That is exactly the point pnbruckner and I concluded. The altitude found on non-local sources have no meaning. Only the observation of the surroundings on the specific location makes sense.
The bug is that the description is plain and simple wrong.
You might call it common understanding, but then common understanding is wrong too.
Trying to argue for the common understanding here would be like arguing the earth is flat again. 
So your suggestion for the right description is ?
I mean , i took an average of what Google, and a ânationalâ map-provider suggested, and came pretty close (only had to change few meters ( in height above sea-level ) 
, to match the Temperature of Met.no (weather-integration) , with my own local temp-Device
Even coordinates are slightly different for these 2 sources (so my house floating around a bit, depending upon which map-source i look at) 
Sorry i canât see the forest because of all the trees , so looking out the window i have no idea of how high up in the sky i am
Temperature?
We are talking about sunrise and sunset. Not temperature.
Trees are actually one of the things you could take into account and the lower your height because of that.
I donât know what astral would do with a negative observer elevation.
Astral does have a second option for specifying observer elevation. Instead of a single number (i.e., height above ground level), it can also accept a tuple of two numbers, distance to, and height of, and obstruction, like mountains, or trees, etc. HA doesnât support that (but Iâm thinking about adding it to my custom Sun2 integration.)
Yeah, the sun integration is a bit lacking.
That is also why I am your excellent Sun2 instead. Thanks for that work!
Yeah, well others use the altitude for temperature as well, as it in fact makes a different
If you stand at the foot of a 300meter high cliff / mountain , you can be sure that the temperature, is not the same at the top , same principle as you use for the âsurroundingsâ when you consider sunrice/sunset, if you live in a valley, or on top
Yes i do also take the trees into account, when i consider sunrice/sunset, obviously
PS: You are not talking about sunrice/sunset, you are talking about cordinates and altitude and their sources in HA, which btw some other than native integrations makes use of
True, the altitude might be used for temperature too, which just makes it a lot more complicated, because the temperature needs to be calculated as altitude over sea level.
This is require some rewriting of HA to make it right or maybe the sun integration in HA should be ignored and Sun2 recommended instead, so the altitude setting in HA can be used for those other things.
Indeed , i even found that the definition of sunset/sunrise ( and sun above/below horizon ) variates, some sources use the tip of Sun , others use the center of the sun, who knows some might even use 2/3 or 3/4 
I donât even remember why i or when i installed Sun2, but it must have been doe to âmissingâ options in the native ( or because it was already dark when my very first automation kicked in 
PS: But speaking about sunrise/sunset , and OPâs question, i admit i have to agree, sun-integration does shows wrong time, for me more than the 8 min OP has âŚ
So i do really wonder where it âpicksâ the data from.
My coordinates are fairly right cording to met.no and google, same is the altitude (height above sea level , cording to met.no and google ) , only thing (data) that i could suspect as misleading is the âtimezoneâ we have to choose, as that places me in Stockholm, 600km north east of here ( and definitely at a level near sea )
Well atleast the timezone is right in HA , guess on just have to be âobservantâ upon which sources various integration make use of, and âadapt/fine-tuneâ best way possible
So, i had to follow your âsuggestionâ , while iâve never had sunrise/sunset correct here ( Living at 132-134m above sea-level (i guess) , so i always type 132 when i have to type in altitude.
Followed your example, more or less, i just typed 0 zero meters, in the Map-settings.
Sunset/sunrise changed to 1 minute off ! , compared to the +8 min off, i had when typing 132m
Now i just wait out the next weeks temperatures, to monitor the ( met.no )variation, compared to my outside sensors
So indeed âaltitudeâ in the context of the HA Map settings, is misleading ( kind of ) , and probably totally irrelevant in 99 of 100 use-cases here.
Go to the Met.no integration entry and click CONFIGURE. Iâm pretty sure its elevation setiing is independent. It probably uses HAâs value as a default when it is first set up, but you can change it to whatever you want afterwards.
