Corrections, Additions, and Changes
Item 1: Modification to program for calculating Br magnetic field values from equivalent source dipoles
In order to make the calculation of Br consistent with the way in which the dipoles were determined, we should only use dipoles within a spherical cap of radius 1500 km when we calculate Br at a given location. This involves adding the following statement: 'If (c.gt.1500.0) go to 57' after the statement: 'c=sqrt(e)'. The new map differs from the old map by from 2.9 nT (in the north) to -7.92 nT (in the south). This difference is less than the minimum contour intervals (-10 and 10 nT) and hence makes little difference in the resulting map. The new Br map, the old Br map, and their difference (old - new) show this clearly. However, this change probably removes the tendency for the weighted residuals to be positive at high southern latitudes and negative in the northern hemisphere and hence obviates the need to invoke an external dipole with q10 of about 4 nT as suggested in the paper.
Item 2: Correction to summation of array err in Equivalent Source inversion program
.
In the calculation of the variable err, the summation should be from 1 to m (number of observations) instead of from 1 to n (number of dipoles). To be specific, the statement 'err = sderr(n,s)' should be replaced with 'err=sderr(m,s)'. When the program is rerun, we find that this quantity continues to decrease, even after Iteration 87, in contrast to what is said in the paper. The implication of this is that we could, if we wished, fit the observed data slightly better at the expense of a rougher equivalent source solution. Such considerations are important when producing a magnetization map, which was not done in this paper. Any changes to the calculated Br map are likely to be very small, probably less than 1 nT at 200 km. Thanks to B. Langlais (IPG-Paris) for pointing out this mistake.
Last updated on Aug. 14, 2000