[time-nuts] Time-zones and World time..

Lux, James P james.p.lux at jpl.nasa.gov
Sat Nov 1 21:55:24 UTC 2008




On 11/1/08 2:28 PM, "Bruce Griffiths" <bruce.griffiths at xtra.co.nz> wrote:

> Hal Murray wrote:
>>> One can, of course, observe meridian passage for a variety of stars at
>>> night, and from that determine the time (given a calendar and the
>>> appropriate almanac data), so you could directly observe midnight.
>>>
>>
>> What does "directly observe" mean?
>>
>> Stars aren't conveniently located where I want them.  I think I'd have to do
>> something like time when a star that's close to midnight crosses a
>> north-south line in the sky that's above my location.  So in addition to the
>> almanac, I also need a (good?) compass.

If you're in the northern hemisphere (sorry Bruce..) one can easily know
where north is, to within a fraction of a degree (because Polaris doesn't
move much).

One could conceivably measure several stars and figure it out, too.



>>
>>
> A compass isn't actually required for determining the meridian.

True.. Presuming one has plenty of time, you could use the sun to determine
meridian (but that makes the whole thing sort of circular..)


> If you have located the Meridian and have identified the star, then its
> not necessary to use one that has a meridan transit at local midnight.
> In any event if you dont know exactly where you are identifying local
> midnight may be problematic.
> A PZT or equivalent may be easier to setup.
>
>> Maybe I don't understand how navigation by sextant works.  They get latitude
>> from the elevation of the sun above the horizon.  Do they also get time
>> (local noon) by noting when the sun is highest?  I think that would be hard
>> to do since you are looking for the peak of a sine curve.  That has a broad
>> flat top.

That's exactly what you do. Actually, you can make several measurements with
a local clock of unknown offset, and fit the elevation angles to the
approximately sinusoidal curve and solve for the peak.  That would solve for
your clock's offset from local solar noon.

You can measure elevation angles to several stars and solve a system of
equations. Still won't give you absolute time, though.  You need to measure
the stars AND something that moves against them (sun or moon) to solve for
everything.  There's a whole literature on doing celestial nav with star
trackers. Check out the papers by Kaplan.

  There isn't any conspicuous peak.  Do the numbers work out?  Can I
>> get the peak close enough to be useful for navigation?

What's useful?  If you're doing shadow and stick style navigation on foot,
and you can wait several hours, you can do bisection of angle sorts of
schemes using only geometry and no clock to determine north to better than a
degree.  The limiting factor for solar measurements without optical assist
is that the sun is 1/2 degree wide, so the shadow is indistinct.  A sextant
has a dark filter so you can pick one edge(limb) of the sun to use as a
reference point.


>>
>>
>>
> An accurate watch or equivalent source of UTC is required.
> Various techniques for determining the time from Astronomical
> observations are somewhat impractical at sea.

Not that impractical, because they were used for centuries.  Granted
Harrison's clock helped a lot, but lots of navigators used other techniques
well into the beginning of the 20th century.

The method of lunar distances is probably the most common "non-clock" means
for determining time. You measure the angular distance from the moon to
various stars (or look for occultations).  Since the moon moves about 15
degrees per day against the sky, one degree is a bit more (24/15ths) than an
hour.  If you can measure to a minute of arc, you're in the minutes of time
area, which corresponds to roughly 1 mile of nav precision (1 minute of
latitude = 1 nautical mile).  Not enough to do a narrow ship channel, but
good enough to not miss the island you're looking for.

You can also measure time by looking at Jupiter's moons.

>>> Don't forget too, that solar noon varies quite a bit (minutes) from
>>> "mean solar time" over the year.  If you're navigating your ship with
>>> noon sun sights, this is pretty important.
>>>
>>
>> What do I google for if I want that correction?

You carry a copy of the Nautical Almanac which sets you back about $20 in
paperback. This is a "motion of the heavens" thing, so it can be calculated
too.  Check out the paper about navigation using the analemma (google for
it).. The author demonstrates navigation using only the analemma and a slide
rule for calculation.
>>
>>
> Googling isnt a reliable technique for marine navigation, at least
> during the journey.
>> There are quite a few time scales.  Which one do navigators use?
>>
>
> UTC is usually required in the calculations, as it is (currently)
> closely tied to the Earth's orientation angle.

Folks use what ever their almanac/emphemeris tables are based on.
Historically GMT (Rule Britannia!) now UTC (because the Cassinis measured
the meter).

---

If you want to spend an interesting $20 or so, get the CD from the Institute
of Navigation (ION) with all the celestial nav papers on it.  All sorts of
stuff about how they go about printing the nautical almanac (an early use of
computer driven phototypesetting), to speculations about Columbus's
navigation, etc.

Another thing is to buy a cheap plastic sextant, an almanac (or use the
online ones) and sight reduction tables, and do some celestial nav in your
backyard. (I got within about 10km the first time I tried it).

A lot of the complexity in celestial nav is in trying to make the conversion
of sight to position doable without anything more than adding and
subtracting, so that's what the tables are. Essentially precomputed
solutions.





More information about the Time-nuts_lists.febo.com mailing list