NavList:

David Iwancio, you wrote:
"this has the potential to be a mess for CelNav."

There is always potential for stupidity out there. But there's no reason it should be anything more than a minor transition. Of course the rabid paranoia of some so-called authorities in "celestial navigation" on this matter has been a weight on the process for decades. It's hard to say how much of this paranoia results from pig-headedness or simple conservatism and how much of it arises from a basic misunderstanding of the tools of celestial navigation. Authorities? ...not understanding the underlying basics of celestial navigation?! Yes, happens all the time in many fields. Note that I am not worried about the almanac teams themselves. Rather I'm talking about other self-appointed authorities with some official status who may try to insert themselves into the process (for a NavList example, see this post from a "Royal Institute of Navigation" authority from almost eight years ago).

You added:
"As noted on page 254 of the Nautical Almanac, "The time argument on the daily pages... is UT1."  The nautical almanac offices will need to decide if they want someone to blow the dust off of the automated routines that produce these almanacs and decide whether to keep the tables at UT1 or change them to New UTC specifically."

Keep UT1 and publish the offset annually. That's the simple way to do it (see my paper below). Whether "they" also choose to publish a more frequently updated value of DUT depends on who "they" are. There is very nearly zero requirement in the 21st century to depend on nautical almanac "authorities". 

And:
"They may be compelled to change to New UTC if that ends up being what radio time broadcasts end up moving to"

Ends up being?? In what possible universe would radio sources of UTC not broadcast the "New UTC"?? Of course they would! But this does not compel listing the Nautical Almanac data in terms of UTC any more so than it has in the past. Leave it at UT1 and publish the offset.

You added:
"(otherwise there'd be an added step of "subtract X seconds before entering the Increments & Corrections Tables")."

Better instruction would be "subtract X seconds from UTC before entering this volume". Print it right on the cover! Print it on the title page. Print it at the head of every page for the first decade of the transition. App developers will happily produce apps that display UT1 directly, and for some navigators that will be a happy solution.

You speculated:
"However, the change itself has the potential to break radio time broadcasts themselves. The signals typically have information on UT1-UTC encoded into the signals (e.g. the "double ticks" I can hear in WWV and CHU), but the encoding systems were never intended for that difference to exceed +/- 0.9 s."

The number of users of those ticks (apart from toys) is nearly zero. They would be dropped and, if desired, replaced by an audio announcement, e.g. "DUT today is 6.4 seconds". The tick system was designed half a century ago for the primitive automated systems of that era. Automated systems can now decipher speech. They could decipher significantly more complicated audio tones if that alternative seems necessary. And yes, a system for this should be designed during the next dozen years. Some governments may realize that shutting down their time signal broadcasts is more efficient than modifying and maintaining them. But in principle this is no problem.

And:
"Humans can adjust to the change in time signal formats, but the hard-wired "atomic clocks" that rely on longwave radio signals can't. Any change in the signal format might cause these timepieces to malfunction in ways that users might not even be aware of."

Yeah, but really, so what? Those clocks are basically toys. In my experience, most owners of such devices spend a lot of time complaining about the fact that they don't work properly much of the time in the first place and never use them for anything except conversation. 

You wrote:
"In the US, I think power plants on the electrical grid still aren't allowed to alter their alternating current timings in a way that might break synchronous clocks, which rely on 1 day being 86,400 s * 60 Hz cycles long, and those clocks went away around the time one could literally get a quartz watch from buying breakfast cereal."

Clocks electrically rated like that didn't quite go away until just a few years ago. The resetting function, which was formerly also part of the electrical power network, disappeared sometime after 1976 (I saw it function once that year). This historical system created an amazing network of exact time synchronization from c.1930 onward, which was useful for businesses (especially in activities like stock trading, etc., that opened and closed at exact times). The synchronization signal, which allowed all clocks with simple circuits to reset themselves to the top of the hour, was apparently dropped decades ago. Synchronization was gone, but rating remained.

