This document is part of the Martian Time Boneyard. It was originally located at
Author: Alan Dechert

Time Keeping on Mars

Time Keeping on Mars

Here are two posts (Aug 14 and 17 1996) I made to the Talk 2000 forum after a bunch of talk about going to Mars.

Can be read in context at the Talk 2000 archives for Aug of 96.

This essay is mentioned at along with many other Mars time-keeping ideas.

It was discussed on Usenet in April 1998 here and there.

On 24 May 1999, I added an intercalation sequence to make the Mars calendar perpetual. I wrote computer code to do the calculations and generate the calendar pages. This code is at the bottom of this page (after the essay). The output from this program can be seen at (600 years worth).

Originally, I set the inauguration point at ../00 00:00:00 UTC = ../00 00:00:00:00 Mars Time. Rather than requiring a new prime meridian on Mars, I opted to inaugurate the calendar as close to the new millennium as possible. So, now it's set at 20:42:04 on New Year's Eve just before the turn of the millennium.

It turns-out that my Mars clock idea (24:60:60 like on earth...others had the same idea before me) is now the leading candidate for use on Mars. An on-line version of this clock, by Michael Allison of NASA, can be seen at where you can watch Martian seconds tick away at a slightly slower rate than Earth seconds.

Although colonization of Mars is not likely to begin any time soon, the possibility of a civilization there does raise some interesting questions about time keeping.

If humans were to live and work there on an on-going basis, after a while, it would seem ridiculous to them to reckon time based on cycles of the Earth. In my opinion, the point at which they broke from Earth time-keeping would be the defining moment of the beginning of a Martian civilization. But what would clocks and calendars look like? How would they be similar to clocks and calendars on Earth? How would they differ from clocks and calendars on Earth?

Mars has its own cycles: years, days, seasons, etc. and they are different than those on Earth. The basic units for keeping time on Earth are the day and the year. Although other models are possible, for a variety of reasons, it is likely that Martian time keeping would follow a similar model.

Fortunate for humans there, the Martian day is very similar to that on Earth. It is about 24.6 Earth-hours. Another important similarity: the angle of the Martian axis to the plane of orbit. This means that there are distinct seasons on Mars just as there are on Earth. These similarities will help greatly to ease human orientation to the Martian environment. An important difference: the Martian year is almost twice as long--about 687 Earth-days.

A day on Earth has 86,400 seconds. A second, then is 1/86400 th (approx.) of the length of one rotation of the Earth on its axis. (Actually, the second is now established as 9,192,631,770 periods of the radiation corresponding to the transition between the two hyperfine levels of cesium-133). Martians could retain the relation of the second as 1/86400 th of a Martian day. So, a Martian day would have 86,400 seconds... just like on Earth. But these would be Martian seconds (about 2 percent longer). The difference would be barely perceptible. You would need to use [slightly slower running] Martian clocks and watches when on Mars, but the advantage would be that you would understand and read them just like clocks and watches on Earth.

With the Martian seconds, minutes, hours and days understood, we can go on to the problem of the Martian calendar. The first question that comes to mind: What year do we begin the Martian calendar with? No doubt, Christian-Martians will want to reckon the year based on the traditional point in time that we count as the birth of Jesus. Since there are about 687 Earth-days in a Martian year, the current year on Mars would be 1061 A.D. (figure the number of days since ../0001 A.D. and divide by 687)

The decision of when the Martian calendar begins is likely to be made by Martians. So it's not clear if the Christian-Martians would win. In fact, it would probably depend on the percentage of Christians there were among the Martians. But it would also depend on the inclination of the Martian leadership . Clearly, Dr. Zubrin would not want to adopt a Christian calendar. He rails agains Christendom, and, it seems that getting away from Christianity is one of his main reasons for wanting to go to Mars (Actually, it's not really necessary to go to Mars to lose one's religion). It may make more sense to begin the Martian calendar with the year zero to mark the beginning of human civilization on Mars.

It could also be part of the Year Zero celebration on Earth to kick off the Year Zero on Mars at the same time. In this case, it would not mark the actual beginning of human civilization on Mars but would symbolize the beginning of our plan to colonize Mars.

Then there is the problem of weeks and months. As I suggested with seconds, minutes and hours, it would probably be best to come up with something similar to what we have on Earth. This will not only ease human orientation to life on Mars but would make it easier to communicate with Earthlings. Also, it would make it easier to adapt computer software for use with the Martian calendar. For these reasons, I suggest retaining the 7-day week. There are about 670 Martian days in the Martian calendar. So, if we retain the 7-day week, the only thing left is the months.

A likely number of months would be 12 or 24. So, you could have 12 months of 8 weeks each or 24 months of 4 weeks each. This doesn't quite work out exactly but it's not bad. Adjustments could be made, as we do with our Gregorian calendar, varying the number of days in a month (I suggest 11months of 56 days and one with 54 days) and with leap years here and there. I would favor a 12 month year. Each month would have about 56 days. The advantage here is that, as with the days of the week, you could use the same names as we use on Earth. Furthermore, the calendar could be started in such a way that the Martian seasons would correspond to the same months as they do on Earth. That is to say, January on the Martian calendar would correspond to mid winter in the northern hemisphere, spring would begin in March, etc., etc.

