Tsunamis: Forewarned is Forearmed
R. Gunasekera
“The high hills and forests shook under the immortal feet of the descending god, the earthshaker, POSEIDON. Dressed in gold, he picked up his splendid golden whip, mounted his chariot and drove across the waves.”
- Homer, The Iliad
Introduction to tsunamis
Ever since 26th of December 2004, Sri Lanka has been hurled not only into a situation of devastation beyond imagination, but also into a maelstrom of confusion. A natural hazard on such a terrifying scale has not impacted on the country in living memory, and consequently, misunderstandings and misconceptions regarding the nature of this phenomenon have run rife in the aftermath. Tsunamis are unpreventable and are bound to recur. However, by understanding tsunamis better (which is what this article aims to achieve), we could reduce the damage done by future tsunamis and reduce human vulnerability. This article represents an initial step in that process of learning and comprehension.
Although tsunamis are often mistakenly referred to as tidal waves, they have nothing to do with tides, which we know are regulated by the gravitational attraction of extraterrestrial objects such as the sun and the moon. To be precise, a tsunami is an impulsive sea wave triggered by a submarine (undersea) disturbance. Tsunami actually means “harbour wave” in Japanese, because of the wave amplification seen in bays and harbours. The actual behaviour of tsunamis and tidal waves differ radically in their sizes, origins, speeds and consequences.
Tsunami: the genesis
What causes a tsunami?
The most common cause of a tsunami is the sudden uplift or subsidence of the seafloor triggered by a submarine earthquake. Most tsunami triggering earthquakes (tsunamigenic earthquakes) occur in the vicinity of “subduction zones”.
A subduction zone is an area where two or more of the Earth?s tectonic plates (fragments of the earth?s crust that are continually in motion) converge and the denser (usually oceanic) plate subsides (slides) under the less dense (usually continental) plate as illustrated in Figure 1. This displaces, tilts or offsets a large area of the ocean floor. On December 26th 2004, the Indian oceanic plate subsiding under the continental Burma plate triggered an earthquake. This released tremendous energy causing the leading edge of the overriding Burma plate to uplift, disturbing the overlying body of water, and resulting in a tsunami (Figure 1). Note that the earthquake and tsunami do not originate at the same location.
It is also noteworthy that most submarine earthquakes do not generate tsunamis: along the west coast of South America, a region where tsunamis are most prominent, only 20 tsunamis have been recorded from a possible 1098 offshore earthquakes (Bryant, E., 1991). This is because earthquake-induced destructive tsunamis require at least 3 conditions: i) a shallow crustal depth (at least < 40 km) at which the earthquake occurs, ii) an earthquake with a magnitude at least greater than 6.5 on the Richter scale, and iii) body mass movement.
Another cause of tsunamis relate to volcanic eruptions. Tsunamis are known to have been generated from volcano-induced earthquakes, large ash flows and partial collapse of volcanoes such as in Santorini, Italy in 1470 BC and Krakatau in Indonesia in 1883 AD. Other causes of tsunamis include large underwater landslides and meteorite impacts.
Where do most tsunamis occur?
As the rim of the Pacific Ocean is almost surrounded by large subduction zones, this area has experienced some of the world’s greatest known earthquakes and hosts some of the most explosive volcanoes. As a result, over 73% of the documented most significant tsunamis have occurred in this region. However, the Mediterranean and Caribbean Seas have hosted tsunamis to a lesser extent (approximately 24%) (Bryant, 1991). The tsunami threat is most acute along the coasts of Alaska, Russia, Philippines, Indonesia, the western seaboard of South and Central America, and Japan. The 1946 Aleutian earthquake triggered a tsunami that recorded a maximum run up of 42 m, in the neighbouring islands (Okal et al., 2003). In comparison, the reported tsunami wave height near in the southern city of Galle in Sri Lanka in 2004 was only approximately 3.5 m.
How does a tsunami operate?
A generated tsunami travels in wave trains that may contain several large waves. A tsunami travelling towards a distant shore will travel across the deep ocean and then across the continental shelf (as shown in Figure 2) before reaching the shoreline. These waves in the deep ocean are difficult to identify, as their wave height is less than 0.4 m.
Crucially, tsunamis have very long wavelengths (distance between two crests of the wave) that range from 10 to 500 km in deep oceans. In comparison, tidal waves and storm surges have wavelengths ranging from 0.04 to 0.4 km. This exceptionally long wavelength allows tsunamis to travel across oceans without losing energy and at great speeds. Velocity of a tsunami is related to the magnitude of the triggering source, ocean depth and coastal morphology. Therefore, with greater ocean depth, the tsunami achieves greater speed – up to 700 km/hr in the deep ocean.
