Source inversion of regional intensity patterns of five earthquakes from south-western Norway

An intensity-based source-inversion technique, that earlier was applied only to larger earthquakes, has been tested on the basis of regional macroseismic intensity patterns from four recent earthquakes in south-western Norway for which high-quality instrumental data have provided independent assessments of magnitudes, locations and fault-plane solutions. In spite of the modest magnitudes of the earthquakes (ML 4.0-5.2) and the asymmetry of the sampled area due to the proximity to the coast, the test was satisfactory enough in that the inversion results for three out of the four events studied came close to the independent solutions from instrumental data. In addition, we inverted the intensity data from a 1954, ML 4.5, earthquake from the same region for which independent instrumental data are lacking, and found results that are consistent with our present understanding of the seismotectonics in this region. The method is based on the inversion of a kinematic function that represents the ground motion, at a given point on the surface, from a line source. Since the problem is non-linear and possibly also bimodal, we used a sharing Niching Genetic Algorithm to perform the inversion. The inversion includes a number of source parametres, the most sensitive of which were the hypocentral coordinates and the fault-plane parametres. The parametres obtained for the south-western Norway earthquakes studied here were generally stable except for the epicentral longitude, due to the partial lack of data towards the offshore (North Sea) region to the west. The present study has shown that it is possible to extend this type of inversion of earthquake intensity data to lower magnitudes, sometimes even in cases when the sampling from the felt area is asymmetric. As such, the results achieved encourage the use of this method also for historical, pre-instrumental earthquakes also from low-seismicity regions. This opens the way for significant advances based on historical data that earlier could be used only for a broader assessment of earthquake locations and magnitudes.