Accuracy of Google Earth vs. GPSr

[+longtrails]

If I want to determine the coordinates for a specific point and can identify that point exactly using Google Earth, should I always trust the Google Earth coordinates over my GPSr?

[+SparkyInCali]

Gpsr coordinates will be more accurate.Google earth is a nice tool but not 100% accurate and some areas are worse than others for the georeferencing in Google earth.

[yogazoo]

I concur with CrazyOn2Wheels, in my area of Montana, Google Earth is consistently a few meters south of actual (not bad). How have I determined this? With post processed, sub meter, Trimble data.

Edited January 14, 2009 by yogazoo

[+trainlove]

I thought that Google intentionally dithered the data by at least a few meters.

 

If I want to determine the coordinates for a specific point and can identify that point exactly using Google Earth, should I always trust the Google Earth coordinates over my GPSr?

The answer is NO, if you are planning on placing a cache using that. The 'rules'

You as the owner of the cache must visit the site and obtain the coordinates with a GPS. GPS usage is an essential element of geocaching. Therefore, although it is possible to find a cache without a GPS, the option of using accurate GPS coordinates as an integral part of the cache hunt must be demonstrated for all physical cache submissions.

Edited January 14, 2009 by trainlove

[+twolpert]

I concur with CrazyOn2Wheels, in my area of Montana, Google Earth is consistently a few meters south of actual (not bad). How have I determined this? With post processed, sub meter, Trimble data.

There have also been a few reports of isolated places where the GE data is farther out than that. Bottom line is that there is no substitue for feet (and GPSr) on the ground.

[timz2]

Everyone will agree that Google Earth can occasionally be better than your GPS-- just by coincidence, if nothing else. I think those coincidences may be more common than you think.

 

Search for survey markers at http://www.ngs.noaa.gov and pick a few good ones that you can get to, then when you get there look around for a nearby object that will be visible in the GE pic. Measure how far away it is and determine its true direction from the known point-- a sun sight will do fine, if nothing else. Use FORWARD at the NGS site to calculate the position of the visible object and compare it to GE's position.

 

No idea what's typical, but when I did that with a few marks in the NY area I found GE to be within a couple of feet. I also tried it in the San Francisco area, where you have to keep crustal motion in the back of your mind, but it still worked pretty well. I'll dig up a few examples in the next day or two.

[+StarBrand]

If I want to determine the coordinates for a specific point and can identify that point exactly using Google Earth, should I always trust the Google Earth coordinates over my GPSr?

In general, any handheld GPS is going to as accurate or more accurate than the Google Earth maps.

 

Some areas near me - the GE maps are off by 60 to 70 feet. I know of areas in Nevada where GE is off by 300 or more feet.

 

On the flip side - I know some local areas where GE spot on agrees with my handheld everytime.

[+Renegade Knight]

If I want to determine the coordinates for a specific point and can identify that point exactly using Google Earth, should I always trust the Google Earth coordinates over my GPSr?

 

You have a couple of issues here.

The GPS is accurate to 20' give or take. No map needed. Stand on the point of interest and you are 20' accurate.

 

Google earth, yeah, you can find the point, but the point and the coordinates assiged to it also have an error. In some locations I see roads 150' off of where they show on the arial photography. The coordinates and where they truly are on the ground as indicated by the satalite photo's can be off some distance.

[XC_Tracker]

If I want to determine the coordinates for a specific point and can identify that point exactly using Google Earth, should I always trust the Google Earth coordinates over my GPSr?

 

No.

[+edscott]

If I want to determine the coordinates for a specific point and can identify that point exactly using Google Earth, should I always trust the Google Earth coordinates over my GPSr?

 

Depends... On most of the US east coast the satellite images are clear and accurate. I find the regular Google image to agree with the published coordinates almost all of the time. When I am off by a bit the cache logs indicate that the GPS users are having the same problem, and the same distance, and in the same direction. However I've been directed to certain areas of the country where the satellite images are out of focus and impossible to read at the scale needed for geocaching accuracy; and in New Zealand I found the satellite image and the road map overprint, which in those areas includes the cache location marker, are out of registration which adds another step in calculating the actual location. The only way to accurately answer your question without any bias is to compare GPS data with Google data in the area that you are interested in. When the Google road map and the Google satellite image differ go with the satellite image data.

[+ghettomedic]

Google earth shows one of my caches over 200 feet off.

[+EScout]

Sometimes the super accurate benchmarks have a big white X painted over them. I assume this was done for coordinating aerial photos.

 

There is one near me that used to have this X, but it has faded. I just put the pointer in Google Earth over the spot where I know the mark to be and got: 33 46.881, -118 17.784.

