Intensity assignments from historical earthquake data: issues of certainty and quality

Abstract

The use of macroseismic data in assessing parameters for historical earthquakes for use in seismic hazard assessment has thrown more attention on the way in which these data are treated. The processes involved in selecting which macroseismic data from a historical earthquake survive to the present day can be modelled as a series of filters, most of which are outside the control of the seismologist/historian, and which cause distortion in the resulting picture of the earthquake. The ways in which the data become distorted should be taken into account when interpreting the data as intensity values. One can usefully discriminate between the certainty of an intensity assignment (how well the data fits the scale) and the quality of an intensity assignment (how well one can trust that the value is a true reflection of what really happened). The expression of uncertainty is usually in the form of ranged intensity values; the expression of quality requires an extra symbol or rating of some sort. A system is presented for three types of quality problems: reliability of intensity assessment, locational certainty or uncertainty, and veracity of the original data. Each of these is treated as a binary variable, giving a final quality code ranging from 0 (best) to 7 (worst). This single integer quality code preserves three types of information which can then be expanded as required by computer programs designed to handle macroseismic data.