Longest line of sight.

[DonatasG_]

As far I'm not familiar with GIS, my question can sound strange 🙂

Is it possible to calculate longest (theoretical) line of sight (in Earth)?

I found many tools which can calculate line of sight, but you must know at least point A, which in my case is unknown.

[XanderBakker]

You could use a DTM of the world and invert the heights (calculate max height - height) and use Sink—Help | ArcGIS Desktop to determine the sink locations (these will be the peaks). With those locations start the process of evaluating the lines of sight and see which one is the longest. This will take a considerable amount of processing capacity, but is feasible. 

[XanderBakker]

BTW, I just stumbled upon this list of The Ultra-Prominences Page that contains a KMZ you can download and use as your locations. This is a list of "ultra prominence" peaks and could be a good start (based on SRTM):

Prominence is the relative elevation of a mountain summit. Prominence is the elevation difference between a summit and the highest point (saddle) that separates that summit from any higher summit.

[DanPatterson_Retired]

You also have to remember that the highest locations need to be surrounded by low elevations in order to see past them.  So even if you were standing on top of the tallest mountain on earth, surrounding mountains, or even plateaus would obscure the ability to see a parcel of land further on.  

[DonatasG_]

You are absolute correct.

Even Everest is not the point with longest line of sight

[DonatasG_]

"Earth's curvature of course and standing on the ground"

Sure

I'm not talking about flatland, even today is Friday

[mbevand]

Hi @DonatasG_ it's 3 years later, but you may be pleased to know that I wrote a C program to do just what you wanted. I have 75GB of DEM data covering the whole planet (downloaded from http://www.viewfinderpanoramas.org/Coverage%20map%20viewfinderpanoramas_org3.htm) and my program can automatically identify the best vantage points (peaks) and the longest sightlines. It takes into account the Earth curvature obviously, and also atmospheric refraction (configurable).

I just finished writing a first version today. It's not optimized yet (single-threaded) but it does correctly identify the exact same longest sightlines as the https://www.udeuschle.de/ panorama generator does when I spot checked it against a few peaks. In particular it does find the 538km sightline from Pik Dankova.

I am going to make the program multi-threaded in the next couple of days. It's not very hard. My performance modeling indicates it should take about 5-10 days of runtime on the PC I'm using (AMD 5 PRO Ryzen 4650G) to scan the whole Earth.

I may not monitor this thread, but you can get in touch with me by email m (at) zorinaq.com

Cheers!

[mbevand]

I'm happy to report that I finished a multi-threaded version of my program. I have been running it for about 1 hour and it already found a previously unknown(!) 510-km sightline:

Viewpoint location: 38.61475,75.18748 (China)

Bearing: 115.6°

Location of end point on horizon: 36.518409,80.331178

Length of sightline: 510.3 km

According to the list at https://web.archive.org/web/20190516221621/https://beyondhorizons.eu/lines-of-sight/ this makes my finding the 2nd longest sightline on the planet

I have decided to enlarge the scope of my search and to analyze 2.3 million viewpoints. This should take multiple weeks to run on a beefier PC.

I have confirmed the 510-km sightline using the udeuschle.de panorama generator (link: https://www.udeuschle.de/panoramas/panqueryfull.aspx?mode=newstandard&data=lon:75.18748$$$lat:38.614... )