@fobos8 I am working on a Survey of a Section of land (square mile) in the Sierra Nevada Mountains, steep, rocky in places, forested. It is open enough to use GNSS. I would expect 4 miles of total station traversing to produce an accuracy of 30cm or so.
We use a lot of redundancy, running up to 6 receivers at once for multiple hours then processing and reviewing the data every evening. TBC flags questionable points, I use the point derivation report to disable problem vectors.
A few monuments can??t be observed directly so we set control pairs then turn angles and distances to the needed monument.
We use RTK but not alone, long static is needed in difficult areas. Hoping to use RTK to set line points for boundary signs, a fire burned it last year so there??s hope.
iIn the redwood forest on the Northern California coast, forget it, traversing is the only way.
One of the keys to RTK is to choose a base location which is as clean and WFO as possible. (Wide F?? Open)
If you are seeing many bad fixes or large errors in your points, you must be working in way more challenging environments than I am or my equipment is more advanced.?ÿ
Using 2" geomax zoom90 robot and geomax zenith40 receivers so not quite the latest from Lecia/Trimble but pretty close (very similar to Lecia TS15 robot and GS16 GNSS). For urban boundary work control points I'm chasing <10mm 2d repeatiblity at ideally 99% CI which is pushing RTK pretty hard I find, static different but not always practical sure to time required. Can only achieve it reliably if no trees/buildings/deep gully/reflective surfaces around, usual good practice things to avoid with GNSS. Braced pole and observations 180?ø rotated pole to mean out any pole run out. Also second fix at least 20 mins later.
?ÿ
Not seeing large errors like from bad initialisations in early GPS days, is either repeat GNSS shots that are 20mm different or occasionally GNSS shots both agreed within 10mm but a total station shot from good control is 20mm off. For rural work boundary it's no problem but urban trickier as always trees on properties and buildings with shiny rooves etc. and tighter precision requirements. All work here has to be done on grid bearings coordinates and recorded so no dodgy scaling GNSS to ground or bearings that change from one side of job to the other.
@lukenz I'm far from an expert. I was looking at the time series for the CORS station nearest me:
It seems the uncertainty in the position of the CORS station is comparable to the uncertainty you're trying to achieve. How would this affect you?
Reference station positions will always vary over time due to geophysical events such as plate tectonics, glacial rebound, and ocean tides. Even the global reference frame has temporal variation which has some impact on that pattern.
That graph is the long-term series, which show the positions fluctuating in a more or less regular pattern on the order of 5mm NEU around the published position.?ÿ IIRC the NGS will observe a point for a minimum of three years before being able to really position it with confidence.
Still, it's good to realize that everything is moving, all the time, and we do need to be aware of it.
I find (and the NGS recommends) that looking at the short-term time series is more beneficial, as it shows more immediate (closer to your date of survey) positional variation as well as a better picture of how far that station is from its published position for the current realization:
Sometimes that warrants using another station, sometimes not.
This is yet another reason why observation epoch absolutely needs to be communicated with survey metadata. In practice, we treat these stations as fixed even though they are really not, because for 99% of typical survey activities, a reference station moving 7mm (combining both N/E variations) over the course of an entire year, then back again over the next year, isn't going to blow a job. An earthquake might though.
We are in more friendly skies for GPS. The stated accuracy of our RTK units is 8mm and in my opinion that is conservative (assuming light multi-path). Often our checks are less than .02' horizontally.
I've been surveying a burned out area with my older units and I didn't run into anything my older units couldn't handle, even under heavy canopy that wasn't burned. One corner located near two unburned yellow pines, checked within .01', kinda shocked me.?ÿ
Every boundary corner got an original location and at least one check usually while running line, but there was never any errors to worry me, I can imagine doing it with a total station but I would have had a crew of sawyers with me like in the old days.?ÿ
This was done with R8-2 units. The R10 units are even better in challenging environments. 20mm will not be surprising in deep timber, up against buildings, ect. But you have to really look for those places here, I wouldn't probably use it around tall buildings and really heavy eastern or NW forest areas.?ÿ
We don't let our surveys float with CORS. There is an awful thread about just that some years ago. Fix your surveys and keep them fixed, that's my policy.?ÿ
It is very possible that the laws/rules for determining boundaries vary from jurisdiction to jurisdiction. That said when the most dignified evidence of a boundary is the monumentation of corners, &c., the most important role of a surveyor doing a dependent survey is to locate the monuments. Then the secondary role is to measure the relative positions of the monuments so that others may use the measurement to more easily find the monuments.
Additional legal requirements in a jurisdiction may call for additional information not directly related to proving the boundary, e.g., flood risk, tax zones, &c.
Depending upon the survey and property involved, it may be judicious to "walk the lines" of the boundary to observe topo calls. I doubt a theodolite is the only means to determine the relationship between the corners. A survey with GNSS is very similar to a radial survey by theodolite, for example.
One danger is when the surveyor puts more trust in the record coordinate geometry than in the evidence of corner monuments. One then risks being labeled a button-pusher, who thinks "my measurements are better than that old stuff used in the original survey; so I will put a new mark where they should have put it originally." That results in several things, none of them good. Property owners spend fortunes to resolve the ambiguities created by the surveyor & Surveyors can't agree on anything - why do we need them? are two examples.
