Search Blogs Blogs

Greenwich Mean Time, revisited

By: jseago
Space-industry professionals, and the general public, commonly refer to our global time reference as Greenwich Mean Time (GMT). But for precise work, we employ the version of GMT synthesized from atomic frequency standards called Coordinated Universal Time (UTC). Because astronomically-determined mean-solar days don’t have quite the same duration as 24x60x60 atomic seconds, the labeling of UTC’s seconds is occasionally adjusted to track solar time on the prime meridian, just as February 29th keeps the calendar year synchronized to the solar year. We’ve experienced 25 of these leap-second adjustments since atomic timekeeping instituted them in 1972. Before 1972, atomic time was maintained even closer to Universal Time, which is GMT by another name. This alternate terminology came about because astronomical almanacs after 1924 changed their astronomical dates to start at midnight instead of noon. The International Astronomical Union (IAU) felt that “GMT” should not imply both and recommended that astronomers use Universal Time to mean “time since midnight.” (Nevermind that non-astronomers were already using “GMT” for time since midnight, and still do today, particularly navigators.) The development of GMT assumed that Earth’s rotation was absolutely steady relative to the fictitious mean sun, an imaginary celestial direction based on the mean longitude of Earth’s orbit, which tries to keep pace with the true Sun by uniformly orbiting the equator once per year. However, by the 1950’s, precise observations affirmed that natural Earth rotation could vary by several milliseconds per day. (By comparison, the blink of an eye is about 100 milliseconds.) Technically then, the assumption of uniform Earth rotation didn’t hold, and scientists realized that Universal Time (UT1) wouldn’t quite track the fictitious mean sun; however, the term “mean (solar) time” was still considered apropos because the violation was so slight. Today, UT1 is defined in terms of the angle between the Earth’s prime meridian and a precisely defined celestial origin which does not rotate like the fictitious mean sun. Nonetheless, UT1 still attempts to preserve GMT; the Earth-rotation angle is multiplied by a scale factor traceable to the fictitious mean sun developed in 1895 by Simon Newcomb of the U.S. Naval Observatory. Because GMT is maintained with different methods than when it was first adopted internationally, some wonder whether “mean time” still exists (and thus, by implication, whether national time services must still coordinate with it via leap seconds). Definitive assessment becomes tricky as one considers the technicalities of “mean solar time at Greenwich;” nevertheless, in a recent paper presented at the 23rd AAS/AIAA Space Flight Mechanics Meeting, the behavior of Newcomb’s fictitious mean sun was compared with UT1 as recorded over the past four decades. To do this, orbital elements were derived from Newcomb’s expression, and AGI’s STK software was used to simulate Newcomb’s mean sun as an artificial satellite. New analysis features in STK 10, known as Analysis Workbench, reported when this object crossed over the prime meridian. Transit times different from noon UT1 thus indicated how far Newcomb’s mean sun had separated from modern UT1.

Earth as a Clock (0h UT, February 29, 2000). The Earth spins at the sidereal rate while the mean-time “dial” rotates once per year with the fictitious mean sun at 12h. Each meridian indicates its own mean solar time.

The analysis affirmed an interesting theoretical result already known by 1960: that Newcomb’s mean solar time at the prime meridian diverges from Universal Time by approximately (1/365) × ΔT, where ΔT equals Terrestrial Time (a theoretically uniform time scale) minus UT1. Because ΔT ≈ 67 s currently, the separation between Newcomb’s mean sun and Universal Time is now about 0.18 s. In principle, a more modern expression should match the true Sun more accurately than Newcomb’s mean sun, but a re-analysis based on an updated theory differed almost negligibly from Newcomb’s century-old determination (about 20 milliseconds presently). By all indications then, UT1 still appears to be synonymous with “mean solar time at the prime meridian” to within a second, and is expected to be for centuries. And UTC, the atomic version of UT1 used for precise civil timekeeping, will remain coordinated with Universal Time as long as leap seconds continue. For technical details on how this question of astronomical timekeeping was assessed with STK 10 and Analysis Workbench, see Paper AAS 13 486 “The Mean-Solar-Time Origin of Universal Time and UTC” by John H. Seago and P. Kenneth Seidelmann.

Dihedral Angle Between the Mean-Sun Satellite Direction and the Prime Meridian Plane, February 29, 2000. Mean solar transit is reached when this angle is zero.

References: Seago, J.H. & P.K. Seidelmann (2013), “The Mean-Solar-Time Origin of Universal Time and UTC.” Paper AAS 13-486, from Tanygin, S., et al. (eds, 2013) Spaceflight Mechanics 2013, Proceedings of the 23rd AAS/AIAA Space Flight Mechanics Meeting, Kauai. Advances in the Astronautical Sciences Series, Univelt, San Diego, CA.
Posted: 3/7/2013 6:01:12 PM