EDIT:
In fact, i know it is. I changed my HA elevation seeing to zero, but the Met.no elevation setting is still 187.
i tried, but couldnât find these setting, even thou i thought that i must have type this info there also âŚsomewhere ⌠iâll check again
Bummer , one have to get use to this âconfigureâ button 
( yep, itâs still 132m ) so i guess im in the right spot there
While i have to get use to the Map Setting in HA , been 132m for about 3 years now !
Where is the WTH of the month, i can click 
Seems my sunrise/sunset times got totally wrong without me changing anything. I use it to trigger blinds go up / go down commands, so we noticed quicklyâŚ
Gone through this thread, checked my location on the map (even the elevetion set to 0mâŚ), everything seems correct. But the only way to get the correct sunrise/-set times is to set my location 1300km south, so instead of Switzerland, my home is set in Tunesia now 
. I would love to see my real home location in HA (not that it really disrupts anything but the weather prediction).
Use the Sun2 integration instead
Has anyone reported this as a bug? It seems like the altitude shouldnât be fed to the integration, since altitude here is height above sea level, not ground.
It seems that there is a lot of confusion here about calculating the position of the sun with respect of the observer.
While the formulas tend to get a bit bulky if you want to do it ârightâ, there is no such thing as âelevation above groundâ. Never. Itâs always elevation with respect to the âgeoidâ or simpler âsea levelâ.
On a smartphone or tablet you donât even need a map for this, as they all have built in GPS. As long as you donât be near an area of military interest, you can rely on that position to about <5m horizontaly and about 10-50m vertically.
You donât even need mobile network access. This works perfectly in open sea and in âairplane modeâ too. You should have a fair share of free sky above you and walking along in NY-City or Grand-Canyon might induce signal echoes that disturb the measurements. Mobile network just might speed up initial satellite fix, as the current trajectories can be fetched over a fast date line instead of decoding the same values from the data broadcasted by the satellites. GPS (and other networks) is a RADIO RECEIVER, utilizing an antenna.
It WILL get complicated, if you want to include the profile of your local horizon to determine, when the sun will climb behind the nearby ridge. This indeed will not only need YOUR 3D coordinates in space, but that of your surrounding too.
There ARE elevation maps that can (and do) be used to calculate that. Results my vary, depending on how thoroughly the country of your position has measured and published these values. Austria for instance has a near perfect elevation map for both ground AND buildings/vegetation. No problem to calculate the time when the sun climbs behind the little forest on the east side of a nearby hill or even the building up there.
You can time it within a second when either upper limb, center or lower limb crosses the line - refraction of the atmosphere included.
But a calculation that does not state what reference it uses, is to be trusted not better than 15 minutes, no matter how many different sources âagreeâ on a value. They may all be wrong. The proof is in the pudding. You might not even get DST correction. Without further notice I would think of âcenter of the sun above mathematical horizon with no air correctionâ. The latter is about a full Sun diameter or 2 minutes in difference that the ârealâ sun is early in the morning and late in the evening compared to âair free earthâ.
I know this, as I am an astronomer and know how to do it. Itâs math, not magic.
If you want the numbers from the penthouse of a sky-scraper, just add the building height to the ground level for your personal âelevationâ - or read the numbers from your smartphone.
No one will ever need elevation above âgroundâ for this. This might be interesting with flying a drone to comply with the regulations, but not for calculating celestial positions.
You keep arguing this everywhere however the built in python function calculates it this way, not HA. So youâre not correct in regards to Python and how python calculates it.
https://astral.readthedocs.io/en/latest/
If you have an issue with this, take it up with the dependency owner, not HA.
Or use GitHub - pnbruckner/ha-sun2: Home Assistant Sun2 Sensor
which can actually take the users position into account
Wait, scrolling months back, Sun2 was already mentioned and even joined by the author.
What petro is arguing against is this sentence.
The elevation is the elevation above ground for the sun library and not the elevation above the sea level.
