NOAA NGS - Coordinate Transformation

[Anonymous]

I'm currently having to convert back to NAD 27 coordinates from NAD83(2011) for projects.  If I use AutoCad to map between drawings I do NOT get the same coordinate that NGS ( https://www.ngs.noaa.gov/NCAT/ ) gets.   Since they are the leader in the area, it make me wonder why CAD doesn't have the most updated parameters/definition for NAD83 (2011)?  I should be within a 0.10' and yet I can't get any closer than 0.5'.  

Also the coordinate system NSRS11.LA-NF doesn't appear to work and I can't only get within a 0.5' using the HARN/LA.LA-NF.  

I've reached out to NGS but they just referred me to the Vertcon/Nadcon BIN files for the coordinate systems which I have no idea how to import that into CAD. 

[Pointdump]

Dustin,
Welcome to the Autodesk Forums.
A little more information please. What version Civil 3D / Map 3D are you using? All updates? I remember 2016 had a little problem with NAD83-2007/2011. Can you post data or a drawing that shows the fumbled transformation?

Dave

[ChicagoLooper]

---

@Anonymous wrote:

Also the coordinate system NSRS11.LA-NF doesn't appear to work and I can't only get within a 0.5' using the HARN/LA.LA-NF.  

---

What does this mean? Are you referring to horizontal discrepancy or vertical discrepancy? Is the X, Y or Z value giving you an error? Or is it a combination of X, Y and/or Z? Vertical datum will not influence the horizontal values.

Converting from NSRS11.LA-NF to LA-N (NAD27) is unusual because it downgrades the accuracy of an up-to-date coordinate system to an outdated coordinate system.

Instead of moving your NSRS11.L.A-NF to a coordinate system referenced against NAD27, why don't you consider moving in the opposite direction, such as from NAD27 to NSRS.11LA-NF?

For a solution to your issues, please upload your drawing.

[Anonymous]

Unfortunately as per our state, they regulate that we submit everything in NAD 27 to keep it paired with their historical Oil/Gas data.   To keep things simple for the field crews, we do all of our work in NAD83(2011).   We are dealing only with horizontal datum and nothing vertical.   NGS is the leader in this area so they will also be assumed to be correct over any other source.  The issue is, I can't map over entire drawings through their NCAT programs only points.   

I've attached a copy of the NCAT conversion which I have to assume is the correct answer.  I've also attached an OPUS sheet showing the conversions written in using various software including Global Mapper and Civil 3D 2016.  

If this is a 2016 issue, then how can I up date the coordinate files?   

This is the response I'm getting from NGS regarding the BIN files aka Grid files.  I'd like to know if Civil 3d utilizes these file from NGS for proper conversions. 

Re: Coordinate Conversions`

Hi Dustin, Under NCAT, the transformations between frames and datums are not definitional per se - there is no set of transformation parameters. Rather, as Krishna stated, the transformations are specified as a grid, with a unique transformation value and transformation error assigned to every lat/long in a supported region. The grid format is described in the NADCON 5.0 report:  In a moment, I will attach a Matlab script which reads both NADCON 5.0 and VERTCON 3.0 grid files. Hopefully this will get you off to a good start with your own code. This Matlab script is not official NGS software, but rather is provided to demonstrate how the grid file can be read into a common coding language. Please feel free to ask me additional questions if the above does not meet your needs. Thanks, Andria

[Pointdump]

Dustin,
I re-traced your steps with NCAT and Civil 3D 2018. I get the same.
As far as I know, Civil 3D uses the LAS/LOS grid shift files provided by NGS, so I don't know what's happening. You might try adding the Autodesk-supplied xml file from >>>Here<<<.
Dave

[Pointdump]

Dustin,
I added that XML file just now to my 2018, and I get NAD27 coords of:
439971.29758338,1733328.88111520
A bit closer.
Dave

[ChicagoLooper]

This issue has been misdiagnosed as an error with either NGS NCAT or as an error with Civil3D. However, when using geographic coordinates, not projected coordinates, to test a point location, Civil3D is consistent with the coordinate transformations. The only error I can see, is the way NGS NCAT is transforming the Lat/Longs. This indicates there could be an error (or poor decimal precision) in the 'input' you are giving NCAT and not an error in the 'output' of NCAT's coordinates.

