3

I use Google Earth Pro (7.1) on a daily basis to assess the acreage of US agricultural fields using the GE ruler.

What is the accuracy we can expect from Google Earth Pro for surface calculation? (the fields don't always have simple shapes and are sometimes in remote areas)

I found a few threads that partially cover this topic. But I didn't find any clear answer.

  • What were the "threads" that you found? – PolyGeo Jan 29 '17 at 23:03
7

I think that might be because there is no single, clear answer. I'll try to summarize my knowledge and the two most related questions:

  1. Google Earth, Google satellite, and Bing aerial accuracy
  2. How accurate are measurements in Google Earth?

Let's start by taking Google Earth and even GIS out of the equation and only consider how accurate a measurement on an image can be. If the resolution of an image is a pixel is 1m, and you are zoomed in far enough to see individual pixels (original, not resampled for display purposes), you can measure to within +/- 1m. Now let's get that into GIS, which requires both orthorectification and georeferencing. Both can introduce error. On top of which, you're looking at errors introduced by (re)projection.

Now let's put that all into Google Earth, which is providing you a seamless interface to multiple original sources. All of which may vary in their original resolution and quality of the above processes, so right off there is no single accuracy measurement. This is why Google has that disclaimer about accuracy:

Google makes no claims as to the accuracy of the coordinates in Google Earth. These are provided for entertainment only and should not be used for any navigational or other purpose requiring any accuracy whatsoever.

Our imagery varies from sub-meter resolution in major cities to 15 meter resolution for most of the earth's surface, with a global base resolution of 1KM. Since our database is constantly being updated, we cannot state a specific resolution for any geographic region.

Google acquires imagery from many different sources with many different file formats, projections and spectral characteristics. All imagery sources are fused into a single global database with a proprietary format that has been developed for the specific purpose of streaming to our client software.

Source, as linked in question 2 above.

All of that is just for the accuracy of the imagery you're measuring from. Then add in the user interface - how far out are you zoomed when you do your measurements? Far enough in to see those individual original pixels, if they're even available to you? Or out far enough to see the entire field, or the field and its surroundings? What display resolution is your computer running, and how accurate/precise is your hand with the cursor as you mark the boundary? What distortions are introduced by the projection(s) Google Earth uses, which may change depending on where you are looking. The best you can do is find something of known distance near your feature and see what it measures at to estimate the error in that area.

So the short answer is, it varies.

  • Thank you. This is a very good answer. What's type of rough accuracy should I have in mind when I calculate a surface with GE? +/- 5%, more, less? – Bap May 7 '14 at 16:47
  • 1
    @bap As explained above, there's really no way to answer that. The best estimation of error you could get would be to find a known dimension nearby feature, measure it, and compare it to the known value. The difference would give you an approximation of error in that area, but that may be more or less depending on how well you define the boundaries. You will be more accurate zooming to 1:20 at each corner (or tracing the entire boundary) than doing so at 1:2000. – Chris W May 10 '14 at 19:12
2

I've measured the lengths of 9 different runways and compared those "ground length" measurements against their lengths given by WikiPedia. The largest error was +0.0336%, the smallest was -0.0013% and the average was +0.0075%.

Runway lengths are measured between the outer edges of their threshold markings as explained here: https://www.faa.gov/airports/southern/airport_safety/part139_cert/media/aso-airfield-standards-quick-reference.pdf

Be sure to familiarize yourself with threshold markings before trying your hand at runway length measurements.

1

To test this I measured the area of an individual farm using the GEP Ruler and compared it to the legally defined area of this property contained on the title deeds (2334,6Ha) at the highest resolution possible on GEP (2299.7Ha).

The "accuracy" in this instance is therefore 98.51%; ie ~1,5% error.

  • 1
    You have assessed the variance in the precision of the value. This is only related to accuracy. The details of how the "Measure" tool in GE is not known to me, but I assume it is using some on the fly projection process. If you use data projected in Web Mercator to calculate distances and areas in ArcGIS the numbers are pretty much meaningless. This topic is constantly revisited as it is a readily available tool to share and discuss spatial data. If you really require accurate areas use an area preserving projection designed for your area of interest, not Google Earth. – Mike May 17 '16 at 17:56
  • I'll also add that title deeds are notoriously inaccurate in terms of stated areas, especially if they're older than the last couple of decades (and that's assuming that a newer one is actually using a new survey and not just restating the same description and area that's been passed down the chain for decades). This is why most of them say 'be it/there more or less'. There are other issues involved with metes and bounds descriptions as well that can affect areas. – Chris W Feb 8 '17 at 20:39

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.