Along these lines how do you best get from a WAAS here position to NAD83?
You know the shifts in your area but that varies in location. I work all over the US.
I always collect data for OPUS but many times I have to start work with a here position and correct it after the fact. The closer I can get the here position the better.
> I always collect data for OPUS but many times I have to start work with a here position and correct it after the fact. The closer I can get the here position the better.
The rule of thumb I recall is that a 10-meter error in base position introduces a 1 ppm error in the vectors. Even a cheap WAAS-corrected handheld should be able to do better than that, often by a factor of 5. Given the other error sources inherent in the RTK approach, I wouldn't think that 1 ppm would be much of a problem.
The programs "work" everywhere I've ever used them. They produce a number, input 83 and they kick out a 27 value. Have you ever checked that 27 value against the actual 27 value? I have, I've done extensive checking of the numbers and they have never matched. The closest has been. 5', but I've seen up to 6'. Other surveyors I've talked to have told me they've seen 15' and someone told me 30', but I never saw that,
Once I saw 6' on a first order 27 monument I quit looking at it. That was years ago. So when I actually need a 27 # I survey in the 27 system and don't use the stretch and fit program.
Vertical is different, they get you closer. My testing of those show from .1' to. 7'. A factor of ten "closer" than the horizontal. When I've tested geiod models against actual geoid slopes, the newer ones preform better with each update.
> Along these lines how do you best get from a WAAS here position to NAD83?
I get a better WGS84 position from the WAAS correction and I know the shift between the NAD83 (2011) and the IGS08 (WGS84) from previous OPUS results in the area. My correction may be off a tenth or two but the correction is applied to a point that is only accurate to 3-5 feet.
I would much rather log a file for OPUS ahead of time but can't always do that.
I can't answer your question in regards to anywhere in the US.
There should be conversions on the NGS website to do that but I don't use them.
James
NADCON is the official datum transformation tool for the U.S. to go between NAD 27 and NAD 83. It has a standard deviation of 0.15m (.5') in the lower 48. I have seen a few occasions where it was slightly higher but rarely in the range of more than about 0.5m (1.6 ft). NGS has created a transformation tool to go from each state's HARN to NAD 83 (2007) called GEOCON and from 2007 to NAD 83 (2011) called GEOCON11. They provide excellent results but they are a pain in the rear to use. -- GEOCON
> > I always collect data for OPUS but many times I have to start work with a here position and correct it after the fact. The closer I can get the here position the better.
>
> The rule of thumb I recall is that a 10-meter error in base position introduces a 1 ppm error in the vectors. Even a cheap WAAS-corrected handheld should be able to do better than that, often by a factor of 5. Given the other error sources inherent in the RTK approach, I wouldn't think that 1 ppm would be much of a problem.
Thanks. That was not really my issue though. I am staking out points. For what I am doing it is not the end of the world if they are not placed exactly where we wanted them as long as we know they are. For example my client may tell me to drill holes starting at X,Y and then every 100' along a NS line. If I use a here base position I may have the whole grid of holes shifted 5-10' from where they want them but it isn't a huge problem as long as I provide the true coordinates of where I drilled them which is easily doable via opus. I just am looking for ways to try and get my here position closer to the true position.
Of course we both know they work. My point was our experiences differ. I would guess the primary reason is location. Here in the northwest I find the conversions usually work well. As this world gets smaller we need to remember not to grab on to local results and typical numbers too hard. It can hamstring you if you find yourself in a different place...
Well, I guess defining "work" for me would be that using the conversion for say a HARN point set on a NAD27 monument would produce the NAD27 value, which it will not do.
I went through this for a FAA project long before getting GPS, the published numbers on monuments onsite would not match CORPSCON conversions, went round and round about it, finally me and the engineer actually discussed it with people who developed NAD83 and they said there is no conversion possible between the two systems so there was a rubber sheet file created.
So for survey value numbers I want surveyed points, not converted points, its' harder but that's why people hire me. Now for mapping quality numbers, the conversions are ok, sometimes not great, but usually ok.
If someone at NGS told you that NADCON was rubbersheeting they didn't know anything about the program. NAD 27 had significant amounts of regional distortion most of which was removed by the original NAD 83 adjustment - that being said every station in a given area had didn't levels of distortion. The process used in NADCON is to created two sets of grids one for changes in latitude and one for changes in longitude. Then the utility that creates the grids looks at all the geodetic control in the country and determines through a least squares process the most likely value for the positional change at each grid node (in the lower 48 that's every 15'). The control closest to the grid node would have the greatest influence and the further away another station was the less influence it has. Even though virtually all of the original NAD 27 station participated in the grid development it's still a model or a best fit and you will never match the original NAD 27 values exactly. The only way to match original values is to either use NAD 27 values in the adjustment, in which case the old distortions will rapidly appear in the adjustment or in the case of an older survey readjust it using the original observations.