From the NGS website, I know the benchmark's coordinates, when converted to DDD MM.MMM and rounded are: 33 46.882, -118 17.784. this is much closer than I expected, given the size of the pointer, and guessing where the middle of it covered.

[timz2]

given the size of the pointer, and guessing where the middle of it covered.

Turn on the ruler-- that will give you a crosshair. Might as well zoom in to 200 ft altitude or less, so the crosshair will be precise to the hundredth of a second that GE gives you.

[timz2]

Four examples from the Bay Area-- I'll get the NY ones later.

 

In Google Earth, go to 37 42 22.15 N 122 23 26.904 W which is the NAD83(2007) lat-lon for AB7679. The white east-west line is painted on the pavement to mark the station, which is in the curb.

 

In the other ones, you can't see the station itself:

 

37 48 23.507 N 122 19 52.441 W is the actual lat-lon of the north corner of the 10-by-13-ft concrete block in which AE5211 is set.

 

37 46 48.171 N 122 17 52.845 W is the actual lat-lon of the intersection of the north crosswalk line and the white line between the southbound lanes on that street. HT0882 is about 50 ft away from that.

 

37 47 44.245 N 122 16 47.386 W is the actual lat-lon of the south edge of the seawall in line with the west edge of the walkway to that yacht (the Potomac). HT0654 is 1.4 ft north of that.

 

As you see, GE can do astonishingly well as long as we stay close to sea level. More to follow.

[timz2]

Another one in SF--

 

The flagpole 20+ meters west of AE5209 is at 37 48 20.458 N 122 28 01.597 W. The flagpole itself may not even be visible on the GE pic-- just its shadow?

 

In NY, try 40 41 26.994 N 74 03 17.782 W which is the drain with the white arrow arcing around it (21 meters east of KV6856).

 

40 41 54.146 N 74 02 17.648 W is the inside corner of the wall on the SE side of Ellis Island, a meter from AB3000.

 

40 42 03.921 N 74 00 53.487 W is the SW corner of that rectangular Coast Guard office bldg east of Battery Park (30+ meters NE of KV0587). GE is off by 3-4 ft in this case.

 

GE is farther off at KV4023, 40 46 12.637 N 74 01 00.652 -- the mark is in the middle of that white stone that's visible west of the east edge of that semicircle. GE is off by 6 feet; I wonder if it's a coincidence that this is the first example that's not close to sea level.

Edited January 16, 2009 by timz2

[timz2]

All the above examples are based on good GPS stations, which means you can't see the station itself, which means you have to take my word for the location of the visible object (unless you go there and do the measurement yourself).

 

But if we're willing to accept third-order buildings and so on, there are thousands of places we can use to check GE's accuracy. How about California's capitol dome, which is JS4124 at 38.576655 n 121.493647 w. Send GE to that lat-lon and you seem to be off a few feet-- but for all we know that could be due to GE's camera not looking vertically down on the dome. For all we know GE has homed in on the correct point on the ground, invisible under the building.

 

A good example of that is the Washington Monument HV4442 at 38.889467 n 77.03524 w, where the GE cursor shows that position being nowhere near the top of the obelisk-- but pretty near the center of the base, which is what it should be. (Incidentally, it seems that station is supposed to be good to a couple of centimeters.)

 

Apparently the camera was looking just about vertically down on the Transamerica Bldg HT3117 at 37.79516 n 122.402772; likewise the Daniels and Fisher tower KK1706 at 39.748098 n 104.995680 w.

 

Just grabbing a few water tanks at random-- ME3155 is 41.796730 n 87.956935 w; the camera didn't see it vertically, but you can see how close the cursor is to the center of the base of the tank. Ditto DG2665 at 33.612797 n 84.449494 w and CS2834 at 32.956925 n 96.763673 w.

 

Maybe GE can't really be that good-- maybe they collected all that NGS control and fitted their images to it. They knew the actual lat-lon for all those water tanks and conceivably they based their lat-lons on those known points. Can't say much about that possibility, but note that they didn't get fooled by KV4906, where the actual tank (presumably a newer replacement for the tank C&GS triangulated) is 50+ feet away from the 40.788858 n 74.254382 w position given for KV4906. And they didn't get fooled by the Ellis Island tanks: KV4307 at 40.700156 n 74.039921 w actually is the one that's been demolished; KV4310 at 40.700014 n 74.040103 w is the one that's still there.

[+tec_64]

http://www.personainternet.com/mile77/3709562.htm

[+tec_64]

I've also done the aerial photo registration mark (+ or X) on a rock thing and got very good and consistant accuracy results in Google Maps vs my ol maggy 210.

[+coggins]

http://www.personainternet.com/mile77/3709562.htm

I have no easy access to the older almanacs.

Let me help you out there...

http://www.navcen.uscg.gov/archives/gps/