Let’s Get Well Adjusted

By Doug Aaberg • December 4th, 2018

It’s December!

Over this past year, my blog has focused on some new products, underutilized software routines, and various other tips and tricks. For this last blog of the year, I am going to circle all the way back to a rudimentary survey issue: Traverse Adjustments. While this article is lengthy, I think you’ll find it provides an easy-to-follow outline of how to perform quality checks and even make traverse adjustments without modifying or adding a lot of additional steps to your current field procedures.

In working with surveyors over the past couple of years, I have discovered that many of them simply do not bother adjusting their field data. Their reasons are not unfounded. With the equipment they use and by following good field procedures, they are performing field checks that depict results within a couple of hundredths of a foot, so any adjustments would produce minimal changes to the raw data.

The other deterrence can be that some sites are not conducive to performing a survey in a manner that is consistent for a traditional compass rule adjustment and/or a lack of understanding of a least squares adjustment. In short, it’s just not worth the effort. Not performing any adjustment at all is something that can cause a very heated debate amongst surveyors. Most state boards and surveyors’ associations endorse the practice of adjusting errors out of any survey. And certainly if you sign an ALTA Survey using the current regulations, you are implying that you have done that very thing.

Let’s step through what I have found to be a very typical type of survey that combines traditional EDM measurements combined with GPS – RTK measurements. Even if this does not match up with your typical survey, I am hoping you will still find some useful information contained herein.

The above diagram shows a typical ground survey performed with an EDM and 4 GPS check points gathered RTK to produce grid coordinates. The vector data was not stored. The field process was to locate control points to establish grid coordinates, set up a total station on one of those GPS points, back sight another, and perform the survey. Then, throughout the survey, perform additional checks on EDM traverse points as well as the other GPS points by locating them using a different point number as they went along. The surveyor would then perform a quick quality check by simply inversing between the two point numbers representing a control point. As long as the result was within reason, the surveyor was “good to go.”

Following that exact work flow actually produces enough data to perform a compass rule adjustment or more preferably, a least squares adjustment.

A very important fact about a survey like this must be understood. GPS-RTK produced coordinates will most likely never yield a measured distance that is as precise as that of a traditional EDM. There will always be some error.

For the purpose of this example, I am not addressing grid to ground distortion, which is another very important part of using GPS for surveys.

1.1.1 Compass Rule Adjustment

1. Let’s begin by examining the field data using Carlson’s raw data editor.

The PT or point records are treated as control points for the survey. They were produced from the GPS portion of the survey. Point number 1 is just a copy of point number 51.

The surveyor then occupied point number 1 and took a back sight on point 52, located it again as point number 2 and moved on with the survey. You will notice the error of 0.032’ between the EDM located point 52 and the GPS coordinate. That is not at all unusual. It does not negate the GPS points; they are just less precise than their EDM counter parts.

The check points located during the survey were uniquely coded as a CHK for a description. For example point 15, which was a check point for point number 5, was coded CHK5. We can make use of those.

2. From the Process (Compute Pts) pulldown menu, select Process Settings

Notice the option to add a user defined Check Point Code. The program will recognize anything entered here preceeding a point number as a Check Point.

Example: Point 3 EDM is also point 53 GPS.

To examine the precision of the field work, from the pulldown menu, select Process No Adjust.

The Process Options dialog box appears.

For the first time through, I am not selecting the option to check a closure result.

3. Click

The program produces a processed data report and at the bottom of that report, the residuals between check points are displayed. The residuals between 5 and 15 and 1 and 11 are much less because they are all EDM located points. At the very least, this is a report that can be filed away that indicates the precision of the survey that includes the GPS points.

4. From the Tools pulldown menu, select Direct-Reverse Report

This report shows all of the distance and angle means for any points located via the set collection method, i.e. doubled angles. It also shows any errors such as distance or angles that exceed the tolerances set in the process settings.
This survey, although not ideal, was performed in a manner that allows for a compass rule adjustment. The surveyor traversed through the control points and took a check shot on the beginning point.

5. From the Process (Compute Pts) pulldown menu, select Compass

6. Enter the closing point number as point number 1.

In this case, an Angle Balance cannot be applied. However, a compass rule adjustment can still be performed.

The program scans through the raw data and finds the traverse legs and places them in order. I always use the option Prompt so that I can check the order of the points to make sure none are neglected or included that shouldn’t be.

A diagram of the traverse and side shots appear with the legs to be adjusted highlighted in red.

7. Click <OK>

The compass rule adjustment report is displayed with the same check points displayed at the end of the report.

1.1.2 SurvNET – Least Squares Adjustments

Of course, the limit to a compass rule adjustment is that the second day of field work outside that shown above has to be processed individually. In addition, the GPS points are held as a static location with no real value to the survey other than establishing approximate grid coordinates. Also, a more linear collection of the field data needs to performed to get the adjustment to work and all sideshots are only adjusted by the value of the control points by which they were located. A least squares adjustment allows the field work to be performed more similar to this example, which is somewhat of a “rats nest” approach to collecting the data. You can also “weight” measurements or control points to better control the adjustment process.

