Can hydrodynamic modelling of tsunami contribute to seismic risk assessment?

In recent years a great effort has been made to quantitatively model major historical events in Europe, searching for models that are compatible with both seismic, hydrodynamic and sedimentological evidences. The possibility to do so, with the available modelling techniques, has been checked against instrumental tsunami and seismological data. Most of this effort has been made within the projects GITEC and GITEC TWO. The quantitavive data that can be gathered from historical sources are: ext ension of the source area, wave heights, first wave polarity and, only rare ly, travel times and time intervals between main waves. With this data the accuracy of source studies is a function of both the model uncertainty (can tsunami waves be accurately modelled ? are seismic parameters compatible w ith tsunami source parameters ?) and the observation errors. These are larg e when we deal with historical documents hundreds of years old. Here we present the conclusions obtained from the study on the 1755 Lisbon earthquake and tsunami. The available historical data are reliable enough t o allow a good source determination and, surprisingly, the results obtained differ from those obtained from previous macroseismic studies. The analysis of the results obtained in the tsunami modelling, checked agai nst the macroseismic data, allow the identification of an important seismo and tsunamigenic area, close to the SW Portuguese coast, that must also be taken into consideration for the evaluation of seismic hazard. (C) 1999 Els evier Science Ltd. All rights reserved.