Here is an eclipse map I've been working on. I still have some work to do, but it's good enough for people to start using now. Has a few features others don't, so might be more useful.
Another eclipse map
#2
Posted 22 December 2023 - 08:34 PM
Nice job, I like it. I used it to measure the size of a clearing I'm planning to use. If I may offer a suggestion, you could add a checkbox to change the measurement from miles to feet.
#3
Posted 23 December 2023 - 03:51 PM
Cool! Nice work.
#4
Posted 25 December 2023 - 12:35 PM
Nice job, I like it. I used it to measure the size of a clearing I'm planning to use. If I may offer a suggestion, you could add a checkbox to change the measurement from miles to feet.
I made it so that anything under 1/4 of a mile is measured in feet.
#5
Posted 25 December 2023 - 08:45 PM
I made it so that anything under 1/4 of a mile is measured in feet.
Perfect. Thanks.
#6
Posted 27 December 2023 - 03:45 PM
For those interested, the code is public domain, and is all self-contained Javascript. To make it easier to browse the code, I've added a Github repo: https://github.com/g...areclipseviewer
- jrussell likes this
#7
Posted 30 December 2023 - 04:01 PM
Very nice, thanks!
#8
Posted 23 January 2024 - 10:14 PM
Here is an eclipse map I've been working on. I still have some work to do, but it's good enough for people to start using now. Has a few features others don't, so might be more useful.
That is awesome! Thanks!
#9
Posted 24 January 2024 - 10:35 PM
Here is an eclipse map I've been working on. I still have some work to do, but it's good enough for people to start using now. Has a few features others don't, so might be more useful.
That is really awesome. I used it in my house (4m19s) and my church (4m23s)
Also very cool to be able to measure a direct line from my point to the center line so easy as well
#10
Posted 25 January 2024 - 09:45 AM
This is a really nice map to have. I could tell from other maps I was going to be pretty close to the center line where I plan on watching from. With this map I was able to measure and I'll be within a mile.
#11
Posted 30 January 2024 - 10:05 PM
#12
Posted 31 January 2024 - 10:18 PM
Pretty cool. How hard would it be to look up the time zone of the center of the map?
It's on the "to do" list with a pretty high priority. Hadn't been working on it for the past few weeks, but plan to start working on it again now.
- Bigred77 likes this
#14
Posted 17 March 2024 - 10:39 AM
Thanks for making this public.
I've got a couple of technical questions. I've been calculating the contact timings myself and agree with your values to within a second or so (and with Jubier) when I use the same elements and Delta t. I see you're using the VSOP87/ELP2000-82 elements from https://eclipse.gsfc...hp?Ecl=20240408 (as does Jubier). It sounds like those are fairly old - have you found any newer ones? The ones used on https://eclipsewise....pr08Tgmapx.html are based on JPL DE430 but given to lower precision. I only see sub-second differences in the contact times using the JPL DE430 values versus VSOP87/ELP2000-82, so maybe the old ones are good enough.
I've only skimmed your code but can't see that you use the observer height in your calculations - is that right? It should be easy to implement that, and potentially geoid to ellipsoid conversion.
#15
Posted 19 March 2024 - 06:19 PM
I have written an article showing how to compute the Besselian Elements using the JPL DE. In the example, I use DE405, but it can be directly replaced with any DE version. The VSOP elements are at the top, and the ones computed with DE405 are at the end, you can see there is very little difference. I ended up using the VSOP elements mainly because that's what other maps used, and I didn't want to have to try to convince anyone that mine was actually more accurate, and they likely wouldn't believe me anyway. But the really significant variable is Delta-T, which I think everyone is using a slightly different value for. I honestly can't remember how I came up with the one I'm using, but the end goal is to use data from the IERS automatically, I just hard coded a value to have something for April 8 so I could focus on other functionality.
Sticking with the VSOP elements should be good enough for any modern time. Last I checked the difference in the position of the Moon between VSOP87 and the latest DE was only about 100 Meters. If you go centuries into the future, or past, you'll likely start to see significant differences.
You're right in that I do not account for observer height. And that's because the algorithms Bessel developed treat the Earth as a spheroid, and provide no mechanism for doing so. But also, the amount of data required to do so would be enormous. In my unfinished article on how to use Besselian Elements, I do start off with the idea that Bessel's algorithms' days are numbered. It will likely be much easier, and more accurate to just use 3D models of the Moon, Earth, and Sun to make predictions in the future. But it's unlikely that the use of altitude for anyone on the ground will make any discernable difference.
