Surveyors often ask, "What are the benefits of network design for a GPS survey, given that the accuracies of individual GPS baselines are a function of satellite geometry and not survey network geometry." The short answer is: "Network design helps to provide a measure of confidence in your future survey." That measure of confidence is a function of your network design.

The purpose of network design is to estimate the confidence of your future survey, before you enter the field. Network design allows you to experiment with different variables so as to meet or exceed the stated survey accuracy requirements.

Design Variables

The number and physical location of survey points.
The number and types of observations to be measured.
The observation standard deviations (standard errors) you expect to achieve in the field.

Altering any one of these variables will change the estimated confidence of your survey. Network design allows you to perform "what-if" analysis on these variables so that you can estimate how you will do in the field.

Network Design Goals

Perform each survey in a cost effective way. If a survey can be performed with fewer points on the ground, while still meeting accuracy requirements, wouldn't it be beneficial? Further, if you could select locations on the ground that were easy to gain access to and make observations from, wouldn't that be beneficial?
Determination of the field procedures and equipment needed to achieve accuracy requirements. This could be something as simple as using a more accurate total station, or perhaps changing your field procedures a bit to achieve better accuracy (for example, making terrestrial measurements during the cooler times of day, better instrument/target setups, making additional measurements, and so on).
Determination of whether you should take on the project. Based on the accuracy requirements, you may decide that given the nature of your equipment and/or crew, you may not be able to meet the requirements and therefore should pass on the survey.
Quick completion of the design. The network design process should require significantly less time than the survey itself; otherwise, the design process may not be worth the effort. For medium-sized projects, an hour or two of "what-if" analysis may be all that is required.

Some Thoughts On The Design Process

Network design allows you to achieve the first three goals above by providing you with estimates of the accuracy that will be achieved given the input observation types, their standard deviations and station locations in the survey. After an initial design, you may discover that the accuracy estimated will not meet the survey requirements. Using an iterative process of changing out the variables mentioned above, you may find a way to satisfy the accuracy requirements.

Before bidding on a new project, you might initially set up an elaborate design with many different observation types built in. After running the design and satisfying the confidence requirements, you might then scale back the network with fewer stations and observations. After running the design again, you may happily discover that you are still within the accuracy requirements of the project, but now the project will cost less to perform.

Next, you might consider using only GPS for the project. However, after running your proposed network through the design process you might discover that a problem has emerged that cannot be fixed through GPS alone. In fact, you may need to add terrestrial observations for some portion of the project in order to stay within accuracy requirements. This might occur in an area in which you have poor satellite visibility or in an area in which the points you need to establish are only a few hundred meters apart. Perhaps only the terrestrial equipment can give you the accuracy you need in these areas.

After the design is completed, you will have created a blueprint (of sorts) for the field crew. That blueprint will tell them roughly where to locate the stations, the types of observations to measure at each station, and the level of accuracy needed for those observations. You could conceivably use GPS in one section of the project, a 10-second total station in another section, and a 1- to 2-second total station in yet another section or the project. Through the use of network design, you can determine how the survey should proceed.

Of course, the most important element to design is achieving "in the field" what you designed in the office. If you are unable to measure angles to +/- 5 seconds or measure distances to +/- 0.004 meters (like you specified in your design), then your project will probably not meet the expectations derived from design. Bottom line: Don't be overly optimistic about what you can achieve in the field.

Having said all this, there is no substitution for experience and intuition from prior projects.

Quick 3D geodetic network design (pre-analysis) setup

To quickly create a 3D geodetic network for use in Pre-Analysis, do the following:

Run COLUMBUS.
From the File menu, select New.
From the Options menu, select Datums and select the datum to be active.
From the Data menu, select Stations.
Select the Geodetic grid and enter the approximate geodetic coordinates for each design station.
Click Keep and Exit.
Right click on a station, then right click on the station to which you want it connected; a line will appear between them. Right click on as many additional stations as are applicable to the current observation types you plan to set up in Step 8 below.
Click the Setup Design Leg toolbar button and select all the observation types you plan to measure between these stations. Edit the Standard Deviations for each observation type if needed, then click the OK button. To clear a route, click the Cancel button, then from the View menu, select Clear Current COGO/Design Route.
Repeat Steps 7 and 8 until all observations are defined.
Select the stations which are to be part of the current pre-analysis by either selecting Network Stations from the Network menu or by clicking the Select Network Stations toolbar button. You can select all stations or a subset of stations.
Select the stations which are to be fixed in 1D, 2D or 3D by either selecting Fixed Stations from the Network menu or by clicking the Select Fixed Network Stations toolbar button.
Select the applicable observations which are to be included by either selecting Observations from the Network menu or by clicking the Select Network Observations toolbar button.
Note: Any observations linked to stations which were not selected in Step 10 will automatically be unavailable to select. This ensures you do not have to filter them out manually.
From the Network menu, select Pre-Analysis. After the pre-analysis is finished, a summary View displays.
From the Results menu, select any of the results views. Each of these views shows different pre-analysis results which you can use to isolate any potential problems in your design (e.g., Distance Errors, ALTA Positional Uncertainty, etc.).

Hint: Before exiting COLUMBUS, save the network to a COLUMBUS data file for future use.

One approach for establishing geodetic coordinates for each station

With COLUMBUS, setting up a theoretical network for Pre-Analysis is relatively easy. The only hurdle is defining the approximate geodetic coordinate for each station in the design. However, with a little forethought, even this task can be automated to some degree. Below is just one example of how it can be done.

Most of us probably have access to various topo map software. Usually, these packages allow you to view USGS topography maps from anywhere in the country (depending on the zones you have CD maps for).

These packages allow you export coordinates to an ASCII (Text) type file. One such package, called "Terrain Navigator", allows you to defined points on the topo map (on screen) and export them to a Lat (decimal degree format), Long and Station Name format. This file type can be directly imported into COLUMBUS by following the steps outlined by the topic, Importing coordinate data from a comma-delimited text file, and setting the import format for Lat/Long to be "DD.ddddddd" format. Some packages, including this one, may even let you export the elevation.

If the package does not export Lat and Long, but rather UTM, State Plane or a Local N,E,Up coordinate system, you can import those coordinates and transform them to geodetic within COLUMBUS.

Once you have your geodetic coordinates loaded/defined, it is just a matter of pointing and clicking to finish the network design.

1D vertical network design (pre-analysis)

With COLUMBUS you can also perform network design for 1D vertical networks. Here is how to do it:

Follow the steps in the Quick 3D geodetic network design (pre-analysis) setup, with the following exceptions.

In Step 5, leave the coordinates set to zero.
In Step 8, select only height difference observations.
In Step 11, select at least one (or more) stations to be fixed in 3D and all remaining stations to be fixed in 2D (so the 2D ones can float in 1D).
In Step 13, you will be prompted with the message that two or more stations have the same coordinates (because all the geodetic stations you provided have zero for coordinates). Click YES to continue processing.