[LEAPSECS] earth speeding up

[Tom Van Baak]

Rob,


> I think you mean many days shorter than 86400 seconds, not longer?


Right. Sign error. Thanks.


> Any betting person would say the plot shows an upward trend over the past 40 years. A simple linear fit suggests the earth will be back to an honest 86400 second day within a few years, around MJD ~59000 (year ~2020).



> http://longbets.org


Perfect. Are you game? Since there's no insider information or human control over LOD, we don't even have to exclude gov't employees from playing. I'll put $20 on 2020.


> Your plot is Steve's (per Morrison & Stephenson), upside down:


Yup. period vs frequency.


> The recent trend is indeed flatter than either historically observed or the value from lunar tides, but you're suggesting that flatter will turn upside-down.  Evidence from the long term trend suggests otherwise.


It's not unusual that really long-term trends dominate over long-term which dominate over mid-term which dominate over short-term, etc. But what's at stake here with leap seconds is not so much what will happen in 100k years even 100 years, but the next 10 or 20 years. It just looked to me that regardless of deep past or deep future, we're heading for a negative leap second -- in about the same slow time-frame as the discussion of possible change to UTC wrt leap seconds.

The possibility of a negative leap second is not in itself an argument for or against changing UTC. If anything, it's an argument for the LSEM (+/- leap second every month) proposal.

/tvb