Image-1. For best results, provide coordinates to NCAT, using coordinates with 6 decimals or more.

Image-2.Global Mapper. The projection being used in LA83-NF.

Image-3. Civil3D. Compare coordinates to image-1.

[ChicagoLooper]

@Anonymous wrote 

....NGS is the leader in this area so they will also be assumed to be correct over any other source....

Whether NGS is the leader in your area is debateable. The interface of the NGS is woefully inadequate because the 'projected coordinates' (x,y) can accomodate only 3 decimal places whereas 'geographic coordinates' (Lat/Longs) will accept 10 in decimal format and 5 decimals in DMS format. The precision in your input makes a difference and with only 3 decimals in your x,y values, the leader in your area is daring you to use geographic coordinates instead of projected. Time to find a new online converter?

The error you are getting is due to your incorrect input. You are using NAD83(2011) epoch. You should use the NAD83(1986) epoch instead. And if you wish to maximize your output, combine the 1986 input with lat/longs instead of x,y feet.

For input use NAD83(1986).

If you wish to get NAD83 State Plane or NAD83 UTMs, both the input and output datums must be NAD83(1986).

If you have ultra-high precision Lat/Longs and you want SPCS, you can try >>EARTHPOINT<<. 

[Anonymous]

I'm still battling this issue and it appears to be even worse in South Louisiana.   No one has been able to explain these differences with CAD yet.   While I agree that NGS is the leader is debatable, I still need to get this resolved or at least get Autodesk and NGS on the same page!     The plat below shows NAD83 (2011) XY which is what we survey in mapped over to NAD 27.   If I convert this same number in NGS and GM Geographic Calculator it's close to 30' OUT !!!!!!    This is absolutely not acceptable.       

[ChicagoLooper]

This is a local or 'site specific' coordinate system created by the surveyor. You can tell by the 'truncated values' in the Cartesian coordinates. The surveyor is using Louisiana South Zone projection with datum specific to the site. If you can make contact with the surveyor and he's willing to give you more information (trust me, there's more info) you could replicate the surveyors coordinate system, add it to your built-in Coordinate System Library then assign this replicated system to your modelspace.

You did not disclosed what you want georeferenced against NAD27. Your uploaded image? Line work? Something else? Full disclosure, please.

To put your image with the yellow boxes into NAD27, do this. (Do not combine these instructions with other instructions that may have been whispered into your ear by self-appointed coordinate Gurus. Combining the instructions with info you've read, heard, watched, tasted or smelled, may cause the instructions below to give inaccurate results.

1. Start from brand new imperial template. MAPCSASSIGN command. In search box type in LA-S-MOD=>select the result from this search=>Click ASSIGN button. You have just assigned Louisiana State Plane NAD27, South Zone, feet to modelspace.
2. Go to Geolocation Tab=>Tools Panel=>Mark Position dropdown=>Lat-Long icon.
3. On command line for Lat type in 28D57'53.5" ENTER. For Long type in 56D22'46.3" ENTER. 
4. Then, in modelspace, at your flashing cursor, type 'BHL' for a label and left click an empty area in modelspace to end this Geomarker command. Zoom extents. You have just placed the northwest geomarker in its correct spot. 
5. Repeat steps 2, 3 and 4 for the southeast point. Label this geomarker 'As-Drilled' when you get to the flashing cursor part.   Select both geomarkers=>Properties Palette=>Scroll to and change SIZE from 1.0 to 25.0. Your Geomarkers will get bigger. (Note: If you had 2 decimals in the seconds you'd get more accurate positions, if you had 4 then even better, but all you have is 1 decimal, too bad. You can blame the lousy position on the precision, or lack of precision, of the Lat/Long coordinates.
6. Turn on Bing Hybrid imagery for a visual background. Insert your image. You can scale and rotate the image using the the geomarkers as reference points.

