4.4. Data Quality Control

Many factors are involved in achieving good baseline control at a magnetic observatory. Some suggestions are presented in the following sections.

A good adopted baseline shows a low scatter of individual baseline determinations, and has small drifts and few offsets. However this is dependent on the stability of the variometer installation, how well its parameters (scale-values, orientation, temperature coefficients, etc.) are known, and other factors as well as the quality of the absolute observations. It should also be remembered that a large long-term variation in a baseline does not mean that the baseline observations were poorly made. The change could be real, the result of seasonal temperature changes or possibly pier tilt, for example. Scatter on a shorter time scale could be an indication of poor observations, but it could also indicate problems with the magnetic environment, poor temperature control, or problems with the magnetometer.

If possible, calibrate the variometer parameters in a controlled environment before installation, if this has not already been done by the manufacturer.

INTERMAGNET recommends to put effort into determining the adopted baselines, e.g. by the use of curve-fitting algorithms to fit time-dependent adopted baselines to the series of absolute observations. A statistical estimation of the quality-of-fit of the adopted baselines should be made. The adoption of the baseline should also take into account known sudden disruptive events, and estimations of the uncertainty of absolute observations (partly derived from the individual observations and corresponding variometer values, and perhaps partly derived from other factors such as an estimate of the experience of the observer).