Link Traverse Survey Closing Error Beyond Allowable

Has anyone faced such kind of problems? 

Of course. I can't say that I've ever done a 120 km long project, but weeding out errors is a part of every project.

Typically a road project would need traverse points every 200-300 meters, so that is maybe 20 traverse points between 5km check in points. That is a lot of opportunities to mess up.  I'd be looking at using RTK to densify your checks between the static points to isolate the errors better. Then you may have to re-observe a portion of the traverse.   

@norman-oklahoma Thank you for your reply. At that time there was a resource limitation of RTK GPS survey instrument. I have done the traverse survey repeatedly by implementing the necessary procedures for total station survey but the error was the same.

Typically a road project would need traverse points every 200-300 meters, so that is maybe 20 traverse points between 5km check in points. That is a lot of opportunities to mess up.

I agree.

Depending on how the static data were observed, I would be expanding my error budget to account for the relative error in the GNSS-positioned values. Short baselines, even with good static procedures, can be problematic.

I have done the traverse survey repeatedly by implementing the necessary procedures for total station survey but the error was the same.

Then it sounds like your traverse data is blunder free. It could be a problem with the static positions, or the project specifications could be too stringent.

Another possibility is not properly accounting for the scaling on the projected coordinate system. What datum and projection is being used here? 

If you are willing to share your raw data files (ie/ the angle/distance/measure up readings) and the static positions perhaps someone would be willing to have a look at it.  

What Norman has said for sure. Is it angular only how does the distance check?

Is the project in ground or grid?  If the project is in grid, you will need to apply the appropriate ground to grid corrections for the total station observations. 

@rover83 I agree with short base line and number of traverse stations determine the misclosure error.

@charlie_wagner The project is in grid. I have applied the necessary factors to change ground to grid.

Strength of traverse figure (long skinny)?

Maybe free up some GPS points and see how the adjustment changes their values ?

Have you tried to isolate the areas with larger than expected errors and run the adjustment between those sets of azimuth pairs separately.  Another idea would be to forget the internal azimuth pairs and just hold on the static GNSS points per pair.

Like Gary said (didn't see his post until I hit send).

If you're still seeing the same misclosure after re-running the conventional work, I'd be running a minimally constrained weighted least-squares adjustment to find out whether any of those observations are statistically significant outliers.

Maximum allowable closure tolerances are ultimately arbitrary values and should not necessarily apply in every scenario. I've worked projects that required horizon closures of less than 5 seconds at every single station...no matter what instrument and no matter what the traverse leg length was.

If you're double checking and your procedures are still good, like Norman said, maybe the standards are just too stringent for this particular project.

This reminds me of a problem I was asked about one time. The surveyor was taking a 400 m backsight and traversing 6 km to the next azimuth pair. They had a 20cm misclosure doing it that way. Hold the two static control points 5 km apart and adjust the traverse to it. Then check azimuth pairs at each end for errors using the adjusted traverse points. Chances are your checks will meet the expectation that way.

Even better, set the needed control doing a static network survey including the existing control.