[opwieurposiu]

As an example of a problem that can occur: Client wants to survey the elevations of a parcel of land for building a house. The surveyor goes out and collects data with gps survey equipment, which produces a .csv file of northing,easting,and elevation values relative to some National Geodetic Survey benchmark. The engineer imports this .csv into his cad software, produces a grading plan .csv file. The earthmoving contractor imports the grading plan into his software and grades the site.

Client gets house built, moves in and discovers that the front yard turns into a swamp after every rain.

Somewhere in the chain of importing and exporting .csv files, software A was using survey feet and software B was using international feet.

This can also cause problems like structures built in the wrong spot, fences built on neighbors land, etc.

[clucas]

I worked as a CAD operator, surveyor, and occasional field crew worker at a civil engineering firm for a few years. We never ran into issues like this - by far the more common issue was the difference between NAVD and NGVD. At the time I was surveying, the FEMA flood insurance maps in our area were only partially updated, so the base flood elevation could be in either datum, and you had to really watch it so you didn't end up about 0.7' off vertically.

[abofh]

The difference is .0002%. your examples would barely add up to an inch over a mile.

[opwieurposiu]

Sometimes the benchmark for northing/easting/elev is many miles away from the actual site. This is especially true with gps surveys.

It only takes a few inches to wind up with a substantial drainage problem.

[beerandt]

There's a ppm error there as well, but that's a usually a different problem. If different units were used, it would be obvious when checking in on the benchmark.

If it's in northing and easting, the baselines are to the South and West of all coordinates covered by the projection, plus an offset. So it might start at 1000000 instead of zero, to ensure format consistency and to catch blunders.

So it's about 2ppm, multiplied by the number of feet between you and the Southern or Western most edge of your state (if single projection state) plus 10^5 or 10^6 or similar.

It's designed to make an error large enough to notice.

Vertical however is rarely a problem, as it's all based off local benchmarks.

[function_seven]

Wouldn't the elevation difference between benchmark and site be the important bit here? If so, then the difference would rarely be more than a mile, and the resulting error would be at most an 1/8"

[opwieurposiu]

It has been 20 years since I worked in this area so likely I have some details wrong. If elevation is distance to the center of the earth 1/8" per mile is 40 feet!

From the link below you can see the noaa benchmark is using both elevation in feet above sea level and in meters from the center of the earth.

https://www.ngs.noaa.gov/cgi-bin/ds_mark.prl?PidBox=aa3449

[beerandt]

If you're not converting H (gps orthometric height) to E (local elevation datum to "sea level") first, you're doing it wrong, for a number of practical reasons.

Usually there's not even a reason to use H directly anymore, since elevations should reference a local benchmark.

GPS has changed things, but surveyors are still obscenely practical when it comes to procedures for eliminating systematic error.

[1-6]

Using an RTK-GNSS requires a gravitational model of the earth to measure mean sea level (MSL). Using your position, you use the GEOID model to figure out the delta of the nav satellite's orbit to the MSL.

[beerandt]

It's that you're in the wrong e, n location, which may be at an entirely different elevation than where you want to be.

[1-6]

We're not even getting to the topic of elevation (yet). The US Survey Foot is 1200/3937 which is a not a clean value that terminates after a few decimal places. With high grid coordinates (X,Y), your origin is based at 0,0 as origin. When you scale between metric and US Imperial, if you're not using enough significant digits it impacts how the model scales. This can cause a difference of 1 foot or more from my experience. Floating point operators also become an issue especially in CAD. Standardizing to one number is a great start.

[dmtroyer]

seems like an issue that surveyors and engineers should be well acquainted with.

[drdeadringer]

How often does a "should" translate into reality?

[1-6]

It's entirely possible that they might not be aware. Surveyors are concerned with measuring bearing and distance from an arbitrary location (5000,5000) in some cases. The distances they traverse are not enough to see differences between US Survey foot and Int'l Feet. Cartographers and GIS folks who tie into a photogrammetric basemap are probably more aware. Most Civil engineers are like 'NAD83? Ok, whatever you say~!'

[Enginerrrd]

This is true. I barely understand coordinate systems enough to get by and somehow I know them better than any other engineer I've talked to. Surveyors are 50/50 on whether they know all that stuff or not.

Really though, there's no excuse for errors caused by that stuff. Survey data should almost always only be used relative to itself.