Hints For Locating Blunders
|Top Previous Next|
The blunder detection routines often have several good candidates. This can make it difficult to isolate the specific measurement that has caused the blunder. In these situations, it helps to have more insight into the finding process. Here are some of the most important issues:
1. Blunder Sensitivity. Blunder location works best where the quality of the rest of the loop is high. On the other hand, if the survey is sloppy, the problem will be difficult to pinpoint. This is because the normal random errors in the loop overwhelm the effects of the blunder. From my experiments, if the loop has a standard deviation of around one, a 10-foot blunder is required before the problem will be visible. The following table shows the sensitivity level vs. error level:
Unblundered Length Azimuth Inclination
Error Level Sensitivity Sensitivity Sensitivity
0 - .3 STD 1 Foot 1 Degree 1 Degree
.3 - 1 STD 5 Feet 5 Degrees 5 Degrees
> 1 STD >10 Feet >10 Degrees >10 Degrees
2. Large Blunders vs. Small Blunders. Large blunders are usually easier to detect than small ones. However, this can be tricky. For example, a 180-degree error may seem large, but if it is on .5 foot shot, the resulting error is less than a foot.
3. Multiple Blunders. If you have two or more blunders in a loop, individual blunders will be very difficult to detect. Two blunders on the same shot are almost impossible to isolate. If there are two blunders on different shots, one blunder tends to dominate and obscure the other. This sometimes make it possible to fix one and then locate the other.
4. Unique Signatures. The more unique the blunder is, the more likely it will be isolated. For example, let’s say you have a blunder on a shot with a compass reading of 180 degrees. In addition, let’s say that there are no other shots that come to within 45 degrees of the blunder. This blunder will be so unique that it can be isolated even in a relatively low quality loop. On the other hand, if you have many shots with a similar azimuth, random errors in the rest of the loop can cause the program to select several shot as the possible blunder shot. If you see several candidates with the similar "Change" values and similar improvement ratios, it may indicate that the blunder is not unique enough to be isolated. You can, however, narrow it down to a few candidates.
5. Blunder location works better on small loops. As the loop grows longer, random errors make a larger contribution to the error. Eventually, the accumulation of random errors becomes large enough so that the effect of the blunder is lost.