Bencmark for Gauging Accuracy?

[deadspot]

Well, this is my first post and I hope I am not asking the obvious. I have spent some time (too much) on the forum learning what I can. I picked up a used Garmin GPSMap 60 and have been playing around with geocaching and finding benchmarks and also mapping (e.g. custom maps, etc).

 

For kicks I wanted to know what others use as a known control point to get a sense of how accurate their units are. From reading I know the accuracy isn't a simple thing and there are various factors involved. At first I thought the answer was to find a local benchmark, take readings, and then see how "close" it was. But in reading about benchmarks I understand some are "adjusted" and some are "scaled". Even then I am not sure if I can assume that the coordinates shown for each benchmarks are in fact the "exact" coordinates to use.

 

Is that the only option? are there better ways? Like picking coordinates off one of the online maps (e.g. a building or structure)? Or is it finding a "GPS Only" benchmark from the Geodetic site's list of benchmarks. My next thought was to find a local surveyor that might have some documented control points they would share. Any thoughts would be appreciated.

[Grasscatcher]

Go to any of your local benchmarks that are described on their data sheet(s) as having "Adjusted" horizontal locations.

 

Those coordinates have been established to a greater accuracy than your GPS receiver can determine.

[deadspot]

Thanks - I appreciate the feedback. I just wanted to confirm I understood. Can one assume at all that the benchmarks that indicate "GPS OBS" are any more accurate? Meaning, is a benchmark indicating both "Adjusted" and "GPS OBS" any more accurate? I realize, as you mentioned, that in any case these adjusted coordinates are better than the accuracy of the unit. Thanks again.

 

Go to any of your local benchmarks that are described on their data sheet(s) as having "Adjusted" horizontal locations.

 

Those coordinates have been established to a greater accuracy than your GPS receiver can determine.

[+Woodstramp]

Thanks - I appreciate the feedback. I just wanted to confirm I understood. Can one assume at all that the benchmarks that indicate "GPS OBS" are any more accurate? Meaning, is a benchmark indicating both "Adjusted" and "GPS OBS" any more accurate? I realize, as you mentioned, that in any case these adjusted coordinates are better than the accuracy of the unit. Thanks again.

 

Go to any of your local benchmarks that are described on their data sheet(s) as having "Adjusted" horizontal locations.

 

Those coordinates have been established to a greater accuracy than your GPS receiver can determine.

 

When I first got in to GCing I failed to find a few and was concerned that my new Explorist 400 was a lemon. Had the same idea as you to use a benchmark to check it.

 

I posted a topic like this in the Benchmark forum. Some good person there asked me my location and they found one of these "adjusted" BM's near my location. Glad they did, because I knew nothing about benchmarks and just assumed they were all exactly where the coordinates said they were. . This one also had a "witness post" for the BM which made finding much easier. When I put my Explorist on the BM it said I was 7' from the BM. That showed me that the unit was accurate. Maybe you have one around like that one. Good luck.

[GeoBobC]

My thoughts: you will need to take a significantly large enough sample size for the result to be meaningful. For example, let's assume you take one GPS out and determine how far it reads from the benchmark. A day or two later, you take a second GPS out and measure the distance it reads. Will you conclude that the GPS with the closer reading is more accurate? If you do, you will almost certainly be at risk of reaching an incorrect conclusion. Satellite configuration varies from day to day; weather conditions can vary and alter accuracy. To demonstrate this, try taking the same GPS out three days in a row, and see what types of reading you get. I can almost guarantee you will get three different distances ("accuracy") with the same unit.

 

That's why a test like this can so often result in inconclusive or misleading results. This forum is full of posts about Unit A outperforming Unit B, or a unit having better or worse results after a firmware upgrade, etc. Most of the time, these results are not differences in the unit or software, but simply differences in environmental factors.

[Grasscatcher]

Thanks - I appreciate the feedback. I just wanted to confirm I understood. Can one assume at all that the benchmarks that indicate "GPS OBS" are any more accurate? Meaning, is a benchmark indicating both "Adjusted" and "GPS OBS" any more accurate? I realize, as you mentioned, that in any case these adjusted coordinates are better than the accuracy of the unit. Thanks again.

 

Go to any of your local benchmarks that are described on their data sheet(s) as having "Adjusted" horizontal locations.

 

Those coordinates have been established to a greater accuracy than your GPS receiver can determine.

