Simulation for frequency and precision of absolute measurements at Memambetsu and Kanoya
Ikuko FUJII, Takeshi OWADA and Yasuhiro MINAMOTO
Abstract
With respect to the absolute geomagnetic observations at Memambetsu and Kanoya, we simulated the effect of extending the observation interval from the current weekly observations to a longer interval. When using past data to evaluate observation values, there are examples where the currently used simple extrapolation of bi-weekly observations causes normal values to be judged as outliers. An attempt at replacing simple extrapolation with prediction by adding the temperatures and gradients of the sensor of the magnetometer used for continuous observation produced no significant improvements in precision.
If the interval between absolute observations is extended and the simple extrapolation method used, then the standard for re-observation needs to be relaxed. We estimated the standard under which the probability of re-observation remained at one in ten observations, and found that bi-weekly observation could reduce uncertainty in a single absolute observation to slightly less than the precision required over the long term (1 nT or 0.1 minute). At an observation frequency of once every four weeks, the re-observation standard itself will exceed the required level of precision.
However, these simulations were carried out using data from a period during which stable baseline values were obtained. If there is a disturbance in magnetic fields due to artificial factors, then an extended interval between absolute observations will make maintenance of precision difficult.
(Received 20 June 2011; received in revised form 26 August 2011; accepted 28 September 2011)
