Accuracy of the final IMAGE data
Final IMAGE magnetometer data have undergone a process during which
The quality of the final data can thus be assumed high. Generally the user can expect that
Timing has been checked.
Clear error peaks have been removed.
Magnetograms have generally been inspected.
Note: due to the increased number of requests on newest events, we have also put preliminary data from some sites on-line. This concerns stations whose data according to our experience seldom include errors. These data will also go through the final verification.
The temporal accuracy is about the same as the (original) sampling interval, presently 10 seconds (20 s until November 1992). In some cases, there are slight discrepancies in the averaging methods, so there may be apparently larger time differences than 10 s between stations. For the time definition, see the description of the IAGA format.
Accuracy of the variation field is 1 nT (0.1 nT at permanent observatories).
Baselines are usually close to "real" values. At permanent observatories they are continuously checked.
Users of the data should always be critical (especially concerning timing), and PI should be informed of suspected errors. Examples of known problems at IMAGE stations are:
Generally such problems do not disturb case studies, but they can be detected especially in statistical studies. Some of the known errors cannot be easily found during the normal checking, which is more focused on single event comparison between stations.
Occasional (extra) daily baseline variation in X at PEL. Amplitude in the worst cases is even some 10 nT (e.g. December 29, 1994). Due to this problem, a new magnetometer was installed on June 9, 2000. It seems that the problem is now absent.
Occasional man-made induction effect at HAN obviously due to other nearby equipment (STARE radar?). Technical improvements have reduced the anomalous effect but not totally removed it. The problem is not considered serious for ionospheric studies.
Single error spikes or baseline shifts. This concerns especially data of 1980's.
Spikes of some nT due to an inaccurate temperature recording.
No liability whatsoever is accepted for any loss or damage of any kind resulting from any errors or inaccuracy in the data that the IMAGE team provides.
Main data page
IMAGE home page
Updated: 12.5.2016 - AV