The elevation in the sun library is used to handle this part of your explanation.
Preamble:
I use the term âsea levelâ for convenience, as it is the term that can be understood by everyone as âthe zero reference from which all geographical elevations are measuredâ. It makes some assumptions about the general shape of the earth, not the actual sea level. When talking about geographical locations, different definitions of this reference surface are in use, most notable the WGS84 reference frame.
When doing astronomical calculations on angles between an earth-based observer and solar system bodies (= nearby objects), the value of interest is the distance to the gravitational center of the earth (that defines the celestial trajectory), which can be a totally different number but mainly make a difference when calculating stellar eclipses by satellites or such.
So there it is:
At the summit of mount Everest (8,848 m) the maximum difference is 0.00338821 degrees.
Due to the very small difference the astral package does not currently adjust the solar elevation for changes in observer elevation.
r+elevation = âreference surfaceâ + elevation = commonly referred as "elevation above âsea levelâ " as it makes no difference whether you sit atop of an 8000m high mountain (whose âheightâ is given as âabove sea levelâ anyway) or an 8000m tall building at the beach.
Sunrise
The time in the morning when the top of the sun breaks the horizon (assuming a location with no obscuring features.)
All examples of âelevation above groundâ within Astral are in fact examples of elevation above sea level. At the summit of Mount Everest, the elevation above ground is certainly zero. Even the ISS sample, when talking about elevation above ground is as of how far âdownâ you see beyond a theoretical, flat horizon due to the curvature of the earth i.e. because of âthe sea level reference geoidâ. Reading this document, I always find âgroundâ noted, where âsea levelâ is meant. I bet the ground reference is the WGS84 definition not the very soil you are standing on.
And according to the docs, this value isnât even used to calculate Sun elevation.
If you sit in a plane flying over a big mountain, the height above ground will change by big numbers. This however does not change the time of sunrise as seen from your seat at a distant horizon. Itâs because your cruise height (elevation above sea level) does not change. Not surprising, the sun is not jumping up and down, mimicking the surface profile below you.
It seems as Astral does âUpper limb, with atmospheric refraction for unobstructed (mathematical, flat) horizonâ. That in fact should be fine within a few seconds of reality.
Even if calculated correctly, the effect on elevation for locations on earth surface (not satellites) should be very, very minuscule, not â8 minutesâ.
Iâll go through the code to determine where this weird dependency comes from.
Dude, quit pulling things out of thin air and read the code please. Hereâs the calculation HA uses:
And hereâs where HA uses it in the code, you can see we are calling a direct getattr on the Location object, i.e. pointing directly to the function in the previous link:
And the observer elevation effects at my location every 100 meters:
| observer_elevation | sunrise | offset in minutes |
|---|---|---|
| 0 | 06:55:36 | 0.0 |
| 100 | 06:53:27 | -2.1476496333333333 |
