NavList:

Yes, that's a weak article about a surprisingly mediocre dissertation. I would have expected better from the hallowed halls of the very famous "Michigan Technological University" in the U.P. of Michigan.

Just as your message posted here, I was writing up a reply on the S&T message boards. I sent it, but it has not appeared there for some reason. I'll add it here, too:

While the original dissertation and this article call attention to some little-known facts about sunsets, the genuinely critical points have been entirely missed

Sunrise and sunset times have been tabulated as key data in calendars and almanacs for centuries and especially since the early 20th century. The numbers that we all quote so routinely are the result of a STANDARDIZED calculation of a largely fictitious pair of events. The calculation assumes 34 minutes of arc of refraction (and 16' for the Sun's semi-diameter yielding a net, fixed value of 50 minutes of arc), and that refraction number is the focus here. But much more importantly, the standard calculation assumes an observer who can see the sea horizon or a large lake horizon (or equivalent, like a dry lake) AT sea level with his or her eyeball resting right at ground level. These are conditions which almost never occur. If an observer is 100 feet above ground, the visible sea horizon is depressed downward by 10 minutes of arc (a correction known as "dip" to navigators). This changes the times by about 60 seconds in mid latitudes. And of course, the vast majority of observers cannot see anything resembling the sea horizon or a lake horizon in the first place! So what does it MEAN when we quote the time of sunset in Denver, for example, based on this standardized model of sunset? There's nothing like the sea horizon. Depression of the horizon can't be quantified. Mountains and other topographic features determine the actual disappearance of the Sun. Atmospheric pressure at altitude is lower and therefore mean refraction is different. The tabulated sunrise/sunset times are quite simply, a standardized, legally-observed estimate, nothing more. We can easily produce more accurate numbers using proper physical models, and some people do so for simulation work and for detailed forensic analysis in accident cases.

Why, then, do we have these standardized numbers for sunrise and sunset? The answer is mundane, earthly: for data sorting and for legal purposes. Legal reasons are obvious enough. All around the globe there are public venues, like parks, that open and close based on the tabulated time of sunrise and sunset. Rather than fussing over astronomical details, the standardized computation provides a simple, unambiguous answer that everyone can agree on.

The data sorting question might be less obvious. When does the night end? Consider this coming Sunday night/Monday morning. It's going to be bitter cold here in New England, and we would like to know what the low temperature for the night will be. It is normal and expected in meteorology that the night is counted from sunset to sunrise and a full "day" is counted from one sunrise to the next. The overnight low temperature this "Sunday night" is currently forecast to be 7 degrees F where I live. That temperature refers to the temperature just before dawn, which by most calendrical reckonings is Monday morning. The standardized computations of sunrise and sunset allow us to create "bins" for data that fit the cycles of the weather. This meteorological sorting of data is the primary origin of the standardized computation of sunrise and sunset times. The exact, observable instant of sunrise is irrelevant.

Finally, the article mentions the celestial navigation consequences of incorrect sunrise and sunset times. This is absurd. At most, celestial navigators use the predicted times of sunrise and sunset for sight planning --so that they know when to be out on deck to shoot the stars in evening or morning twilight. Clearly the author of the dissertation, Teresa Wilson, is engaging in a bit of armchair speculation when she suggests that these times could affect navigational accuracy. Only a navigator of very low competence or perhaps a navigator trying to describe an unwieldy so-called "emergency navigation" scheme would suggest using the tabulated times of sunrise and sunset for celestial navigation.

Frank Reed
ReedNavigation.com/aboutfer/

To avoid confusion, this is a comment posted as a reply to the article on the Sky & Telescope here:
https://www.skyandtelescope.com/astronomy-news/we-dont-really-know-when-the-sun-rises/

Frank Reed
Clockwork Mapping / ReedNavigation.com
Conanicut Island USA