NOTES:

•  This method will work whether you start a brand new drawing with an assigned CoordSys of NAD27 or NAD83. The Lat/Long coordinates will be the same and accurate either way. What won't be the same are the x,y coordinates when comparing x,y NAD27 vs. x,y NAD83.
• Don't expect the geomarker to grow legs and move to its correct spot if you change the assigned CS from NAD27 to NAD83. If you change the modelspace's CS, then the reference is broken and you'll need to manually reestablish the geomarker's position from scratch.
• You can't, and shouldn't, expect the x,y coordinates of Louisiana State Plane South to match the coordinates listed on the survey. The surveyor has created a site-specific CS where he has truncated, or shortened, the x,y values to be consistent with a system referenced against a psuedo-State-Plane CS. Had he not used a site specific CS, then the x,y values of the geomarker would match the x,y values shown in the survey.
• To view a geomarker's coordinates in DMS (deg-min-sec), type LATLONGFORMAT on command line and make this variable =1 then select a marker and view coordinates at bottom of Properties Palette in the Locations section.

[Anonymous]

I’m familiar with the surveyor and the firm since I worked there for years.  All they are doing is using the query function in cad to map from a NAD 27 referenced DWG to a NAD83 referenced DWG and vice versa.   We do all our survey work in NAD83 LA South in this instance but have to report in NAD27.   I can not replicate NGS’s NCAT program in auto cad.  This appears to be the worst case where it is 26’     This will be an absolute nightmare if CAD can’t match NGS or at most be within a 1’.     If you can try using my Work flow in CAD then compare on NGS NCAT site.   

[ChicagoLooper]

---

---

---

@Anonymous wrote:

........This appears to be the worst case where it is 26’.........

Huh? You're definitely doing something wrong. You can do all of this rudimentary transform stuff in  Map3D so forget about NGS. The idea that NGS if 'The Leader in the Area' is laughable.  

If you know the surveyor is using a query, then what's stopping you from using one too? It took me less than two minutes to query the points known as BHL and As-Drilled into a drawing with NAD27 assigned. Your insistence to use your workflow and not follow the surveyors method of using a query is giving you grief.

The problem is further complicated with the surveyor's Lat/Longs. They are in DMS and the seconds have only one decimal place. Why? You need more. You'll have wiggle room when you plop down a point using lat/long with only one decimal in the DMS seconds. If you want accuracy, then use high precision Lat/Longs.

I placed both of the surveyed points in two separate drawings, one with NAD27 and another with NAD83, then compared the NAD27 points withe the NAD83 points. Then QA/QC with the surveyed coordinates shown on the survey sheet. The margin of error is +/-5 feet.

Your number of 26 feet is, is,is, well......I don't even wanna go there. Get over it. Drop NGS stuff. Run a query in Map and you're done.

[Anonymous]

So I meant to say my work flow is identical to his since I worked there and learned under them.   The reason they only show one decimal is really irrelevant since this is simply a choice in a custom tagging program they have.  

the issue is, if I have a discrepancy with ANOTHER survey company we will always have to revert to NGS as the final answer.   I agree that their resources are weak but in the survey community they have the final say.    So to sum up, my coordinates and work flow will match the C&C plat I attached but because NGS doesn’t give the same answer I’m forced to change my workflow and am not able to rely on the transformation that CAD is giving me because these don’t match with NGS.  Repeatability is KING and if I can’t repeat these numbers through a second source, NGS, then I can’t rely on it.    Here lies the problem.  Matching NGS has to be part of the solution and I can’t just disregard it since it will be the arguing point when I don’t match up with another company.  

[Pointdump]

Dustin,
NCAT admits that NAD27 accuracy is +/- 1 Meter in some spots of US.
But if NCAT is your standard, then the solution is obvious. Import NCAT points into your drawing, then use the RUBBERSHEET or ALIGN command to match up Linework and Maps.

