Second Order Differential Leveling

linked below are the FGCS digital leveling specifications:

https://geodesy.noaa.gov/FGCS/tech_pub/Fgcsvert.v41.specs.pdf

We do a lot of second order leveling, mostly class II but some class I. First of all, you MUST use an invar rod. And the level must be capable of second order. We limit sight lengths to 70 meters for class II and 60 meters for class I, as per the spec. If you want to submit the leveling to NGS, there are two other differences between the two classes: for class I the rod must be calibrated, and temperature profiles are required. The latter means that you must have two thermistors on the tripod, high and low, and record the values for each sight. To me this requirement only makes sense for a long linear line, where the error can accumulate if going uphill or downhill. But we have to do it even for small area leveling if it is to be bluebooked, and we have a monitoring job we do every year over a 2000 foot deep salt mine where we have to bluebook it every fifth year.?ÿ

And yes, a line of sight skimming the ground IS definitely a problem, the heat coming off of the ground will bend the line of sight upwards a bit.

You can be fooled with loop closures if you close back to the POB, errors can accumulate going up hill and then de-accumulate coming back down. This is greatly minimized when using a one piece invar rod.?ÿ

Thanks for the reply, read the specs.

Thanks for the reply. I don't see a correlation between closing back to the POB and errors caused by hills. What length invar rod do you have? How would you fit a 3m

rod in a work truck? What about tripods? We have one rigid tripod, which is the prefered leveling tripod of our most junior instrumentman.

Posted by: @field-dog

Thanks for the reply. I don't see a correlation between closing back to the POB and errors caused by hills. [?ÿ .?ÿ .?ÿ .?ÿ ]

I loved levelling and the pinnacle was a three year contract (every time they filled or drained the drydock, about every six months) for first order Class I work at a Navy drydock using a parallel plate micrometer level, boxed calibrated invar rods, umbrellas, etc.?ÿ Since the drydock was in sand with oil extraction subsidence nearby they had a system of salt water injection wells surrounding it to keep it plumb.?ÿ To do it right it was difficult work, move as fast as you can to ameliorate changing conditions, pick your weather window (cool & cloudy with light winds) and accept the fact that 30% (T&M contract, so no big deal) of the runs would be rejected, I think mostly because of tidal effects, machinery movement (locomotives & cranes, etc.) and vibration from nearby giant pumps, etc. Fun and lucrative times.

But I digress. Concerning levelling in hilly country if you do an out and back loop on a hot and calm day the uphill foresights are at the bottom of the rod and the backsights are at the top, therefore travelling through different air densities (caused by a temperature gradient near the ground) and the diffraction will bias the run. On the return run the foresights are high and the backsights are near the bottom so *assuming* no change in the weather, etc., the bias will cancel out and you think you have excellent closure for the loop.?ÿ Maybe not concerning the elevation of the forward benchmark.?ÿ That's why NGS insists on minimum sight lengths and minimum ground clearance for high accuracy level work along with other strictures concerning equipment and procedures.?ÿ I quote:?ÿ "Refraction is generally the greatest source of error in leveling."

Posted by: @john-hamilton

You can be fooled with loop closures if you close back to the POB, errors can accumulate going up hill and then de-accumulate coming back down. This is greatly minimized when using a one piece invar rod.?ÿ

I beg to differ. The accuracy and stability of the rod does not ameliorate diffraction effects when shooting low then high on the rod in hilly country in bad diffraction atmospherics.?ÿ Can you expound concerning why a highly stable Invar rod (accurate scalar) eliminates such problems??ÿ

@field-dog

I was referring to using a standard folding bar code rod, usually 2, 3, or 4 pieces. Each joint in the rod can introduce a small error. Going up hill the bs shots are high on the rod, fs shots are low, coming back down they are the opposite. I do most of my leveling work in PA, which is quite hilly. I have seen errors accumulate to as high as 0.25 feet at the top of a large hill. Before we had invar we would buy a new rod each year because the joints on Leica rods tended to get loose over time. The Zeiss/Trimble design seemed to be much better. Of course, this can be calibrated by a simple procedure (take shots on each section and compare against leveling done using a single section or invar, and then a correction can be applied based on which section each shot was made on. 

We use 2 m invar when leveling on flat lock walls and dams. 3 m in open country. Rods and struts fit in a large box which can be strapped to the roof rack of a truck. HOWEVER, it is difficult to ship 3 m rods for calibration to CA, they must be sent by common carrier, too big for UPS or fedex. We also have a 0.5 m invar strip for leveling to BM's set vertical (much easier to hold on a vertical mark than a heavy invar rod. 

I bought a one piece illuminated rod last year for a project along a highway that had to be done at night, It was not invar. We then used it on a deformation project that we do 6 times a year that had a 14 m difference in elevation in about 100 m. The error was very apparent. I talked to the manufacturer, they said it was within spec, but it was not good enough for our project, so I sent it back.  

The NGS offers a class on geodetic leveling to FGCS standards. Lots of good presentations are linked from the class web page here:

https://geodesy.noaa.gov/corbin/class_description/leveling_1020.shtml

Screen capture of class page with details about class:

?ÿ

There is a class scheduled for December 2020 at the Instrument and Methodologies Branch near Corbin, VA. Beautiful area. Everyone I know that has taken the class spoke highly of it.

Thanks for the reply and for clarifying leveling on hills.

Good info on rod joints as a possible source of errors.

Thanks for posting the class URL. Good info available in PDF files.

Posted by: @john-hamilton

had a 14 m difference in elevation in about 100 m

Was that error perhaps 14 mm?

@notsomuch

That was NOT an error, it was a DIFFERENCE IN ELEVATION, implying that it was a steep run

@field-dog Do not forget that you need a specific rod type or your work will not meet the minimum standards of what you're looking for.

Also, you're not doing Class I or II work if you're not adhering to the minimum standards set by that document. The lower limit is to prevent reflection of the digital level when sighting low.

@john-hamiltonAhhh, my mistake.  Yes, that is a steep run.

Thanks for the reply. Concerning balancing backsights and foresights, do you think it would be a good idea to use a percentage of the backsight distance to put the foresight distance within the minimum standards? I calculated 10m (32.8084') is 11.11% of 90m (295.276'). It doesn't seem right to take, for example, a 30.0' backsight and then go 62.8' for a foresight.

Also, a peg test routine on surveying software vs. a routine on a digital level. Which is best? Have you ever seen any large discrepencies between the two?

Posted by: @field-dog

It doesn't seem right to take, for example, a 30.0' backsight and then go 62.8' for a foresight.

Definitely?ÿ not good.?ÿ If you study the standards, they want individual legs to have balanced backsight and foresight within a pretty close match and also the sum of backsights over the whole traverse to match the sum of foresights withing a tolerance.

With a digital level it's pretty easy to balance turns, at least on flat ground.  A good rodman will pace pretty close, then you take a test shot and give him the move, then he sets the turtle or pin.  Balancing a run within a foot or so becomes no big deal.

@field-dogif you're using a digital level, you can program in the standards. The instrument will ask you to accept a shot that is breaking your parameters.