The alternating current rules requiring exactly the correct number of "thirds" in a day served to "rate" clocks and the regulatory requirement for that in the US lasted much longer than the synchronization "reset" signals (thirds are the next sexagesimal division after "seconds" and since the US electrical grid is 60 Hz, it turns out that our wall clocks for decades were counting thirds). They were still useful since any clock plugged into wall power could use that electrical signal as its source of timing (rate) if not "time", and once set correctly and not disconnected such clocks would keep excellent time for years. Examples of clocks like this were found in common kitchen appliances until very recently. If you have to reset the time on your kitchen stove or your microwave oven after a power outage, then you probably have one of these clocks. The rating requirement for the electrical current power grid was finally relaxed just a few years ago (three years? five years?). I have waited out of geeky curiosity to see if my 15 and 30-years old kitchen appliance clocks would become less accurate, but we have enough power failures here (island living!) that there's been no reliable evidence of it.

You concluded with a little whimsy:
"How I Learned to Stop Worrying and Love Stark's Tables"

First a minor quibble: I object most strenuously (that's a joke!) to using Bruce Stark tables as a shorthand for Lunars. He did not improve the math or science of lunars though he contributed greatly to publicizing them, and any non-broken method from the historical period can produce results just as good as his idiosyncratic modern tables. Besides, "How I Learned to Stop Worrying and Love Lunars" has better alliteration!! :)

But since you mentioned lunars, that's a good example of the things that can go wrong when almanac time standards change.

From 1767 through 1833 the official British nautical almanac --and other publications derived from it including various un-licensed American copies-- employed GAT as the time variable for lunars. That's "Greenwich Apparent Time" or "Greenwich Sundial Time". An observer would shoot and clear a lunar and then look up the predicted lunar distances in GAT. A little interpolation would then produce the GAT for the observation. Next the observer would use the altitude of the Sun (either taken as part of the lunar observation or separately at some more appropriate time of day) and calculate the Sun's hour angle. That hour angle, either added to or subtracted from 12:00:00 would be identical to the Local Apparent Time or LAT at ship. The difference in time is the longitude: subtract LAT from GAT, and you have your longitude in time units. Easy, right? And notice I didn't say anything about the "Equation of Time" since both times are "Apparent" time, equivalent to unadjusted sundial time.

Beginning in 1834 many antiquated features of the official British nautical almanac were replaced or revised and some modern tools were added. In particular the lunar distances were published for GMT instead of GAT. From this date a navigator could shoot a lunar and simultaneously note the GMT on the chronometer. Then open the almanac, do the interpolation, and get the GMT found from the lunar. That could then be compared directly with the chron. GMT recorded at the time of the lunar sight. Any difference would be noted, and if large enough, the chronometer error might be adjusted. Easy and fast and no need to measure a separate altitude. Alternatively, if a navigator felt compelled (by education and tradition) to observe the Sun's altitude and do the traditional calculation ending with longitude, then the hour angle of the Sun also had to be corrected by the Equation of Time. This was an extra step in lunars from 1834 onward. 

One might think that every navigator in 1834 would have heard the news that the tables have changed. Surely it would have been posted to all the internet sites! Ok, no internet sites, but surely they would have received some "notice to mariners"? Nope. Nothing like that existed either outside formal services like the Royal Navy. Navigators had to dig to discover the required modifications to the process of working lunars. Some certainly gave up on lunars after the transition, and that helped accelerate the decline of lunars, which were already fading by 1834. Others, without doubt, puzzled it out in their calculations. There is primary-source evidence of this, in fact.

The key factor to ensure no collapse of celestial navigation capabilities following the elimination of leap seconds is to ensure that navigators are properly informed. This assumes, getting back to your first point about a "potential...mess", that almanac "authorities" don't screw it all up. No matter what happens, at least some celestial navigators will bail out. Like those lunarians who gave up after the transition from GAT to GMT, there will be some 21st century navigators, traditionally trained, who will decide that it doesn't work anymore --and they will rant about it online, too. But this is just a matter of education. If we get the word out --much easier now than in 1834-- there should be no problems.

Frank Reed
Clockwork Mapping / ReedNavigation.com
Conanicut Island USA

File: