Indian tsunami early warning system
The end-to-end performance of the system was tested for the first time during the earthquake and tsunami event on September 12, 2007 off the west coast of Sumatra. Earthquake information is provided by monitoring real time data from 17 land-based seismic stations operated by the India Meteorological Department (IMD) and nearly 300 seismic stations from international sources within 15 minutes of occurrence.
In order to confirm whether an earthquake has actually triggered a tsunami, it is essential to accurately measure the changes in water level near the fault zone. Bottom pressure recorders (BPRs) are used to detect the propagation of tsunami waves in the open ocean and consequent sea-level changes. A network of BPRs has been installed close to the tsunami-genic source regions to detect tsunamis, by the NIOT. These BPRs can detect changes of 1 cm at water depths of up to 6 km. A network of tidal gauges along the coast helps to monitor the progress of a tsunami and to validate the model scenarios. Near-real time data from national and international centres is received through V-SAT communication and internet respectively.
Working of the ITEWS
Tsunamigenic zones that threaten the Indian Coast have been identified by considering the historical tsunamis, earthquakes, their magnitudes, location of the area relative to a fault, and also by tsunami modelling.
The east and west coasts of India and the island regions are likely to be affected by tsunamis generated mainly by subduction zone related earthquakes from the two potential source regions, viz., the Andaman-Nicobar-Sumatra island arc and the Makran subduction zone north of the Arabian Sea. ITEWS system detects all earthquake events of more than 6 magnitude occurring in the Indian Ocean in the less than 20 minutes of occurrence. BPRs installed in the deep ocean are the key sensors to confirm the triggering of a tsunami.
The NIOT has installed four BPRs in the Bay of Bengal and two BPRs in Arabian Sea. In addition, NIOT and SoI have installed 30 tide gauges to monitor the progress of tsunami waves.
ICMAM has customised and run the tsunami model for five historical earthquakes and the predicted inundation areas. The inundated areas are being overlaid on cadastral level maps of 1:5000 scale. These community level inundation maps are extremely useful for assessing the population and infrastructure at risk. High-resolution coastal topography data required for modelling is generated by the National Remote Sensing Agency (NRSA) using Airborne Laser Terrain Mapper (ALTM) and Cartosat Data. INCOIS has also generated a large database of model scenarios for different earthquakes that are being used for operational tsunami early warning. Communication of real-time data from seismic stations, tide gauges and BPR’s to the early warning centre is very critical for generating timely tsunami warnings. A host of communication methods are employed for timely reception of data from the sensors as well as for dissemination of alerts. ISRO has made an end-to-end communication plan using INSAT.
A high level of redundancy is being built into the communication system to avoid single point failures. INCOIS has established a state-of-the-art early warning centre with all the necessary computational and communication infrastructure to enable reception of real-time data from all sensors, data analysis, as also generation and dissemination of tsunami advisories.
Seismic and sea-level data are continuously monitored in the ITEWS using a custom-built software application that generates alarms/alerts whenever a pre-set threshold is crossed. Tsunami warnings/watches are then generated based on pre-set decision support rules and disseminated to the concerned authorities for action. The efficiency of the end-to-end system was proved during the large under-sea earthquake of 8.4 M that occurred on September 12, 2007 in the Indian Ocean.
The Sumatran tsunami of April 11, 2012 was another major event that proved the efficacy of the system. A tsunami was generated by an earthquake of magnitude 8.5 and its largest aftershock of magnitude 8.2, off the western coast of Sumatra. The earthquake generated a small ocean-wide tsunami that was recorded by various tide gauges and tsunami buoys located in the Indian Ocean region. ITEWS detected the earthquake within 3 min 52 seconds and issued six advisories (bulletins). Based on its SOP, ITEWS issued Tsunami Warning to only 3 regions of the Andaman & Nicobar Islands, while keeping all other regions under alert/watch. Thus, false alarms and unnecessary public evacuations were avoided, especially in mainland India.
Challenges of false alarms
ITEWS follows a unique SOP for selection of closest scenario from the scenario database. The new decision support tool scales up/down the scenario results based on magnitude. This approach narrows down the areas to which tsunami information has to be disseminated, such as warnings for the regions close to the earthquake region and alert/watch for regions farther up.
Warnings for far regions are confirmed only after sea-level change is confirmed by tide gauges and/or BPRs thereby increasing the accuracy and reducing the false alarms. Since its inception in October 2007 till May 2016, the Warning Centre has successfully monitored 511 earthquakes of magnitudes above 6.5 (M > 6.5), out of which 85 were in the Indian Ocean region (both on land and under-sea). The end-to-end capabilities of this warning system have been well proven during all the major tsunamigenic earthquake events in the Indian Ocean.
It may be noted that only on seven occasions a tsunami warning/watch/alert was issued by ITEWC, that too only for selected near-source areas in the Andaman and Nicobar Islands and east coast of India. If warnings were generated, solely based on earthquake parameters, as is the case with many traditional warning systems, it would have called for an Indian-ocean wide tsunami warning for several events. Thus, the timely advisories generated based on estimated wave arrivals and amplitudes and revised bulletins based on water level observations avoided false alarms and unnecessary public evacuations for all major events in the Indian Ocean.
Since it became operational, ITEWS has improved significantly in its functionality. Its observation network has improved with better reception of data from multiple sensors through multiple modes of communication. With new decision support tools, there have been improved tsunami forecast results by scaling up/down existing scenario results based on magnitude. Earlier this was done on the basis of pre-run scenario database of pre-defined magnitude.
A fail-safe satellite based communication system, V-SAT aided emergency communication system (VECS) has now been established connecting seven emergency operation centres in the Andaman and Nicobar Islands. Also, computation time has dropped with advanced information and communications technology (ICT) facilities. The tsunami website—tsunami.incois.gov.in, has been redesigned and developed with the world wide web consortium (W3C) standards and usability guidelines in open web platform.
Real time storm surge modelling and generation of advisories done using IBM-high-performance computing (HPC) facility at INCOIS and a national central data pool has been created for real time acquisition of seismic and GPS data, real time processing, data sharing, storing and archival.
The way ahead
Although India has come a long way in establishing an efficient tsunami early warning system, there are several limitations to the present system that need to be overcome. For one, the present system is suited for mainland forecasting where the wave travel time is about 3 to 4 hours from the epicenter. Tsunami alert is confirmed based on water level information from deep ocean pressure recorders (DOPR) which are placed at 3000-4000 m water depth. The DOPR along with surface buoys are suited for warning where sufficient time to process information is available. This suits the mainland but, DOPR systems connected by a submarine cable are ideally suited for Andaman Islands where the tsunami wave travel time is very short. The cable based system is being used in Japan where the fault is very close to the land. Also, the decision support system at the Warning Center is based on pre-run scenario database of pre-defined magnitudes. Next generation decision support with real time inputs from bottom pressure recorders will substantially improve the forecast.