When the tsunami wave crosses the continental shelf and approaches the shore, the shape of the waveform alters from a sinusoidal (normal) waveform to a wave with a sharp peak (mathematically known as a Stokes wave) and the bottom part of the wave flattens due to the large wave period (the time between two crests of the wave). Tsunamis tend to have wave periods of usually 10 to 45 minutes. The significance of this change in wave shape is that when the wave approaches the coastline, the bottom or back portion of the wave slows down on the increasingly shallow ocean floor.
At this stage, the front portion of the wave peak of the wave sharpens and continues at a much higher speed, usually around 50 to 60 km/hr near the shore. To accommodate difference in speed, the wave increases in height and has the effect of drawing the shoreline out to sea. This is why in shallow shores the sea initially appears to retreat. Due to the large distances in wavelength, the bottom part of the wave lags further behind to form a wave travelling wholly above sea level and a series of breaking waves, that has the potential to carry large amounts of water. It also creates the effect of the incoming wave seeming like “a wall of water” or “tide-like floods”. However, it is not clear which of the several waves will have the most destructive impact on the coast.
Destructive power of a tsunami
The destructiveness of a tsunami is related to area inundated, the severity of wave impact and area eroded. The severity of the wave impact is a result of several factors that include topography of the seabed, morphology of the coastline, which relates to the dissipation of the energy of the tsunami and direction of travel of the tsunami. Areas that would bear the greatest impact of a tsunami are shores perpendicular to the direction of the wave front of the tsunami, with long gentle slopes and concave in shape like bays or harbours. Bays and harbours tend to concentrate the energy of a tsunami. Also, if the tsunami wave period tends to approximate to some harmonic of the natural frequency of that basin, it could cause resonance and increase the destructiveness of the tsunami. Tsunami waves that arrive at high tide are known to have refracted and boosted destructiveness.
Although mangroves and barrier reefs reduce the degree of hazard, if there is a gap in the reef, the shore becomes exposed and that locality would be at a much higher risk due to the concentration of the wave energy. At its peak, the December 26th 2004 tsunami would have deposited approximately 100,000 tonnes of seawater for every 1.5 m of coastline. Therefore, the most important parameter for hazard evaluation from tsunamis is the extent of “run up”, which is the maximum height of a tsunami when it reaches the shore. It is accepted within the scientific community that tsunamis with greater than at least 1 m of run up are deemed very dangerous.
In broad terms the cross-sectional area inundated by a tsunami on land is equivalent to the cross-sectional area of water carried under the wave crest of the tsunami in the proximity of the shore (Figure 3). This allows scientists to relate the maximum distance of flooding with the run up height of a tsunami. This is a very useful indicator for civil defence authorities. The after effects of a tsunami involve both receding and invasive phases, much like water sloshing to and fro in a bowl. On average, the danger period of a single tsunami could last from to 4 to 5 hours.
The turbulent swirl patterns of the tsunami observed near on the west coast of Sri Lanka near the town of Kalutara on December 26th 2004 are not totally understood by the scientific community. A possible explanation might be that they were an artefact of hydro-dynamical action between the interaction of the tsunami and an oblique wave leading to flow impingement, refraction and semi-vortex formation.
Potential future threats of tsunamis to Sri Lanka
Having established the dynamics of tsunamis, we can now turn our attention to taking concrete steps to defend ourselves from their destructive power. To implement this, at least four stages of preparation are required: i) monitoring, ii) warning, iii) land-use planning and iv) education.
According to the National Geophysical Data Centre (NGDC) that catalogues tsunami events and the Indian Meteorological Department, within the last 250 years there have been in excess of 60 tsunamis in the Indian Ocean region. Therefore, tsunamis are not that infrequent in the Indian Ocean.
The source locations of these tsunamis, as seen in Figure 4, show that their spatial distribution is strongly related to the tectonic activity of the region. The number of possible run ups that each of these tsunamis incurred illustrates their respective spatial extent and intensities. Out of these 60 events, eight tsunamis affected Sri Lanka including the December 26th 2004 tsunami. There have been two tsunamis triggered off the coast of Pakistan in 1819 and 1945. In the Bay of Bengal there are at least 4 tsunamis that affected the eastern coast of Sri Lanka in 1762, 1847, 1881 and 1946. The latter three relate to earthquakes in the vicinity of the Nicobar Islands, India. It is also probable that in 1882, a small tsunami was triggered off the north east coast of Sri Lanka and subsequently a tsunami run up was registered in the northeastern town of Trincomalee. However, prior to 2004, the most significant tsunami to affect Sri Lanka within the last 250 years was a result of the highly explosive eruption of Krakatau volcano in Indonesia on August 27th 1883. Tsunami run ups of greater than 1 m have been documented in Trincomalee, Colombo in the West and Galle in the south coast of Sri Lanka (National Geophysical Data Centre, Boulder Colorado, USA).