- R Botero likes this
#16
Posted 20 March 2024 - 04:54 PM
Thanks for the reply. I see a difference of 2.5-3s at C2 an C3 with VSOP87 vs your DE405 elements for a point near the centreline. The difference is symmetrical around mid eclipse with DE405 giving about 5 seconds longer totality, which suggests the difference may be due to moon or sun size, rather than position as you point out. Of course these differences are comparable to larger lunar limb corrections, so arguably aren't too big a deal.
Your Delta t is very close to the latest measurements. Fortunately Delta t is fairly stable recently, but preditions for 2024 from around 5 years ago were around 5 seconds high, so unless various websites or software have been updated recently predictions would be biased early.
You may have noticed the flurry of questions about Delta t in another thread. It would help if people listed the various assumptions used - ephemeris, sizes, Delta t, limb correcctions or not, elevation used or not. Maybe you could include a link to such a summary on your page?
About elevation, a 2000m height difference would shift the contact times by around 2s for an eclipse early or late in the day, so arguably worth implementing. Including height is trivial for the local circumstances (contact times, obscuration) - you just include it in the conversion from geodetic to geocentric. Maybe you mean calculating the various curves will be harder with elevations? But I'd say the contact times are where precision is most important, so it makes sense to include elevation there.
When you say "the amount of data required to do so would be enormous", wouldn't elevation data be included with the map data you use?
- gmiller123456 likes this
#17
Posted 22 March 2024 - 10:16 AM
Thanks for the feedback. I would not expect the difference to be anywhere near 3 seconds, I'll have to go back and look. Mine does not do limb corrections, I believe Xavier Jubier's does. I doubt any use elevation, as the map data does not include it, though there are services and data available for it. Even though, as you pointed out, someone on Mt Everest, or even a small mountain at sunrise or sunset would see a difference of a few seconds, such cases are rare, most eclipses start and end over water. My code does allow supplying the elevation for computing the local circumstances, but not for any of the curves or outlines. I wouldn't expect the different sizes of the Sun, Earth, and Moon that have been used in the past to account for any significant differences, I haven't actually done any comparisons, but they're likely just tiny fractions of a percent. I should have noted where the values I used came from in the source code, but don't see a reference, so I'll need to add them.
Edited by gmiller123456, 22 March 2024 - 10:17 AM.
#18
Posted 23 March 2024 - 12:16 PM
Jubier's map website does both: the contact times listed in the local circumstances popup are uncorrected, but he includes the lunar limb corrections in the "LC" column. The popup also displays the elevation of the site. (I wonder where he gets the elevation data.) I agree with his values to his stated precision (0.1s) when I use the same elements, Delta t, and elevation.
Jubier's map site elements use the IAU 1976 solar radius, 959''.63. If you used instead 959''.98, I estimate totality would be about 1.6s shorter at a spot near the centreline.
Edited by timelapser, 23 March 2024 - 12:17 PM.
#19
Posted 24 March 2024 - 05:51 PM
There are times where one would not want to have the slight 1 or 2 second error corrections made that the "lunar limb" or "elevation" provide — especially when comparing the differences in the event circumstances of various locations, as is a key feature of this map.
Stick with the elements you think are the best — VSOP87 I think you said. Either way, a shift in the Moon's location of 100 meters would amount to less than one-tenth of a second difference in time.
The variables that can make the biggest difference in the timing are the ΔT and the ΔUTC. It can make a difference of 5 seconds or more if the wrong ones are used. ΔUTC is 69.184 seconds, and only changes when a leap second is inserted. ΔT = ΔUTC - DUT1. DUT1 is currently -0.010918 seconds and can be found here from the IERS: https://hpiers.obspm...op-pc/index.php . It won't change much between now and the eclipse. That makes ΔT currently 69.195 seconds.
I like this map a lot. You can set the umbra and penumbra shadows to any point in time during the eclipse.
Having the time zone adjust automatically is a big plus.
Edited by MrSyzygy, 25 March 2024 - 04:29 PM.
#20
Posted 02 April 2024 - 11:23 AM
I'm seeing a John Irwin calculated solar radius of 959.95 arc sec. Is that more accruate for this eclipse?