"Tom Mchugh" <[EMAIL PROTECTED]> writes: > One thing which doesn't seem to have surfaced in the discussion > yet, is the imponderable effect of plate tectonics upon the accuracy > of any type of sundial over a period of 10,000 years, which effect > would cause both a latitude and longitude change in the position > of any fixed dial. Since different plates move in different directions > and at different speeds, one would have to compute a correction table > based upon the particular plate upon which the dial is located.
If you only use the sun through a N-S slit to synchronize the clock, then latitude shouldn't matter. The longitude correction will be small, but perhaps not negligible for some locations. > Another possible source of error is that therre is a good likelyhood > that many parts of the Northern hemisphere will be under a mile or > two of ice in 10,000 years. There is substantial geological opinion > to the effect that we are now enjoying the balmy climate of an interglacial > period. And, if one builds a dial in an area not likely to be crushed > under a glacier, there is still the problem of changes in rotation rate > due to the shifting of thousands of cubic miles of water from the deep > ocean basins to northerly land glaciers. I don't know whether anyone > hascome up with an accurate model of the effect of glaciation on > the rate of change of earth's rotation and nutation &c. Danny Hillis is thinking of a desert. Everything lasts longer in a dry climate (see the pyramids). You are right that no potential for glaciers should be a site criterion. I wouldn't expect the change in rotation rate due to ice to be a problem for the same reason that the secular slowing shouldn't be. The error cannot accumulate because you are constantly synchronizing the clock to the sun. I do worry about keeping the clock synchronized if the sun disappears for several months due to a very bad turn of weather, a nuclear winter, or a meteorite impact. Art Carlson
