ERROR CAUSES
There seems to be a misconception that modern methods utilising various adaptions of GNSS have all but eliminated errors from the acquisition of positioning data. This is false. While certain types of errors are almost eliminated with the use of GNSS, others have been introduced, or become more prominent. Traditional survey methods using theodolites were particularly prone to operator error, especially where conditions dictated that many short setups were used. Using GNSS is much less prone to operator error, and even when conventional surveying is still required, such as in jungles, modern total stations reduce many of the possible errors through data recording and electronic distance measurement. However, at the same time, some errors remain the same while others have become common due to:
- Far more data must now be processed with less personnel.
- Positioning personnel are less qualified than in the past. It was once common to have more than one degree-qualified surveyor on a land crew. Now, with contract prices so low and production so high, most positioning personnel are low paid and rarely have formal qualifications. Many are trained on the crew, and only understand the basics of how to acquire and/or process the data using “black box” technology. In conventional surveying, with errors so easy to make, it was common practice to always assume that you had made errors and only accept measurements once it was proven that you hadn’t. Redundancy, and many checks and balances were built into the procedures to trap errors. Now, many positioning personnel simply assume that the technology will prevent them from making errors. With no (or little) error detection built into their procedures, they often don’t know when they have made an error.
- Some newer systems that automate positioning (especially on vibrators) are not sufficiently understood by the people that operate them. While these systems have quality control information in their data that can warn of unreliable measurements, it is sometimes not passed on by the intermediate processing software, or it is ignored. GNSS sometimes produces wrong measurements, usually due to signal or correction loss. You must be aware of how to detect them and continuously monitor for them.
- Some highly automated systems, especially in shallow water positioning, are known to occasionally either over-smooth positions and thus artificially straighten bends or not filter raw data correctly and allow spurious measurements to affect the final positions.
- With larger and more complex surveys (especially TZ), there is more manual editing of data to patch together measurements from different systems, and to inject revised data from moved or remeasured points. Every time a manual edit is made, there is a potential for an error. When these are performed on a dataset which has already been through existing quality control, who is going to discover the error?
- Then of course, some problems never go away. Drillers sometimes don’t drill on the flag and forget to tell anyone. Nodes and cables in shallow water can move. There are many more examples.
On the next page we discuss error Detection Methods.