>>>Corpscon<<< is another "official" source of conversions. Using their HPGN-to-NAD27 conversion, Map 3D is pretty close:

Another way to accomplish what you want, and I think the best way, is to use an >>>Affine Post Processor<<<. It will handle translation, scale, and rotation with ease. You need at least 3 common points.
Dave

[Anonymous]

David, 

I'm not sure where you're getting your number for NAD 83.  

Let's try this exercise with 

LA South NAD83 (2011)

X: 3,906,002.00'

Y: 173,857.00' 

I continue to get these number in LA South NAD27  

MAP 3d 2018 

X: 2,625,198'

Y: 113,152'

NGS NCAT

X: 2,625,200'

Y:  113,179' 

I'm trying to avoid doing any rubbersheeting or additional steps that will only allow for more error to be introduced.  

I would take these number being under a foot difference ALL DAY.  But there not.  It's 26' and change.   

We need a solution not a work around.   AutoCAD is to expensive for me to still have to do work arounds.  This is an issue that needs to be resolved.  

[Pointdump]

Dustin,
"I'm not sure where you're getting your number for NAD 83."
Your post #4.

For your second data point, yes, I'm getting same:
             Northing Easting
NCAT    113178.851, 2625199.717
Map3D 113152.354, 2625196.618
Difference N:26.497, E:3.099

Curiously, I could not check Corpscon. It said "Out of Range". Which makes you wonder if the NGS Binary Grid Files are out of range at Burwood Bayou.

You should open a support ticket with Autodesk.

If you could provide some data with at least 3 points, I'd like to try the Affine Post-Processor route. Or post the Binary Grid Files you got from NGS for checking against the MAPCSLIBRARY files.
Dave

[Anonymous]

So we were able to use Corps Con and it doesn’t match NCAT program either by the same amount.  Starting to think maybe NGS has an issue on their end.  I’ll keep you informed.  

[ChicagoLooper]

Are yiu sure?  I can’t remember where, but I read on the internet somewhere NGS NCAT is the Leader in the Area. 

Also read the results you get in Corpscon are the exact same results you’ll get from Map3D. 

[ChicagoLooper]

Verified. I can confirm the surveyor's coordinates shown in Post 9 of this thread using his geographic coordinates (Lat/Longs) as the source data.

Image-1. Surveyed point 'As-Drilled' (from Post 9).

In image-2, notice the Lat/Longs taken from the survey have only 1 decimal in the 'seconds' portion of DMS. Also the X,Y's in columns B and C are rounded to whole numbers. You be the judge, does the survey have the precision you expect? Failure to acknowledge the degree of precision may cause issues.

Image-2. Columns B and C taken from survey in Post 9. Columns D and E taken from Map3D and used for verification.

To demonstrate the significance of Lat/Long precision when inputting your points, consider the effect of rounding decimals. The surveyor provided coordinates in DMS format, 28-57-48.2N and 089-22-42.6W, latitude and longitude respectively. The survey provides the Lat = 48.2 seconds. If the seconds collected in the field are 2 decimals or more then 48.15 and 48.24 will also be treated as 48.2 when rounding to 1 decimal. Furthermore, the survey provides Long = 42.6 seconds and if the field collected data is actually 42.64 or 42.55 then those seconds will be treated as 42.6 also due to rounding.   

Back in AutoCad you can illustrate the consequences of the max/min Latitude and the max/min Longitude that occurs due to rounding. For the survey point named 'As-Drilled', see image-3.

Image-3. For 'As-Drilled' any point within the red bounding box will conform to the DMS coordinates shown in the survey due to rounding.

If your Lats/Longs coordinates put a point on the outer edge of the bounding box it'll be treated as though it's in the middle when rounding the 'seconds' of DMS to 1 decimal. This, of course, is true not only true for points on the perimeter, but for any point within the box. Depending on your drawing, rounding can play a significant role in this +72 sq-ft box. Whether that role is negligible or adverse will vary from drawing-to-drawing.