Tsunamis in the World

The Fifteenth International Tsunami Symposium was held during the XX General Assembly of the International Union of Geodesy and Geophysics (IUGG) in Vienna, Austria, on August 19 and 20, 1991. The Symposium was sponsored jointly by the IUGG Tsunami Commission, International Association of Seismology and Physics of the Earth’s Interiors, and the International Association for Physical Sciences of the Ocean. The coconvenors of the Symposium were E. Bernard (United States), M. El Sabh (Canada), N. Shuto (Japan), and S. Tinti (Italy). Thirty-nine presentations on tsunami research were made by scientists from 13 countries during the 2-day Symposium on Tsunami Disaster Reduction. The sessions were organized into Observations, Physical Processes (i.e., generation, propagation, run-up), and Hazard Reduction.

The amplitude and frequency modulation observed in bottom pressure records of the 6 March 1988 Alaskan Bight tsunami are shown to be due to dispersion as predicted by linear wave theory. The simple wave model developed for comparison with the data is also consistent with an important qualitative feature of the sea floor displacement pattern which is predicted by a seismic fault plane deformation model, i.e. the existence of a western-subsidence/eastern-uplift dipole.

The Historical Tsunami Database for the Kuril-Kamchatka region has been developed at the Tsunami Laboratory of Novosibirsk Computing Center as a part of the “Integrated Tsunami Research and Information System” Project. This database covers the region within 41° to 64° N and 130° to 168° E and consists of two main parts: the earthquake database and the tsunami database, embedded inside a graphic shell that has been especially developed to provide the ability for fast search, retrieval and mapping data. The earthquake database contains the source data of almost 8000 events occurring within the region from 1737 to 1990. Source information includes date, time, epicentral coordinates, depth, magnitude, and seismic intensity followed by indexing to data sources. The tsunami database covers the same period and contains 124 events, with 109 of them having regional and 15 distant sources. The tsunami data set consists of four main blocks: detailed source data, coastal observations of tsunami wave heights, original descriptions of tsunamis and bibliography. As supporting software for data compilation and storage, DBASE-IV plus was used. For data retrieval and visualization, a special graphic shell has been developed which can be run on any IBM-compatible PC under MS-DOS. It is based on a menu-driven system with on-screen buttons for process management and on-screen windows for data input and output. The specially developed background mapping software provides the ability to display the observational data on an actual geographical basis.

Giant submarine landslides in the Storegga area on the continental slope west of Norway took place on at least three occassions during the Late Quaternary. This paper provides a summary of present knowledge regarding tsunamis generated as a result of the Storegga Slides. Most attention, however, is given to the tsunami generated by the Second Storegga Slide that took place circa 7,000 years ago. The tsunami generated by this landslide is believed to have struck most coastlines bordering the eastern North Atlantic. The paper summarises the geological evidence for the former occurrence of this tsunami. These results are compared with the results of recent mathematical modelling of the landslide and tsunami. Remarkably, there is relatively good agreement between estimates of tsunami run-up derived from the sedimentary evidence and run-up values obtained from the modelling experiments.

Tsunamis generated by volcanic eruptions are much less frequent than tsunamis produced by submarine earthquakes. In the Mediterranean Sea only 2 percent of the observed tsunamis were caused by volcanoes according to a recent study by Soloviev (1990). In Italy the percentage of events related to volcanic activity is distinctly higher than in any other country of the Mediterranean, which is expected since most of the European active volcanoes and volcanic areas may be found in southern Italy. Of the 21 cases of which some information is available, 11 were observed in the Campania coasts and are related to Vesuvius activity, 7 are due to volcanic activity in the Aeolian Islands some 50 km northwest of the Messina Straits, while the others are related to Etna and to volcanic activity in the Sicily Channel. The most significant of these tsunamis seems to be that occurred in December 1631 during the great Vesuvian eruption (16–18 December) that was observed in the entire Gulf of Naples from Ischia to Sorrento and that began with a remarkable water withdrawal of more than 1000 yards followed by waves invading the coasts deep inland. Most cases are small tsunamis observed as anomalous waves on the coasts and causing no victims. This paper is the first systematic attempt to analyse tsunamis generated by volcanic eruptions in Italy and to discuss the tsunamigenic potential of the Italian volcanic sources. Most of the attention is focussed on Vesuvius and the Campanian volcanic area, that will be thoroughly examined. A similarly deep investigation on the Aeolian Islands tsunamis is postponed to a following paper.

The Portuguese, Spanish, and Moroccan coasts are subjected to large tsunamis generated by earthquakes with epicenters located in the Azores-Gibraltar fault. The 382 and 1755 tsunamis were generated in this area. The 1755 tsunami originated large waves causing considerable damage elsewhere along the southern coast of Portugal and Spain and northern coast of Morocco. The data already available, together with data on modern offshore seismicity, are very important to assess tsunami risk and for the establishment of environmental protection policies.

The research carried out till now shows that the A.D. 382 tsunami was generated in the same area as that of 1755. The 1761 tsunami was probably generated in a different area. The generation area of the 60 B.C. tsunami is still doubtful, and the so-called 1731 tsunami, believed to have been originated on the coast of Morocco, appears not to have occurred. The research shows also that tsunamis generated in the Azores region are rare, weak and local.

