Search

A performance measure for a DART® tsunami buoy network has been developed. DART® buoys are used to detect tsunamis, but the full potential of the data they collect is realized through accurate forecasts of inundations caused by the tsunamis. The performance measure assesses how well the network achieves its full potential through a statistical analysis of simulated forecasts of wave amplitudes outside an impact site and a consideration of how much the forecasts are degraded in accuracy when one or more buoys are inoperative. The analysis uses simulated tsunami amplitude time series collected at each buoy from selected source segments in the Short-term Inundation Forecast for Tsunamis database and involves a set for 1000 forecasts for each buoy/segment pair at sites just offshore of selected impact communities. Random error-producing scatter in the time series is induced by uncertainties in the source location, addition of real oceanic noise, and imperfect tidal removal. Comparison with an error-free standard leads to root-mean-square errors (RMSEs) for DART® buoys located near a subduction zone. The RMSEs indicate which buoy provides the best forecast (lowest RMSE) for sections of the zone, under a warning-time constraint for the forecasts of 3 h. The analysis also shows how the forecasts are degraded (larger minimum RMSE among the remaining buoys) when one or more buoys become inoperative. The RMSEs provide a way to assess array augmentation or redesign such as moving buoys to more optimal locations. Examples are shown for buoys off the Aleutian Islands and off the West Coast of South America for impact sites at Hilo HI and along the US West Coast (Crescent City CA and Port San Luis CA, USA). A simple measure (coded green, yellow or red) of the current status of the network’s ability to deliver accurate forecasts is proposed to flag the urgency of buoy repair.

U.S. Tsunami Warning Centers use real-time bottom pressure (BP) data transmitted from a network of buoys deployed in the Pacific and Atlantic Oceans to tune source coefficients of tsunami forecast models. For accurate coefficients and therefore forecasts, tides at the buoys must be accounted for. In this study, five methods for coefficient estimation are compared, each of which accounts for tides differently. The first three subtract off a tidal prediction based on (1) a localized harmonic analysis involving 29 days of data immediately preceding the tsunami event, (2) 68 pre-existing harmonic constituents specific to each buoy, and (3) an empirical orthogonal function fit to the previous 25 hrs of data. Method (4) is a Kalman smoother that uses method (1) as its input. These four methods estimate source coefficients after detiding. Method (5) estimates the coefficients simultaneously with a two-component harmonic model that accounts for the tides. The five methods are evaluated using archived data from eleven DART buoys, to which selected artificial tsunami signals are superimposed. These buoys represent a full range of observed tidal conditions and background BP noise in the Pacific and Atlantic, and the artificial signals have a variety of patterns and induce varying signal-to-noise ratios. The root-mean-square errors (RMSEs) of least squares estimates of sources coefficients using varying amounts of data are used to compare the five detiding methods. The RMSE varies over two orders of magnitude between detiding methods, generally decreasing in the order listed, with method (5) yielding the most accurate estimate of source coefficient. The RMSE is substantially reduced by waiting for the first full wave of the tsunami signal to arrive. As a case study, the five method are compared using data recorded from the devastating 2011 Japan tsunami., 20 pages; 3 tables; 15 figures

Low, neap tides occurred along the West Coast of the United States and Canada around 0500 UT January 27, 1700, a time for the last major Cascadia earthquake/tsunami based on teletsunami arrival times in Japan [Satake et al., 1996]. However, high-range perigean spring tides occurred only a week later, illustrating how accurately occurrence times must be known to constrain the region's background water levels during such events. The ranges of background water levels increase northward from 3 m at Monterey to 8.5 m at Queen Charlotte, as estimated from the amplitudes of perigean spring tides and subtidal winter fluctuations.

[1] We have developed a method to compute the total energy transmitted by tsunami waves, to the case where the earthquake source is unknown, by using deep-ocean pressure measurements and numerical models (tsunami source functions). Based on the first wave recorded at the two closest tsunameters (Deep-Ocean Assessment and Reporting of Tsunamis (DART)), our analysis suggests that the March 11, 2011 Tohoku-Oki tsunami generated off Japan originated from a 300–400 km long and 100 km wide area, and the total propagated energy is 3 × 1015J (with 6% uncertainty). Measurements from 30 tsunameters and 32 coastal tide stations show excellent agreement with the forecasts obtained in real time. Our study indicates that the propagated energy and the source location are the most important source characteristics for predicting tsunami impacts. Interactions of tsunami waves with seafloor topography delay and redirect the energy flux, posing hazards from delayed and amplified waves with long duration. Seafloor topography also gives its spectral imprint to tsunami waves. Travel time forecast errors are path-specific and correlated to the major wave scatterers in the Pacific. Numerical dissipation in the propagation modeling highlights the need of high-resolution inundation models for accurate coastal predictions. On the other hand, it also can be used to account for physical dissipation to achieve efficiency. Our results provide guidelines for the earliest reliable tsunami forecast, warnings of long duration tsunami waves signals and enhancement of the experimental tsunami forecast system. We apply the method to quantify the energy of 15 past tsunamis, independently from earthquake magnitudes. The small tsunami to seismic radiation energy ratios, and their variability (0.01–0.8%), reinforce the importance of using deep-ocean tsunami data, the direct measures of tsunamis, for estimates of tsunami energy and accurate forecasting.

