# What time is NASA using for Horizons web output?

Here's an example of some output from JPL's Horizons web interface.

NASA does not (appear to) specify which GMT is at those hours that the specific transits occurred, or at least I can't find it.

Should I assume its UTC 0 or GMT 0? Thanks!

2019-Dec-19 06:52:31.872     223.20921 -16.31128 223.20931 -16.31131 340.624206  37.520282
2019-Dec-19 12:26:29.760     223.86907 -16.58967 223.86916 -16.58970 341.077681  36.930691
2019-Dec-19 18:00:27.648     224.52893 -16.86521 224.52902 -16.86524 341.522894  36.341088


See the FAQ of the JPL's website:

Coordinate time is the time used in the development of ephemerides for solar system objects. Under General Relativity, the rate at which actual clocks tick is called proper time. The rate of proper time depends on the location and motion of the clock, so there is no single proper time for the solar system as a whole. So for representing the motions of he solar system bodies using numerical integration, a time scale called 'coordinate time' is used, which is not the rate of any physical clock, but a parameter for which the equations of motion are simply expressed.

The integrated ephemerides for solar system bodies are stored as tables of positions and velocities as a function of coordinate time. In order to evaluate a measurement, such as the round-trip light time from Earth to Mars, or the direction to Saturn as seen from Earth, the proper times of the measurement must be converted to coordinate time before looking up the positions.

Currently the International Astronomical Union (IAU) has defined two coordinate times for the solar system and both are equally accurate. The coordinate time used for the JPL ephemerides, as the independent variable in the solar system, barycentric relativistic, equations of motion, is Barycentric Dynamical Time (TDB) or, in French, Temps Dynamique Barycentrique. This coordinate time is defined such that in the vicinity of the Earth the difference in coordinate time and international atomic time (TAI) is 32.184 seconds plus a small variation that is less than 3 milliseconds. TAI differs from Universal Coordinate Time (UTC) by an integer number of seconds (34 as of July 1, 2012), which changes only when leap seconds are added. UTC is the basis for civil time (e.g., Pacific Standard Time equals UTC - 8 hours).

So the time is neither UTC nor GMT, it is TDB to avoid the problems with leap seconds.

Many thanks to Christiano for finding my error about TAI.

• Are you sure? I read: "The coordinate time used for the JPL ephemerides, [...], is Barycentric Dynamical Time (TDB)" not TAI. AFAIK, TAI is never used and all the JPL ephemerides use TDB. That's true also for EPM and INPOP ephemerides. – Cristiano Sep 20 '19 at 16:28