 

Go strictly by the "Adjusted" coordinates.....GPS OBS coordinates may be coordinates submitted by other than official agencies.

 

Also, read GeoBobC's comments above. Different days satellite positioning etc can produce different results.

In other words,...."one in a row does not make a trend" or prove anything.

 

The experimenting yourself and learning is the fun part.

[+user13371]

...Is that the only option? are there better ways? Like picking coordinates off one of the online maps (e.g. a building or structure)? Or is it finding a "GPS Only" benchmark from the Geodetic site's list of benchmarks. My next thought was to find a local surveyor that might have some documented control points they would share. Any thoughts would be appreciated.

Too complicated.

 

Mark waypoints for any number of fixed locations on your own -- they don't have to be previously surveyed or marked by anyone else. Your mailbox, the entrance to a parking lot, etc... Return to each regularly and see how close your GPS thinks you are to the waypoint.

 

Even more fun: Sit on your back porch and punch GPS reading into Google Earth. Then zoom in and use the measuring tool in GE to see how close it is.

 

See also: http://forums.Groundspeak.com/GC/index.php?showtopic=204098

Edited January 1, 2009 by lee_rimar

[Grasscatcher]

I can't believe that ANYONE would recommend GE for checking accuracy on ANYTHING.

Too many places where the georeferencing of their (GE) imaging is notoriously and visually in error.

GE is "really really neat" but it's really just for the "oh looky, there's my house crowd"....but if you like it, fine.

[+twolpert]

At the risk of stirring up the pro- and anti-averaging bigots on the forum, you might also try averaging your waypoint at the benchmark (on each of several visits, as suggested by GeoBobC). I don't know if the 60 has averaging, but being a slightly older Garmin unit, I assume it does. You probably need to average for at least 5 minutes in order for it to do much good.

 

You should also observe the EPE (estimated error, or estimated accuracy, or EPE, or whatever it's called on the satellite page on the 60) when you take these readings. This is a measure of how accurate the unit thinks it is. By comparing that with your averaged waypoints on successive visits, you'll get a feel for how EPE correlates with accuracy in the field. That's really the point here -- for you to get a feel for relative accuracy. You're going to find that your unit, like every other recreational handheld, is plenty good enough for hiking, geocaching, and so forth.

 

I agree with Grasscatcher about Google Earth. There are enough referencing errors that you don't want to use it to determine unit accuracy without independently verifying that the location in question is properly referenced in GE.

 

OK, let the flaming begin.

[+EScout]

Deadspot:

You are in the USA, so go directly to the NGS and choose the correct benchmarks. This is the best way to get confidence in the performance of your GPSr. Go to this site: NGS Datasheet County Form

 

Choose your State, then your county, GPS sites only. Do a sort by Long. or Lat., and then highlight some local sites, and click Get. Read the Station Descriptions, and choose disks in open public locations. (some of these GPS Sites are not accessible, or not in a public location.) Enter the coordinates in your GPSr. You will have to do some rounding, and by rounding and imputing DDD MM.MMM, you will be within a 3 foot circle of the disk. If you want to be more precise you can figure where the rounded coordinates are compared to the disk, and stand over that spot.

 

Look at your GO TO distance when standing over the disk, and look at your accuracy reading (or EPE.) Go to more than one site, and on several days to get confidence in the accuracy of your GPSr. You might be surprised of the accuracy of most GPSrs. I have consistently found the actual accuracy better than what is shown as the "accuracy" or "EPE" on the unit's screen.

[+twolpert]

Coordinates on the datasheet will be for the NAD83 datum. For all practical purposes, this is equivalent to the WGS84 datum used by your GPSr by default.

[+user13371]

I can't believe that ANYONE would recommend GE for checking accuracy on ANYTHING....

...I agree with Grasscatcher about Google Earth. There are enough referencing errors that you don't want to use it to determine unit accuracy without independently verifying that the location in question is properly referenced in GE....

Really? It doesn't seem too far off for my purposes, at least in the area around Portland Oregon.

 

I can sit on my back porch with my GPS plugged in my computer - and GE shows the correct position (at least within the reported EPE on the GPS). Likewise for tracks and waypoints I've imported -- if I can see which side of a road I was biking along or where I found the best mushrooms, that's good enough for me.

[+twolpert]

I can't believe that ANYONE would recommend GE for checking accuracy on ANYTHING....

