Right now, I think the winter solstice is around December 21. Has it always been around this date constantly?

Since the earth's axis's 23.45° tilt will be on the exact opposite side in 13000 years (or was 13000 years ago) due to precession, it'd seem like the seasons would flip every 13000 years. So, +/- 13000 years from now, it would've turned summer around December 21 in the Northern Hemisphere, right?

  • 4
    $\begingroup$ Yes. I could type a long answer, but it's already been done: en.wikipedia.org/wiki/Precession_of_the_equinoxes $\endgroup$
    – Marc
    Mar 5, 2014 at 3:21
  • $\begingroup$ The 2003 and 2007 December solstice occurred on December 22. The 2044 and 2048 December solstice will occur on December 20. Source: timeanddate.com/calendar/seasons.html $\endgroup$
    – David H
    Mar 8, 2014 at 6:26
  • $\begingroup$ @mickael-caruso you've got 3 answers now, how about ticking the one you think best answers the question? $\endgroup$
    – Jeremy
    Mar 13, 2014 at 20:09
  • 1
    $\begingroup$ @Marc except "yes" isn't the correct answer to the question of the seasons inverting every 13000 years. Perhaps you want to edit your comment? $\endgroup$
    – Jeremy
    Mar 13, 2014 at 20:12
  • $\begingroup$ @david-h the time and date site you point to shows the dancing around of the solstice date, but it doesn't show a seasonal drift because the calendar is regularly readjusting for it. en.wikipedia.org/wiki/… $\endgroup$
    – Jeremy
    Mar 13, 2014 at 20:21

4 Answers 4


The Gregorian Calendar was created so that annual astronomical events, specifically the vernal equinox (used to determine when Easter is), would on average keep their places in the calendar year over time. It is the best official approximation to the definition of the tropical year, which is defined as "the length of time that the Sun takes to return to the same position in the cycle of seasons". Because this calendar describes 97 leap years out of every 400 years, it defines the average year as exactly 365.2425 solar days, or exactly 365 days, 5 hours, 49 minutes, and 12 seconds.

However, the mean tropical year is in reality about 365 days, 5 hours, 48 minutes, and 45 seconds, or 27 seconds shorter.

Because the Gregorian Calendar is based on the tropical year, the calendar dates of the year will keep up with the solstices and equinoxes, and thus the seasons. If this calendar were exactly the length of the tropical year, then the calendar would keep the vernal (northward) equinox around March 20th for all time.

But because of the slight inaccuracy, it will take about 3,200 years (60 s/min * 60 min/hr * 24 hr/day / 27 s/year) for these 27 seconds to add up every year to be 1 full day, and that will result in the solstices and equinoxes marching backwards in the calendar by 1 day every 3,200 years or so, depending on the accuracy of the 27 seconds difference. This very slow shift is due to the slight inaccuracy in the Gregorian calendar in, on average, matching the tropical year, not because of the precession of the equinoxes.

3,200 years from now, if the Gregorian Calendar is still used, the date of the vernal (northward) equinox will be on average one day earlier in March. The precession of the equinoxes will still occur, so the Earth's axis tilt will be significantly different from today. The Earth will be at a noticeably different position with respect to the Sun on the vernal (northward) equinox from where it is today, in 2014, on the vernal (northward) equinox, but it will still be in March.

This inaccuracy may very slowly increase over time, because according to the same Wikipedia page for the tropical year, the tropical year is very slowly getting shorter, and the mean solar day is even more slowly getting longer. But for 10,000 years to come, the Gregorian Calendar will keep the vernal (northward) equinox in March, even if it slowly shifts earlier in the month.

This is in contrast to the scenario that you imply, where the calendar date would correspond to the relative position of the Earth in its orbit around the Sun. That describes the sidereal year, the time taken for the Sun to reach the same spot in the sky relative to the stars, which is 365 days, 6 hours, 9 minutes, and 10 seconds. A sidereal calendar would explain why you might think that precession would cause the dates of equinoxes and solstices to change in the calendar year. That would result in a shift in the calendar of one full month in 1/12th the cycle length of the precession of the equinoxes, or about 1 full month about every 2,000 years.

  • $\begingroup$ Re: your last paragraph... Where does he talk about a correspondence to the position of the earth around the sun? $\endgroup$
    – Jeremy
    Mar 13, 2014 at 20:26
  • $\begingroup$ @Jeremy If precession changed the calendar dates of solstices and equinoxes, then that implies that the calendar is based on the Earth's position relative to the Sun (a sidereal calendar) instead of a tropical calendar. It was implied, but not talked about directly, but that's root the question: on what is the calendar based? $\endgroup$
    – rgettman
    Mar 13, 2014 at 20:45
  • 1
    $\begingroup$ Ok, I see what you're saying... A slight rework of that paragraph could make it clear that there is an implied presumption in the question that the seasons relate to orbital position that is incorrect. Do you want to give it a crack, or shall I? $\endgroup$
    – Jeremy
    Mar 13, 2014 at 20:50
  • 1
    $\begingroup$ See also: astronomy.stackexchange.com/questions/13008/… $\endgroup$
    – user21
    Jan 8, 2016 at 6:29

Dates are set by progress through the currently used calendar.

Season, solstices and equinoxes are set by the direction of tilt of the earth towards/away from the sun, based on the earth's orbital position and the absolute direction of that tilt.

The date of seasons, solstices and equinoxes will shift as long as the calendar system does not match the reality of the earth's orbital motion.

Dates of seasons, solstices and equinoxes had shifted by about 14 days by the time the Gregorian calendar began to replace the Julian calendar.

The whole purpose of a particular pattern of leap-years is to keep the seasons/solstices/equinoxes in long-term step with the calendar. Over four years, there will be small wiggles back and forth...


Yes, and no. yes, the equinox will drift over time, but the seasons will not invert every 13000 years.

The reason is the calendar we use is already adjusting for most of the drift due to the precession of the equinoxes. In 10,000 years the drift in seasons will be only around 10 days. http://en.wikipedia.org/wiki/Gregorian_calendar#Accuracy


Heres my take on whats going on. As per precession, in 13,000 years from now, june should be winter for us in the northern hemisphere. But, being on spaceship earth, our season is moving along with its month. So in june which is currently aperihelion will actually still be dec 20 or 21 in 13,000 years but at aperihelion. Currently the winter solstice falls around perihelion. Actually, and I forgot where i read it from, but the last time the winter solstice and perihelion fell on the same day was back in the year 1246. But due to precession, perihelion is now occurring around Jan 3. So using the perihelion point as a reference,and according to my calculations, perihelion is going to go through every day of the month in approx. 2150 years time. The perihelion will then jump into Feb...


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .