Nº 7 2013 > The future of time
To abolish or not to abolish the leap second?
The past and future of Coordinated Universal Time
Ronald Beard, Chairman, ITU–R Working Party 7A
Until the mid-1960s, the rotation of the Earth was the basis for determining the length of a day and for defining time-scales. But the rotation of the Earth is irregular leading to increasingly complicated versions of rotational time-scales — including the short-lived Greenwich Mean Sidereal Time — to be created in an attempt to produce a uniform time-scale. Finally, the search for a uniform time-scale led to a change from Earth rotational time to atomic time-scales.
International Atomic Time (TAI) was introduced as a continuous reference time-scale in 1970. It is based on the readings of atomic clocks, and is independent of the irregularities of the Earth’s rotation. For the purposes of celestial navigation, however, users needing to determine the rotational angle of the Earth still required access to a time-scale related to rotational time, with an uncertainty of less than a second. This led in 1971 to the adoption of the present Coordinated Universal Time (UTC) system. UTC is a stepped atomic time-scale, defined by ITU’s Radiocommunication Sector (ITU–R), formerly the International Radio Consultative Committee (CCIR), in Recommendation ITU–R TF.460.
The steps are known as leap seconds, and were introduced in UTC to reconcile the difference between the uniform atomic reference time, TAI, and rotational time. The maximum difference between UTC and TAI is limited to plus or minus 0.9 second.
Nature’s changing course
Apparent solar time, as indicated by a sundial, or more precisely determined by the altitude of the Sun, is the local time defined by the actual diurnal motion of the Sun. However, because of the tilt of the Earth’s axis and the elliptical shape of the Earth’s orbit, the time interval between successive passages of the Sun over a given meridian is not constant.
The difference between mean and apparent solar time is called the equation of time, and the variation between apparent and mean noon can amount to up to 16.5 minutes. Until the early nineteenth century, voyagers relied on apparent solar time, backed up by astronomical ephemerides (tables giving the calculated positions of celestial objects at regular intervals throughout a period). But as clocks improved — and their use by ships at sea and by railroads grew — apparent solar time was gradually replaced by mean solar time.
Mean solar time is the measure of astronomical time defined by the rotation of the Earth with respect to the Sun, and takes account of the orbital motion of the Earth around the Sun. When referred to the meridian of Greenwich it was called Greenwich Mean Time (GMT) but it is now known as Universal Time (UT) and, when adjusted for the Earth’s polar motion, it is known as UT1.
The mean solar day is traditionally described as the time interval between successive transits of the fictitious mean Sun over a given meridian. Historically, the unit of time, the second, was defined as 1/86 400 of a mean solar day. Ephemeris Time (ET) replaced UT1 as the independent variable of astronomical ephemerides in 1960. This was, in turn, replaced by relativistic time-scales in 1984 and resulted in the current Terrestrial Time (TT) as the geocentric time-scale used for astronomical ephemerides.
Overtaking celestial navigation
Since the late 1980s, electronic navigation and communication systems have significantly overtaken celestial navigation. In order to operate, these global systems require a continuous time reference, and several continuous time-scales have been established for internal use for that purpose.
It turns out that these internal continuous time-scales are also ideal for comparisons among precision time centres, as well as for precision time applications and the dissemination of precise time in general. The ease of using these continuous time systems contrasts with the complexity of dealing with a coordinated universal time-scale that involves leap seconds. The application of leap seconds is known to cause difficulties to various networks that use precise time, whether distributed locally or internationally.
These ad hoc system times emanating from continuous internal time-scales are currently being used as a reference in many applications — such as global navigation satellite systems — in order to avoid the use of the discontinuous UTC time-scale. This has unfortunately led to a proliferation of “pseudo” time-scales, calling the current definition of UTC into question.
Standard time and frequency signals
ITU–R Study Group 7 on “Science services” set up its Working Party 7A to deal with “Time signals and frequency standard emissions”. ITU–R Working Party 7A is thus responsible for standard time and frequency signal (STFS) services, both terrestrial and satellite. The scope of the working party includes the dissemination, reception and exchange of STFS services, and the coordination of these services, including satellite techniques, on a worldwide basis.
