Page 28 - Marine Magnetometer Processing

7. Further Processing Introduction Basic processing of magnetometer data involves identifying anomalies, plotting the positions of targets from those anomalies and estimating the mass of iron in the target. Further processing of the magnetometer data can be done to get better results or extract more information. This section discusses how the anomalies we want can be identified within a number of different noise sources that affect magnetometer surveys. Dealing with Spikes and Dropout The ability to detect the smallest magnetic anomalies is usually limited by noise in the measurements. Some noise can be removed by filtering but it is always better to minimise the noise recorded during the survey rather than having to deal with it in post-processing. Spikes in the data are short duration jumps in the measurements that are considerably bigger or smaller in amplitude than the background magnetic field value. Spikes may be caused by a number of sources including electrical noise and excessive motion of the magnetometer towfish. Dropout also causes spikes in the data and this is caused by intermittent communications with the instrument (Fig. 24). Removing the spikes, or ‘despiking’ can be done by removing any measurement larger or smaller than some given value. This can only be used reliably when the spikes are much bigger in value than the change in magnetic field across the site. A smarter method involves applying a gate or window around the average of the last few measurements and rejecting any subsequent measurements outside this ‘gate’. Here the rejection value ‘tracks’ the average value of the measurements so a narrower rejection range can be used. Care is needed in ensuring that despiking does not actually remove useful anomalies. Towfish Turns Magnetometer data is likely to be unreliable when the vessel, and thus the towfish, is turning sharply. The measurements from the instrument may increase or decrease sharply as the towfish turns but resume the same background level when the track straightens out (Fig. 25). In some cases the data gets very noisy during the turn with large spikes being recorded. The problem may be caused by wobble in the towfish as it turns because the data is usually affected less for wider turns, whilst rougher sea states can make the problem worse. Care is needed when processing data showing these symptoms as any anomalies seen in the turn may be caused by noise and should be discarded. Figure 24: Time series plot showing dropout Figure 25: Chart plot showing change of magnetic signal during turns Marine Magnetometer Processing \[27\] © 3H Consulting Ltd