Reciprocal Leveling

> Anyone ever use this method for transferring elevations across rivers or other bodies of water.

The closest I've come was last year when we did 3 crossings of the Sacramento River, but we used 1" total stations for the reciprocal work rather than levels. The length of the crossings were in the 600'-700' range. The 2-sigma vertical errors were under 0.01' for two of the crossings, and a shade over 0.02' for the third.

Joshua,

I have a hard time reading a rod past 200 feet. I take it you're not lucky enough to have a nice, low-traffic bridge anywhere near, where you could run a level loop. I like Jim's method.

Dave

Dave,

Yes, This is for a bridge replacement project. I actually shut the existing bridge down at 15 minute intervals and ran 3 wire one way between existing control monuments. The bridge is about 4,000' in length and very is unstable.

I was concerned about the accuracy of my circuit due to bridge's condition. So I ran under the bridge upon the old piers. The problem is there is about 1,200' - 1,500' that I can not occupy for surveying. I really want to close my circuit without going back over the old rickety bridge. So my reciprocal leveling would be 1,500' shots.

I do plan to run static control as a check also. My error budget is about 0.04' vertical though.

I think Jim's method will work for you, if not only for the simple reason that the total station/theodolite will generally have a higher magnification. oops sorry I take that back, just checked the NAK-2 is 32x and Trimble S6 (for example) is 30x

500m sounds like a long long way to read a staff at.
A 1" instrument with autolock would be perfect.

How About OPUS-RS ?

Set multiple points on each side and do simultaneous GPS observations side to side.

OPUS-RS should work along the Indiana/Kentucky border and the Missouri/Kentucky. Along Illinois there are not sufficient/any CORS area. However even simple static GPS should get you 0.02' side to side. OPUS-RS being a means of confirming your precision.

Paul in PA

I observed (watched and stayed out of the way) while our NGS Advisor and the DOT did a river crossing on the Arkansas River here in Little Rock. They checked it by running 1st Order levels across the river on a pedestrian foot bridge. The length of the bridge is 1.2 Kilometers. The River Crossing procedure was quite involved, 4 Zeiss NA2 levels with long run micrometers and 4 targets. It took about 8 hours to do but this was the first river crossing for the DOT so there was some learning curve involved. They followed this up with trig levels and GPS OBSERVATIONS, RTK and Static. I have heard that NGS is going to develop a River Crossing procedure using Trig levels that can be 1st Order with 0.5 second total station and 2nd order with a 1 second instrument. Their procedure will be interesting to study. By following their procedures our work will have more validity.

For updates to the NGS River Crossing investigations you should contact either John Ellingson (john.ellingson@noaa.gov) NGS Wisconsin state geodetic advisor or Dave Zenk (dave.zenk@noaa.gov), NGS Minnesota state geodetic advisor.

trig-levels for sure!!!

If you can get a 1" gun.

Observe both directions, apply curvature and refraction and mean the results, be sure you observe under the same conditions, best if you can do it at the same time, don't do it in the middle of the day. If you can wait till night even better.

> I think Jim's method will work for you, if not only for the simple reason that the total station/theodolite will generally have a higher magnification. oops sorry I take that back, just checked the NAK-2 is 32x and Trimble S6 (for example) is 30x
>
>
> 500m sounds like a long long way to read a staff at.
> A 1" instrument with autolock would be perfect.

I do have S6 but it is a 2" gun. I am still going to give it a shot, pun intended, with autolock. My NAK² has the optional 40 power eyepiece.

> trig-levels for sure!!!
>
> If you can get a 1" gun.
>
> Observe both directions, apply curvature and refraction and mean the results, be sure you observe under the same conditions, best if you can do it at the same time, don't do it in the middle of the day. If you can wait till night even better.

:good: :good:

Thank you for the contact information.

> Observe both directions, apply curvature and refraction and mean the results

By taking simultaneous reciprocal observations you eliminate the effects of C&R, so applying those corrections is an exercise in cancellation.

Well, not exactly, I always apply them, because if you don't you won't see the two match which they may well do.

The refraction can be quite different looking one way compared to another, and I like to get the "real" number. I know, I know it will be subjective.

You are correct meaning them with or without the C&R applied should give the same number, but after doing it for so long I can't give it up;-)

Doing it at night will make many issues with refraction go away.

Manual of Geodetic Leveling

There is coverage of this issue in the NGS Manual of Geodetic Leveling available here: http://www.ngs.noaa.gov/PUBS_LIB/Geodeticleveling_nos_3.pdf

See chapter 4.

I look forward to the work mentioned in Arkansas being incorporated into a new document.

2 secs over that distance is 5mm (or 0.015") so if the rest of your setup is good you may be able to stay within your error budget.
Although I have heard it said that angular accuracy is roughly doubled on the vertical angles compared to the specified accuracy of the instrument. (ie less accuurate). Not sure why that is though. Only reason I can think of is if you were doing multiple sets using different parts of the horizontal circle you might get more accuracy but you can't do that with the vertical circle.

C&R cancel exactly only if the two backsight distances were equal and the two foresight distances were equal. They will probably be close, but why not go through the exercise to see if there is any effect?

> I actually shut the existing bridge down at 15 minute intervals and ran 3 wire one way between existing control monuments. The bridge is about 4,000' in length and very is unstable.

I was turning east a few days back from a subdivision. If I had not actually seen the traffic heading west was backed up, the "very unstable" would have given away the location. At least I think I know where you are talking about although there are a number of bridges meeting that description.

I am very happy to see some work going on at that location as I drive over that "very unstable" bridge almost daily.

Just south of that location, due to a sheer cliff face, Chris and I closed a traverse by

shooting across to the other side,
driving around to set-up,
shooting back over on the north side of the property,
driving back around to set-up again,
driving back over for a back sight,
and back around again.

It took a while, but was much less time than the chopping that would have been required. The water was not too bad to look across. Just pick the right time of day to minimize heat waves or fog! We ended up with a very good (although not ideal traverse figure) closure.

Trig leveling might be worth a shot. Very easy to make multiple observations from each side.

> C&R cancel exactly only if the two backsight distances were equal and the two foresight distances were equal.

In my case, I ran closed level loops to one point on each side of the river, then setup up tripods and tribrachs over each point. Using some custom-made adapters and Trimble height sticks, measurements to the tribrach contact planes were recorded. A total station (Geodimeter GDM640) was locked into the tribrach over one point, and a prism with target into the tribrach over the other. (Distances from the tribrach contact plane to the instrument horizontal axes and prism/target centers were previously determined.) A taped measurement to the gun's horizontal axis was made and recorded for blunder-detection purposes. Multiple zenith angles and distances to the prism/target were recorded in both faces. Then the total station was replaced by a target, and the prism/target on the other side was replaced by a theodolite (Leica T2000). Multiple zenith angles to the target were recorded in both faces. All data was hand-entered onto a custom form and later reduced in the office. I believe this procedure effectively cancels any C&R effects.

By knowing HI at each end, and using only the two points, then C&R must cancel to the degree they are truly symmetrical.

The form is probably the one posted about 3/4 down in [msg=270327]this thread[/msg]. I found it a little hard to decipher but eventually made a sketch that let me figure it out (I'd have to do it over now). It certainly does the job very well once you understand it.