ITU–R Working Party 7A develops and maintains Questions, ITU–R Recommendations in the TF Series, Reports, Opinions and Handbooks relevant to standard time and frequency signal activities, covering the fundamentals of STFS generation, measurements and data processing. The related ITU–R Recommendations are of importance to telecommunication administrations and industry. They also have major consequences for other fields, such as radionavigation, electric power generation, space technology, and scientific and metrological activities. They cover the following topics: terrestrial SFTS transmissions, including high-frequency, very-high frequency and ultra-high frequency broadcasts; television broadcasts; microwave links; coaxial and optical cables; space-based SFTS transmissions, including navigation satellites; communication satellites; meteorological satellites; time and frequency technology, including frequency standards and clocks; measurement systems; performance characterization; time-scales; and time codes.
A major standard administered by ITU–R Working Party 7A is the definition of UTC in ITU–R Recommendation TF.460. This gives ITU — as one of the international organizations involved in the dissemination and coordination of time and frequency services, and in standards development — a central role in the definition, determination and maintenance of UTC.
The definition of UTC is more than a simple statement. Rather, it is a comprehensive process of incorporating recommendations into myriad standards and applications throughout the telecommunication and navigation communities.
Although the original purpose of UTC was not to be the standard for civil time, it has been adopted as the basis of official or legal time in most of the world, and as the standard reference time and basis of the time zones.
The actual value of UTC is calculated at the International Bureau of Weights and Measures (Bureau International des Poids et Mesures — BIPM) from about 420 atomic clocks operated in some 70 time-standards laboratories around the world. UTC is based upon the International System of Units (SI) second, and the inclusion of highly accurate primary clocks in centres around the world as data sources for calculating UTC ensures that there is only about one second of deviation from calculated ideal uniform time in several million years.
UTC is the only time realized by local approximations maintained in timing centres and laboratories designated as UTC(k), with k being the timing centre or laboratory’s designation. These UTC(k) realizations are used in disseminating time signals to users of precise time and those who need to know the current time or real-time values. International Atomic Time is the metrological reference used as the basis for the calculation of UTC, and provides a reference in frequency only.
The future of UTC
In October 2000, a new Question provided the impetus for initiating studies on the possible revision of Recommendation ITU–R TF.460-6. Question ITU–R 236/7 on “The future of the UTC time-scale” originated in response to matters raised by the Consultative Committee for Time and Frequency of the International Committee for Weights and Measures (CIPM).
A Special Rapporteur Group on the future of UTC was established to stimulate studies by ITU Member States and ITU–R Sector Members, and to gather information as a basis for possible modifications to related Recommendations. Creating a Special Rapporteur Group was thought to be necessary because any change to the UTC time-scale — or the identification of an alternative time-scale — would have a significant impact on radiocommunication, telecommunication, satellite navigation and computer systems, and could even affect the social perception of time.
Representatives of BIPM, the Interna-tional Earth Rotation and Reference Systems Service (IERS), the International Union of Radio Science (URSI) and the International Astronomical Union (IAU) participate in the Special Rapporteur Group as well as in ITU–R Working Party 7A. These organizations have also set up their own working groups to investigate the matter. Reports from these working groups indicate that there is no strong consensus within their organizations either for or against changing the definition of UTC.
Leap seconds and length of day
Leap seconds are currently added to UTC to limit its divergence from UT1 to no more than 0.9 second. In other words, the current practice of using leap seconds to adjust UTC maintains length of day— the difference between the astronomically determined duration of the day and 86 400 SI seconds — at no more than 0.9 second.
A change to the leap second method is being considered. This would make UTC a continuous atomic time-scale, which would gradually diverge from UT1 (which depends on the Earth’s rotation angle). The divergence would result not only from the irregular rate of rotation of the Earth, but also from the fact that the defined duration of the SI second does not perfectly match the duration of the second determined as a fraction of the mean solar day.
Over the past 50 years, the UT1 second has been 2×10–8 s longer than the SI second on average, causing about 35 seconds difference between TAI and UT1 today. The rate of the Earth’s rotation is also predicted to gradually slow down further, so that in the near future more than one leap second per year would be needed.
Preparing for WRC‑15
The various discussions and studies that had been taking place on the question of establishing UTC as a continuous atomic time-scale led to this matter being submitted to the World Radiocommunication Conference (WRC) in 2012 for decision. The topic was discussed at the conference, but many participants felt that more information was needed before a decision could be reached. WRC‑12 therefore adopted Resolution 653 (WRC‑12), which reflected their agreement to bring the question to the attention of relevant outside organizations, to have ITU–R Working Party 7A carry out further studies, and to include the topic as an agenda item for WRC‑15.
As part of the preparatory efforts for WRC‑15, a special workshop is being held at ITU headquarters in Geneva on 19-20 September 2013 to provide more information to interested parties and perhaps to stimulate additional studies.