Skip to main content
U.S. flag

An official website of the United States government

Official websites use .gov
A .gov website belongs to an official government organization in the United States.

Secure .gov websites use HTTPS
A lock ( ) or https:// means you’ve safely connected to the .gov website. Share sensitive information only on official, secure websites.

Performance of UTC(NIST) and UTC(NIST) Adjustments

The BIPM publication Circular-T is the official record of UTC – UTC(NIST) time differences. Figure 11 is a graph of the four year period ending in February 2020, showing the UTCr – UTC(NIST) differences in addition to the UTC calculations. The graph indicates that the time difference between UTC and UTC(NIST) has seldom exceeded 5 ns during the past several years.

UTC – UTC(NIST) graph
Figure 11. Approximately four years of recent data, UTC – UTC(NIST) (computed over 5-day intervals) and UTCr – UTC(NIST) (computed over 1-day intervals). UTC – UTC(NIST) (blue) represent the realized time offset of UTC(NIST) with respect to UTC, as observed by BIPM.

NIST scientists decide when to make adjustments to UTC(NIST). They do so by analyzing all of the scientific information they have available. This information includes BIPM data, direct comparisons with other national metrology institutes via two-way satellite time and frequency transfer or GPS carrier-phase observations, and local or remote comparisons with atomic frequency references. The NIST scientists meet weekly to decide what frequency offset ought to be applied between TA(NIST) and UTC(NIST). They attempt to satisfy two abstract goals: to reduce the absolute magnitude of the future time error UTC – UTC(NIST), and to preserve or improve the frequency stability of UTC(NIST). In general, due to the nature of clock noise and time transfer noise, both goals cannot be simultaneously achieved over all averaging intervals. Nonetheless, the NIST steering adjustments typically keep the root-mean-square deviation of UTC – UTC(NIST) to within < 3 ns, and the frequency instability of UTC(NIST) to below \( 2\ \times{10}^{-15} \). Over long averaging intervals, such as years, both deviation measures are much lower. 

To illustrate this, Figures 12 and 13 show the time deviation (TDEV) statistic and overlapping Allan deviation (a measure of frequency instability) of these data, respectively. Both graphs show UTC and UTCr data.  Over almost all averaging intervals, the mean variation of the UTC – UTC(NIST) time offset is held to 2 ns or less, and the UTC – UTC(NIST) frequency instability is \( \le2\ \times{10}^{-15} \). Over averaging intervals longer than 1 year, the rate instability of UTC – UTC(NIST) < 10-16, which is near the absolute uncertainty limits of the world’s best primary frequency references.

time deviation statistic graph
Figure 12. The time deviation statistic is a measure of mean variation over a variable averaging interval \( \tau \). For intervals shorter than about 1 month, the increasing trend is a consequence of random-walk phase noise characteristic of uncorrelated atomic clocks and time scales. For intervals longer than about 1 month, the time deviation tends to decrease as a result of successful steering adjustments, and UTC(NIST) approximates UTC well on average.
Allan deviation graph
Figure 13. The overlapping Allan deviation is a measure of (fractional) frequency instability as a function of averaging interval \( \tau \). At intervals shorter than 1 month, the excess noise present in the series UTCr over UTC is observable. At intervals longer than about 1 month, the trend is decreasing showing that UTC(NIST) realizes a stable rate with respect to UTC.


Technical Contact

Created November 19, 2020, Updated November 20, 2020