-- Alan Dechert

After I posted this idea, someone responded with an alternate plan. I responded...

Some minor quibbles with Bob McClenon:

Although your system for marking the hours and minutes on Mars seems workable, I still think my scheme is better. At one point, you say, "There is no inherent reason why the day should be divided into 24 hours of 60 minutes." While it is true that there is "no inherent reason," there are legacy concerns.

There is no inherent reason to call the first month of the year January either. However, imagine a newsarticle in the New York times: "On Tuesday, November 53rd on Mars, Jay Gary fell attempting to scale Olympus Mons and had to be rescued by the robotic rover..." While "November 53rd" sounds funny, everyone will instantly comprehend it. Now suppose they had 20 months and that the date was called Quindecaber 22 [instead of November 53rd]. Any newsarticle about Mars using Martian dating would require a footnote explaining that the Martian calendar has 20 months and that, for example Quindecaber is the 15th month and that this would be late fall in the northern hemisphere, and so on.

No one on earth would ever remember how the Martian calendar works and so it would need to be explained over and over forever. Some Earthlings would think, "those Martians are W-I-E-R-D !" so they just won't want to hear about it [and be reminded how dumb they are because they can't remember the Martian calendar]. This is no small matter either since Martians will be forever trying to milk Earthlings for more shipments of Dove Bars, Pepsi, Elvis memorabilia, etc.

Similarly, I think it best to have clocks, watches, etc. look and work the same [except they would run a bit slower] as they do on Earth. You say, "Redefining the second would cause too many problems in the exchange of scientific data... " It's a trade-off. Your scheme would not be without data exchange problems. With your 20 hr scheme [where minutes are not an whole number of seconds] there will be plenty of problems ["using one system for precise measurement of time using a standard unit and another for practical measurement of time using a local unit"]. For example, if you need to figure the time between 13:46.546 and 16:33.91 Martian time in to Earth hr., min, and secs you will still have a hassle and most Earthlings will never understand it [What? They don't have seconds?].

Then there's the problem with computer software... Clearly, it would be a while before Martians would have much time for writing software code. As much as possible, they would need to use software requiring little or no modification. My scheme requires no modification for handling hrs, mins, and secs and only slight modification to deal with the number of days per month. It is very easy to modify clocks so they run a bit slower. It's not so easy to change all the time-handling code in all the software for use on Mars.

And you have slightly simplified the relationship of Earth-seconds to Martian-days. You have taken the figure 24 hrs 37min and calculated the number of seconds (88620) from that. It's messier than this. A more precise figure is 88642.44 (no doubt, there is a better number but this is the best I found on the net). The fact that the Martian day is not a whole number of Earth-seconds will uglify any scheme like you propose.

You say, "In any case, the second, which is the standard international unit of time..." No problem. I'm not proposing to change that. It's worth noting that it is no longer Earth-based. It used to be defined as 1/86400 of the rotational period of Earth but it is now based on Cesium-133 characteristics. The Martian second I am proposing is for common time-keeping purposes: watches, clocks, computers, bookkeeping, etc. Scientific instruments would not need to be recalibrated.

I am mildly disappointed that you agree with me about not using a Christian era calendar for Mars. Since when does God stop at the Ionosphere? If one believes that Jesus was God incarnate, how do you justify ignoring the date of Christ's manifestation in this solar system? Perhaps you are not a good enough Christian. I wonder if other Christians would agree with you. I am sure some scheme could be worked out for holy days, etc.

