Rewiring the Tsunami Early Warning System

By: M V Sunanda, T Srinivasa Kumar, Dipankar Saikia, S S C Shenoi, Shailesh Nayak
The Japan Meteorological Agency (JMA) on 11 March 2011 warned that a 3-m-plus tsunami would hit north-eastern Japan. In fact, the wave that came ashore stood more than 10 m high — reaching 40 m in some places.
Planning n Mitigation

The Indian Tsunami Early Warning System (ITEWS) can successfully issue timely and accurate warnings. However, the 2011 earthquake in Japan and 2012 in Northern Sumatra once again emphasised the limitations in traditional approaches that necessitate improvements. Thus the Earth System Science Organisation (ESSO) has initiated improvements in the ITEWS that includes establishment of co-located broadband, GPS and strong motion sensors for measuring displacements and ground accelerations in real-time.

The Indian mainland and islands are located in a zone of significant seismic activity, where many earthquakes and accompanying tsunamis have been observed and recorded. The two subduction zones, the Andaman-Nicobar-Sumatra island arc and the Makran region have been identified as tsunamigenic zones in the Indian Ocean based on historical records.

A great shallow foci earthquake of magnitude Mw 9.2 occurred on 26 December 2004 on Andaman-Sumatra subduction zone and generated a massive tsunami that caused extensive destruction all along the Indian mainland and Andaman & Nicobar Islands. In response ESSO successfully set up a tsunami warning centre for Indian Ocean at Indian National Centre for Ocean Information Services (INCOIS), Hyderabad. Since its inception in October 2007 till June 2013, the Indian Tsunami Early Warning Centre (ITEWC) successfully monitored 356 earthquakes of M > 6.5, out of which 60 were in the Indian Ocean region (both on land and under-sea). For all these major events in the Indian Ocean, timely advisories were generated based on estimated time of wave arrivals and wave heights and the stakeholders were informed through bulletins. This avoided false alarms and unnecessary evacuation from the coastal areas. One of the most critical aspects of tsunami warning system is the quick estimation of earthquake parameters with reasonable accuracy in the shortest possible time. The consequences of diametrically opposite behaviours of large earthquakes, in terms of tsunamis, in the recent times—Tohoku-oki (2011) in the Pacific and Northern Sumatra (2012) in the Indian Ocean, demand the improvement in tsunami early warning systems.


Limitations of the current warning procedures

Traditionally, the estimation of tsunamigenic potential of an earthquake relies on the measurement of the magnitude of the earthquake, which is not reliable to the extent that it should be. For example, the Tohoku-oki earthquake of magnitude Mw 9.0 was initially underestimated by the Japan Meteorological Agency (JMA), which inarguably is one of the finest centres for earthquake and tsunami early warnings. The earthquake detection centres elsewhere too estimated much lower magnitudes (Mw 7.9 – 8.0) initially. The magnitude of the earthquake was underestimated at least by an order one to two (M 7.2 after 8.6 seconds and revised to M 8.1 after 116.8 sec) [M. Hoshiba, et al., 2011, ‘Outline of the 2011 off the Pacific coast of Tohoku earthquake – Earthquake early warning and observed seismic intensity’, Earth Planets and Space], which in turn underestimated the expected tsunami wave height as 3–6 m. But in reality, the sudden sea-floor displacement generated a massive tsunami that overtopped the tsunami protection walls and broke through as far as 10 km inland along the coast. Though JMA could issue the warnings within 3 minutes, unfortunately, that was based on the gross underestimation of the earthquake magnitude (8.6 M instead of 9 M). The accurate estimate of size of the earthquake could have resulted in an accurate estimate of tsunami wave height. The actual wave height reported from the adjacent areas of Tohoku was 39.7 m (at Miyako).

On the other hand, in the case of Northern Sumatra earthquake of magnitude Mw 8.5 on 11 April 2012 only a small ocean-wide tsunami (~30 cm at Sabang, Indonesia) was generated in contrast to the estimated wave heights of 6-8 m initially. Later when more data became available, it was realised that it was a strike-slip earthquake which generated very little or no vertical motion in the ocean floor hence avoiding the sudden disturbance of water column essential for the generation of a tsunami. Similar was the case with the 2005 Nias earthquake (8.6 M), Indonesia. That event too did not generate a sizable tsunami as expected [A O Konca, et al., 2007 ‘Rupture Kinematics of the 2005 Mw 8.6 Nias–Simeulue Earthquake from the Joint Inversion of Seismic and Geodetic Data’, Bulletin of the Seismological Society of America].