...I agree with Grasscatcher about Google Earth. There are enough referencing errors that you don't want to use it to determine unit accuracy without independently verifying that the location in question is properly referenced in GE....

Really? It doesn't seem too far off for my purposes, at least in the area around Portland Oregon.

 

I can sit on my back porch with my GPS plugged in my computer - and GE shows the correct position (at least within the reported EPE on the GPS). Likewise for tracks and waypoints I've imported -- if I can see which side of a road I was biking along or where I found the best mushrooms, that's good enough for me.

Some folks have run across places where the georeferencing is off. From what I've seen, it's localized and random. I've only seen it once personally. OP was asking about determining GPS accuracy. Seems like a bad idea to point him in the direction of something that might be off.

[+user13371]

Some folks have run across places where the georeferencing is off. From what I've seen, it's localized and random. I've only seen it once personally.

Ah, at first you and Grasscatcher made it sound universal, like nobody anywhere should trust GE...

OP was asking about determining GPS accuracy. Seems like a bad idea to point him in the direction of something that might be off.

Sure, it WOULD be if I had suggested it as his sole reference. Go back and read my original post -- I suggested marking points he could verify himself first, and then for "fun" look at some (again, locations he COULD verify, like his own back porch) in GE.

 

And just to stay on topic, I still think most of you are making "gauging accuracy" too complicated.

 

Mark any known spot (doesn't have to be a USGS benchmark), use averaging if you like, and return to it several times (over a period of days/weeks, different times of day, etc) to see how close the GPS thinks it is on subsequent visits.

 

Can't get much simpler than that - and don't need to!

[+Redwoods Mtn Biker]

Running with Lee's idea of not complicating things:

 

I like using benchmarks for calibrating altitude. I try to find any along hiking routes that I'm following, just for this purpose.

 

For horizontal accuracy, I also use aerial photos, typically USGS DOQQs, as well as previously collected tracklogs.

[MtnHermit]

I like using benchmarks for calibrating altitude.

Like this:

 

 

I just stumbled onto this one, South Rim of the Grand Canyon. I was very impressed with the humble eTrex.

[GeoBobC]

Lee Rimar has the simplest and perhaps the best idea: simply find an easily accessible spot (like your backyard), mark with a waypoint, come back multiple times and see how far off from the original waypoint you are. Regardless of where the spot is, repeatability is the key: will the GPS record the same location, day after day? I've never heard of a bias in position accuracy, where a GPS consistently errs in a certain direction. Do this 10 or 20 times and you will probably begin to see some trends that are meaningful. I think it is far more important to take a large number of readings than it is to start with a known location.

[+twolpert]

Some folks have run across places where the georeferencing is off. From what I've seen, it's localized and random. I've only seen it once personally.

Ah, at first you and Grasscatcher made it sound universal, like nobody anywhere should trust GE...

OP was asking about determining GPS accuracy. Seems like a bad idea to point him in the direction of something that might be off.

Sure, it WOULD be if I had suggested it as his sole reference. Go back and read my original post -- I suggested marking points he could verify himself first, and then for "fun" look at some (again, locations he COULD verify, like his own back porch) in GE.

Sorry, I apparently misunderstood the intent of your original post. And I certainly did not mean to imply that Google Earth is univerally hosed. But I can certainly see how it came across that way.

 

Edited to add: My only point was that he might be unnecessarily concerned if he happens to be in one of GE's poorly georeferenced areas. For example, if he's sitting on the back porch but GE shows him across the street or in the next block. It's the imaging -- not the GPS. But there's no easy way for him to know that. Similar phenomenon to drivers following their automotive units down the embankment and into the creek because the unit shows a bridge

Edited January 2, 2009 by twolpert

[+user13371]

I just stumbled onto this one, South Rim of the Grand Canyon. I was very impressed with the humble eTrex.

Cool pic. But is that repeatable accuracy?

[deadspot]

Well - a quick thank you to everyone that posted replies - it was helpful discussion. If anything, it was reassuring that I wasn't going crazy. I did go out today and took readings from a few adjusted benchmarks. This discussion gives me some better context for understanding the results. Thanks again!

[+user13371]

... helpful discussion...reassuring that I wasn't going crazy....

A completely unrelated matter

 

... did go out today and took readings from a few adjusted benchmarks. This discussion gives me some better context for understanding the results

What kind of results did you get?