Unfortunately all of the GPS receivers also have the transformations developed by what was then the Defense Mapping Agencey (DMA) and they are at best horrible (+/- 7 to 10m). For those that consider WAGS 84 and the original NAD 83 to be the same and use these transformation to get back to NAD 27 the results can be really stinky.
herman cain time
I do not have the facts to back this up, but.... I think you have the data you seek.
At the time in the survey when the base station fixes it's , record the LLH. Send a couple of hours and test it to see if it works. Test under similar conditions to your normal work
1 Locate the base via autonomous position
2 record the LLH
3 collect a set of rtk positions
4 stop the survey
for the test to work properly there probably needs to be enough time between the setups for the constellation to have changed in a meaningful fashion.
1 locate the base using the previous LLH
2 collect the same set
3 compare these values with the previous values
The autonomous position is an accurate position, but the question is what datum do you want to use. Opus is the best bet. It also affords a check on you base station location.
If someone at NGS told you that NADCON was rubbersheeting they didn't know anything about the program. NAD 27 had significant amounts of regional distortion most of which was removed by the original NAD 83 adjustment - that being said every station in a given area had didn't levels of distortion. The process used in NADCON is to created two sets of grids one for changes in latitude and one for changes in longitude. Then the utility that creates the grids looks at all the geodetic control in the country and determines through a least squares process the most likely value for the positional change at each grid node (in the lower 48 that's every 15'). The control closest to the grid node would have the greatest influence and the further away another station was the less influence it has. Even though virtually all of the original NAD 27 station participated in the grid development it's still a model or a best fit and you will never match the original NAD 27 values exactly. The only way to match original values is to either use NAD 27 values in the adjustment, in which case the old distortions will rapidly appear in the adjustment or in the case of an older survey readjust it using the original observations.
Yes this is exactly what they told me, rubbersheeting is my word for a more complex calculation, but don't really know a better one word description for it. It was really quickly obvious that the conversions weren't good enough to use from a surveying perspective. To get 27 data it was necessary to survey in 27, not survey in 83 and convert. It was very important to me back in the day to have 27 numbers and still today it is at times.
There are mine projects in NAD27 that CAN'T be converted because of permitting issues, and there are older DOT projects that pop up at times. Also oil & gas data is often on 27, now that's usually not real accurate and more mapping grade anyway.
My point is that as the NGS told me is there is no real calculation between the systems. And the conversion program WILL NOT give a 27 number.
As Dave said, the transformations in NADCON are the best available, but not good enough for survey. If I want good NAD27, I tie in NAD27 stations and do an adjusttment.
I recently looked at 4 first order triangulation stations in Montana. None of these has ever been GPS'd, but they do of course have NAD27, NAD83(86), and NAD83(92) coordinates.
The NADCON shifts (HARN-27) had a range of 1.151 m in northing and 1.127 m in easting, with standard deviations of 0.52 m and 0.51 m. MUCH larger than I would have expected for a given area. I don't have a triangulation diagram for that area, but there must be some bias in the NAD27 that wasn't there in the NAD83 adjustment.
As Dave said, the transformations in NADCON are the best available, but not good enough for survey. If I want good NAD27, I tie in NAD27 stations and do an adjusttment.
Exactly!!!
There are HARN points on tri stations out here, but mostly they occupied first order bench marks
I should add that I am more accustomed to seeing an accuracy of 0.15 m in the NADCON transformation here in the east (western PA).
I started doing GPS before NAD83(86) was released, NAD27 was all that we had to work with for control (and of course no CORS or HARN conveniently located along the highway). I did a lot of adjustments of NAD27 GPS networks, I would always solve for scale and azimuth bias, which could get quite large, relatively speaking. And of course back then there was no geoid model. We had to model the geoid by tying in benchmarks around the project, and computing tilts, sometimes various sets of tilts if the project was extensive or in an area of disturbed geoid. I would compute tilts, and then compare against astro stations in the area. Ya'll have it so easy nowadays..............
What NADCON does give is an approximation of the 27 values. My point is those values vary in quality and fitness for various purposes. I use them on 1A certificates once I've proven they meet the standard. I use them to align FIRM panels as well. They are also handy for search values when attempting to recover points from differing maps and eras. The fact they aren't perfect does not make them useless...
