Task B - Probabilistic analysis of macroseismic data to forecast damage scenarios

Abstract

DESCRIPTION OF TASK B The Task B “Probabilistic analysis of macroseismic data to forecast damage scenarios” includes the following seven Actions: Action B.1 Setting-up of aWorking Group on the macroseismic database (INGV); Action B.2 Exploratory analysis of macroseismic fields for their clustering in classes of attenuation decay (CNR-IMATI); Action B.3 Estimation of the probability distribution of the site-intensity Is , given epicentral intensity and distance from epicentre (isotropic case) (CNR-IMATI); Action B.4 Exploratory analysis of observed and synthetic macroseismic fields for their clustering in classes of anisotropic attenuation decay (CNR-IMATI); Action B.5 Estimation of the probability distribution of the site-intensity Is , given epicentral intensity and distance from epicentre (anisotropic case) (CNR-IMATI); Action B.6 Application of scoring rules (CNR-IMATI); Action B.7 Probabilistic analysis of macroseismic data using the logistic model (INGV) This report is basically divided into two parts. The former (Chapter I) deals with the methodological aspects of the Actions B2 - B6; it describes a clustering procedure to select classes of macroseismic fields with similar attenuation trend within a database, it proposes a probability model and an estimation method to evaluate the distribution of the intensity at site in both isotropic and anisotropic cases, and finally it presents three validation criteria by which to compare observed and estimated intensity at sites. The latter part of the report (Chapters II - V) focuses on the results obtained and provides solutions to specific issues arosen in applying the above-mentioned methods to the test areas considered in the project (Mt. Etna (Italy), Portugal and Azores Islands, Alicante-Murcia (Spain), Iceland). In addition to the collection of the data related to the test areas, Action B1 has also dealt with the digitalization in a GIS of isoseismal images drawn from the literature, and related to two historical earthquakes occurred in North Iceland (Chapter V). Action B7 was planned to compare the beta-binomial probability model analyzed in the project with the other probability model (logistic model) for the seismic attenuation given in the literature. The discussions had with the partners during the course of the project led us to prefer the comparison with the deterministic attenuation relationships currently used in the hazard evaluation of the test areas. The results of these comparisons are illustrated in Deliverable B6.1.