rosh Hashanah | Lunisolar | Tishrei 1, kept in autumn by leap months | Sep 11 sunset | End of dry summer (Jerusalem) | Close of grain, grape & olive harvest |
| Nowruz | Solar, equinox | Exact instant of the March equinox | Mar 20 | Start of spring (Tehran) | Start of the planting season |
| Islamic New Year | Lunar, uncorrected | New moon; no leap month; drifts ~11 days/yr | Jun 16 | Any — Mecca has little seasonality | None, by design |
| Enkutatash | Solar, fixed | Meskerem 1, fixed Gregorian equivalent | Sep 11 | End of kiremt rains (Addis Ababa) | Harvest begins; Meskel daisies bloom |
| Baisakhi | Solar, sidereal | Mesha Sankranti — sun's sidereal entry into Aries | Apr 14 | Pre-monsoon, hot & dry (Amritsar) | Rabi wheat harvest |