Tsunami Warning Systems Efficient use of Tide Gauge

Tsunami Warning Systems Efficient use of Tide Gauge Stations I. V. Fine 1, 2, F. E. Stephenson 3, A. B. Rabinovich 1, 4 and R. E. Thomson 1 1 Institute of Ocean Sciences, Sidney, B. C. Canada 2 Heat and Mass Transfer Institute, Minsk, Belarus 3 Canadian Hydrographic Service, Sidney, B. C. Canada 4 P. P. Shirshov Institute of Oceanology, Moscow, Russia

The Sumatra-Andaman Tsunami of December 26, 2004 (Mw = 9. 3) Phi Island (Photos by J. T. and Carolina Malatesta)

As a result of the Sumatra. Andaman Tsunami … § There has been a strong international commitment to install or upgrade many tide stations and seismic stations. § This is an excellent start, but what can we learn from past experiences?

What will this new network of tide gauges look like in 30 -40 years? What does past experience teach us? § Canadian Arctic § GLOSS program in the Indian Ocean

In all probability… § “Disaster response” funding will be used to install many instruments and related communication systems § Many nations / agencies will lack the means or the commitments to maintain these networks in the long-term § The ability of these networks to provide effective tsunami warning will decrease

An alternate strategy § Before installing new coastal stations and offshore buoys, try to assess the hazards and the warning requirements § Plan the network of stations to provide the required warning times § Implement a network which is sustainable

Tsunami Detection and Warning For an arbitrary tsunami source location, we can estimate a safe warning time (DT) -- for any specified coastal site and known warning station – as the time delay between the arrival time at the station and the coastal site

Tsunami sources in the Indian Ocean (416 AD-2005) 1945 2004 1861 1797 (courtesy of V. K. Gusiakov [2005]) Most of tsunami source regions in the Indian Ocean are along the Sunda Trench subduction zone (along Burma, Andaman and Nicobar islands, Sumatra and Java). However, there are sources (e. g. 1945) in the NW part of the ocean (Murray Ridge)

General seismicity of the Indian Ocean 416 AD - 2005 (courtesy of V. K. Gusiakov [2005]) The main seismic area is the extensive subduction zone near the NE margin of the Indian Ocean. However, there also other zones: the SW, Central, Carlsberg and Murrey ridges

Tsunami Detection and Warning For an arbitrary tsunami source location, we can estimate a safe warning time (DT) -- for any specified coastal site and known warning station – as the time delay between the arrival time at the station and the coastal site

Delay (minutes) between a “Nicobar” warning station and the City of Vishakhapatnam A “Nicobar” site would provide an optimal warning time (> 2 hours) for a tsunami source located in the SE Indian Ocean (similar to the Dec. 26, 2004 earthquake). However, such a site would not be optimal for tsunami sources located in the West Indian Ocean.

Delay (minutes) between an “Andaman” warning station and the City of Vishakhapatnam An “Andaman” site would provide a less optimal warning time (about 1. 5 hrs) for sources in the SE Indian Ocean compared with a “Nicobar” site, but would improve warning time (> 2 hrs) for sources in the eastern part of the Indian Ocean.

Delay (minutes) between a “Minicoy” warning station and the City of Vishakhapatnam A “Minicoy” site would provide excellent warning time (2 -4 hrs) for sources in the West Indian Ocean but poor warning for sources in the SE Indian Ocean (such as the Dec. 26, 2004 earthquake).

Delay (minutes) for Vishakhapatnam for all three warning stations combined Simultaneous operation of all three warning stations would remarkably improve the safety (time delay) for Vishakhapatnam. Distant waves, excluding those that came from the south, would be recorded 2 -4 hours prior to the arrival of the waves.

Delay (minutes) between a “Nikobar” warning station and Phuket (Thailand) A “Nicobar” site would provide an optimal warning time (> 1. 5 hours) for a tsunami source located off the western coast of Sumatra (similar to the Dec. 26, 2004 earthquake) and for West, North, and SE Indian Ocean. However, such a site would not be useful for tsunami sources located between the Nicobar/Andaman Islands and the maincoast (in the Strait of Malacca)

Delay (minutes) between an “Andaman” warning station and Phuket (Thailand) An “Andaman” site will provide better warning time (> 2. 0 hours) than a “Nicobar” site for a tsunami source located in the northern part of the Andaman Sea, but will be much less useful for a wave source near Sumatra (similar to the Dec. 26, 2004 earthquake). Such a site would also not be useful for tsunami sources located between the Nicobar/Andaman Islands and the coast of Thailand (in the Strait of Malacca)

Delay (minutes) between an “open ocean” warning station and Phuket (Thailand) An “open ocean” station deployed in the Strait of Malacca between Nicobar Islands and Phuket would significantly improve the situation for sources located between the Nicobar/ Andaman Islands and the mainland coast (Malacca Peninsula). It would provide warning times > 1 hour.

Some examples for Chile

Several scenarios for Valparaiso

Conclusions § Modeling can be used to provide useful information on safe warning times § This supports decision making about the optimum placement of gauges § The effect of inoperative stations can be immediately accessed § Time delays for data transmission can also be modeled