To overcome such difficulties and to understand the fault geometry that governs tsunamis, it is essential to estimate the seismic moment tensor solutions. However, the estimate requires a larger amount of data that becomes available only after a certain amount of time. Longer wait for sufficient data to make a decision at the warning centre is unfeasible as the warning centre is expected to provide warnings at the earliest. Often the procedure to predict tsunami wave height and travel time depends on the worst cases and there might be overestimates in those that deviate from such scenarios, especially the strike-slip ones.


New Approaches

As illustrated above, traditional methods of earthquake magnitude estimation only based on seismic data and the prediction of tsunami wave heights can go wrong if the earthquake mechanism is not taken in account in addition to its magnitude. This is more serious for near source regions like Andaman and Sumatra coast as they lie very close to the subduction zone and the available time for warnings and response is too short. This was precisely the limitation faced during the 2011 Tohoku earthquake and the 2012 Northern Sumatra earthquake that necessitated the development of new tools and techniques for determining the true size of an undersea earthquake and the actual ground displacement. Such techniques call for receiving and analysing data from multiple sensors like seismometers, GPS sensors, strong motion sensors, etc. in real-time.

Use of GPS technology: Monitoring the crustal deformation in real-time makes it feasible to achieve rapid estimation of actual earthquake scales, since the measured permanent displacement directly gives us the true size of the earthquake by seismic moment, which in turn, can be used for tsunami warning. The real-time deformation monitoring technique is based on near-field global position system (GPS). Using coastal GPS stations’ data near the epicentre, the new method estimates the energy transferred by undersea earthquake to the ocean to generate a tsunami [S V Sobolev, et al., 2006, ‘Towards Real-time Tsunami Amplitude Prediction’, Eos Transactions, American Geophysical Union]. Recent analysis showed that by using GPS displacements, it is possible to calculate how far the stations moved because of the quake and that in turn helped in deriving an earthquake’s true size, called moment magnitude. This magnitude is directly related to earthquake’s potential for generating tsunamis [S K Singh, et al., 2012, ‘A Method for Rapid Estimation of Moment Magnitude for Early Tsunami Warning Based on Coastal GPS Networks’, Seismological Research Letters]. This method allows the rapid estimation of seismic moment tensor solutions and the earthquake source determination in a shortest possible time compared to the traditional approaches.

Using Strong Motion Sensors: During large earthquakes the broadband seismometers at the near source region often get clipped due to saturation. To overcome this the stations are generally augmented with strong motion accelerometer. The near source dense network of strong motion sensors provides unsaturated recordings of moderate to large earthquakes and early peak amplitudes from these records can be used to estimate the magnitude of an earthquake [A Zollo, et al., 2006, ‘Earthquake magnitude estimation from peak amplitudes of very early seismic signals on strong motion records’, Geophysical Research Letters]. When even few seconds are critical for a local tsunami warning, the quick estimation of earthquake magnitude, source parameters using accelerometers, broadband seismometers and GPS receivers could result in improved tsunami early warning system.

Rapid estimation of fault parameters using W-phase: After the occurrence of an earthquake it is very important to determine the fault geometry to estimate its tsunamigenic potential. However, it is difficult to calculate these parameters at an initial stage as we need surface wave data over longer period for accurate estimation. To compute and interpret the kinematics of deformation using moment tensor the W-phase is used, since it carries long period information of the source at a much faster speed than the traditional surface waves [H Kanamori, 2008, ‘Source inversion of W-phase: speeding up seismic tsunami warning’, Geophysical Journal International]. Hence the W-phase inversion method can be used for rapid and robust determination of seismic source parameters with sufficient accuracy for tsunami warning.



In order to provide accurate and rapid tsunami warnings, the most critical part lies in estimating the earthquake magnitude and source parameters accurately and as fast as possible. The advantage of such joint network of broadband seismometers, accelerometers and GPS receivers is that it will provide a continuous update on accurate source parameters with higher accuracies.

The ESSO-INCOIS has recently implemented a nation-wide project to connect all standalone remote seismic as well as GPS stations established under various projects all over the country through VSAT to fetch real-time data. A parallel data centre has been made operational at India Meteorological Department, New Delhi along with the ESSO-INCOIS to acquire, process and utilise the dataset and made operational. The data received from broadband seismometers, accelerometers and GPS receivers is being processed in real-time using VRS3Net and SeisComP software by which improved earthquake parameters and displacements are being calculated. The users can also download this data from the website for research. In addition, ESSO-INCOIS has taken up a project on establishing an integrated network of 35 GPS receivers and strong motion accelerometers at seismically active Andaman and Nicobar Islands. The data from these stations will also be acquired and processed through well established data centre for enhancing the timeliness and accuracies of tsunami early warnings. The site selection is underway for installation of the sensors and by end of December, 2013 the Andaman and Nicobar GPS and strong motion network would be made operational.

Leave a Reply

Your email address will not be published. Required fields are marked *