Benchmarks

BTW, score one for Kent.

RTK for transfer of elevation is not a great idea, generally, and in your case specifically. If you dont know what the difference between GEIOD12a, NAVD88, and the various NAD83 realizations, you are not prepared to effectively use RTK. The NGS website is a good resource.

Are you a licensed surveyor? I admire your willingness to ask for help.

NAVD88 is a vertical datum. NAD83(xxx) is a horizontal datum. The biggest issues I know of would be which geoid model your software is using, and the generally poor repeatability of vertical measurements with RTK.

It probably doesn't matter a lot what the base station is on if what you are doing is comparing the elevations of two rover shots.

I'm not experienced with this, but it's my impression that if you want NAVD88 elevations, your software has to have that geoid model, but if you want the most accurate true relative elevations you want the latest and greatest geoid model.

Who cares and why worry about it for a building lot? The important thing is to leave a good, solid, stable BM on or near the site so that everyone involved in the project will be working from the same height. You could be on '29, '88, assumed of 1000, or just close to any of them. Probably really important is to make sure utility tie ins are all properly related on the same datum.

dmyhill,

"Are you a licensed surveyor?"

No sir. I'm a carpenter. I work for a living. 🙂

I know what Geoids, Horizontal Datums, and Vertical Datums are, thanks to lots of very helpful folks on this excellent Forum. I just don't know particulars. I can't rattle off the minute differences between datums, or which one is currently in effect, or how to "convert" between them. Point in question: If the water district base is using NAD 83 (2011) epoch 2010.00 as a horizontal datum, how would I go about discovering what vertical datum they're using? If they are using NAVD88 as a vertical datum, then my results might be "close enough". Otherwise, it was just a coincidence that my results were as close as they were. I don't like to operate randomly. That's why I'm asking. :-S

Dave

Bill,

"The biggest issues I know of would be which geoid model your software is using, and the generally poor repeatability of vertical measurements with RTK."

Poor repeatability is definitely a problem. And you make a good point about it just being a comparison between two rover shots.

Are NAVD88 elevations Ellipsoidal or Orthometric? Do NAVD88 Elevations depend on what particular Geoid was in effect at that time?

Dave

NAD 83 (2011) epoch 2010.00 is rendering ellipsoid height.
You need the cipher between the ellipsoid and geoid height to determine an "elevation" BVUT that is the least of your problem.

To really establish/transfer an elevation there are GPS static surveys protocols established by NGS

Dave(A good name),

"Who cares and why worry about it for a building lot?"

That was my first thought and the reason I wanted to save my employer the expense of a 3-mile level loop. But I am concerned about the utility tie-ins. This topo is a "preliminary" for a "rough draft" civil design, and I was just wondering if anyone else was using RTK to transfer elevations or whether I'll have to tighten things up with a level loop.

Dave

Project datum is project datum; does your project require it to be on a specific datum?

If it is; what is the error tolerance? If you can't be confident that RTK will achieve what they want, then you better get out the rod and level. The right tools for the right job.

If you can use RTK and still meet the requirement; I try to use as many known points as I can; throw out the outliers and take the average. The state DOT here has many points all along it's routes; I can easily find 4 or 5 and tie them, in less than an hour (I use RTN). Just put the gear in the van and drive to the next one.

Close enough means a lot of different things to a lot of different people: from horseshoes and hand grenades to setting wing patterns on a Boeing 747.

Like Dave Ingram said; make sure you leave a good bench-mark on site and be sure to document what you've done on your site plan.

Radar

Robert,

Thank you. A well put clarification of concepts.

Dave

> ... I hit 2 different published county benchmarks. ...my elevation disagreed with the published elevation by 0.15 feet. At the other benchmark which was in the clear, again averaging (6) 1-minute shots, I was within 0.02 feet.
It may very well be that the benchmarks themselves are not that tight. I can name lots of cases were I've found that much difference, using closed loop digital levelling, between monuments of a benchmark network.

Averaging 6 shots is one thing, Averaging shots taken over a span of time is another. The average of 2 shots 6 hours apart would be better than 6 shots in a 10 minute span. And one minute is not particularly long for 3 mile vectors.

