NavList:

Mark Coady, you wrote:
"It dawned on me that I don't have a 5000 year old air or nautical almanac handy, and was curious how the bearing of the sun in a particular geographic position on say the winter solstice might be different then than today. "

As a matter of fact, you do have a source of almanac data for 5000 years ago. It's called "Stellarium" (there are other products with similar data). Stellarium does, in fact, accurately depict the sky thousands of years in the past. If you enter a year like -5000, it accurately shows the sky for that year (5001 BC). There are limits, but it's plenty good enough for archeo-astronomy. You can simulate sunrise and sunsets on different dates for sites around the world including Stonehenge and other very ancient archeological sites, and you'll find that there are only small differences compared to today.

Human history is just the right length so that things are relatively simple, and we can sum up the issue almost entirely with one phrase: precession of the equinoxes. The history of human civilization, depending on how we use that slippery word, extends back not much more than ten thousand years. Before that is pre-history. All constructed sites, all ruins, are younger than that. There's an amazing site in southeast Turkey, very close to the border of Syria, called "Potbelly Hill" or Göbekli Tepe that's as much older than Stonehenge as Stonhenge is older than my house (but Stonehenge is warmer in winter). Its oldest elements date to about 9500 years ago. That's a long, long time ago, but you'll notice that it's less than half a precession cycle. The zodiac signs today are out of alignment with the constellations they were named after some 3000 years ago, but otherwise there are no major changes. That's good for archeo-astronomy because things get rapidly more complicated when we go up an order of magnitude. 

Over 10,000 years, we really only have to worry about precession, and the obvious non-repeating motions of the planets and the Moon, but on longer timescales, we would have to consider the changing inclination of the Earth's axis, the precession of the Earth's orbital perihelion and the changing eccentricity of the Earth's orbit. The changes are so large that you might suspect they would even alter the climate, much as Milankovic did nearly a century ago. There's a reasonably good Wikipedia article on those cycles and their apparent effect on climate. And this is also the time scale where the stars themselves begin to change. The constellations we see today are only slightly different from those that were seen a thousand years ago. A few nearby stars, like Alpha Centauri and Arcturus and one of my favorites, 70 Ophiuchi, have moved noticeably in a thousand years relative to the other stars, but when you get to time periods that are tens and then hundreds of thousands of years from the present, the constellations become difficult to recognize. If your time machine landed on Earth 50,000 years ago, you would recognize Orion, but few other star patterns would be identifiable.

As Paul Saffo has already indicated, the data available for archeo-astronomy are reliable and even too complete ...profuse in detail. Many alignments are probably real. Others fall into that category of scholars in love with their theories, fooling themselves. It happens to the best of us.

I mentioned that there are uncertainties and "limits" on our modern tools like "Stellarium". One of those limits you can understand by thinking about lunars. The system of longitude by lunar distances depends on tying the Moon's exact position relative to the celestial sphere, which is governed by "absolute time", to the local apparent time on the Earth, which is dependent on the exact rate of the Earth's rotation. There are small corrections that have to be made over decades to maintain this connection between absolute time and time based on the Earth's rotation and therefore its orientation. These "delta-T" offsets were actually discovered by examining records from ancient history. Edmund Halley and others noticed that solar eclipses extrapolated back into ancient history occurred at the wrong longitudes according to the historical record. But those observations do place strong limits on delta-T thousands of years ago. Unfortunately if we go back twice as far, there are no more records of eclipses and without those records we simply cannot talk about synchronization of events in the sky with events on the Earth without some hours of uncertainty. We know where the Earth was 10,000 years ago. We know where the Moon and the planets were. We know with high accuracy how the Earth was tilted and how its axis was oriented. But we only guess at delta-T. We do not know which way the Earth was facing at a given instant of time. This means that we can predict the exact circumstances of solar eclipses twenty or even fifty thousand years ago, but we cannot say whether a given eclipse was visible in Europe or the Americas. We can only say that it occurred. The connection between celestial events and local time on the ground that is the basis of longitude by lunar distances and related methods breaks down when we step into pre-history...

Frank Reed
PS: Before I forget to put this in print:
Dear People/Entities of the Distant Future, I require a time machine for certain experiments. Please send. Thank you.