Working towards early warning

The Indian Ocean Tsunami Warning System that is taking shape will be a coordinated network of national systems, owned and operated by member-states.

AFTER devastating floods struck North Bengal in 1922, engineers embarked upon an expensive scheme of building retarding basins to control floods. But Prashanta Chandra Mahalanobis, the famous Indian statistician, undertook a statistical analysis of rainfall and floods in the region over 50 years and advised against the prevailing engineering wisdom for what was a low probability event.

A tsunami in the Indian Ocean, unlike one in the Pacific, is also a very low probability event. Compared to 790 tsunamis in the Pacific, the tsunami of December 26, 2004, triggered by a 9.3-magnitude earthquake off the Sumatra coast, was only the second one in the same period. The last one was in June 1941 as a result of an 8.1-magnitude earthquake in the Andaman Sea. But the scale of devastation of the 2004 tsunami was such that the total number of deaths it caused - around three lakh in the countries of the Indian Ocean region - was more than the cumulative casualties of all tsunamis until then worldwide. This unprecedented catastrophe demonstrated that while a tsunami in the Indian Ocean is a very low probability event, it can be an extremely high impact one when it does occur.

This makes the establishment of a regional Indian Ocean Tsunami Warning System (IOTWS) an imperative. Francois Schindele, chairman of the International Coordination Group for Tsunami Warning System in the Pacific (ICG/ITSU) stated: "Considering that every ocean and sea basin can be impacted by a tsunami and that the next tsunami can occur anywhere and at anytime, there is a need to design a global tsunami warning and mitigation system."

At the same time it places a stringent requirement on a proposed warning system in the regions outside the Pacific. As Keith Alverson, head of section, Operational Observing System Section, Intergovernmental Oceanographic Commission (IOC) of UNESCO, observed: "Putting in place a system without accounting for the equally plausible scenario that no large tsunami occurs in the Indian Ocean for centuries, or that the next large tsunami occurs in some other unprotected area, would be an enormous mistake. Just as people living in coastal regions of the Indian Ocean today must be protected against the eventuality of a similar event occurring in the near future, future generations must be protected against the eventuality that the next event may occur centuries from now. We must avoid developing a rapidly installed system that is no longer working when it is called upon for use."

A proposal to set up a global tsunami warning system by mid-2007 under the aegis of the IOC was mooted by UNESCO Director-General Koichiro Matsuura on January 13, 2005. The Indian Ocean component of this system, the first regional component, has been taking shape over the past one year. There is already the Hawaii-based Pacific Tsunami Warning Centre (PTWC), which came into existence in 1968 following the creation of the ICG/ITSU, and the International Tsunami Information Centre (ITIC) in 1965. The IOC itself was established in 1960.

While the June 2006 target envisaged by Matsuura for the Indian Ocean network may be a bit optimistic, considerable progress has been made towards putting in place an effective tsunami warning system. UNESCO, on its part, is still maintaining a deadline of July 2006. An ICG for the IOTWS, chaired by India's P.S. Goel, Secretary to the Department of Ocean Development, met in Perth in August 2005 and laid down the basic components and parameters for a reliable IOTWS.

An improved seismographic network, a network of real-time sea-level gauges in the Indian Ocean and deep-sea pressure sensors have been proposed, along with national tsunami warning centres (NTWCs), for a reliable warning and mitigation network for the region.

The ICG said "it will be a coordinated network of national systems, owned and operated by member-states hosting or otherwise taking responsibility for them". A network of Regional Tsunami Advisory Centres (RTACs) and NTWCs is envisaged, with the regional centres capable of disseminating seismic data, oceanic responses and advisory messages across the region, but not warnings. This is the key difference between the PTWC and the IOTWS. The RTACs will not issue warnings outside their own nations, whereas the PTWC does. The national autonomy of the NTWCs will be respected.

Twenty-five countries (all the countries of the region except Somalia) have already established communication centres to receive advisories issued by the PTWC in Hawaii and the Japan Meteorological Agency (JMA) in Tokyo. At this stage, these would be based on seismological data on earthquakes above a certain threshold of magnitude. Since only a small fraction of such earthquakes generate tsunamis, this interim system is prone to a high rate of false alarms.

