What's the skinny about doing elevations with gps? Are you guys doing construction stakeout with it? When I used to do elevations with Gpsurvey it used to work pretty well but when I switched to TGO I use it mostly for 2d surveying. So what are the rest of you guys doing?
It depends on what I’m doing. I have VRS and use RTK with a base station when I can’t get cell service at a site. The manufacture’s spec on vertical accuracy is +/- 0.02 cm. I’ve pretty much found that to be the case, and when I’m experiencing something greater than that—something is wrong. I’ll use GPS for shooting panel stations for an aerial survey. I’ll use it for stakeout or for setting control where the elevations are not critical. For curb and gutter or very flat sewer, or for a building I’ll use either the robot with trig leveling or break out the level and do differential leveling.
it depends on two separate things 1. Vertical GPS vector accuracy and 2. Accuracy of geoid.
It is very important to know how accurate geoid is in your area.
but if the distances between at least 3 vertical control points are short, say < 5?m than geoid is not needed because gps can be calibrated to them. In mountainous area you would probably need more dense vertical control because geoid might be changing more there.
Redundancy!!! It all depends on what we are doing. We are using a VRS network most of the time. If we are setting control for a topo then we shoot the CP three time three different times of the day. We rarely see the Horz. Coords move more then 0.10'. Elevations are all over the board so we average them and see how they check with the BM. Then we make a decision on what to do from there. I do not trust he elevations that much so I almost always pull out a level as a check. The firm I work for only uses GPS for topo in open fields and we never use it for layout.
The reason I started the thread is when I attended gps training sessions (ten years ago) it was always stated that one had to have conventional levels surrounding your project. I was wondering if that is still the case, i.e., does one have to run conventional levels on the perimeter of your project before using gps for "good elevations"? From these answers it appears still the case. Of course if someone has some other ideas feel free to chime in.
We typically localize on the a project BM. Most of the time our checks are very good (.03'). We do live in a flat state, with open sky for the most part. If there are any obstructions like trees or buildings the accuracy will go way down. I couldn't comment on any state with steep verticle relief. State Highway projects require Total Stations for paving, or topo of hard surfaces, but only makes a couple hundredths difference. Hardly worth the extra time it takes. On a flat sewer line, and buildings we will use a Total Station for tighter accuracy. We do several checks on both horizontal and elevations at the beginning of a project, and always check each day on at least one BM at the start and end of the day. I like to do a check anytime we pass by any control, or previously staked point during the day also.
if geoid model for the area is not accurate enough you will still need to tie in reliable benchmarks with GPS (run levels to density the vertical control if necessary)
calibration to such control assumes the sea level is flat, so it works for small areas, I would say 5km x 5km square or so.
Depends on what you're doing & size of the site. I'll bring control in from known BM's a couple miles away (shooting several) to a site, establish my project control BM. Then do everything robotically. But on sites where I use GPS for elevations, I'll make sure I have a good calibration and usually will run my robot through them all to double check.
Hammy - Love the Avatar dude!! Good luck with those Coyotes & don't underestimate them. Look out for Korpikowski, not to mention Mike Smith standing on his head.
The NGS now relies on GPS for vertical determinations almost exclusively. We 2cm vert reliably using the commercial network here. If you keep an eye on your local conditions like muiltipath and sky view, check in, and do multiple observations to check the repatability then it should be fine. Is the local network in good shape, are people basing thier conclusions about vertical based on checks with some old published values that might be form a completely different system or datum than the network. Are they relying on a geoid taht might not be from the same time frame as the network positions? Did you adjust to whatever old datum you want to compare to? If all of thoser things are taken into consideration, then good vertical with GPS is easy and can save a lot of time and money.
Rule of thumb:
Dirt-OK
Concrete, pavement, metal-not OK
i use it in Kalimantan "Indonesia" for drill hole and hauling road.. we get around 10 cm difference between GPS elevation and Total Station elevation for each 1 km..
I agree. Topo shots on dirt, no big deal. Hard surface shots, expect accuracy and precision issues.
Depends on the percent of fall.
..."manufacture’s spec on vertical accuracy is +/- 0.02 cm."
!!!!!REALLY!!!!! ±0.02 cm?! Unless I'm thinking wrong that's about 0.01 inches. Who promised you that?
Because Louisiana is in a crustal motion area, benchmarks go stale and out of spec within a few years (3-5 years). GPS is how we obtain current, up-to-date elevations in South Louisiana with the aid of our Real Time Network of CORS sits operated by LSU. When benchmarks are "updated" by NGS, they use our LSU Real Time Network of CORS sites.
The geoid is pretty good for Louisiana. (Highest "mountain" here is 356 feet.)
MightyMoe - generally agree with your philosophy - we do make exceptions where we fix to a digital level-run loop and site BM's occasionally - seems to work for projects under a couple miles in length. Geoid 09 has given good/repeatable results.
you got some heavy reading to do:
what does height really mean? t h meyer, d r roman, d b zilkoski
NOAA technical memorandum NOS NGS 58 guidelines for establiching gps derived ellisoid heights (two and 5 cm confidence level, two sigma) d b zilkoski j d d'onofrio s f frakes
two good choices to start with...
Sorry my decimal point is in the wrong place. Should have been 2 cm
I've been getting good results between level runs and Geiod09.
However, if I'm setting building, curb and gutter, blue tops, bridge or manhole stakes it's going to be the robot, total station, chains and levels. We combine GPS with other instruments when setting some stakes. Such as, we will layout blue tops with GPS but drive them down to grade using a level. Of course some blue tops will be set just with a stringline and a tape.
I don't see that changing anytime soon. It's just the way RTK works. Too much slop for tight grading.
Mo> The reason I started the thread is when I attended gps training sessions (ten years ago) it was always stated that one had to have conventional levels surrounding your project. I was wondering if that is still the case, i.e., does one have to run conventional levels on the perimeter of your project before using gps for "good elevations"? From these answers it appears still the case. Of course if someone has some other ideas feel free to chime in.
The tighter your vertical control, the better. Surround your site, but include internal benchmarks as well.
We are currently working a site where our vertical checks with GNSS is at 0.01' everytime, but personally, I get out the level for our sewer work.