> The published elevations are NAVD88. My GPS base was a water district base which uses NAD 83 (2011) epoch 2010.00.
Saying that elevations are "NAD83" implies that the elevations are ellipsoidal elevations. NAVD88 elevation will typically differ from ellipsoidal by about 70 feet.

>Is it acceptable practice to use GPS to transfer an elevation?
I've done it. There are NGS specifications for transferring elevations by GPS. Far greater concern here is your understanding of the datums and processes involved. 0.17' in 3 miles isn't very much for most ordinary topographic work.

Radar,

The "final" civil drawing will, of course, have to include the information of which benchmark the project is based on. I don't have much experience with RTK, so I'm unsure of accuracy expectations. Every civil drawing I've ever seen shows the reference (published) benchmark and the vertical datum it's based on, but I've never seen any mention of what method was used to transfer the elevation. If I had a closer benchmark, I'd just run a level loop.

I have another one of these topos to do where the closest benchmark is 8 miles away.

Dave

You've gotten some good advice and information so far. A few more random thoughts..

For the time spent you could have set one unit on the bench in a good environment and done rapid static on the other bench and your site control. You could follow that up with an occupation of site control with rapid static to the other 2. This would up your quality and free you up to do other things while the units cook.

The thing to drill into your head is that RTK from one base or observations under similar constellations are nothing more than side-shots. As for the heights, geoid separation varies widely. Around here it's 56 feet. Once you get a handle on the terms and what normal local numbers look like, never forget those are local numbers...

Good Luck, Tom

Mark,

"Far greater concern here is your understanding of the datums and processes involved."

Yup. Agreed. That's why I'm here. This is the best place to go.

"...average of 2 shots 6 hours apart would be better ..."

I would definitely grab some afternoon shots to check if I end up sticking with RTK.

Dave

Tom,

"...while the units cook."

Units? Ha ha! I've got ONE limping 5700 with iffy batteries. And my knowledge of GPS is about on a par with my equipment. But I continue to learn. Thanks for your response. I greatly appreciate all the help I get here.

Dave

Didn't realize you didn't own a base...
After investing more time in learning, you may consider picking up an older base and radio to up your game. Alternatively just grab another used rover (some under a grand right now). It seems the number of units is an exponent to productivity, not a multiplier...
You've already made a great first step by stepping in and asking for help. Now you are in phase two where you push past any belligerent comments that may come up:^)
Given your attitude I suspect you'll be ready for the purchase at some point.
Best of luck, Tom

>Are NAVD88 elevations Ellipsoidal or Orthometric? Do NAVD88 Elevations depend on what particular Geoid was in effect at that time?

NAVD88 is Orthometric. Specifically Helmert Orthometric, which means that they applied the geoid model available at that time to estimate where "sea level" was underground. As others have said below, there will be many feet of separation between ellipsoidal and orthometric numbers just about anywhere in the USA.

It will depend on the situation at your location whether the tilt of the geoid is significant in 3 miles, and whether a newer model gives a significantly different answer than the NAVD88 model. The newer model may be more accurate to the world, but probably is not be what everything else around was built with.

Tom,

Using company tools; they're not mine. No spares and no money for anything not absolutely necessary. The water district bases are a Godsend when I need to use GPS. I do a lot of creative workarounds every day with all the tools at my disposal.

Here's this weekend's reading:
http://gpstraining.com/downloads/MANUALS-QUICK%20GUIDES/RTKTrainingRevD.pdf

It's deadly dull, and certainly written by Trimble's lawyers, but I'm plowing through it.

Dave

> "...average of 2 shots 6 hours apart would be better ..."
>
> I would definitely grab some afternoon shots to check if I end up sticking with RTK.
The idea is to get shots under different satellite geometry, and therefore under different multipath conditions.

I strongly recommend this from the NGS.

Dave,
We use a local network from the blueprint company. When things are rough or cell service unreliable, we switch to rapid static. I download the data from any of their close base stations and process the data.
Remember, there are dozens if not hundreds of users nearby. Some of them run a tight ship and share data freely. It's like having another rover or rovers (and someone to babysit each one).
You seem like a 'think outside the box' kind of guy. As you read and learn do a little research on the local free sources of data.
Good Luck, Tom

btw- May be heading to Casa Grande for a week in a few months. I seem to recall signs for Mojave Valley on the route (but its been 15 years). I may hit you up for a coffee break along the way if you're up for it...