According to Matsuura, individual countries should try to develop their own detection networks and their own risk assessment and preparedness plans for an effective warning system. Australia, India, Indonesia, Iran, Malaysia, Pakistan and Thailand have made plans to set up national systems capable of detecting, analysing and transmitting tsunami alerts. The Indian system, to be set up at a cost of Rs.125 crores, is expected to become operational by September 2007.

The sea-level network in the Indian Ocean has been upgraded with the establishment of 23 real-time stations (with near-real-time high frequency sampling gauges suitable for tsunami early warning). These stations form part of the Global Sea Level Observation System (GLOSS) set up in 1985 and transmit data every hour to the PTWC and the JMA through the Global Telecommunication System (GTS) of the World Meteorological Organisation (WMO).

Unlike the Pacific, where the entire Pacific Rim is tsunamigenic (the `Ring of Fire' as it is called), in the Indian Ocean tsunamigenic quakes are known to occur only in two regions. One is in the East, where a fault runs in a long arc from Myanmar to Sumatra and encompasses the Andaman-Nicobar Islands. The other is the Makran fault in Pakistan.

The Perth meeting concluded that to issue a warning to the vulnerable spots within 30 minutes, a quake's location and magnitude had to be determined within minutes of its occurrence. Towards this, improvements have been made in the GTS to enable it to carry data and information relevant to tsunamis as well.

IN an earthquake of high magnitude, values from seismometers situated close to it are not accurate enough to determine the parameters of the quake whereas data from distant detectors can lead to an accurate determination. For an early warning system, real-time data would be required. But India is not part of the global real-time network under the Global Seismic Network (GSN) and the Federation of Digital Seismic Network (FDSN) because Indian policy does not permit sharing of seismic data in real time because of unspecified national security reasons.

If India has to become part of the early warning system, this policy has to change. According to the Department of Science and Technology, which monitors seismic events with its network of seismic stations under the India Meteorological Department (IMD), and the Department of Ocean Development (DOD), a policy change to make available in near real time data of events with magnitude beyond a threshold of, say, 5 or 6, is being considered (Frontline, February 11, 2005). However, the document titled "Technical Aspects of an IOTWS", arrived at in the Paris meeting of the ICG in March 2005, states: "The critical need is the availability of real-time data... . Data must be made available in real-time to all NTWCs and centres designated for processing and analysis."

The document acknowledges that there are national commitments towards building capabilities, which, it says, will together provide the technical and scientific basis of an IOTWS. Also, existing national, regional and international seismic networks will be essential components of the total system. As regards data from the observation systems, the document says: "Immediate, free and open distribution of raw data from the observing systems in real time must be acknowledged as a founding principle for all national, regional and global tsunami warning systems."

The deep-sea sensors for the Indian Ocean system would be of the type developed by the U.S., six of which are in operation in the Pacific as part of the PTWS (Frontline, February 11). Called the Deep-ocean Assessment and Reporting of Tsunamis (DART) system, it consists of a seafloor bottom pressure recording (BPR) system, also called `tsunameter', capable of detecting tsunamis as small as 1 cm, and a moored surface buoy for real-time communication. The BPR monitors the water pressure with a resolution of approximately 1mm of sea water, with 15-second averaged samples, which is converted into "sea level height". An acoustic link is used to transmit data from the `tsunameter' to the buoy. The data are then relayed via the Geostationary Operational Environmental Satellite to ground stations, which disseminate them immediately to tsunami warning centres.

The technical working group of the ICG noted that the multipurpose national ocean buoy networks in the Indian Ocean - both coastal and deep-ocean - operated by India, Malaysia, Indonesia, Thailand, Australia and Japan could be equipped with BPRs to measure tsunami height in open seas. According to it, sensors installed in buoy arrays 100-200 km offshore have been shown to detect tsunami waves in advance of their arrival on the coast.

According to the Pacific Marine Environmental Laboratory (PMEL) of the National Oceanic and Atmospheric Administration (NOAA) in the U.S., the first-generation DART system (DART-I) that it developed has proven to be robust and reliable, with a cumulative data return reliability of over 91 per cent in the 1997-2003 period. The system features an automatic detection and reporting algorithm triggered by a threshold wave-height value. DART systems have reacted to and reported six seismic-induced tsunami events that contributed to appropriate operational decisions.

