Steve, I think this is quite an improvement. I agree with you about the median, especially as many of the predictions are not necessarily independent. I haven't reproduced the curves of your new plots, and I don't have a missing estimate of DeltaT, but I have some suggestions:
1. The plot of "Year 2100 Estimates" has a warning: "Estimates which are good during 20th/21st centuries are generally poor over the long term history". The same is repeated in the text, and this gives the impression that the predictions would be invalid over the long term future. Considering the New York Stock Exchange as an example, it makes little sense to judge predictions on any scale by how they extrapolate back to the 1817. Another way to look at it is that estimates made with old data would be expected to apply better to old data - just like estimates based on recent data would be expect to apply better to the near future. So I suggest you rephrase your comments something like this: "Estimates based on recent data are optimized for the future and not intended to be extrapolated to the past". (My own personal opinion is that there will be decadal fluctuations, and the long-term slowdown will revert to a more rapid deceleration after/if the ice caps stop melting and t he centuries-long glacial rebound stops - this is why I show three cases of a deceleration rate but use only one value for the LOD as of now. Estimates for sea level rise are 1 meter by 2100, and we have 70 meters to go - but that's even harder to predict.) 2. With one exception, I suspect the difference between the many predictions boils down to just two parameters: the assumed value of LOD at some fiducial time, and the value of the parabolic term starting at some fiducial time. To make your plots, you probably have created a table with just these parameters, and converted the fiducial time to be 2020 for all curves. If you provided a table of these two parameters, it would eliminate a lot of confusion about the differences between predictions. You could even have entries showing the time of the last datum used for the prediction (or publication date), and also how much each of the two parameters would contribute to UTC-UT1 by 2100. This trivial with a spread sheet and will help people who are trying to figure things out in-depth. 3. I believe what appears as a thick red curve on your "Various Estimates" figure is the actual data from the IERS, which appears thick on my screen because the vertical bars are compressed. It's great to accentuate the actual data, so I suggest you could make this more clear by adding something to the legend after "IERS EOP C04" - perhaps you could add in parenthesis: "(the observations)" Good show! -----Original Message----- From: LEAPSECS [mailto:[email protected]] On Behalf Of Steve Allen Sent: Monday, February 02, 2015 9:17 PM To: Leap Second Discussion List Subject: [LEAPSECS] the year 2100 On Wed 2015-01-28T17:05:16 +0000, Poul-Henning Kamp hath writ: > So why don't you simply list all the estimates as we find them ? > > As far as I can tell, these guesses/estimates are not inferior to the > ones you already list ? > > Listing them all would be sound science, and maybe the good ol' > "Trust the median" will eventually help us, given enough estimates ? I'm not so sure about that median. I remember that science by vote said the spiral nebulae were inside our own galaxy. Predicting the evolution of Delta T remains akin to predicting the weather in the core of the earth, the amount of sea level rise, etc. I now have a framework for presenting the Delta T estimates that I know, and for continuing the discussion. This web page lays them out http://www.ucolick.org/~sla/leapsecs/year2100.html If anyone's favorite estimate for Delta T is missing then kindly point out the citation for its values. -- Steve Allen <[email protected]> WGS-84 (GPS) UCO/Lick Observatory--ISB Natural Sciences II, Room 165 Lat +36.99855 1156 High Street Voice: +1 831 459 3046 Lng -122.06015 Santa Cruz, CA 95064 http://www.ucolick.org/~sla/ Hgt +250 m _______________________________________________ LEAPSECS mailing list [email protected] https://pairlist6.pair.net/mailman/listinfo/leapsecs _______________________________________________ LEAPSECS mailing list [email protected] https://pairlist6.pair.net/mailman/listinfo/leapsecs
