Effects of changing second definition in atomic clocks to compensate leap seconds

Question

The International Telecommunication Union (ITU) is an Institution to standardize the Universal Time Coordinated (UTC). The ITU has been discussing to get rid of the so-called leap seconds for various reasons. A leap second is added to accomodate for the slowing rotation of earth and the resulting difference between our earth's time and the astronomical time. By current definition, a leap second is added, if the Bureau International des Poids et Mesures (BIPM) find out, that earth's atomic clocks have a difference to the astronomical time of 0.9 seconds. Half a year later, the leap second is subtracted from earth's time by stopping all clocks for a second.

As this process is not deterministic, the ITU discussed to get rid of the leap second, as there always are quite some computer systems, which are unable to cope with leap seconds without crashing (e.g. the russian GLANOSS system.

I started a discussion with one my professors about simply changing the way our atomic clocks count a second, which is currently defined by the quantum changes of caesium atoms - one second equals to 9.192.631.770 quantum changes. I suggested to simply redefine this definition to a few quantum changes more per second, making a second effectively longer and thus making the reason for a leap second irrelevant. My professor contradicted, that this would a huge mistake, where we unfortunatly had to stop the discussion.

Beside the obvious organizational problems of changing this definition, I cannot think of any technical or physical problems, as the mapping of an arbitrary number of quantum changes to any time interval is not something physically or mathematically derived (as for example PI). Also, some computer systems needing a very high time resolution (as GPS or Galileo e.g.) already do not take leap seconds into account at all - so to summarize, I do not see any problems with my argumentation so far.

I have a test tomorrow and got the strange feeling, that this discussion will go on then, changing my final mark. Can you think of any technical, mathematical or physical problems I did not think of yet? This really drives me mad.

Answer 1

okay, I've talked to another professor, and it basically comes down to the Système international d’unités (SI) or Common Unit System. Most of the defined Units are based upon a very small set of variables like the weight of an atom or the definition of second. If you change the length of a second to accomondate something "mundane" like the position of the earth, you have to redefine nearly all of the SI units as well, which would be a so massive impact on all scientific areas, that nobody would dare to change it.

So, in conclusion it is as I suspected: The reason for not changing it is an organisational one, not a physical. There's no physical reason to not change earth's second definition and leave the astronomical system intact as it is.

Answer 2

Extremely later answer, sorry. Your suggestion:

"I suggested to simply redefine this definition to a few quantum changes more per second, making a second effectively longer and thus making the reason for a leap second irrelevant."

The physical problem to redefine SI-second this way is the mixture of two completely different procedures how to measure time. On the one side you still want to use atomic clocks, but on the other side you want to couple it directly to astronomical observations. This coupling would require permanent redefinitions of SI-seconds because the rotation speed of earth is irregular and changes. Totally impracticable from a scientific point of view.

It is the core idea of SI-seconds and UTC to have seconds with constant length in time - a valid physical motivation! Personally I find the idea of leap seconds okay. Nobody has made a better suggestion for coupling of astronomy and atomic time until now. The ITU discussion is only about if we shall remove this coupling and accept our calendar day as atomic day independent on astronomical observations or not (in extreme case some hundred years later midnight would turn to noon).

Else you are also right when you say that a redefinition of SI-second implies a big organizational problem.