As some of you old timers may remember, I took up learning the basics of surveying, beginning around 10 years ago, primarily relying on these forums, Ghilani and Wolf, and practiced with a 5" total station on our 24 acres of wooded land.
I created a control network of about 50 points; made thousands of observations between them; adjusted them all in Starnet. For azimuth, I made hundreds of individual celestial observations and reduced those using Larry Scott's phenomenal spreadsheet.
Recently I contracted a license surveyor to survey the property (as we're contemplating a subdivision). He did this using a Leica GPS system (in a few hours I might add, vs. my 10 years lol). I asked him to share his coordinate file. So now I'd like to compare my work with his and move and rotate my survey to the proper location on the surface of the earth and to the "right" azimuth.
I'm uncertain as to the proper way to do this. I'm completely unfamiliar with GPS. He used NAD 83, as did I, but his survey is in "Survey Feet"; mine's in international feet. I'm also uncertain whether the project combined factor plays a role with GPS surveys.
Comparing the two surveys, The average distance of my four corners to his four corners is 9.4325', which given my uncertainty as to the location of my survey on earth sounds reasonable; I can't remember how I originally placed the survey on the grid...could have been Google Earth. The largest residual in distance is .0475'. The average azimuth variation is 20.25". I'm confident my azimuth is good to probably around 10" or 12", so I'm not yet certain why the difference between surveys would be as much as 20".
So, to do this do I move and rotate my survey by the average displacements and azimuth differences? I'd love to put his observations into Starnet but have no clue what precision and residuals are associated with modern GPS observations.
Advice most welcome.
The difference between the US survey foot and the International foot is insignificant unless you run it out for miles. Unless things have changed Google Earth uses WGS 84 coordinates, not State Plane grid coordinates, WGS 84 is close, but not the same.
Using a 5" gun and multiple star shots are going to create their own error ellipses while using GPS, if processed properly will be tighter than your methods if it is processed properly. If what you are doing is just a learning exercise, I would suggest constraining your control network using your surveyors coordinates to rotate and translate into grid coordinates and elevations. I'm not sure if you can do that with Star Net as I have never used that program.
I did some research. I may have gotten the coordinates from the original survey done around 1980 that referenced "Vermont Grid Coordinates". They don't match what I've got though...particularly in easting. May not be important now.
In 1980 the coordinates would have been in NAD27, and would have been much different from Vermont SPC for NAD83. If the sets of coordinates are within feet, they are likely on the same system, so the older ones were perhaps later 1980's.
There would be some small difference between NAD83(86) and NAD83(2011). The NGS toolkit has a program to do conversiobs, which I haven't used recently enough to recall the name off the top of my head.
BTW, welcome back. We had some interesting exchanges back then.
One way to compare surveys would be to enter the new survey coordinate data in Starnet with fairly tight tolerances, and your angles and distances (not coordinates or azimuths) with your best estimates of your standard errors.
Use one set of point numbers for one survey and those plus 1000 or 10000 or whatever for the other survey so it is easy to compare.
Make only one point in your data to have the same point number as the other survey so they are tied at that point.
Never try to use a zero standard error to tie them together, as that can make the least squares calculations blow up.
There may be other ways, but this one should give you some insight.
I know I need to move my entire survey some distance (about 9'). Is there any advantage to using the point that the surveyor set his base station on during his survey? It's the same point that I did all the celestial observations from. It's a rock solid steel post anchored to ledge. Once moved I can then compare the azimuth difference and rotate the survey accordingly.
@chris-bouffard mentioned that the GPS survey "if processed properly"...I think the coordinates my surveyor gave me were raw...right from his data collector. Is there some post processing needed for GPS surveys that would make them any better?
And we did have some very insightful discussions back then. I learned a lot thanks to you and others here.
Looking at your map I see 3 longer distances that compare very well with each other but I do not know which is GPS or TS. The short 239 footer needs to be checked and the correct distance verified. I have had my trusty GPS give me a 0.2' error now and then for no apparent reason, a tree that I thought was no problem, a powerline, a loss or gain of a satellite, who knows but when I set up my TS15 to do some checking and start staking there it will be. I suspect that the TS distance will check and not the GPS.
After the TS distance checks you will have to decide which GPS point is wrong on that 239 foot line. If there is a tree close by a point I might start there, either way I would hold the TS angles at each end and see which GPS point needs to move to make the GPS angles fit the best. After that I would translate and rotate your traverse to that 2042 foot line (that might have a new bearing now) and see how the points compare.
At this point a scale factor might become apparent but I am old school and will always hold what appears to be a well run traverse like this over GPS.
After the TS distance checks you will have to decide which GPS point is wrong on that 239 foot line. If there is a tree close by a point I might start there, either way I would hold the TS angles at each end and see which GPS point needs to move to make the GPS angles fit the best.
