25.01.2015 Views

IASPEI - Picture Gallery

IASPEI - Picture Gallery

IASPEI - Picture Gallery

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

IUGG XXIV General Assembly July 2-13, 2007 Perugia, Italy<br />

(S) - <strong>IASPEI</strong> - International Association of Seismology and Physics of the Earth's<br />

Interior<br />

JSS002 Oral Presentation 1771<br />

Standards and benchmarks for tsunami modeling<br />

Prof. Utku Kanoglu<br />

Department of Engineering Sciences Middle East Technical University <strong>IASPEI</strong><br />

Costas Synolakis, Eddie Bernard, Vasily Titov, Frank Gonzlez<br />

A substantial number of new tsunami codes, more than 20 by some accounts, have been employed in<br />

the aftermath of the December 26, 2004 Boxing Day tsunami. Predictions from these new untested<br />

models are being presented at international meetings even though results differ from established<br />

paleotsunami measurements by factors as large as four. These substantial differences have attracted<br />

press attention, requiring hazard mitigation professionals to provide public explanations of the<br />

discrepancies. Since an increased number of nations in tsunami-prone regions around the world must<br />

develop tsunami mitigation plans, both validation and verification are essential for all numerical models<br />

used in planning. Validation refers to the process of ensuring that the model solves the parent equations<br />

of motion accurately, and is accomplished through comparison with analytical solutions. Verification<br />

refers to the process of ensuring that the model represents geophysical reality, and is accomplished<br />

through comparison with both laboratory measurements and field measurements. The gold standard for<br />

inundation numerical codes must include validation, verification, and the essential step of formal<br />

documentation in the peer-reviewed scientific literature. This process must be considered continuous; it<br />

is essential that even proven models be routinely reviewed and subjected to additional testing and<br />

documentation as new knowledge and data are acquired. So far, only a few existing numerical models<br />

have met these standards, and such models must be the only choice for inundation forecast modeling;<br />

the consequences of failure are too serious to risk a lower standard. A forecast system also requires<br />

accurate tsunami measurements and well-tested numerical procedures for inverting the measurements<br />

to produce predictions with a proven model. Clearly, however, there is no absolute certainty that even a<br />

properly validated and verified numerical code will produce accurate, real-time inundation forecasts; this<br />

is why multiple models should be implemented to improve forecast system robustness. We discuss<br />

analytical, laboratory, and field benchmark tests by which tsunami models can be validated, verified and<br />

formally documented in peer-reviewed journals indexed by ISI. This painstaking process may appear<br />

onerous, but is the only defensible methodology when human lives are at stake. Model standards and<br />

procedures that reflect this approach are being adopted for implementation in the U.S. Tsunami<br />

Forecasting System that is under development by the National Oceanic and Atmospheric Administration.<br />

Keywords: tsunami modeling, standards benchmarks, tested model

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!