Tsunami risk assessment for the Tongan Islands

Introduction

The Pacific Tsunami Warning Center (http://ptwc.weather.gov/), a branch of the NOAA’s National Weather Service operating out of Honolulu, issues warnings to nations in and around the Pacific Ocean pertaining to the threat of tsunamis. The warning bulletins are issued in two forms: a Tsunami Information Bulletin (TIB) is based on seismological measurements, and indicates that earthquakes have occurred which had the potential to generate tsunamis; a Tsunami Warning Bulletin (TWB) is based on evidence from ocean buoy detections and indicates that a tsunami has indeed been generated.

The PTWC bulletins are picked up by the Tongan Meteorological Services (http://www.met.gov.to/). Upon receipt of a TIB-type bulletin, the Tongan Meteorological Service will enter an state of alert and track developments through heightened communication with the PTWC. Upon receipt of a TWB-type bulletin (including when a TIB is elevated to a TWB) the warning is promptly propagated through local channels, including radio, TV, police, and emergency services, to the Tongan population. The procedure is documented here.

In addition to warnings originating from PTWC, the Tongan Meteorological Service may issue public tsunami warnings on the basis of local earthquakes of magnitude 6.7 or greater.

Between 2004 and 2013 a total of XX TIBs and YY TWBs were issued for Tonga by the PTWC (an archive of all notifications is kept here http://ptwc.weather.gov/ptwc/archive.php?basin=pacific). Of these warnings, 100% have been false alarms. The only tsunami to have hit the Tongan Islands during that period did so without warning on the morning of the 30th of September 2009. It claimed 9 lives in Niuatoputapu, and a total of 180 including those in the Samoas. Thus in spite of our efforts, we are 0 for 1 in terms of forecasting actual events in Tonga, and 0 for YY in terms of false positives.

As much as we want to correctly forecast tsunami events, we should also strive to avoid issuing false alarms, which foster complacency.

I’m interested in why false alarms occur. In particular, why, after a wave has been measured to be real by an oceanic buoy, does it not show up on the Tongan shores? In this work I would like to explore the vulnerability (or invulnerability, as it seems) of the Tongan islands to tsunamis such as those recently observed. It is not my intention to declare that any islands are risk-free, but to explore the reasons for so many false alarms, in the light that they are expensive and foster complacency amongst locals (which itself is dangerous).

It may be that some islands receive significant shelter from neighboring islands and reefs. We will explore this possibility.

Another, more curious, reason that some islands are spared may stem from their smallness. There is a rule-of-thumb in physics (normally raised in reference to electromagnetic waves or sound waves) that a wave will only interact with objects larger than its wavelength. Is it possible that some atolls, whose underwater footprints are narrower than a tsunami’s deep-water wavelength (typically several hundred kilometers), are simply invisible to tsunamis?

Previous Work, and the Aims of this Work

–

Testing the Waters

How big does an island need to be for a tsunami to see it?

Data Collection

Modelling platform

http://depts.washington.edu/clawpack/users/geoclaw.html

http://depts.washington.edu/clawpack/users/quick_tsunami.html

Bathymetry

http://www.niwa.co.nz/our-science/oceans/bathymetry/download

http://www.ngdc.noaa.gov/mgg/inundation/tsunami/inundation.html

http://www.ngdc.noaa.gov/mgg/global/global.html

Tsunami Wave Templates

–

Methodology and Results

Collect wave-buoy data related to historical tsunami warnings. Calibrate models of each of those events (the location of the event, magnitude, polarity, orientation, asymmetries, will all be fitted parameters). The models should reproduce the observed waves, or rather observed non-waves at Tongatapu. Speak to Tonga MET about this. The historical models will then help explain why the waves were not observed.

Would they have been observed if the local bathymetry was different? Try this.

Would they have been observed if the waves were different (location, orientation, etc)? Try this.

–

Conclusions

–

References

Hebert, Schindele, and Heinrich (2001) “Tsunami risk assessment in the Marquesas Islands (French Polynesia) through numerical modelling of generic far-field events”, Natural Hazards and Earth System Sciences, 1, pp233-42. Download pdf here.

Latu (2009) http://pacific.scoop.co.nz/2009/10/out-of-hours-tongan-tsunami-silence-troubles-nation-in-pacific-scare/