In Part I of this article, a new method of earthquake source location and velocity structure determination was formulated based on the Bayes theory. In Part II, stability and uncertainty analyses on synthetic data are carried out.
The synthetic tests show that, with the type of arrival-time data found in Western Norway, the following can be resolved: the P and S velocities of the second and fourth layers (Moho is the third interface), the dip of the third interface, and the source parameters for the earthquakes. For these parameters initial estimates can be in error of 10 per cent of velocities, 100 per cent of dips, 10 km for hypocenters, and 10 sec for origin times. The remaining parameters had to be fixed within 0.3 per cent.
Using arrival times from 23 earthquakes in the Northern North sea, an inversion was made with constraints as found necessary from the synthetic test.
The results were as follows: the Mohorovičoč discontinuity has an eastward dip of 3.3 ± 0.6 per cent in the region between the latitude of 59° and 61° and 1.1 ± 0.5 per cent in the region between 59° and 62° north. The Moho dips northward about 1 per cent in the southern region. The upper mantle velocities are 7.94 ± 0.05 km/sec in the southern and 8.05 ± 0.03 km/sec in the northern region. In the source layer (layer 2), the velocity is 6.62 ± 0.03 km/sec and 6.55 ± 0.02 km/sec in the southern and northern regions, respectively. The epicenters moved (in average) 2 km and 5 km to the west and north, respectively, and 2 km toward shallower depths, as a result of the inversion. The average change of origin time was 0.5 sec toward earlier occurrence.