Yes, that bothered me. My dozens of observations (forward and reverse; multiple setups, from both ends) make it 239.35; The GPS observation is 239.04. There is some canopy near both 3102 and 2, and when his survey was done, the leaves had not yet fallen off the trees. Still I would have thought the observations would have been closer.
If I translated my survey 9.4325' south (the average delta of all four points) the delta between mine and his (3102 and 1:2) is .022'. The delta between 2 and 2:2 is .228'!, which is huge considering the distance. I think I used 3PPM with a .009' constant in Starnet (I'll check). Grasshopper tendencies aside, I have to believe my distance would be more reliable than his.
The difference between the US survey foot and the International foot is insignificant unless you run it out for miles.
While this is true for distances between discreet points, that falls flat when dealing with coordinate systems using large false northings and/or eastings. Here in northwestern Oregon the difference in USFT & IFT coordinates is around 14' when using OR83N.
While the different units will match to the hundredth at these distances, they could account for some of the shift.
If you are using coordinates in the millions. for example 1,000,000 northing 2,000,000 easting for US Survey feet then International feet will be 1,000,002 northing and 2,000,004 easting. You should convert your coordinates to the same system before trying to match the surveys.
Firstly, I wouldn't be overly concerned about US feet vs international feet for a boundary survey. While it is certainly relevant when it comes to coordinates or long linear projects (pipelines/railways/highways), as far as distance measurements, you are not going to see any significance between observations until you get into lengths of 10000' or more.
Again, with azimuths, the 20" difference you are seeing is likely just the difference between your astronomic azimuths (per celestial observations), and grid azimuths (per local projection... likely a state plane system). If you know the projection used by the surveyor, the azimuthal difference (convergence) can be determined by NGS, via their NCAT online tool. To simplify it..... state plane grids are big squares, plopped onto the round earth surface.....the further you go away from any assumed correct center azimuth, the greater the azimuthal difference appears to be.
I am not a starnet expert, but with your data having unknown precision and residuals, I would propose that you are in the realm of helmert's transformation, which ignores the observations and adjusts just using the resulting coordinates. I am pretty sure starnet has this capability, but I made a Helmert's program in Excel to accomplish this. Helmert's is the most common adjustment used by surveyors..... whether they know it or not.
Happy to share my excel program, but it maxes out at 20 control points.
If you are using coordinates in the millions. for example 1,000,000 northing 2,000,000 easting for US Survey feet then International feet will be 1,000,002 northing and 2,000,004 easting. You should convert your coordinates to the same system before trying to match the surveys.
The two surveys ARE using the same coordinate system Vermont State Plane. I just could not remember how I initially placed the survey in the first place. I think I took a single point (probably "2" in my survey), zoomed way in on Google maps, and took it from there.
As for azimuth, my celestial observations corrected for grid convergence. I used UTM Zone 18, which renders a 1.66929652 convergence factor. That matches NCAT, which I just looked up. Also did LaPlace correction and Xi and Eta corrections, but don't ask me to explain them now. i forgot a lot of stuff I once knew!
What I don't understand is that in NCAT, they show an entirely different convergence (-00 03 25.12) for SPC than for UTM, so I may have made that mistake. But in any case, my astro azimuth delta is off by 20", not minutes or degrees, so I don't think convergence explains the difference.
@rfc Vermont State Plane coordinates expressed in International feet will put you in a different place on the ground than Vermont State Plane coordinates expressed in US Survey feet. So I wasn't precise earlier when I said you should be using the same coordinate system. I should have said you should convert from US survey feet to International feet or vice versa if your coordinates can be materially impacted by a change of 2 parts per million.
@rfc Those coordinates shown in the picture would about 3.33' away from the same coordinates in US survey feet if they were placed on the ground. I'm assuming you are showing a picture or your coordinates in International feet.
@lurker Have no idea whether Google Earth uses International or Survey feet.
@rfc This is just a google result. I have no idea if it is correct or not.
international foot (0.3048 meters) for its internal calculations, which can cause a shift when data is imported from systems that use US Survey Feet (1200/3937 meters). Users importing survey data must account for this difference, as the slight variation between the two units can become a significant error over long distances, as noted by bentleysystems.service-now.com and Reddit users.
Mentioning Google Earth may have been a red herring. I've not used GE for anything other than locating one point on my survey years ago to "get it in the neighborhood". I'm not declaring my coordinates to be correct; quite the opposite, I'm assuming that the coordinates of the GPS survey are correct (and about 9' away from mine). Therefore I'm going to 1. move my entire survey by some amount (TBD) to bring the two closer together, then 2. analyze distance discrepancies (such as the distance between 1:2 and 2:2 in the GPS survey as well as the 20" azimuth discrepancy I'm seeing.
FYI, I've changed to US Survey feet in my survey and have noted no difference.
Ignore the 20" difference that's like a tenth in 1000' and just part of the normal scatter