Needless to say, the December 26th 2004 tsunami was the largest documented in the Indian Ocean. The earthquake that triggered this tsunami was the 5th largest earthquake ever recorded and the energy it dissipated was equivalent to 30,000 Hiroshima type bombs. Prior to the December 26th 2004 tsunami, the last destructive tsunami in the Indian Ocean occurred on June 2nd 1994, which inundated several villages in the islands of Java and Bali, Indonesia and led to over 250 casualties.
Where will the next destructive tsunami originate?
In the Indian ocean, tsunamigenic earthquakes could occur in any of the possible 3 regions: a) mostly off the coast of Sumatra b) 70-100 km south of the Kutch region of the Arabian Sea and c) the bay of Bengal Figure 4 also shows the epicentres of events greater than Richter magnitude 6.5 (earthquakes that could cause tsunamis) in and around the central and eastern parts of the Indian Ocean as recorded by the National Earthquake Information Centre (NEIC), US Geological Survey since 1973.
The 88 earthquakes recorded are mostly concentrated off the shore of Sumatra, Indonesia, along the area called the Sundra trench. Since 1973, there have been 8 tsunamis in the Indian Ocean region. As most tsunamis in the Indian Ocean are earthquake induced, it is possible that the next destructive tsunami that might strike Sri Lanka would also originate off the coast of Sumatra.
The potential for future destructive tsunamis from this region is still great, as the December 26th 2004 Aceh earthquake did not relieve all the stress along the faults at the Sundra trench. This fault system is still highly stressed. Furthermore, the large displacement caused by the earthquake has heightened the stress of the neighbouring faults. Past global seismic activity does suggest after such powerful earthquakes, other large earthquakes do follow (Kagan and Jackson, 1991; Goes, 1996). Therefore, it would be very useful to conduct a quantitative study in order to evaluate whether the statistical probability of another big earthquake in the region has increased since December 2004.
The other distinct possible source for a destructive tsunami in the Indian Ocean might be the collapse of the unstable south flank of Krakatau volcano, Indonesia. However, broad global tsunami patterns do indicate that each region of the world has its own cycle of frequency and pattern in generating tsunamis. These tsunamis could vary significantly in size, from small to large and highly destructive ones.
What will a tsunami monitoring system achieve?
The role of a tsunami monitoring centre is to notify and alert local warning centres to prompt civil defence actions against an oncoming tsunami. To achieve this purpose, a monitoring centre must rely on information derived from geophysical instruments such as seismic sensors, ocean-bottom pressure gauges and tide gauges. From global and regional arrays, and seismic and oceanographic stations, a tsunami monitoring centre can accurately determine the location and magnitude of a tsunamigenic event and evaluate the probability of a tsunami.
If a tsunami has been generated, tsunami warnings are issued to all possible areas of inundation. These warnings include seismic data, potential tsunami wave heights (run ups), and projected impact (travel) times. In turn, local warning stations along coasts could enhance the effectiveness of their warning capabilities, produce tsunami hazard inundation area maps and initial risk assessments, and develop localized tsunami hazard mitigation programs including public and community awareness. However, for the Indian Ocean, as destructive tsunamis are infrequent and human memory is surprisingly short, complacency and inefficiency could be the bane of such a system.
Although warning systems would do little to reduce the vulnerability to the near-source area of tsunamis, which might be inundated within minutes of the earthquake occurring, they would have a profound impact in far source areas such as Sri Lanka. Had such a system existed, the death toll from the December 26th tsunami in Sri Lanka would probably have been no more than a few 10s of people. It took approximately 110 minutes from the occurrence of the earthquake to the time when the tsunami inundated the coast of Kalmunai, the first region of the impact, in the east coast of Sri Lanka (Figure 5).
The key purpose of an Early Warning System (EWS) is to reduce vulnerability from a hazard. A good EWS should include tools such as monitoring, scenario development and forecasting. An EWS should also be mindful of its integration into the larger social system. However, as reducing vulnerability is essentially a social process, EWS should be used only as technological tools in conjunction with other disaster preparedness steps. To eliminate future loss of life on a similar scale, it might perhaps be prudent to contemplate tsunami drills similar to earthquake drills in Japan, siren warning systems, integration with Information Communication Technologies (ICT) and education.
