NEVER use the Molodensky 3-parameter transformations. These were all computed by the old Defense Mapping Agency (DMA), now the National Geo spatial-Intelliegence Agency (NGA). In their own publication -- file:///C:/Users/User/Documents/PUBLICATIONS/WGS/NGA-STND-0036-1-0-0-WGS84.pdf on pages 156 & 157 of 207 you can see that they break the lower 48 states into CONUS, East and West and their estimates the 3-D 1 sigma (68% confidence) uncertainty ranges from 6.5 to almost 10 meters!! Regrettably these estimates never show up in the array of tools that many developers provide. ALWAYS use NADCON/NCAT from NGS. This is well defined by a Federal Register Notice from 1990
If it helps I did a short YouTube video on the subject of transformations.
NEVER use the Molodensky 3-parameter transformations.
Joel is using Trimble-speak. "Molodensky" where NAD83 job coordinate systems are concerned = shorthand for null transformation in Trimble-World. Not technically correct but functionally correct.
@base9geodesy yeppers. One of the very reasons i used ncat as a check before proceeding on. I know all about the 3 parameters issue. I always do a check for anything in usa by using NGS tools. I know the issues of software developers being behind getting things updated. I truly think the way the new datum is being handled is going to be much better than the past. Now the manufacturers will not have an excuse lol
@norman-oklahoma yes that was my first guess but it was not. It was TBC. I have confirmed that. Ncat proved that. NCAT gives you the error estimates from any transformation. I trust the designer's of both datums more for transformations vs off the shelf software. So any time I even go from one datum to another i use the ngs tools for those. Outside of conus thats a different story. But ngs is very solid and if a issue did arise i would rather go to the source.
@mightymoe tbc with new update nailed ncat. It was the older version of tbc coordinate system manager that was bad. They even stated that in the latest coordinates system manager update. It has not been out long at all.
The issue is that the conversions don't actually "get on" the point. It's been my experience that they get you close, if .5 to 1' is good enough then that's probably what you'll find. The real way to "get on" NAD27 is to get on the monuments. It depends what you're trying to do. Old highway jobs-get on that control. Old city work-get on that control. Oil and gas geological data maps-doesn't really matter since they aren't on great control anyway.
@mightymoe i see what you are saying now. That most definitely could be true of points that have moved or whatever. The point was only as good as the last time it was measured. In my case once I reverse engineered there scaling the coordinates it was bang on less than .15 ft most were less than a tenth. Only one was about.15 ft . So ncat was correct and tbc now is. At least in my area. But I ended up not being able to hold the true values as its all relative to the scaled coordinates. But I promise you everything i saved has a big note stating it. It site control not nad 27 or nad 83. Lol.
3 feet? Check on the US feet, Int'l feet thing before going further.
Ellipsoid, US feet, and why aren't you just doing a local calibration anyway? Who knows how well they fit the actual datum.
@dmyhill I was almost ready to do that. But its very linear and once i started investigating and following what all had been done prior it was easy to figure out. And now its all cleared up and in the clients hands. Hopefully i will get to redo all the control so its correct. I tied it all down and did what was needed.
I'd be skeptical of creating a custom calibration or transformation unless I had plenty of monuments to go off of. As a general rule if a someone wants NAD27 they are getting NCAT- or TBC-converted values based off of a good, solid NSRS survey, plus a narrative explaining the pitfalls of NAD27.
It's rare that clients actually, truly need NAD27, and if we can get them in a meeting prior to finalizing the contract, it's not uncommon to convince them to go with a more reliable datum, or at the very least deliver both current NSRS as well as NAD27.
