Help with understanding how distance is calculated in Earth

Hello all! Apologies for not introducing myself properly, but I came here wondering if someone here may have the answer to one question that's been bugging me for a long time seeing as no one else really has answered it elsewhere (or I'm just stupid  ). Hope you guys don't mind that, err rash entry . Anyhow, I'd like to learn how Google Earth measures distances (not in the mathematical way, strictly). Let me explain. If you had a straight 100m road with no changes in elevation, it would read 100m, no problem. Then, say there's a road that climbs up 100m at, say, a 10% gradient (so a roughly 6 degree incline), will it say the road is 100m long by measuring along the surface or will it say the 2D distance and the elevation separately, so the road would read as 99.45m long and 10m high at the end?

Welcome, this has been asked before and I know a few members can help you, so keep checking back for the answer, members are from around the globe so could take a day or two.

Thank you! I will try searching around and will wait, thanks for pointing that out. If anyone finds a relevant thread a link would be greatly appreciated!   Thanks again

This is hardly a rash question, ramthecowy. It's been my experience that there are a lot of things about how Google Earth is constructed that are deliberately not shared with the public. I, for one, have always been curious about how GE maps and models elevations, and indeed, if the same formats are employed in the new 3D cities imagery. There are some members of this forum who developed the original program, and they could no doubt answer our questions with certainty, but their membership is I suppose purely honorary, and as far as I know none of them have put in an appearance here since the day they enrolled. Most of them, I think, no longer even work for Google.



Other members may have better answers, and some may even have some certain knowledge, but for me (and most other users, I suspect), the only recourse is to use something akin to the scientific method: form a hypothesis, devise experiments to test it, and then form a working theory until it doesn't match observations or until better evidence in provided (like Seer or Opus Penguin might log on and explain how they constructed the program).



My hypothesis is that GE measures distance only on the horizontal plain. I assume this based on the fact that land surveyors must have done it that way, or square mile sections in mountainous terrain would be anything but square. Two experiments I devised:

• I traced route 308, the ancient road through Kurogari Pass between Osaka and Nara. Looking at Properties (right click in Temporary Places) and the Measurement tab, I get a distance of 2.37 km. Copying that same path, and changing the Altitude to 1000 meters absolute, the distance measurement is the same. (One gets similar results by using the Ruler tool from the bottom to the top of a mountain.)
  
  
• I used the Ruler tool to measure 1 side of a section of mountainous terrain in western Oregon, where forest management practices create large areas of clear cutting, usually following surveyed lines. GE measures the distance up and down over the mountains the same way the survey crews did long ago, as one mile.

Now here's an experiment for you to try. Use the Ruler tool to trace the road to the top of Pike's Peak (or some other road where you can find the odometer distance online). If my "theory" is correct, the two should differ at least slightly.

Attachments:

Two Distance Experiments.kmz (2.48 KB)

Thank you so much washi, you're a legend! As you can see I tried the Pikes Peak experiment as you suggested. Despite giving it some extra distance after the finish line and before the start point, it's still about 0.4 miles off. I'm convinced to support the hypothesis that Earth measures distance along the horizontal plane too    then stores the elevation data seperately. Really appreciate the time you guys took for me, I couldn't have have done it on my own! <3

I'm impressed that you went to the trouble to do the experiment, ramthecowy.



Actually, I have a confession to make. Before I posted the reply, I did a quick bottom-to-top-of-mountain test, and I did indeed get the same results, but I was using kilometers, and there was some rounding involved.



After I posted the reply, I did a more careful check, using meters as the unit. The results did not comply with my expectations. I made two lines, one with the Ruler tool and one with the Path tool. Because I "eyeballed" the second line, I was not surprised when the distances were a little different. What was surprising to me was how the measured distances differed when I float the two lines to become horizontal. (I've added the 4 lines to the first folder and attached it here.)



I still think that GE measures distance on a horizontal plane, but I think it's a bit more complicated, probably as a result of the secret ways the program calculates distance, and how it handles operations like conversion and rounding in each of the complex sub-programs that we visualize as tools. One has to recall the often heard caveat that one should never use GE for navigation or any other purpose where an error would put life, limb, or property at risk. When we measure distance in GE, we get remarkable accuracy, but it was never advertized or intended to be precise. And besides, the price is right!

Attachments:

Two Distance Experiments plus One More.kmz (2.85 KB)

Aw, it was just some 15 minutes of my time, well worth it considering it'll help others. And to be honest, I ought to admit the reason for why I ask. This information was crucial to me because I create a lot of tracks, racing circuits for a game/simulation (consider it how you will) where elevation is factored into the lap time a car puts down. In order to create these to some degree of realism, I'd need to know if the elevation data I'm grabbing off Google Earth is playing a role in distance and whether I'd need to factor that in by calculating backwards and having to go through all that nuisance   turns out I don't . While of course perfection is impossible it would be better to have it at least resemble the actual thing. Within the game calculations itself there is a margin of error as the simulated driver does not match the driving style of a human, doesn't make mistakes, isn't affected by weather conditions, etc Either way, as there are others like me in the game community I figured it would be a good idea to jump into the GE forums and ask as it's a tool lots of us rely on over there.

So yes, it won't be likely that these circuits will match reality very well, but it's better than nothing, and it's better to know rather than spend ages agonizing and guessing.

And considering it is a free tool, we've to be thankful for what we get indeed