Some Features of Finnish Earthquake Data

摘要

Some features of the Finnish earthquake catalog are investigated. A general look is taken at the data available with the help of administrative information. Suitable stochastic models to describe the sequence of occurrence times and strengths of earthquakes are tested against actual data. The gamma, two-parameter lognormal, normal, and Weibull distributions are fitted to data sets of the largest events. For complete subcatalogs, the general assumption of earthquake occurrence following a Poisson distribution seems to be justified for the data available despite the difficulty of establishing the criterion to identify dependent events. The testing procedure employed does not reject the central assumption of this model when dependent events are included in the sample, but the results improve when these events are removed. The binomial distribution may be used parallel to the Poisson distribution. The details of the testing of the stochastic models are discussed.