Decision analysis for a tsunami detection system – case study: Sri Lanka

Abstract

The massive devastation of the Indian Ocean mega tsunami in 2004, clearly calls for an efficient early warning system for the Indian Ocean countries. The 'Regional Watch Provider' system currently planned makes some countries entirely dependent on the watch providers. We discuss a feasible and reliable tsunami detection mechanism for Sri Lanka that would give it some independent information for decision making purposes. It contemplates Sri Lanka deploying its own additional detection tools. A stand-alone tide gauge system, a buoy system and a combined gauge and buoy system are tested with the status quo or 'do nothing' alternative. A decision analysis exercise investigates the probability of tsunamis, probability of failures of each detection system, the probability of successful completion of an evacuation and the probability of false warnings which have their own costs. This paper concludes with the derived recommendations for a preferred tsunami detection system for Sri Lanka.

An update of proposed Sri Lanka warning system for east and west coast tsunamis

“…A reassessed reliability eight years after the original study remains the same. With respect to tsunami buoys, Wickramaratne et al (2011a, 2011b) presented a range of 79-92 per cent. With the increased number of buoys and heightened technology associated with DART (Deep-ocean Assessment and Reporting of Tsunamis) buoys, the reliability is reassessed based on available data from the US National Data Buoy Center (NDBC).…”

“…The activities involved in each of the above elements can be made to form a network which then can be simulated to derive stochastic estimates of time taken to evacuate a particular location from the point of triggering the earthquake/tsunami. From the initial work of Fernando et al (2008), there has been a progressive development of simulating the TWE network pertaining to the Indian Ocean region (Ruwanpura et al , 2009; Wickramaratne et al , 2009, 2011a, 2011b).…”

“…Mechanical systems, whether they be tide gauges or buoys, often experience failures where supply of critical sea level information is interrupted. Wickramaratne et al (2011a, 2011b) assessed the reliability of tide gauges to be in the order of 80-99 per cent based on historic records of gauges in the Indian Ocean. Reliability is computed as the percentage of data availability in the given time period.…”

“…In spite of the proven and unmatched benefit tsunami buoys provide, their characteristic drawbacks have led buoy owners and warning authorities to reconsider future deployments. Tsunami buoy systems are necessarily expensive: nearly US$700,000 per unit (production and deployment) and have a short lifespan of 4-5 years (Wickramaratne et al , 2011a, 2011b). In addition, a considerable cost (∼US$50,000 per trip) is incurred for any repair of the unit, which requires rental of chartered ships.…”

Planning of Taking Parks as Temporary Emergency Shelters

“…[21] Wickramaratne et al (2011) indicated that "successful evacuation is the ultimate decisive factor of the whole warning process since an incomplete evacuation could cause thousands of lives to be lost." [22] In Taiwan, disaster relief roads are defined in the following manner: those 8 m or more in width are for emergency evacuation; those 15 m or more in width are for firefighting, emergency response, and rescue; and those 20 m or more in width are for material supplies, external communications, support, and emergency shelter, if necessary. [19] The outward access conditions of a shelter determine the disaster management functions.…”