How well do the NGS tools predict the difference between these datum realizations in your area? I'm finding about a foot, mostly east, in my area but the tools (if I ran them properly) are saying NAD83(96) + 0.030 ft east 0.024 north = NAD83(2011).
I noticed a discrepancy when playing with my sighting network across the north end of town. My measurements of a church spire didn't fit the NGS data sheet nearly well enough. I added an OPUS session on the ground near the church and close-up triangulation sights and found my optical network was within 0.3 ft of the OPUS result, mostly off to the south. I was reasonably satisfied with that difference because it was out at the end of the network with less-than-ideal sighting angles. But it made me wonder how well the old realization fit the later one.
So I got GPS results on a reference mark for the nearest triangulation station (second order) and next closest surviving tri-station (first order) for comparison. Thanks to Kent for processing some pretty poor data on the RM for me.
NJ0775 MORRIS, first order, a fairly decent OPUS report, difference 0.96 ft at 110 degrees.
NJ0769 HAAS second order, static session L1 on RM3, difference 1.01 ft at 124 degrees
NJ0770 Marion First Pres Church spire, 3rd order intersection station, 1.7 ft at 106 degrees
It looks from this like the model isn't so great in this region.
I hope to measure some additional triangulation stations for confirmation. There is also another good intersection station that I'd have to get permission to set up where I'm close enough and have a view for an azimuth reference, and one that is hopelessly surrounded. The rest of the classically derived horizontal control around here is gone. The only other tri-station in town is long destroyed, and its last RM went away a couple years ago. None of the tri stations within tens of miles have GPS data on the data sheet. The nearest HARN was an elevation bench mark. So I'll be looking quite a few miles out.
-----
Side comment: when I was working on MORRIS, a neighbor told me that some company dug down and set up a tripod on it a few years ago. They left a railroad spike and short iron rod, but those didn't help me locate it because it was next to an iron gate. I left a longer iron rod. I wonder what they were up to. Those guys didn't file an NGS recovery report. I will. Here's my draft.
Those NADCON tools aren't intended to be survey grade transformation tools. The NGS hasn't included the measurements to your triangulation monuments in their 2011 adjustment because they felt that the measurements were not of sufficient quality. Whats the sense in adjusting something +/-0.03' when the measurement is +/-1'?
Were you using NADCON or GEOCON?
Http://beta.ngs,NOAA.gov/GEOCON or GEOCON11
This tool was developed for transformations between the most recent versions of NAD83.
The horizontal coordinates of the monuments you observed were established by classical geodetic methods i.e never GPS observed. If you observed CBN or FBN (HARN) network monuments it would check your calculated difference much better. BTW, the foot difference you are seeing is typical for your area on the classic monuments.
They occupied a number of monuments to develop the HARN network, I don't know any that are intervisible so there will be no check on their angular accuracy.
Near my area they are generally 40 to 70 miles apart, probably in more populated areas they are closer together.
Other monuments are usually NAD27 with a NAD83 number calculated, and those don't check very well.
Most HARN points were Bench marks, this gave them the ability to see the error in the ellipsoid compared to the elevation.
It is more rare to see a HARN point on an old tri station than on a Bench, at least in my neighborhood.
Any monument that isn't a HARN I would consider untrustworthy as a geodetic point unless they say it's been occupied on a datasheet.
linebender, post: 370222, member: 449 wrote: the foot difference you are seeing is typical for your area on the classic monuments.
Thanks for the confirmation. I think you are from Wisconson and also work in Iowa. Have you measured many classical tri-stations down in this area?
GeeOddMike, post: 370206, member: 677 wrote: Were you using NADCON or GEOCON?
For this position, GEOCON2.0 using FBN (=1996) to 2011 gives coordinates that are different by 0.029 ft E, 0.024 ft N or radial 0.038 ft at 49.5 degrees.
As the GEOCON documentation states:
GEOCON employs high resolution grids (1 arcminute by 1 arcminute) to obtain unprecedented fidelity in modeling coordinate differences. Frequently, one may see that the reported quality is extremely good (e.g. 1 cm or better), but this is not comparable to a geodetic readjustment of survey measurements. The National Geodetic Survey considers actual readjustment of survey measurements, and not coordinate transformations, as best practice.
