Though not as large as many of the historical tsunamis to affect Peru, the February 21, 1996, Chimbote tsunami is important for the following three reasons. First, this tsunami was the first in Peru's history subjected to an extensive post-tsunami field survey by members of the international tsunami research community, including several from USC (For more details regarding this field survey, please refer to the reference for Bourgeois et al, 1996, under the references section). Secondly, this event is interesting because it occurred off of Peru's north central coast, north of 10º S latitude. In these latitudes the Peru-Chile subduction zone is relatively quite, and does not produce the large tsunamigenic earthquakes common to the south. Thirdly, the tsunami was disproportionally large (runup in excess of 5 meters at some locations) with respect to the magnitude of the source earthquake. These three peculiarities make this tsunami an interesting case study.

THE EARTHQUAKE

The earthquake that generated the tsunami (estimated moment magnitude 7.5) did little if any damage to the mainland, where most people felt only a moderate shaking. For this reason, most Peruvians were not alarmed by the quake, and continued about their business as usual. According to Peruvian Civil Defense the citizens of Chimbote (located closest to the epicenter) evacuated the waterfront area during the first withdrawal of water from the shore. At the smaller communities along Peru's coast however, it is unclear whether people recognized the warning signs of the impending tsunami, or even knew to look for them at all. A common misconception is that only a very large earthquake can produce a tsunami, and most people associate the size of an earthquake with the amount of shaking felt on the land. This might be true some of the time, but it is a big mistake to stake ones life on it.

The efficiency of tsunami generation by crustal uplift during an earthquake is controlled by the speed at which the rupture (the actual break in the fault) propagates down the fault. Research indicates that earthquakes with slow rupture velocities are the most efficient tsunami generators, and earthquakes with the slowest rupture velocities are actually called tsunami earthquakes. Due to their tsunamigenic efficiency, a relatively low magnitude tsunami earthquake is capable of generating a very large tsunami. This fact makes tsunami earthquakes particularly dangerous to coastal communities located near areas of subduction zones prone to earthquake of this type.

The rupture velocity of the 1996 quake was classified as moderately slow (Newman and Okal, 1996). Not quite as slow as that of a tsunami earthquake, but slow enough to generate a disproportionately large tsunami with respect to the surface waves felt on land.

THE TSUNAMI

According to eyewitnesses, the tsunami's arrival at all locations was marked by an initial withdrawal of water, followed by a series of crests. Most people remember two or three crests occurring in quick succession, the second of which was the largest. Eyewitness stated that the actual wave crests did not break as they arrived at the shore. Runup measurements, high water marks, and other data collected during the field survey support eyewitness observations, indicating that the tsunami struck the coast as a steady flow of water, rather than a violent bore.

Plots of regional and local tsunami runup distributions are presented in the figures above. The greatest runup value of 5.14 meters occurred at the port of Chimbote, located on the north side of Chimbote Bay. At this location the tsunami inundated a pier, beached several small boats, and destroyed a few smaller structures; but caused little other damage. Runup also exceed 5 meters at Coishco, where the tsunami damaged several waterfront structures. At Samanco Bay, immediately south of Chimbote, the tsunami transported boats as far as 300 meters inland. The most dramatic runup effect occurred at the tomblo (low strip of land connecting a headland to the mainland) separating Chimbote and Samanco Bays, where the tsunami inundated the entire 1.5-kilometer wide by 4.5-kilometer long tombolo. In all, the tsunami destroyed or damaged several houses and boats, caused some damaged to port and harbor facilities, ruined crops, and damaged the beachfront property of several resorts in the area.

A total of 12 people were killed, and 57 injured by the tsunami. The dead included 6 fishermen near Coishco, 4 people gathering firewood near the mouth of the Santa River, and 2 children looking for gold on the beach at Cameo Santa. All tsunami fatalities occurred in remote areas, where it is likely that people were unaware of the impending danger until it was too late. In more populated areas, such as Chimbote, people recognized the signs of the approaching tsunami, and were able to evacuate the coast in time.

The relatively light damage and death toll resulting from the tsunami can also be attributed to the presence of a heavy storm swell affecting the Peruvian coast at the time of the tsunami. As a matter of precaution against the storm swell, officials had closed the regional ports for five day prior to the tsunami, and evacuated several larger ships to deeper water. Thus, by pure luck, waterfront activity was at a minimum when the tsunami arrived on the morning of February 21.

A WORD OF WARNING

Though unusual, the 1996 tsunami was not unique. A similar tsunami struck Peru in 1960, with runup reportedly reaching 9 meters at some locations. Though this maximum runup value is questionable, the tsunami did kill three people, and cause damage to various coastal communities in north-central Peru. Interestingly, the earthquake that generated the 1960 tsunami was very similar in magnitude and rupture speed to that which generated the 1996 tsunami. Could it be that the subduction zone north of 10º south latitude typically generates slow, earthquakes that produce large tsunamis? The data suggest that this is possible, but the data are sparse, and will not support such a broad conclusion at this time. Regardless, the residents of northern and central Peru should heed the warning sent by the 1960 and 1996 tsunamis, that the correlation between felt shaking, and tsunami height is sometimes not very direct.