Best thread on this list in quite a while, thank you so much, both Jean and Eric.
I’m interested to see any info you could provide on the decision-making and resulting mplementation! Thank You, Troy > On Jun 20, 2020, at 11:27 PM, Eric Jensen <ejens...@swarthmore.edu> wrote: > > Hi Jean, > > As an astronomer, this message caught my eye, so I thought I’d make a few > comments. Before I do, I want to say that (a) I appreciate people working on > astronomical code, which can be tricky, and (b) my goal here is to be helpful > rather than annoyingly pedantic. If I appear to fail at (b), I hope you’ll > take it as a failure to convey the right tone by email rather than anything > else. Some of the background below is likely to be things you already know, > but I’m trying to be complete. > > That said, I’d encourage you *not* to make the selection of this parameter a > user choice, simply because it can’t be the case that both values are correct > - so you’re putting the user in the position of trying to guess which is the > right answer, which isn’t helpful to people who want to use your code. > > As you note, 15 degrees per hour (where “hour” here is a mean solar hour, > i.e. the time we normally use) is the speed that the Earth rotates compared > to the average position of the Sun. (The “mean Sun” is a convenient > fictitious object that moves at a constant rate throughout the year, whereas > the real Sun doesn’t, due to the Earth’s slightly elliptical orbit and tilted > axis.) But that rate combines both the Earth’s rotation on its axis and its > motion around the Sun - because it moves in its orbit each day, it has to > rotate a little more than 360 degrees (relative to some distant reference > frame) to get the Sun back to the same position in the sky. So the Earth’s > rotation rate compared to a distant reference frame (e.g. the stars) is > 15.04107 degrees per hour, so it rotates 360 degrees in 23 hours and 56 > minutes (a “sidereal day”), and rotates about 361 degrees in a mean solar day > (24 hours). (To easily derive that “magic value”, you can note that the > Earth will have completed 365.25 rotations with respect to the Sun in one > year, but one extra rotation with respect to the stars; the ratio > 366.25/365.25 is the same as the ratio of your two factors.) > > So if you’re calculating the Sun’s position based on some coordinate system > fixed to the (distant) sky rather than the Sun (which I assume must be what > you’re doing), you need to use the Earth’s rotation speed compared to that > coordinate system, which is the larger value (15.04107). > > The question still remains of why you’re getting different answers from > different other codes you’re comparing to. While it’s possible that some are > simply wrong, my best guess would be that some are taking into account > refraction by the Earth’s atmosphere and some are not. Because the path of a > light ray is bent as it travels through air, it turns out that when you see > the bottom limb of the Sun touch the horizon, the whole Sun is already below > the horizon! The exact amount of refraction depends on atmospheric pressure > and temperature, but on average it’s about 0.6 degrees (which coincidentally > is about the angular diameter of the Sun). That slightly more than half a > degree translates to about two minutes of time, so if that’s about the > difference you’re seeing, that could be the cause. > > The other reason could be different definitions of sunset, i.e. is it defined > by when the top limb of the Sun is at the horizon? Or the center? Or is it > some other distance below the horizon? (You list 0.833 degrees below, which > is a combination of the average refraction and the Sun’s angular radius, i.e. > it’s when the upper limb of the Sun would appear to go below the horizon, > given average refraction.) But other other codes might make other > assumptions that could lead to differences, i.e. they could be separately > accounting for refraction on top of a specified angle of the Sun below the > horizon. > > I suspect that I haven’t told you much (if anything!) that you didn’t know > already, but maybe the background is useful for others reading the thread. > But the bottom line is that again I encourage you to dig a little deeper to > try to find the source of the differences rather than exposing this as a user > choice. > > As for authoritative sources to compare to, as Bill Ricker notes, the USNO is > generally my go-to source. Unfortunately, their website has been down for > months as they reorganize, and unfortunately looks like it will still be a > while. (https://datacenter.iers.org/data/2/message_400.txt - that’s about > the IERS but the same is true for the astronomical parts of the USNO > website.) So without access to USNO, I’d be inclined to compare to NOAA, who > at least are clear on their website that they are including refraction - they > say they assume 0.833 degrees of refraction, though as noted above I think of > that value as including both 0.6 degrees of refraction *and* the angular > radius of the Sun, so maybe there’s still a little ambiguity there. > > If you want to compare against a different Perl solution, in my own > astro-related code I make use of Astro::Coords and Astro::Telescope, which > depend on Astro::PAL, a Perl interface to a widely-used positional astronomy > library. I’d be curious to hear what you find if you compare your code to > those. (Or perhaps I can try to make a comparison later, but I need to go to > bed now, and this message is too long already!) > > Hope some of this is at least a little helpful, and I’m curious to see where > you end up - > > With best wishes, > > Eric > > >> On Jun 19, 2020, at 12:03 PM, Jean Forget <j2n-for...@orange.fr> wrote: >> >> Following the creation of RT ticket >> https://rt.cpan.org/Public/Bug/Display.html?id=132336 >> I have decided to implement the proper precise algorithm in >> DateTime::Event::Sunrise (already implemented in Astro::Sunrise) and >> to check the values with NOAA's solar calculator and with Stellarium. >> >> After I have implemented the proper precise algorithm, there are still >> differences between the values obtained by Stellarium and the NOAA on >> one side and by DT::E::S on the other side. I have found that these >> differences are the caused by the "magic" values 15 and 15.04107 used >> in the module. >> >> 15.04107 is the angular speed of Earth's spin, while 15 is the >> combination of Earth's spin with the orbital speed of the >> Sun-Earth system. According to Paul Schlyter 's description of >> the precise algorithm, we should use 15.04107, but I have found >> that using 15 instead gives results closer to Stellarium. >> >> More abstractly, the problem is: when you are an low-level >> amateur astronomer, who is an authoritative source for you? >> Whom do you trust? Paul Schlyter or the Stellarium team or >> the NOAA? >> >> I have decided to defer this question to the module user. There is a >> new parameter, "algo", which lets the user choose between trusting >> Paul Schlyter more than Stellarium or trusting Stellarium more >> than Paul Schlyter. >> >> Now you may code either >> >> my $sunriseset = DateTime::Event::Sunrise->new( >> latitude => 42.3358, >> longitude => -71.2044, >> altitude => '-0.833', >> precise => 1, >> algo => 'Schlyter'); >> >> or >> >> my $sunriseset = DateTime::Event::Sunrise->new( >> latitude => 42.3358, >> longitude => -71.2044, >> altitude => '-0.833', >> precise => 1, >> algo => 'Stellarium'); >> >> Can you think of a better name for the new option? The creator of >> ticket 132336 suggests >> spin => 15.04107, >> or >> spin => 15, >> >> The problem is that the "magic" value 15.04107 is not easy to remember, >> especially if the module user knows even less astronomy than myself. >> While "Schlyter" and "Stellarium" are easier to remember. >> >> I will release the new version of DT::E::S to CPAN next week. >> If you have a better idea for the name of the new parameter, please >> tell me before this release. >> >> In the meantime, the updated code is available on Github. >> >> Thank you for your advice, >> >> By the way, this is the second attempt to write this message >> to the datetime mailing list. As it happened already, my first >> attempt was lost somewhere and did not reach you or the websites >> https://www.nntp.perl.org/group/perl.datetime/ >> and https://www.mail-archive.com/datetime@perl.org/. >> >> Jean Forget >