The 7.3 magnitude earthquake in the Aleutian Islands on November 17, 2003, triggered a `tsunami watch' in Hawaii and Alaska, but on the basis of data received from three tsunameters, which showed that the wave was not significant, no warning was issued, thereby saving huge evacuation costs. However, post-event data showed that a 2 cm tsunami wave observed by a tsunameter was found to have grown to a surprisingly large 40 cm wave on the north shore of Oahu, Hawaii. But the trigger threshold in the buoy for reporting an event was 3 cm. This showed that even a small tsunami was potentially dangerous.

Second generation DART systems (DART-II) that are under development will allow bi-directional communication, which would enable transmission of tsunami data on demand, independent of the automatic algorithm. This would ensure the measurement and reporting of tsunamis with wave amplitude below the automatic reporting threshold.

The 32 DART systems that the U.S. plans to deploy in the Pacific, Atlantic and the Caribbean, will have the DART-II technology. A NOAA study in the aftermath of the December 2004 tsunami evaluated six different technologies for tsunami detection and concluded that a DART-type system was the most effective. In fact, the number of DART or DART-like systems deployed globally is likely to shoot up to 80, with over 20 alone in the Indian Ocean region.

According to ICG's working group on tsunami detection for the IOTWS, in the arc extending from Bangladesh to New Guinea and in the northern Arabian Sea, the other hotspot, about 15 DART-like systems would be required. Another ring of DART systems, to give another level of protection, has been envisaged on the basis of a suggestion from Harsh Gupta, former secretary of the DOD and chairman of this working group. An estimated 20-plus DART-like systems would be an optimum number, according to the working group. The first such buoy was deployed in November in Indonesia's Sunda Strait as part of a German assistance of $50 million for a tsunami early warning system and tsunami hazard mitigation programme in Indonesia. Altogether 10 such buoys will be in place by 2007.

Many other countries, including India, have evinced interest in DART-like systems for the national tsunami warning systems. The NOAA, according to the U.S. government, plans to make DART technology available to countries that want it. And to coordinate acquisition, deployment and operation of these systems, a DART Partnership has been established amongst DART or DART-like system users. The group will be chaired by Australia and the members include Australia, Indonesia, India, Malaysia, Thailand, Germany and the U.S.

According to David McKinnee, NOAA coordinator in the ICG/IOTWS working group, this DART operators' group could eventually become the global coordinating group for DART technology, handling everything from setting standards for tsunami-detection technology and sharing R&D to producing deployment plans and even coordinating the production of tsunami-detection technology outside the U.S.

The Indian early warning system will include 10-12 DART-type systems. (It is not clear whether this would be integrated with the IOTWS network to be established around the Indian subcontinent or not.) These will be deployed in the two zones with tsunamigenic potential. According to the DOD, some of these may have to be imported initially, but eventually they would be developed indigenously. The first Indian DART system is expected to be deployed by mid-2006. All data from the various sensors would be sent to a dedicated Indian Tsunami Warning Centre (ITWC), which will be operated by the Hyderabad-based National Centre for Ocean Information Services (NCOIS) of the DOD.

As part of the proposed system, the Department of Science and Technology has established a dozen digital broadband seismic stations and networked to make data available on a near real-time basis for strong earthquakes. To achieve this, the system will be further augmented to be able to carry out automatic analysis of location and magnitude of any earthquake much quicker than the present 10 minute-delay for transmitting the relevant seismic parameters. The system will include 20-25 automated sea-level gauges capable of transmitting data in real time along the Indian coast as well as on the Andaman and Nicobar islands. The gauges on the islands would be able to provide an hour's advance notice to the mainland in the event of a tsunamigenic earthquake in that region. Tide gauge and data buoys (with and without BPRs) will also be networked to validate the arrival of tsunami waves at the coast.

The Indian early warning system is also intended to monitor storm surges that occur due to the strong cyclonic winds. Storm surges can increase sea levels by a couple of metres and destructive storm surges can, in fact, cause greater increases. India's east coast experiences a high rate of occurrence of cyclones and the associated storm surges. On the face of it, clubbing tsunamis and storm surges in the same warning system may seem inappropriate. But advances in radar technology apparently enable shore-based high-frequency radars that can detect surface currents and wave characteristics at distances of over 200 km, thus being suitable for tsunami warning as well. However, the radar's ability to distinguish a tsunami from a storm surge depends on the waves growing in amplitude as they approach the shallow waters near the coast. Even so, the DOD has procured such radars from CODAR Ocean Sensors, a U. S. company for installation along the Indian coast.