Grid vs. Ground, again

The DOT's I work with all use the same conversion system.

All coordinates for a project are done using surface coordinates and are calculated by multiplying the State Plane coordinate by the combined scale factor. In effect this is using 0,0 and scaling the project from there.

This results in state plane bearings and surface distances at the chosen elevation and longitude.

This is not a perfect system, but remember what you are trying to accomplish is to get the project correctly built on the ground and it works well for that.

Even if you do a LDP you will need to account for your elevation. So still you have a scale factor to obtain surface distances.

Many surveyors hate this system-I've never liked it all that much myself-but it does work well for the DOT's and they are used to it.

I have a spreadsheet that outlines metadata for these types of projects that makes it simple for anyone who wants to use them.

> 1) Supply project on grid. Use average combined scale factor in field to reduce ground distances to grid distances. Try to persuade engineer to modify his practices so that he draws on ground instead of grid. He's using Civil 3D. Can/does this happen automatically? Some have suggested that project can be scaled in CAD from grid to ground for the purposes of design and then reduced back to grid for purposes of stakeout, but there seem to be many opportunities for error with this workflow.
>

I get nervous about scaling CAD drawings. particularly 3D CAD drawings, because it will eff up the elevations by said scale factor (unless there is some 2D only scaling routine I'm unaware of).

> 2) Pick one point on the project and hold the grid coordinates of it as fixed, and run everything on ground from there. Engineer does not need to concern himself with scaling the work from grid to ground. Trouble is, someone else getting the data might assume based on the coordinates that it is on the grid, and different troubles might ensue.
>
> 3) Same as 2, but add an offset -10000, -10000 for X,Y
>
3 is better than 2, but both require metadata that will ride along with the project. 3 at least forces someone to look at the metadata or completely discount the coordinates all together when a future user realizes the coordinates put them in the wrong county. An LDP can actually be created that supports a modifrickified (Loyal Olson's term) State Plane System that allows for working Geodetically in a scaled system. I took a couple of days to figure one out a few months back after Loyal mentioned it. You'll need some good projection software to make it work, but it isn't too bad once you figure out how it's done.

> 4) Low Distortion Projection. Seems like it might avoid some of the problems identified above, but if the engineer isn't getting the idea of grid vs. ground, then I can't imagine trying to explain to me the intricacies of a low distortion projection......
>
It doesn't sound like the engineer cares about geodetic coordinates anyway (or am I wrong?). So using an LDP might be just fine - he may never even realize it. Unless there is some government agency forcing the use of a particular projection, I'd go for it. For local work for cities and other small infrastructure managers I've been using LDP's. The coordinates don't mean anything to anyone without the metadata (unlike modifrickified coordinates that suggest a meaning intrinsically), so no worries about abuse. You can also select various points of significance to label with the LDP coordinates as well as the true SPC coordinates. Anyone with the ability to rotate, translate and scale could then put them on SPC from the LDP with minimal effort.

Not tooting my own horn, but there will be a two part article in American Surveyor that I wrote with Loyal's help about this very topic. I believe the dilemma you find yourself in is evidence that the SPCS is no longer the most viable solution to many surveying problems.

Moe has a good solution if you do two things:
1. Put the metadata (conversion procedure) on the drawings
2. Subtract off the millions and hundred thousands from the surface coordinates so no one can mistake them for SPC.

It's still easy to convert back and forth, and the engineer gets to work at the surface.

"Even if you do a LDP you will need to account for your elevation. So still you have a scale factor to obtain surface distances."

Welll...yes and no.

The 'developed surface' of a properly designed LDP would be AT the average "elevation" of the project. So in the case above, that would be ~1500 ft. 'elevation.' So the only "elevation factor" that you have to account for, is the variance between 1500 ft. and any given point (or line) on the project.

Any point within 100 vertical feet of 1500 ft., would be +/- ~5ppm (grid-ground), which is only 0.005 ft. per 1000 feet horizontal.

Loyal

>..... The 'developed surface' of a properly designed LDP would be AT the average "elevation" of the project......
:good
The whole point of an LDP is to have such minimal differences between grid and ground that you can ignore them.

> The DOT's I work with all use the same conversion system.
>
> All coordinates for a project are done using surface coordinates and are calculated by multiplying the State Plane coordinate by the combined scale factor. In effect this is using 0,0 and scaling the project from there.
Oregon used this system before instituting it's statewide LDP system.

