NavList:
A Community Devoted to the Preservation and Practice of Celestial Navigation and Other Methods of Traditional Wayfinding
Re: Leap second today
From: Paul Hirose
Date: 2012 Jul 03, 20:16 -0700
From: Paul Hirose
Date: 2012 Jul 03, 20:16 -0700
Old surveying books say quite a bit about celestial. It even gets a full chapter in a 1989 textbook I own. By that time the main purpose of celestial was to find true north. The book lists some useful ephemerides, including the Nautical Almanac (but not the Astronomical), and an almanac sold by the Leitz company. An interesting accessory is the Roelof solar prism, which fits over the instrument's objective lens. It dims the Sun and splits it into four images in a square pattern. The separation is such that there's a small diamond shaped area in the center where the images don't touch. That's where you put the crosshairs. It's hard to simultaneously manipulate both slow motions and stay precisely on target, so a common trick is to set one crosshair ahead of the body, track with the other, and mark the time when both crosshairs are "on". Polaris can produce an azimuth accurate to 1" by averaging several observations from a first order theodolite, says the book. Sun observations are more convenient, but accuracy is about 10". In the optical era the U.S. Coast & Geodetic Survey used elaborate celestial instruments and methods for latitude, longitude, and azimuth determinations at selected stations in their trianulation networks. Their old manual on geodetic astronomy (Special Pub. 237) is now online in the NOAA Library. Some photos of geodetic astronomers working in the field are online too: http://www.photolib.noaa.gov/cgs/astro1.html Almanacs and Time Scales. Modern fundamental ephemerides, such as the JPL DE series, produce positions of solar system bodies as a function of TDB (Barycentric Dynamical Time). Because TDB is always within 2 milliseconds of TT (Terrestrial Time), for most applications these time scales are identical. UT1 is unsuitable for a precise ephemeris due to unpredictable variations in rate. However, in addition to Earth's position, we are interested in its orientation. That requires UT1. The different roles for these time scales are seen in The Astronomical Almanac tables, where RA and dec. are a function of TT, and GHA Aries a function of UT1. It's the user's responsibility to know the time in both scales, and compute hour angle from RA and GHA Aries. For a navigator that would be too complicated, so the Nautial Almanac uses only UT1 and directly shows the GHA of solar system bodies. A ∆T estimate is necessary to create tables in this format. However, it need not be very accurate. The GHA of the Moon is 30 times less sensitive to TT than UT1. Snce UT1 is the almanac time scale, its exact value is known when the tables are computed. The only time error is in TT, via the ∆T estimate. But a 1 second error would change the Moon GHA only half an arcsecond. George Kaplan has speculated that the NA time scale could change to UTC if leap seconds are abolished. In that case the 30:1 ratio works against you. I.e., TT (via the fixed offset from UTC) is known exactly, with UT1 computed from a ∆T estimate. "In the case of Earth-rotation-dependent data, however, the convenience [of an almanac in UTC] would come at the price of some degradation in accuracy. It seems likely that the increased error (for tabulations computed about two years in advance) would not be worse than that from assuming UT1=UTC now, which is probably a common assumption. Furthermore, if necessary, the error could be removed by application of corrections based on the measured value of UT1–UTC for the date on which the data are needed." --