A new procedure modeling the probability distribution of earthquake size

The probability distribution of earthquake size is needed as input data for some earthquake analyses. A common procedure is to calibrate the so-called b-value in the Gutenberg-Richter relationship and to use it as the best-estimate model parameter in an algorithm to simulate the observed earthquake-size distribution. This paper introduces a new procedure for such a simulation, on the basis of performing optimization to search for the optimum model parameter. The new option and an existing method are then both utilized to model the earthquake-size distribution around Taiwan since 1978. Owing to the nature and the power of optimization, three case studies presented in this paper all indicate that the new optimization procedure can indeed improve such a simulation over the existing procedure. Moreover, with a proper tool such as Excel Solver, practicing the new method to model the observed earthquake-size distribution is as effortless as using the existing procedure.