Please be kind and truncate.

> The DOT's I work with all use the same conversion system.
>
> All coordinates for a project are done using surface coordinates and are calculated by multiplying the State Plane coordinate by the combined scale factor. In effect this is using 0,0 and scaling the project from there.
>
> This results in state plane bearings and surface distances at the chosen elevation and longitude.

Moe, can you elaborate a little more on what the process is that the DOT's are using?

In the areas where I work, the grid is below the ground, and the combined scale factor is a number less than 1 (as published on an OPUS report, for example).

So, if I ran some GNSS sessions and submitted my observation files to OPUS and got state plane coordinates for a few points on a project as well as the associated combined scale factors for each point, are you saying that I would multiply the X and Y values of each SPC point by the combined scale factor to arrive at the surface coordinate?

For some reason I'm having trouble wrapping my head around that. If the scale factor is less than 1, wouldn't that have a tendency to make the ground distances less than the grid distances?

In this case, I'll suggest that you can have your cake and eat it too.

The global spatial data model (GSDM) allows you to preserve the "big picture" geometrical integrity while simultaneously supporting local "flat-earth" surveying.

The P.O.B. datum point is the project point you and the engineer agree upon and those local eastings/northings can be used in CAD and/or design as you wish.

But, by publishing ECEF coordinates for the rest of the world, everyone knows where you surveyed it to be.

Among many souces, I'd start with #20 on the following list of papers.

I'm willing to consider a low distortion projection for this project. Can someone suggest a good resource explaining how to do it?

See #7 and #54 on the same list as above.

> It doesn't sound like the engineer cares about geodetic coordinates anyway (or am I wrong?)

I think he's used to working with ground coordinates, and doesn't want to hear anything about these minor details.....

> Not tooting my own horn, but there will be a two part article in American Surveyor that I wrote with Loyal's help about this very topic. I believe the dilemma you find yourself in is evidence that the SPCS is no longer the most viable solution to many surveying problems.

I look forward to the article. Thanks for your suggestions.

Earl,

I've read your paper #20 in the past, and it is very interesting. My hesitancy with it is that I have an impression that the equipment that I have (data collector, etc.) does not support the model you describe. Am I wrong?

Al

Big Al

I ain't Moe, but...

The values published on the OPUS Report (and returned by most software) are expressed as the GROUND to GRID relationship, NOT the grid to ground relationship.

edit:

More specifically... the 'k' (scale factor) gets you ellipsoid to grid or vise versa, the Combined factor (caf) gets you Ground to Grid or vise versa. So the "elevation factor" (actually ellipsoid not orthometric) gets you from ellipsoid to ground or vise versa.

So you DIVIDE the Grid Distance by said "factor(s)" to get the Ellipsoial (k factor) or Ground (Combined factor) distances.

Loyal

Ask your vendor. The GPS community is well aware of all the GSDM concepts. Those concepts are inherent in the underlying computational process. But, as I understand it, their marketing focus is on what people will buy. Best wishes!

email me. [email protected]

For instance a combined scale factor of 1.0003 and a northing of 1,000,000 will result in a surface northing of 1,000,300.

Be sure not to truncate the coordinates, it makes it too difficult to "get back" to the state plane value. Also it is much easier to sample in Quads, photos and such into a cad drawing and then "scale up" the quad or photo by the scale factor around 0,0 to get on your coordinate system.

You first need to be sure that State Plane won't work. At low elevations and small scale factors I would just go with State Plane. The scale factor shown above is pretty typical where I am and is 0.3' per 1000' which isn't acceptable for most applications. You really don't want there to be over 1.5' per mile of difference in grid and ground distances. Figure out what kind of scale factor there is in the area and if its really small just use state plane.

I'm not sure where you are but often a computer system will express the combined factor in ground to grid terms the 1.0003 will be shown as 0.9997 but what it's really telling you is your grid distance is smaller than your ground distance and you need to multiply the grid distance by a number larger than 1 to get a ground distance.

One thing I would suggest is keeping Grid North on your LDP the same as SPC Grid North.

It's just a pet peeve of mine.
James

Big Al

OK, that's what I was thinking. Thanks, Loyal.

Interesting to me is that you write about the scale factor being used to convert a DISTANCE from grid to ground or vice versa.

Whereas, Moe seems to be talking about scaling a coordinate, which is different, right?