| 200 | 06:53:17 | -2.3217881166666667 |
| 300 | 06:53:00 | -2.59045255 |
| 400 | 06:52:43 | -2.875768133333333 |
| 500 | 06:52:26 | -3.1597126666666666 |
| 600 | 06:52:10 | -3.43689495 |
| 700 | 06:51:54 | -3.705733133333333 |
| 800 | 06:51:38 | -3.965997633333333 |
| 900 | 06:51:23 | -4.217969666666667 |
| 1000 | 06:51:08 | -4.462115966666667 |
| 1100 | 06:50:54 | -4.698954083333334 |
| 1200 | 06:50:40 | -4.928995183333333 |
| 1300 | 06:50:27 | -5.15272065 |
| 1400 | 06:50:14 | -5.370573733333333 |
| 1500 | 06:50:01 | -5.58295825 |
| 1600 | 06:49:49 | -5.790240266666667 |
| 1700 | 06:49:36 | -5.99275115 |
| 1800 | 06:49:24 | -6.190790833333334 |
| 1900 | 06:49:13 | -6.3846312 |
| 2000 | 06:49:01 | -6.574519183333334 |
| 2100 | 06:48:50 | -6.760679499999999 |
| 2200 | 06:48:39 | -6.943317133333333 |
| 2300 | 06:48:29 | -7.122619566666667 |
| 2400 | 06:48:18 | -7.298758716666667 |
| 2500 | 06:48:08 | -7.4718925333333335 |
| 2600 | 06:47:57 | -7.642166516666666 |
| 2700 | 06:47:47 | -7.809715016666667 |
| 2800 | 06:47:37 | -7.974662366666666 |
| 2900 | 06:47:28 | -8.13712385 |
| 3000 | 06:47:18 | -8.29720655 |
| 3100 | 06:47:09 | -8.455010216666667 |
| 3200 | 06:46:59 | -8.610627849999998 |
| 3300 | 06:46:50 | -8.764146366666667 |
| 3400 | 06:46:41 | -8.915647116666666 |
| 3500 | 06:46:32 | -9.065206366666667 |
| 3600 | 06:46:23 | -9.212895716666667 |
| 3700 | 06:46:14 | -9.3587825 |
| 3800 | 06:46:06 | -9.502930133333333 |
| 3900 | 06:45:57 | -9.645398433333334 |
| 4000 | 06:45:49 | -9.786243833333334 |
| 4100 | 06:45:40 | -9.925519733333335 |
| 4200 | 06:45:32 | -10.06327665 |
| 4300 | 06:45:24 | -10.199562483333333 |
| 4400 | 06:45:16 | -10.334422633333332 |
| 4500 | 06:45:08 | -10.467900250000001 |
| 4600 | 06:45:00 | -10.600036366666666 |
| 4700 | 06:44:52 | -10.730869983333333 |
| 4800 | 06:44:44 | -10.8604383 |
| 4900 | 06:44:37 | -10.988776766666666 |
| 5000 | 06:44:29 | -11.115919183333334 |
| 5100 | 06:44:21 | -11.241897866666667 |
| 5200 | 06:44:14 | -11.366743683333334 |
| 5300 | 06:44:06 | -11.490486183333333 |
| 5400 | 06:43:59 | -11.613153633333333 |
| 5500 | 06:43:52 | -11.734773133333334 |
| 5600 | 06:43:45 | -11.85537065 |
| 5700 | 06:43:37 | -11.9749711 |
| 5800 | 06:43:30 | -12.0935984 |
| 5900 | 06:43:23 | -12.21127555 |
| 6000 | 06:43:16 | -12.328024616666665 |
| 6100 | 06:43:09 | -12.443866883333333 |
| 6200 | 06:43:02 | -12.55882275 |
| 6300 | 06:42:56 | -12.672911966666668 |
| 6400 | 06:42:49 | -12.786153466666667 |
| 6500 | 06:42:42 | -12.89856555 |
| 6600 | 06:42:35 | -13.010165866666666 |
| 6700 | 06:42:29 | -13.12097145 |
| 6800 | 06:42:22 | -13.230998733333333 |
| 6900 | 06:42:16 | -13.340263599999998 |
| 7000 | 06:42:09 | -13.4487814 |
| 7100 | 06:42:03 | -13.556566983333333 |
| 7200 | 06:41:56 | -13.663634683333333 |
| 7300 | 06:41:50 | -13.769998399999999 |
| 7400 | 06:41:43 | -13.875671583333334 |
| 7500 | 06:41:37 | -13.980667266666668 |
| 7600 | 06:41:31 | -14.084998083333334 |
| 7700 | 06:41:25 | -14.188676233333332 |
| 7800 | 06:41:18 | -14.291713616666668 |
| 7900 | 06:41:12 | -14.394121716666666 |
| 8000 | 06:41:06 | -14.495911733333333 |
Thatâs an 15 minute swing over 8k meter elevation change. The elevation makes a difference and itâs not calculated by HA. Thereâs really no other proof you need. If you donât like this calculation, take it up with astral. Thanks.
