NetCDF Time Convention

Time is stored in the NetCDF files in Julian days. It is a slight modification of the Julian day used by astronomers as we start at midnight, rather than 12 noon. The following is an explanation of Modified Julian Day that was pulled off of the Naval Observatory web site.

Modified Julian Date

The Modified Julian Day (MJD) is an abbreviated version of the old Julian Day (JD) dating method which has been in use for centuries by astronomers, geophysicists, chronologers, and others who needed to have an unambiguous dating system based on continuing day counts.

The JD counts have very little to do with the Julian calendar, which was introduced by Julius Caesar (46 BC) and in force until 1582 when Pope Gregory XIII directed the use of an improved calendar, now known as the Gregorian Calendar. In the case of the Julian day count, the name was given because at the time, the Julian calendar was in use and, therefore, the epoch of the day count was fixed in respect to it. The JD counts days within one Julian Period of exactly 7980 Julian years of 365.25 days.

Start of the JD count is from 0 at 12 noon 1 JAN -4712 (4713 BC), Julian proleptic calendar. Note that this day count conforms with the astronomical convention starting the day at noon, in contrast with the civil practice where the day starts with midnight (in popular use the belief is widespread that the day ends with midnight, but this is not the proper scientific use).

The Julian Period is given by the time it takes from one coincidence to the next of a solar cycle (28), a lunar cycle (19), and the ancient Roman Indiction (a tax cycle of 15 years). At any rate, this period is of interest only in regard to the adoption of the start, at which time all periods counted backwards were in coincidence.

The Modified Julian Day, on the other hand, was introduced by space scientists in the late 1950's. It is defined as

MJD = JD - 2400000.5 
        

The half day is subtracted so that the day starts at midnight in conformance with civil time reckoning. This MJD has been sanctioned by various international commissions such as IAU, CCIR, and others who recommend it as a decimal day count which is independent of the civil calendar in use. To give dates in this system is convenient in all cases where data are collected over long periods of time. Examples are double star and variable star observations, the computation of time differences over long periods of time such as in the computation of small rate differences of atomic clocks, etc.

The MJD is a convenient dating system with only 5 digits, sufficient for most modern purposes. The days of the week can easily be computed because the same weekday is obtained for the same remainder of the MJD after division by 7.

EXAMPLE: MJD 49987 = MON., 27 SEPT, 1995 
        

Division of the MJD by 7 gives a remainder of 0. All Mondays in 1995 have this same remainder of 0.

Note that for 1993 the MJD = 48987 + DOY
For 1994 the MJD = 49352 + DOY
For 1995 the MJD = 49717 + DOY
For 1996 the MJD = 50082 + DOY
For 1997 the MJD = 50448 + DOY
        
where DOY is the Day of the respective Year.

The MJD (and even more so the JD) has to be well distinguished from this day of the year (DOY). This is also often but erroneously called Julian Date, when in fact it is a Gregorian Date expressed as number of days in the year. This is a grossly misleading practice that was introduced by some who were simply ignorant and too careless to learn the proper terminology. It creates a confusion which should not be taken lightly. Moreover, a continuation of the use of expressions "Julian" or "J" day in the sense of a Gregorian Date will make matters even worse. It will inevitably lead to dangerous mistakes, increased confusion, and it will eventually destroy whatever standard practices exist.

The MJD has been officially recognized by the International Astronomical Union (IAU), and by the Consultative Committee for Radio (CCIR), the advisory committee to the International Telecommunications Union (ITU). The pertinent document is

CCIR RECOMMENDATION 457-1, USE OF THE MODIFIED JULIAN DATE BY THE STANDARD FREQUENCY AND TIME-SIGNAL SERVICES.

This document is contained in the CCIR "Green Book," Volume VII. Additional, extensive documentation regarding the JD is contained in the Explanatory Supplement to the Astronomical Ephemeris and the Nautical Almanac , and in the yearbooks themselves, now called The Astronomical Almanac . The Almanac for Computers also provides information on such matters.

NOTE: The MJD is always referred to as a time reckoned in Universal Time (UT) or the closely related Coordinated Universal Time (UTC) , International Atomic Time (TAI), or Terrestrial Dynamic Time (TDT). The same is not true for the DOY. This is usually meant in a local time sense, but in all data which are given here at the observatory, we refer the DOY to UT also, except where specifically noted. One could call it then something like Universal Day of the Year to emphasize the point. However, this would introduce a completely new term, not authorized by any convention. Moreover, it is not really necessary to use a different term because we simply follow logically the same practice of extending a time and date measure to the UT reference as we do when we give any date or hour.