An analytical model of tsunami generation by an earthquake has been built by taking into consideration some parameters of the source and the properties of the Earth’s crust. This model allows to get a solution of the problem in a simple form whose study is performed only through analytical means. The following phenomena are discussed: the tsunami height is studied in a 2D space as a function of the seismic focus depth for vertical-fault point- and line-sources and the maximum of this function is examined; the concept of “continuation” of a tsunami in the Earth’s crust is illustrated; the dependence of the tsunami height and energy on the rise time and on the duration time of the dislocation at the source is faced. Eventually some considerarions are made on the “tsunami earthquakes” and on the differences between the Mediterranean and the Pacific Ocean tsunamis.

A method to reduce the noises in a OBS raw data by using the band-pass filter is presented. Tsunami records extracted from the raw data obtained in the deep ocean are compared with the computed results of the linear long wave theory. Although the computed and measured tsunami give the same maximum height of the order of 1cm, there are many differences between the two. Transformation of the tsunami on the continental slope determined from computed and observed data indicates that computed wave paths or directivity may be different from the observed one and that the physical dispersion is needed in the simulation to reproduce tsunami traveling along trenches and ridges.

The influence of the local bathymetry and the shape of the contour in the resonant response of the Gulf of California (Mexico) to incident tsunami waves of various approach directions and frequencies is analyzed.

We conduct indoor experiments of the oblique reflection of solitary waves, paying special attention to the 2-dimensional features of the wave crests, examined by measuring 2-dimensional water surface displacements. The critical angle of incidence between regular and Mach reflection is found to be about 50 degrees. Our experiment also suggests that the difference between the angles of incidence and reflection depends on the incident wave height. The growth rate of the stem wave depends on both the amplitude and the angle of incidence.

Present practice to analyze tsunami forces on coastal structures is generally based on static loading analysis. This compares maximum predicted forces with the resistance to overturning or sliding, but does not account for the fact that some structural motion could occur without failure. While applied forces at some threshold level will give incipient motion, actual motion will be determined by the force exceeding the threshold level and the related acceleration. Three different levels of forces must then be considered: a force that will not exceed the threshold causing incipient motion; a magnitude and duration which will cause motion but not failure; and a magnitude and duration which will cause structural failure. It is also necessary to investigate the axis of rotation for structural failure. Once motion is initiated there would be foundation failure under the toe, moving the axis of rotation inward under the structure. This increases the acceleration and causes increased motion.

Assessment of tsunami hazard is a fundamental element of natural hazards evaluation and should be routinely performed for all coastal areas known to be exposed to tsunami attacks. Italy has been invested by tsunamis, mostly generated by local earthquakes with focal region close to the coasts. In previous preliminary investigations by the author the most hazardous Italian area was shown to be that embracing eastern Sicily and Calabria in South Italy. This paper addresses again the problem of the tsunami hazard evaluation in this region: previous calculations are improved, since new estimates are computed by considering a more complete basic data set, by using a slightly more sophisticated statistical approach and by introducing some account for the efficiency of the tsunami generation mechanism in the evaluation procedure. It is confirmed that this region has a substantial potential of being hit by local tsunamis: the zone with the highest expected rate is shown to be the Messina Straits, i.e. the narrow channel separating Calabria from Sicily, where the greatest event ever occurred in the whole region was observed, following the 28 December 1908 Messina earthquake.

The seismotectonic characteristics of the Anatolian fault zones are investigated, and the discontinuous crack propagations along these zones are discussed by examining macro earthquake activities and the historical earthquake data. Some of the tsunami observations in the surrounding seas and lakes of Anatolia and their relations to the fault systems are discussed.

We have developed an integrated system to estimate tsunami risk quantitatively. It relies merely on a broad-band long-period three components seismic station linked to a personnal computer. The algorithm includes automatic detection of P,S, Rayleigh and Love waves and location of the epicenter. The latter is estimated exclusively from long period data: epicentral distance is obtained from S — P delay times and source azimuth from the polarity of P waves in the horizontal plane. The seismic moment is calculated through the mantle magnitude Mm computed from spectral amplitudes of Rayleigh and Love waves over a broad range of periods (50 to 300 s.). The seismic moment is then used to compute an expected tsunami height, taking into account corrections due to epicentral distance. It can also be ultimately corrected after previous study for site effects such as run-up, resonance of bays, etc.

Both the mantle magnitude Mm and the relationship between seismic moment and tsunami amplitude on the high seas are fully justified on theoretical bases, and have been verified experimentally on extensive datasets. This system has been fully operational and running satisfactorily since 1987 at the Polynesian Tsunami Warning center (CPPT). It can be used either in far-field or in near-field, at distances as small as 1.5°.

A tsunami intensity is defined as the logarithm of the local tsunami height to the base two. In terms of this tsunami intensity, important tsunami aspects are classified such as tsunami profiles near the shoreline, damage to individual houses made of wood, stone or reinforced concrete, damage percentage of wooden houses in a flooded area, damage percentage of fishing boats, damage to aquaculture rafts and effectiveness of tsunami control forests, based upon data from old documents of the past tsunamis.

Gorringe Bank, 37 °N, 11 °W, in the Atlantic Ocean, SW Portugal, is an important tsunamigenic region, as illustrated by the tsunami following the 1 ^st November 1755 earthquake.