We use a scattering index to identify the regions in the Pacific Ocean where topographic features scatter significant tsunami energy. Based on linear wave theory, the index is computed from the Smith/Sandwell topography. Consistent with numerical simulations, it shows that there is a narrow band of strong scatterers running across the ocean from the northwest (Emperor Seamount Chain) to the southeast (Easter Island Fracture Zone). The eastern Pacific is nearly devoid of scatterers, except for this band and isolated features along the eastern margin. To the west of the band lies a region with moderate scattering; the strongest scattering occurs in the southwestern Pacific. The Pacific is rimmed by island arcs and shallow continental shelves that also trap and scatter tsunamis. These results show that numerical models of trans-Pacific tsunamis must resolve the effects of the small-scale topography in order to accurately simulate their wave patterns and amplitudes.

[1] Recent paleoseismic and geologic studies indicate that a tsunami occurred 1100 years ago in Puget Sound. This study aims to reproduce and validate the model tsunami, using a finite difference model based on nonlinear shallow water theory and abathymetry/topography computational grid that takes into account the 1-meter rise in sea level that has occurred in the region during the past 1100 years. Estimates of tsunami height, the extent of inundation, and the current velocity pattern are provided at the northern head of Cultus Bay, Whidbey Island, where paleotsunami sand deposits have been found. The model demonstrates that a tsunami generated when the background water level was at mean high water or above could have surged across the then-existing coastal marsh, penetrated the full length of the shallow bay, and deposited the observed sand layers.

A cross-spectral analysis is made of subtidal (0.01-0.5 cpd) sea level (SSL), atmospheric pressure and wind fluctuations observed during 1966-1988 in the southern reaches of the fjord system formed by the Straits of Juan de Fuca/Georgia and Puget Sound. It shows that observed SSL can be largely explained by quasi-steady inverse barometer compensation to atmospheric pressure and the propagation of adjusted sea level fluctuations from the nearby continental shelf, with some local weather forcing in summer. Variations in coherence, amplitudes and relative phase between seasons, locations and frequencies support previous work that both barotropic and baroclinic processes have important influences in SSL within the fjord system.

An analytic theory is presented for free, barotropic Kelvin waves subject to constant, vertical eddy viscosity on a semi-infinite, f-plane shelf of constant depth bordering a straight, vertical coast. A coastal boundary condition closes the solutions of Sverdrup (1927) and Fjeldstad (1929) by requiring that the component of vertically integrated, volume transport normal to the coast be zero just seaward of coastal boundary layers. This condition on offshore transport is then found to hold everywhere seaward of the coastal layers. One condition on the complex wavenumber components is the dispersion relation in which frictional bottom stress appears in an equivalent depth. The coastal boundary condition then fixes the wavenumber components. The effects of vertical viscosity are found to depend on the frequency ω′ relative to the inertial frequency and on the Ekman number E or, more conveniently, the vertical scale E½ of steady, bottom Ekman layers. Below a frequency dependent on E½, low-frequency K...

We study theoretically the effects of linear (Rayleigh) friction on free internal waves in the equatorial wayeguide. The waves may be vertically propagating or in standing vertical modes. Analytic solutions on a beta-plane show that the meridional scale of the wayeguide becomes significantly larger than the inviscid value only when the frequency is much less than the friction coefficient. In the limit of zero frequency, the meridional scale grows without bound. The amount of zonal damping depends on the wave type. Kelvin, high-frequency Yanai (mixed Rossby-gravity), inertial-gravity and low-wavenumber (non-dispersive) Rossby waves decay relatively slowly while low-frequency Yanai and high-wavenumber (short) Rossby waves are much more strongly attenuated. With friction, transitions to evanescence are spread over frequency bands; waves that had zero group velocity without friction are zonally evanescent with friction. Off the equator, friction changes amplitude nodes into non-zero minima and smooth...

Bottom pressure observations (November 1981 to August 1982) show that the diurnal tides dominate the outer reaches of the northeastern Bering Sea shelf. The diurnal tides decrease inshore from the shelf break (maximum K1 amplitude of 38 cm) and appear to be evanescent Sverdrup (Poincare) waves on the outer shelf. The K1 amplitude (2.3 cm) in the Bering Strait is 6% of that at the shelf break. The diurnal phase lags vary little (322°–359°G for the K1 tide) between the stations. North and eastward from the shelf break, the tides become smaller and more semidiurnal. The semidiurnal amplitudes vary considerably between stations; these observations show the effects of semidiurnal amphidromic systems predicted by numerical models. The M2 amplitude (7.6 cm) in the Bering Strait is about a third of that at the shelf break (M2 amplitude of 23 cm). The semidiurnal phase lags are earliest at the shelf break (87°G for M2) and appear to increase toward St. Lawrence Island (336°G for M2 in Shpanberg Strait) and the Bering Strait (202°G for M2). The tidal harmonic constants fluctuated seasonally, possibly in response to the variations in ice cover. The diurnal amplitudes generally increased during winter, while the semidiurnal amplitudes decreased. At the shelf break the K1 amplitude increased 2.5 cm or 7%, but the O1 and semidiurnal M2 amplitudes (0.4 cm or 2% for M2) showed little variation. In Bering Strait the variations in M2 amplitude (up to 3.0 cm) amounted to 37% of the mean amplitude; the decrease may be due to attenuation of the M2 tide from the Arctic Ocean in addition to damping of the M2 tide on the northeastern Bering Sea shelf. Shorter-term fluctuations were also observed in the tides. The tidal fluctuations may contribute to those observed in the deep sea. Numerical models show a set of amphidromic systems (including a semidiurnal system off Cape Navarin) and rapid phase transitions in the vicinity of some of the stations; these models are in relatively good agreement with the observations.