After the devastating Indian Ocean tsunami of December 2004, there has been an increased emphasis on providing timely and accurate tsunami warnings to all U.S. coasts and U.S. interests around the world. In response to this need, the National Oceanic and Atmospheric Administration (NOAA) plans to increase the number of its deep-ocean tsunameters (tsunami detectors, also known as Deep-ocean Assessment and Reporting of Tsunamis, or DART, systems) to a total of 39. There are currently six "first generation" tsunameters: two off the Pacific Northwest, three off the Aleutian Islands, and one near the Equator. The new tsunameters will be distributed throughout the Pacific Ocean, the Caribbean Sea, the Atlantic Ocean, and the Gulf of Mexico. (For the latest map of NOAA tsunameter locations, visit the National Data Buoy Center.) A joint workshop attended by representatives of NOAA and the U.S. Geological Survey (USGS) was convened on July 6-7, 2005, in Seattle, WA, to develop a deployment plan for the new tsunameters. Frank González, Tsunami Research Program Leader of NOAA's Pacific Marine Environmental Laboratory (PMEL), set out the central goal of the deployment plan: to determine an optimal network configuration that meets multiple operational, logistical, and mitigation objectives.
NOAA's tsunameters consist of a bottom-pressure recorder (BPR) that measures the tsunami and a surface buoy that communicates the information to Tsunami Warning Centers. At the heart of the BPR is a quartz-crystal pressure transducer that can detect sea-level changes of less than 1 mm for waves with periods greater than 1 minute. Tsunami waves generally have periods ranging from 5 to 60 minutes, whereas surface waves and swells produced by wind generally have periods ranging from 1 to 30 seconds. The BPR is positioned in water several thousands of meters deep to eliminate pressure changes induced by wind-generated, short-period waves that rapidly attenuate with water depth. The surface buoy receives data from the BPR through acoustic telemetry and sends the information to ground stations by way of a Geostationary Operational Environmental Satellite (GOES) link. The next-generation DART II systems developed by PMEL engineers include a two-way communications capability that allows ground stations to remotely trigger DART transmission of the BPR data. For more information about sampling rates, detection algorithms, and other specifications, visit the DART Web site.
Following introductory comments by Eddie Bernard, PMEL director, and Sam Johnson, chief scientist of the USGS Western Coastal and Marine Geology Team, workshop participants set out to define the set of criteria and objectives for the new DART network. First and foremost, these objectives include the operational objectives of the Tsunami Warning Centers, which were described by Paul Whitmore, director of the National Weather Service's West Coast & Alaska Tsunami Warning Center. Logistical constraints, such as ship availability and long-term maintenance of the almost-global DART network, were described by Shannon MacArthur of NOAA's National Data Buoy Center. Tools to aid in optimizing the DART deployment plan were presented by Hal Mofjeld (PMEL). These tools include the Nonlinear Optimization for Mixed vAriables and Derivatives (NOMAD) software developed, in part, by John Dennis, an optimization expert from Rice University. Graphical tools that display normalized tsunami energy, traveltimes, and other information, developed by Mick Spillane (PMEL), also proved to be extremely useful during the course of the workshop deliberations.
The primary role of the USGS participants was to specify tsunami sourcesgeologic phenomena likely to trigger tsunamisfor optimal positioning of DART stations. Tsunami triggers that were considered included earthquakes, landslides, and volcanoes. The primary source parameters used in the optimization process are location, potential size of the tsunami (in terms of both wave amplitude and aerial distribution of tsunami energy), and relative likelihood of occurrence. For subduction-zone earthquakes, which are the main source of far-field tsunamis, a necessary constraint is to place the DART station at a sufficient distance from the earthquake source to avoid interference between the tsunami signal and apparent changes in bottom pressure induced by Rayleigh waves (a type of seismic surface wave). On the other hand, the DART station needs to be placed close enough to the source to minimize the time to acquire the first direct evidence of a tsunami from the DART buoy. The goal is to allow sufficient time for evacuation should a damaging tsunami be forecast, and for cancellation of a tsunami warning should a tsunami be assessed as nondestructive.
Background on subduction-zone earthquakes and a generalized probability model were presented by Eric Geist (USGS). The long-term rate at which a given subduction zone produces tsunamis of all sizes is linked to the rate of convergence. Efforts to compile convergence rates at subduction zones on a worldwide basis were described by Ken Hudnut, Southern California Project Chief of the USGS Earthquake Hazards Program. Workshop participants then discussed tsunami sources in specific regions, including the North Pacific subduction zonesoff southern Alaska, the Aleutian Islands, Kamchatka, the Kuril Islands, and Japanwhich have a long history of generating destructive tsunamis. In contrast, the Cascadia subduction zone, off the Pacific Northwest, has only recently been recognized as having the potential to generate oceanwide tsunamis, as elegantly described by Brian Atwater (USGS).
A particularly difficult region to assess is the Caribbean, as described by Uri ten Brink (USGS) and Aurelio Mercado (University of Puerto Rico). All the likely types of tsunami sources occur in the Caribbean, including subduction-zone and backarc earthquakes, landslides from the tilted carbonate platform offshore northern Puerto Rico, and active volcanoes, such as Kick'em Jenny in the Lesser Antilles. Although fewer tsunamis have occurred in the Caribbean than in the Pacific, more fatalities due to tsunamis have occurred in the more densely populated Caribbean than in the U.S. West Coast and Alaska combined.
The discussion of tsunami sources concluded with an enlightening presentation by Chris Newhall of the USGS Volcano Hazards Program. He presented an overview of historical volcanogenic tsunamis with references to such processes as pyroclastic flows, submarine caldera collapses and explosions, and even catastrophic flank failures that are the specific triggers for tsunami generation. This information, in combination with (1) monitoring and further research into these physical mechanisms and (2) hydrodynamic modeling, will be important for evaluating the tsunamigenic potential of volcanoes such as Cumbre Vieja on La Palma in the Canary Islands.
One way to see how all this information fits together is to look at specific tsunami events from the perspective of a watchstander in a tsunami-warning center. Several cases described by Vasily Titov (PMEL) were particularly illuminating. For example, when an earthquake occurs in the Aleutians, someone on watch can use the initial earthquake information (epicenter and magnitude) to "forward-model" the tsunami using a precomputed database of tsunami-propagation scenarios that are built on source characterizations such as those described at the workshop. Then, once the data from DART stations are received, the occurrence of a tsunami can be verified, and the source can be updated to produce the best fit to the observed tsunami waveform. The propagation calculations are then carried forward to interface with a high-resolution inundation model for vulnerable coastal communities. This entire process can be carried out by the operational Tsunami Forecast System currently under development by PMEL.
Doug Luther (University of Hawai'i) discussed the importance of developing multiuse strategies to garner the support needed from other scientific communities to ensure the future sustainability of the DART network. Not only will there be manifold oceanographic applications of the DART data, but collaborations with the broader oceanographic community are also needed to fully utilize and share the platform and ship-time resources of DART with other large oceanographic projects. Participants recognized that this workshop might provide a good first step in connecting the tsunami and broader oceanographic communities, as well as in determining how DART networks might be established as part of emerging tsunami-warning systems around the world. Results from the workshop will soon be published as a joint NOAA/USGS Special Report.
For more information on tsunami warnings and preparedness, visit the NOAA Tsunami Web site.
in this issue:
Workshop on DART Network for Tsunami Forecasting