Another alternative is to construct a risk-informed alert system using global tsunami databases based on a Bayesian belief network to reduce intrinsic uncertainties in tsunami forecasting (Woo and Aspinall, 2005). Belief networks are probabilistic inference graphs providing a logical basis for reasoning under uncertainty, and aid decision-making by accounting jointly for uncertainties associated with accumulated experience, and with processed statistical data (Woo and Aspinall, 2005). However, to construct a Bayesian network with little data particularly in the Indian Ocean and as broad tsunami genesis patterns differ with geographic regions, a series of false alarms might adversely affect the credibility of the warning mechanism. A better approach might be to include a Bayesian belief network to an EWS.
Scientists are also examining the past history of tsunamis (paleo-tsunami studies) by looking at the geological record. However, there is occasional debate as to how to distinguish storm surge deposits from tsunami deposits. In other studies, scientists have established a linear relationship between earthquake moment magnitude, Mw (which relates to the energy released by an earthquake), and the run up of tsunamis. However, these linear relationships too differ depending on geographic region and the shape of the coastline.
It is only by understanding the behaviour of a tsunami that we are able to take the most appropriate actions to ensure that should such a natural disaster reoccur, all possible steps would have been taken to minimise death and destruction. More scientific investigations into the patterns of tsunami hazard in the Indian Ocean region could prove extremely valuable in mitigating future threats from this up-until-now under-estimated natural hazard.
References:
Atwater, B. F., Cisternas, M. V., Bourgeois J., Dudley W. C., Hendley II, J. W., and Stauffer, P H., (1999), Surviving a Tsunami-Lessons from Chile, Hawaii, and Japan, U.S. Geological Survey, Circular 1187, Version 1.0.
Bryant, E., (1991), Natural Hazards, Cambridge University Press, Cambridge, UK.
Bryant, E., (2001), Tsunami the Underrated Hazard, Cambridge University Press, Cambridge, UK.
Geist, E. L., (1997), Local Tsunamis and Earthquake Source Parameters, Advances in Geophysics, 39, 117-209.
Goes, S. D. B., (1996), Irregular Recurrence of Large Earthquakes: An Analysis of Historic and Paleoseismic Catalogues, Journal of Geophysical Research-Solid Earth, 101 (B3): 5739-5749.
Kagan, Y. Y., and Jackson, D. D., (1991), Long-term earthquake clustering, Geophyiscal Journal International, 104 (1), 117-133.
Okal, E. A., Plafker, G., Synolakis, C. E., and Borrero, J. C., (2003), Near-field survey of the 1946 Aleutian Tsunami on Unimak and Sanak Islands, Bulletin of the Seismological Society of America, 93 (3) 1226-1234.
Woo, G., and Aspinall, W., (2005), Need for a Risk-Informed Tsunami Alert System, Nature, 433, 457.
Figure captions
Figure 1: Stages of a tsunami generation. Top) Stuck to the subducting plate, the overriding plate gets squeezed. Its leading edge is dragged down, while an area behind bulges upward. This movement goes on for decades or centuries, slowly building up stress. Middle) An earthquake along a subduction zone happens when the leading edge of the overriding plate breaks free and springs seaward, raising the sea floor and the water above it. This uplift initiates a tsunami. Bottom) Dispersion of the tsunami wave to the proximal and distant coasts (Atwater et al., 1999)
Figure 2: The various phases of a tsunami. Stage I) In the deep ocean travelling as a sinusoidal wave, Stage II) Bottom part of the wave slowing down giving rise to a conical shape, and Stage III) At the coast. The tsunami hits the coast as a singular wave travelling above mean sea level. The arrow points in the direction of travel of the tsunami wave (Adopted from Geist, (1997); Bryant, E. (2001)).
Figure 3: Relationship between volume and cross-sectional area of the tsunami and the area inundated. This also shows the large volume carried by the wave as a result of the large wavelength Source: Bryant, E. (2001)).
Figure 4: Map of tsunami source locations in the Indian Ocean for the past 250 years (constructed from data provided by the US Geophysical Data Centre). Note that some tsunami source locations overlie one another. Red circles show the epicentre locations of earthquakes greater than magnitude 6.0 on the Richter scale since 1973 as recorded in the central and eastern parts of the Indian Ocean (Data provided by the NEIC, US Geological Survey).
Figure 5: Adopted animated simulation of the path and time of the December 26th 2004 tsunami on its approach and impact on the Sri Lankan coastline. This clearly indicates the directional path of the tsunami. (Source: National Institute of Advanced Industrial Science and Technology, Japan).
