Exploring Calendars and long-term changes in the night sky

M.T. Cranium

This may not be directly weather-related, but plenty of weather forum regulars are probably interested in questions about the calendar and the night sky.

As human understanding of the complexity of the earth's orbit has been refined over the ages, our calendar has evolved so that it does what it is expected to do, match up the twelve months with the seasons and the year.

But what exactly is a year? The first approximation is to say that a year is the time between when the earth passes a given point in its orbit around the Sun, and the next time it does so. But the earth and the solar system are making a very leisurely trip around the galaxy, so by fixed point we then mean a fixed point relative to that motion. However, what we really mean is a fixed angle of inclination of our axis, because it is that factor which determines the seasons and those determine the year.

Let's say we could get in a spaceship, stay more or less stationary over the north pole of the Sun (not the earth) and point the ship towards the earth's position in early April. Then when the earth was off to our right, it would be in early January. When it was behind our view, it would be in early October. Off to the left, it would be in July.

If we then adjusted slightly and pointed the craft towards the 21 March position, then the northern hemisphere winter solstice would be off to our right. We might also notice that the earth was a bit closer to us as its perihelion occurs on 4 January. And we would see the south polar regions all the time for several months, while never seeing (assuming we are not very far above the Sun) the north polar regions. Once a day we might notice southern Greenland rotating past our view.

Six months later, it would be the opposite story. The only part of Antarctica we might ever see in June or July would be that peninsula (Graham land) which points towards South America. And that would probably be under cloud almost all the time.

So in other words, in this historical era, we would be seeing northern winter off to our right. But if we waited around 13,000 years, we would see it off to our left. In between, we would see it moving forward around the earth's orbit as we had come to know it, assuming that we maintained our "fixed" position with respect to some distant stars (Spica comes to mind since it is out beyond the earth's April position as seen from the Sun). If we chose a nearby star it might wander a bit against that more fixed background -- eventually they all change relative position, but over a few hundred thousand years only a handful would be vastly different in their location in our skies.

Now, am I sure that this winter solstice in the northern hemisphere would move forward? Might it not move backwards?

The Skyglobe program answers that question, indirectly. In our modern era, on days of low winter sun, we look back at night towards Orion and (in darker locations) we can see the Milky Way bisecting the midnight sky. If there happens to be a full moon, that sits roughly where the midsummer Sun would be sitting at solar noon. I say roughly because the Moon has an 18.6 year cycle of declination that takes it as high as 5 degrees of celestial latitude above the Sun's path (ecliptic) and as low as 5 degrees below it. This year, it reaches that minimum, so the winter full moons have been a bit below where the corresponding summer suns might be positioned.

So if I run my Skyglobe program back to the Roman era, about one-twelfth of the way back to the last time the earth was in the modern orientation to its orbit (I will get to a discussion of exactly when that was later), the people alive then would be looking out at the portion of the sky behind the modern 21November position of the earth. At the winter solstice, Aldebaran in Taurus, and the Pleiades would be crossing the southern sky at midnight. About two hours later, those people would see the winter sky that we see at midnight.

Go back further to the days of the early Egyptians or when Stonehenge and other Neolithic sites were being built, and the winter solstice would be happening almost behind our fixed point of view, where the earth is now in late October. Back to about 7,000 B.C.E. we would find the northern winter solstice directly behind us. The midnight sky would be rather devoid of prominent stars or constellations, and what we are used to seeing now would just be rising towards morning.

So going back in time, the motion is against the orbit, so going forward in time it must be "prograde" or with the orbital direction. By about 4500 A.D. the winter solstice would be seen off in the current direction of 21 January. However, it would not be 21 January, it would be 21 December. This is because the calendar will be keeping lock-step with these developments so that the winter season, which our ancestors believed started with the solstice, will stay fixed. In reality, the winter season is somewhat more like 1 December to end of February, or some other arbitrary quarter of the year. The coldest quarter of the year is approximately 10 December to 10 March. But whatever we choose, the winter solstice is early in the winter season.

