I understand(I think), that VRS has an awful lot of redundancy built into the results.
But I also understand(I think), that VRS is generally less accurate than STATIC or even shorter range RTK.
Am I right?
A buddy of mine ran elevation following a survey firm that ran a site with VRS.
He didn't trust the elevations(his nature), and so ran more elevation of the county-wide(non-bluebooked), GPS network that was set in 2000.
The requirements of the 2000 GPS was dual frequency, constrained by available USGS monuments. Elevations were constrained by monuments with better elevation and horizontal by monuments with better locations.
GPS monuments(bluebooked as set by other agencies in neighboring counties) were used as checks and sometimes as constraints.
He missed the VRS elevations by about 0.25 feet.
To me 0.25' is quite a bit . . . but what I don't know . . . is 0.25 feet quite a bit for VRS?
Would you use VRS for elevations?
No, I would use it for ellipsoidal height to compute the orthometric height from.
I would use it for dirt surfaces only.
is 0.25 feet quite a bit for VRS?
Relative to local benchmarks, No. Relative to itself, yes and no. If you are basing the 0.25 on one observation no. If you are basing the 0.25 on averaging several observations, yes.
0.25 in vertical is not surprising for any kind of RTK.
The WSRN here is probably the tightest RTK network around, and it's usually plus or minus a tenth as the day goes on, and that's with good DOPs and clear sky.
0.25' would not surprise me at all. It's never going to get better without redundancy or an independent method.
Set 5 VRS points and run a level through them and see what you learn.
Linebender
I understand the ellipsoid thingy . . .
But are you saying that you actually get ellipsoid results then manually add the height?
...just curious
Linebender
No, I was referencing computing the Ortho Height using a Geoid model. The reference stations in the VRS/RTN we work in publish Ellispoid heights not Ortho heights.
Linebender
Around here, I typically see about 0.10' between published NGS elevations and VRS derived orthometric heights. It could be up to 0.2', but rarely. It's pretty consistent at around 0.10'.
> I understand(I think), that VRS has an awful lot of redundancy built into the results.
>
> But I also understand(I think), that VRS is generally less accurate than STATIC or even shorter range RTK.
>
> Am I right?
>
> A buddy of mine ran elevation following a survey firm that ran a site with VRS.
>
> He didn't trust the elevations(his nature), and so ran more elevation of the county-wide(non-bluebooked), GPS network that was set in 2000.
>
> The requirements of the 2000 GPS was dual frequency, constrained by available USGS monuments. Elevations were constrained by monuments with better elevation and horizontal by monuments with better locations.
>
> GPS monuments(bluebooked as set by other agencies in neighboring counties) were used as checks and sometimes as constraints.
>
> He missed the VRS elevations by about 0.25 feet.
>
> To me 0.25' is quite a bit . . . but what I don't know . . . is 0.25 feet quite a bit for VRS?
In Florida, I believe that a lot of the elevation certificates filed (especially the ones under $100) are done with VRS elevations. I don't have VRS service, so I don't know. But I have heard from those that had it, that it is pretty good (around 0.1' level accuracy). I would check it with some nearby benchmarks before trying to set an elevation with it. But even then, RTK measurements are so moment to moment. One moment its good the next its bad (because of multipath, a key satellite dropped out, sun spots, etc.)
The Bow Tie Surveyor
That would be a tenth either way. One could be high by a tenth and the other low by a tenth, so when you level between them from a published station there's your 2 tenths. Have to know your limits and what is the end result being used for.
The big question is not really about VRS but rather HOW reliable is the current geoid in your area of practice?
Proper use of VRS in a Real Time Network requires check occupations with different satellite geometries. One "wang-bang thank you Ma'am" occupation is not an acceptable practice.
Two centimeter ellipsoid heights are relatively straightforward to achieve ... but what do you have after you use the geoid in an attempt to get an elevation?
It depends ...
I've been using Ohio's ODOT VRS Network for 3 or 4 years now and always seem to match my local counties control network monuments within 0.1' vertically. I've had shots with a spread of up to 0.25' in more challenging enviroments. For anything I want to trust I observe twice and average, which always seems to achieve results that I'm comfortable with. As of late I've been doing a 15-30 min. static session on one of my control points and submitting to OPUS-RS, which has also been agreeing with my averaged VRS solution within a 0.1' vertically and even better horizontally.
Yes.
I use the same basic method. Occupy 2 times at the minimum usually 4-6 hours apart so the satellite geometry has changed (better a day or two apart). If 2 or more sats are near directly OH then even better results for vertical. I have received OPUS-RS results usually well within .1' between occupations. But the trick is knowing what the geoid shift to local datum is. Then a reasonable valuation can be applied.
After that it can be used for simple topos, nothing else. Otherwise, you start to introduce systemic errors that can be hard to explain later. :-O