-- Alan Dechert

* Program        : mars.prg                                    *
* Version        : 1.0                                         *
* Date           :../1999                                   *
* Language       : dBASE IV                                    *
* Author         : Alan Dechert 916.791.0456, *
* Purpose        : Generates table of days-per-year for the    *
*                  Dechert Mars calendar based on the          *
*                  intercalation sequence devised by Dechert,  *
*                  incorporating a suggestion from Paul Hill   *
*                  to use centennial years to distribute       *
*                  remainders in order to reduce drift.        *
*                                                              *
* Description    : The program uses the table described below. *
*                  It empties the table (ZAP) then populates   *
*                  the table with number of days for each year *
*                  (defined by cycle_span variable) as if they *
*                  were all common years (668 days).           *
*                                                              *
*                  As Paul Hill put it,                        *
*                                                              *
*   So while the Gregorian leap year rule can be summarized as:*
*   every       Y MOD 4   = 0,                                 *
*   excluding   Y MOD 100 = 0,                                 *
*   but include Y MOD 400 = 0.                                 *
*                                                              *
*   The Dechert Martian calendar leap year rule with the above *
*   correction can be summarized as:                           *
*   Every     Y MOD 2   = 1;                                   *
*   including Y MOD 10  = 0;                                   *
*   excluding Y MOD 100 = 0;                                   *
*   including Y MOD 500 = 0;                                   *
*                                                              *
*   Or the common year rules can be summarized as:             *
*   every     Y MOD 2   = 0;                                   *
*   excluding Y MOD 10  = 0;                                   *
*   including Y MOD 100 = 0;                                   *
*   excluding Y MOD 500 = 0;                                   *
*                                                              *
*                                                              *
*   Structure for database: D:\MARS\MARS.DBF                   *
*   Field  Field Name  Type       Width    Dec    Index        *
*      1  YR          Numeric        7               N         *
*      2  DAYS        Numeric        5               N         *
*  ** Total **                      11                         *
*                                                              *
*   After the table is filled, the values in the days field    *
*   are replaced based on the leap year rules.  Then all the   *
*   days are summed and compared with the length of the        *
*   natural year given as 668.592 days days per year.          *
*                                                              *
*   This table is then used by the marscal.prg program         *
*   to generate the individual months for each year            *
*                                                              *


cycle_span = 5000
ctr = 1
DO WHILE ctr <= cycle_span
   REPLACE yr WITH ctr - 1
   REPLACE days WITH 668
   ctr = ctr + 1
REPLACE ALL days WITH 669 FOR MOD(yr,2)   = 1
REPLACE ALL days WITH 669 FOR MOD(yr,10)  = 0
REPLACE ALL days WITH 668 FOR MOD(yr,100) = 0
REPLACE ALL days WITH 669 FOR MOD(yr,500) = 0

SUM days TO totaldays

? totaldays
? 668.592 * cycle_span
?IIF(totaldays = 668.592 * cycle_span,"right on", "off the mark")

* Program        : marscal.prg                                 *
* Version        : 1.0                                         *
* Date           :../1999                                   *
* Language       : dBASE IV                                    *
* Author         : Alan Dechert 916.791.0456, *
* Purpose        : Generates months for each year of the       *
*                  Dechert Mars calendar based on the values   *
*                  found in the mars.dbf table.  Each month    *
*                  has 56 days except February which has 52    *
*                  (in "common years") or 53  (in leap years). *
*                                                              *
* Description    : The program output goes to a text file      *
*                  (temp1.txt) defined by the SET ALTERNATE TO *
*                  command.                                    *
*                                                              *
*                  A cumulative day count is maintained and    *
*                  printed for each month.                     *
*                                                              *
*   Structure for database: D:\MARS\MARS.DBF                   *
*   Field  Field Name  Type       Width    Dec    Index        *
*      1  YR          Numeric        7               N         *
*      2  DAYS        Numeric        5               N         *
*  ** Total **                      11                         *
*                                                              *
* Revision History :../99 simplified code removing one      *
*                    function and one procedure.               *
*                                                              *
*                  :../01 added code (3 lines, the first 2  *
*                    and one other) to display approximate     *
*                    Gregorian date equivalent for the 1st of  *
*                    each Martian month                        *
*                                                              *
*                  : 2/6/01 corrected days/sols factor         *
gdate = {../2000}

USE mars

daytrack        = 1
daycount        = 1
cycle_span      = 500
cumulative_days = 1

DECLARE ordmonth[12]
ordmonth[1]  = "January"
ordmonth[2]  = "February"
ordmonth[3]  = "March"
ordmonth[4]  = "April"
ordmonth[5]  = "May"
ordmonth[6]  = "June"
ordmonth[7]  = "July"
ordmonth[8]  = "August"
ordmonth[9]  = "September"
ordmonth[10] = "October"
ordmonth[11] = "November"
ordmonth[12] = "December"


DO WHILE yr < cycle_span
   cntr = 1
   DO WHILE cntr <= 12
      ? ordmonth[cntr] + " " + LTRIM(TRANSFORM(yr,"999999"))   ;
      AT 23 - LEN(ordmonth[cntr] + " " +                       ;
      ? "Beginning with day "+ LTRIM(TRANSFORM(cumulative_days,;
      "999,999,999")) AT 23 - LEN("Beginning with day " +      ;
      ? "of the Mars Epoch" AT 14
      ? "(" +  dtoc(gdate + (cumulative_days-1)/.9732442977) + " Gregorian)" AT 12
      DO PrintDays
      cntr = cntr + 1


? "Done."

   ? "      Sun  Mon  Tue  Wed  Thu  Fri  Sat"
   DO WHILE daycount <= IIF(cntr = 2, IIF(days = 668, 52, 53), 56)
      ?? STR(daycount,4) AT daytrack * 5
      daytrack = IIF(daytrack = 7,1, daytrack + 1)
      IF daytrack = 1
      daycount = daycount + 1
   daycount = 1
   cumulative_days = cumulative_days + IIF(cntr = 2, IIF(days = 668, 52, 53), 56)