The Romans knew enough to construct a 365.25 day calendar that became known as the Julian calendar. But in reality, the time between solstices is about 365.2422 years. By the late medieval period, it was becoming fairly apparent that the calendar was not keeping up to the seasons. It was running too early and ten (or eleven) days had to be taken out of one year (it was 1582 in Europe, at the direction of Pope Gregory (hence the Gregorian calendar) and 1752 in the British empire). After that, to prevent the same problem in the future, leap years were dropped in candidate years if they ended in 00 and were not divisible by 400. Thus around the world, 1800 and 1900 were not leap years but 2000 was. This system should keep the winter solstice and the other three key points of the orbital year (vernal equinox, summer solstice, autumnal equinox) at similar dates. They will always drift around slightly because of the leap years but they keep resetting to similar dates and times every 400 years.

I will continue this in a second post, if you have comments or questions perhaps you could wait to see that second part of the discussion and then post away.

M.T. Cranium

So what is the exact period of this cycle during which the earth's rotational axis performs a circle relative to the fixed stars towards which it may be pointing -- this resulting in a change of the best choice for the "north star" as well?

The skyglobe program shows that the year closest to 2015 in terms of the midnight sky when the ecliptic peaks at about the mid-point of the Milky Way, or when the mid-summer Sun is in the same position relative to either Orion or the Milky Way, would be about 24725 B.C.E. or 26,740 years ago. This is not meant to be an exact timing, as closely as I can align the program this is the most similar looking solstice.

The date shown on the program is 18 December for the summer solstice. Why is it not closer to 22 June? This is because the Skyglobe program shows dates before the Gregorian era in Julian calendar dates. This reset their program at sky-related dates around zero A.D. but the further back in time you go, the more extra days are in that calendar (3 too many every 400 years). Since 24725/400 = 61.81, this tells us there are 185 surplus days in the program, days that should not actually be there (the Julian calendar itself was not in existence very long before zero A.D. and so this is all academic, whoever was alive way back in pre-history had other calendars or ways of marking the solar year). So if there were 185 surplus days, then 18 December 22725 B.C. should be 185 days earlier or 16 June. I may revisit this and get it more precise, but clearly the cycle is not very different from 26,740 years.

Now, halfway through the cycle, or around 13,400 B.C.E., near the last stages of the most recent Ice Age, winter in the northern hemisphere would have been much more brutal than it is today. One could perhaps visualize our distant ancestors, huddled around campfires and probably this would be in the Mediterranean or central American regions more so than our higher ice-covered latitudes. But if we visualize them looking out at Orion and the harshly twinkling Sirius rising after it, then we are dead wrong because their winter sky would have been our summer sky. Our winter (at least Dec-early Jan) sky would have been observed by them in their early summer. But there would have been one startling difference.

The summer night sky has the eclipic plane low in the sky, like the winter daytime sky. The ecliptic is theorized to have remained in virtually its modern orientation to the fixed stars. So in the distant prehistoric summer night skies, only the top half of Orion would even be visible at 35 deg north latitude. The southern half and Sirius would never make it above the southern horizon.

Now as the winter solstice moved forward and Sirius (and Orion) attained higher altitudes relative to the southern horizon, there would have come along some very specific times in the fifth or sixth millenia BCE when Sirius would have made its first, likely very startling to the night-sky-conscious peoples alive then, appearance for one night, then for longer intervals, always getting a bit higher each year. The difference might have been more perceptible over generations than years, and the oral traditions might have preserved this event and remarked on how it was slowly changing with time. By the time of early Stonehenge (presumed to be around 3000 B.C.E.) Sirius would have risen to a maximum azimuth of about two-thirds what we see today, so by then it would be a familiar visitor, albeit more in the autumn sky than the winter.

The point of all this? We should not presume that only the climate changes over time. The night sky also changes in appearance and there are many less significant variables that we could consider too. One is the axial tilt -- that is not fixed but slowly changes over long intervals. The range is something like 22 to 25 degrees, we are near the mid-point of that range nowadays. A larger axial tilt is one of the Milankovitch factors that we associate with larger subpolar glaciation episodes. The slightly warmer summers are not enough to compensate for the colder winters associated with that greater tilt.