Don't know, find the local network RTK not anywhere near as good as your own base <1-2km from where you are working, only use the local CORS site for 0.1m work as usually 10km+ from it. That timeseries plot is only 1 sigma too, I'd be doubling at least, doesn't seem like a very good place to have put a CORS site with that level of movement.
?ÿ
NZ is a skinny country running north/south with lots of ranges of hills/mountains and the NRTK base spacings don't seem close enough to model the atmospheric changes well enough, or our communications networks have too much latency, or too much tectonic movement between bases, whatever it is results here don't match the published studies I've seen for Europe/Japan/Australia
I love answering for a question to someone else 🙂 At the risk of presumption I would say floating is referring to the phenomenon that if you are seeking 10 mm or so of repeatability that in general it is not possible in an active control network without considering and applying time adjustments. Thus the end user is wise to also tie their survey to a passive control network for future reference. Even in areas where there is hardly any plate motion the CORS and as a byproduct the RTN's occasionally update their reference station coordinates usually by more than 10 mm. When that happens obtaining precise repeatability is lost without doing time transformations.?ÿ ?ÿ
Don't know, find the local network RTK not anywhere near as good as your own base <1-2km from where you are working, only use the local CORS site for 0.1m work as usually 10km+ from it.
If you're running with a single-base solution from an RTN station, for sure it's not going to be as good as a network solution or base-rover. And not all networks are created equal, as you mentioned.
That timeseries plot is only 1 sigma too, I'd be doubling at least, doesn't seem like a very good place to have put a CORS site with that level of movement.
Those are pretty typical values. I doubt you're going to find a ground-based station on Earth that doesn't fluctuate relative to the global reference frame by at least a few millimeters per year, some more than others. Most CORS, even the ones that have been operating for decades, have estimated network accuracies (FGDC 95%) of around 1-5mm horizontal with 1-3mm standard deviation. That does vary a bit, but I've never seen one with zero values.
It's also worthwhile to point out that the short-term time series should be evaluated in two ways. First by (a) the "minus" value i.e., comparison of the tracked position versus the published position, which is going to drift more and more the further we get from the epoch date; and (b) the STD or standard deviation values, which gives you an idea of how much fluctuation there is on a daily basis, and thus how repeatable that position is regardless of its drift away from published values.
All that being said, 4D coordinates and time-dependent transformations are going to be critical once we move to an NSRS defined by active stations.
@mightymoe We can get a decent solution if it is reasonably open sky and we are up out of a hole.?ÿ We've had pretty good success this year on monuments in the trees if I can see blue sky and we are up on a hillside, on a ridge or mountaintop.?ÿ I would say our repeatability is a little worse than 1 or 2 hundredths but not much.
Three monuments down in a hole with more canopy we are having to tie from a GNSS control pair but two can be done with just one pair within vision and one is going to require a little bit of traversing, easy though since there is a road, no clearing.?ÿ The fire last year burned out all the underbrush.
I have a survey 4 years ago that is based on an OPUS solution at the south end.?ÿ This year we are using an OPUS solution near the north end.?ÿ The OPUS points are three miles apart.?ÿ We tied in two control points and a monument at the north end of the 4 year old survey and match horizontal about 0.05' and vertically a little worse, about 8 tenths but the differences if I hold the 4 year old values is about 2 tenths.?ÿ The two control points were used to tie in a monument which was un-GNSSable 4 years ago but now it's burned out enough that it can be observed.?ÿ There was a huge 15' diameter log there which is now just some charcoal except the root ball lying on it's side somehow survived.
If you're running with a single-base solution from an RTN station, for sure it's not going to be as good as a network solution or base-rover. And not all networks are created equal, as you mention
It's because the NRTK results are poorer than base/rover (in <2km range) I use base/rover if accuracy required otherwise just use the nearest CORS (in 10-20km range) for rougher work. In fact the CORS is the best metal dector I've ever had as it is broadcasting correct coordinates for our datum, turn on, set out and dig hole. Makes getting started so simple.
?ÿ
The local CORS is state funded and 4 constellation (as is my gear) whereas NRTK is private and subscription based and in our area only GPS/GLO which is also a factor in not using NRTK.
?ÿ
NRTK here based on other stations 70km+ away where weather can be quite different (and lots of NZ close to coast and weather changes blow through) therefore can't assume linear interpolation between station and survey location, think we'd need more like 20-40km spacings here which wouldn't be economic, at least until all cars are relying on it to keep them on the road. Plus a lot of NZ is rural and cellular coverage not good enough for NRTK as latency rates too high.
?ÿ
Also was more talking about the longer time series chart first posted not the daily one where the numbers are pretty tight. Agree that even excellent CORS sites move around in geological and tectonically stable locations at mm level and even cm/dm level over a year which isn't an issue for day to day work.
?ÿ
We have one half of NZ on one plate moving 50mm north a year compared to the other and also where I am we are geothermally active area so earth's crust is heaving up and down over the year.
?ÿ
4d coordinates scare me, find some local surveyor struggle with using correct datum let alone getting time correct as well. Be lots of messes to sort out!