Your result (0.038 ft == 1.1 cm) is extremely good. Looking at the control in your area I am surprised the agreement was as good. As the document states, a geodetic adjustment of the measurements is best practice. Using old measurements combined with new coordinates for control points runs the risk of failing to account for local movement. I would be especially wary of using lower-order control especially intersection stations.
i hope you read Dr Dru Smith's report in the development of GEOCON. Lots of work involved. Your results show that good results can be obtained. The question for me is: for what uses is it appropriate?
Link to the report: http://beta.ngs.noaa.gov/geocon2/GEOCON%20v2%20Technical%20Report%20v03-public.pdf
Cheers,
DMM
GeeOddMike, post: 370695, member: 677 wrote: Looking at the control in your area I am surprised the agreement was as good.
I probably wasn't clear in my last post and some may misinterpret it. GEOCON 2.0 may think it has a model at the cm level, but what I'm finding is more like a foot discrepancy as noted in the original post.
The discrepancy is nearly the same for a first-order tri station and a RM for a 2nd order station, and vaguely similar for an intersection station. I hope to check more old tri-stations for confirmation of the trend. So the comments about needing to be alert to local variations is on target here.
I think I'm 40 miles from a station with both classical triangulation and GPS data (I need to check more carefully but I know it's more than 25). What is typical in other areas for the prediction accuracy of GEOCON in that situation?
You are right I do not understand your posting.
My understanding was that you were comparing positions you derived for points in the NGS database with their published values and noted a one-foot discrepancy. You further indicate that the differences were all in more or less the same direction.
You were comparing three points two triangulation stations (one first the other second order) and one third-order intersection station. All points are within an area less than 20 kilometers.
I got confused when NADCON and GEOCON were brought into the conversation. The practice of the NGS in adjustments from the HARN surveys to date is to use only GPS observations. Conventional terrestrial observations are not used but the points positioned using these "classical geodetic techniques" are fitted to the network. The positions on the datasheet should reflect this fitting. Running either NADCON or GEOCON would obviously be inappropriate.
Reading your first post seems to indicate that you are trying to compare the current realization (NAD83 2011) to the HARN realization. The OPUS and GPS results are in the current realization. Are you comparing making your comparison using the superseded values on the datasheets? How were you able to determine a position of the intersection for comparison?
BTW, Mr Schrock who contributes to this board wrote a nice article which includes an interesting figure showing the network accuracy of NADs since NAD 1927. Note that network accuracy numbers. See: http://www.amerisurv.com/PDF/TheAmericanSurveyor_Schrock-RTN101-Part9_September2007.pdf
The points in your comparisons are all 1934 vintage. While the work to position them met standards of their day what has happened to them since? Reading into your posting seems to indicate that their relative precision would be good but they are not accurate with respect to the national network.
The differences between the HARN and current realization are due to both more and better data as well as different constraints especially when compared to 1996. As you note there are not any data points for the GEOCON grid creation in your area. Only points common to both systems are used to create the grids.
Hopefully you will clarify my points of confusion. Perhaps someone more current on these issues will chime in.
Cheers,
DMM
I start with the published NAD83(1996) latest coordinates from the data sheets for those points (or the RM3 position computed from them). I run GEOCON2.0 to predict NAD83(2011) coordinates. The difference between input and output values is less than 0.04 foot.
Then I look at the difference, my GPS measurements in NAD83(2011) minus the ones out of GEOCON in NAD83(2011). That difference is about one foot.
The tentative conclusion is that the grid model used by GEOCON does not know about a foot of local warpage in the classical triangulation network when you get 40 miles from the HARN points.
Does that make sense?
The comparison in Gavin's article makes it seem quite likely.
Bill93, post: 370829, member: 87 wrote: Then I look at the difference, my GPS measurements in NAD83(2011) minus the ones out of GEOCON in NAD83(2011). That difference is about one foot.
I was actually quite impressed that the positions computed from triangulation were only in error by about a foot. In West Texas, the typical discrepancies are significantly larger. I was dealing with a station (BP0840, HEN 1918, a First-order triangulation station) recently where the published position was in error by nearly five feet. That position is, of course, perfectly good for station recovery.
In thinking it over, I realize that GEOCON couldn't have done anything about that 1 foot. What was needed was a closer point with both triangulation and GPS values in the 1996 adjustment. The closest HARN and many others in the area were vertical bench marks with only scaled horizontal coordinates.