1. The program SurvNET can be launched directly from Carlson’s main survey menu or out of the raw data editor.

2. Create a new project

The program prompts for you to name the SurvNET adjustment project file. Once you save the file, the SurvNET Settings dialog box appears with several tabs. This may look intimidating at first but in actuality, it is very straightforward and will normally only have to be set up once depending on how many different types of instruments you are using.

Coordinate System

This tab allows you to set a projection if needed as well as user defined scale factors and the Geoid to be used. The coordinate system Adjustment Model is 2D, 1D for traditional surveys. The 3D model option is when using GPS vector data.

Input Files

One of the main reasons why I prefer to process data in SurvNET is because I can process all of the raw data files at once. As illustrated above, both the previously explored raw file as well as the second day of work is added to the program at once. Note how you can combine a traditional ground survey with a GPS vector file. This will allow you to actually perform post processing GNSS surveys as well as EDM surveys and level data as part of a total project.


The defaults for these are generally acceptable but I will come back to the Pt Number Substitution String later.

Standard Errors

The measurement errors should be consistent with the instruments you are using. Note the option to save and load these settings. I believe targeting and setup errors should be reasonable based on your survey practices and filed conditions. I personally am not a fan of tweaking those types of settings in order to get passing results. The Coordinate Standard Errors is however applicable to this survey.


Note the ability to create a specific ALTA report.

Output Options

I prefer to not output to a coordinate file until I’m satisfied with the results. Note also the option to create a scaled ground coordinate file for grid to ground conversions.

Once all settings have been made, you can view the survey data.

The graphics are user controlled. Shown above, the control points are green squares while the traverse lines are displayed in purple. The side shots are turned off.

One of the routine things I do with a survey is to perform a blunder detection to look for an errors.

3. From the Process pulldown menu, select Blunder Detection

The program creates a report showing measurement errors that fall outside the tolerances as well as a proximity report that shows any points horizontally located within a preset tolerance. Note the proximity of the GPS points and several other points that are actually just topographic points that fall near each other.

Let’s process the data without using the GPS points and examine the outcome.

4. From the Process pulldown menu, select Network Adjustment.

The program processes the data and displays the results in a single dialog box with several tabs.

This dialog box contains a vast amount of data to examine. Those familiar with least squares will be used to seeing the values displayed. I am going to focus on just a couple values.

5. Select the Adjusted Obs. Tab and scroll to the bottom.

You will see that the survey passes the Chi-Square test, however there is only a couple degrees of freedom. That means there is little redundancy. As a matter of fact, the only redundancy is from the doubling of angles. At this point, the results would be very similar to that of the process without adjustment shown above, but with the added value of being able to process multiple days of field work.

To add more redundancy, we can make use of the check points in SurvNET as well.

6. Back in the Project Settings menu, select the Preprocessing tab and add CHK to the Pt. Number substitution String.

7. Process the network again.

This increases the number of degrees of freedom but the Chi-Square test now fails. Recall that the GPS points are now included in the network and the tolerances as measure in the field were only within .02’ – .05’, while the EDM measured points fell within the coordinate standard error of 0.01’ horizontally and 0.02’ vertically. We could eliminate the GPS points from the survey, which will produce essentially the same results as the compass rule adjustment. Or, at least as an exercise, we can apply a different standard error to the GPS points.

8. From the Tools pulldown menu, select Edit Raw Files

9. Select the Day 1.RW5 file

10. Place your cursor at the line above the list of control points and from the Add pulldown menu, select Control Standard Error.

The program first introduces a CSE line containing all * symbols. This indicates that it is using the default control standard error.

11. Select each field, northing, easting, and elevation and enter the desired control standard error.
In this case I used 0.03’ horizontal and 0.10’ vertically for the GPS points.

12. Repeat the process after the GPS points and return the control standard errors back to the defaults.

13. Save the file and reprocess the data

This time the Chi-Square test passes. Note the standard errors for the adjusted coordinates.
Now before I get too many emails from surveyors who are very knowledgeable in least squares adjustments, I am not necessarily recommending that you include GPS points into a survey this size. It is very small and including points that you know are well outside your normal tolerance may not make sense. However, on larger surveys, the GPS-RTK points may very well be used as part of the adjustment. And if you add enough redundancy to the survey, you should get very good results with a lot more control. In the above example, the end results for the EDM points were within 0.002’ of their relative locations as that of the compass rule adjustment, but the GPS points now fell within 0.02’ of their EDM locations.

My hope is to prod you a bit into adjusting your surveys and experiment with the programs in Carlson that you likely already own. Start with a survey that you are very familiar with and then perhaps compare the results of compass rule adjustment vs least squares. I feel that once you get used to these programs and get traverse adjustments back into your normal work flow you will be very glad you did.

Download Doug's Field-to-Finish Guide

Register for the 2019 Carlson User Conference


Let’s Grow Together.

Douglas L. Aaberg, PLS
Survey Product Manager

« Previous | Home | Next »