DART Stations Research Articles

Modeling the 2022 Tonga Eruption Tsunami Recorded on Ocean Bottom Pressure and Tide Gauges Around the Pacific

AbstractTsunamis generated by the Hunga Tonga–Hunga Ha’apai volcanic eruption on January 15, 2022 were recorded on ocean bottom pressure and tide gauges around the Pacific Ocean, earlier than the expected arrival times calculated by tsunami propagation speed. Atmospheric waves from the eruption were also recorded globally with propagation speeds of ~ 310 m/s (Lamb wave) and 200–250 m/s (Pekeris wave). Previous studies have suggested that these propagating atmospheric waves caused at least the initial part of the observed tsunami. We simulated the tsunamis generated by the propagation of the Lamb and Pekeris waves by adding concentric atmospheric pressure changes. The concentric sources are parameterized by their propagation speeds, initial atmospheric wave amplitudes that decay with the distance, and a rise time. For the Lamb wave, inversions of the observed tsunami waveforms at 14 U.S. and nine New Zealand DART stations indicate the start of the positive rise at 4:16 UTC, the peak amplitude of 383 hPa, and the propagation speed of 310 m/s, assuming a rise time of 10 min. The later phases of the observed tsunami waveforms can be better reproduced by adding another propagating concentric wave (Pekeris wave) with a negative amplitude (− 50 hPa) and propagation speeds of 200–250 m/s. The DART records around the Pacific indicate that the Pekeris wave speed is faster toward the northwest and slightly slower toward the northeast. The synthetic waveforms roughly reproduced the far-field tsunami waveforms recorded at tide gauge stations, including the later phases, suggesting that the large amplitude in the later phase may be due to the coupling of the Pekeris wave and the tsunami, as well as resonance around tide gauge stations.

Potential tsunami hazard of the southern Vanuatu subduction zone: tectonics, case study of the Matthew Island tsunami of 10 February 2021 and implication in regional hazard assessment

Abstract. The Vanuatu subduction zone (VSZ) is known to be seismically very active, due to the high convergence rate between the Australian and Pacific tectonic plates for the majority of the margin. However, this is not the case on its southernmost part south of latitude 22.5∘ S and east of longitude 170∘ E, which is neither highly tectonically active nor has it produced large tsunamis over the past 150 years. It has also not been widely studied. On the 11 February 2021 (10 February UTC), a magnitude Mw 7.7 earthquake triggered a tsunami warning in New Caledonia and Vanuatu 20 min after midnight (local time). With an epicentre located close to the volcanic islands of Matthew and Hunter, this shallow reverse-faulting rupture (< 30 km depth) was able to deform the seabed and produce a tsunami. This was confirmed 45 min later by the coastal gauges of the Loyalty and the south Vanuatu islands, which recorded the first tsunami waves. Showing a typical recorded amplitude of less than 1 m, with a maximum of ∼ 1.5 m in Lenakel (Tanna, Vanuatu), it was observed on most coastal gauges and DART stations in the south-west Pacific region as far as Tasmania to the south and Tuvalu to the north at distances of ∼ 3000 and ∼ 1800 km from the epicentre. In this study, the tsunamigenic potential of the southernmost part of the VSZ and the implications in terms of regional hazard assessment are discussed through (1) the presentation of the complex tectonic settings of this “transition zone” between the Solomon–Vanuatu and the Tonga–Kermadec trenches, (2) the case study of the 10 February 2021 tsunami at a south-west Pacific regional scale using three different tsunami generation scenarios computed with the COMCOT modelling code on a set of 48 nested bathymetric grids, and (3) the simulation of a plausible Mw 8.2 scenario encompassing the active part of this “transition zone”. The validation of the Mw 7.7 parameters for tsunami modelling provides the means to further assess the hazard from potential tsunamis triggered by higher magnitude earthquakes in this region. Tsunami records highlight that > 28 cm wave amplitudes were recorded at eight different coastal gauges, including one with an amplitude of more than 1 m (Lenakel, Tanna, Vanuatu). The tsunami threat at that location would be large enough to warrant an onshore evacuation. Finally, it helps to highlight the significant role played by the numerous submarine features in the region, the Norfolk Ridge being the most important, which acts like a waveguide from the north to the south.

The Volcanogenic Tsunami on January 15, 2022, Based on the Records of Deep-Ocean DART Stations

The Volcanogenic Tsunami on January 15, 2022, Based on the Records of Deep-Ocean DART Stations

Optimizing a Model of Coseismic Rupture for the 22 July 2020MW7.8 Simeonof Earthquake by Exploiting Acute Sensitivity of Tsunami Excitation Across the Shelf Break

AbstractThe Shumagin seismic gap along the Alaska Peninsula experienced a major,MW7.8, interplate thrust earthquake on 22 July 2020. Several available finite‐fault inversions indicate patchy slip of up to 4 m at 8–48 km depth. There are differences among the models in peak slip and absolute placement of slip on the plate boundary, resulting from differences in data distributions, model parameterizations, and inversion algorithms. Two representative slip models obtained from inversions of large seismic and geodetic data sets produce very different tsunami predictions at tide gauges and deep‐water pressure sensors (DART stations), despite having only secondary differences in slip distribution. This is found to be the result of the acute sensitivity of the tsunami excitation for rupture below the continental shelf in proximity to an abrupt shelf break. Iteratively perturbing seismic and geodetic inversions by constraining fault model extent along dip and strike, we obtain an optimal rupture model compatible with teleseismicPandSHwaves, regional three‐component broadband and strong‐motion seismic recordings, hr‐GNSS time series and static offsets, as well as tsunami recordings at DART stations and regional and remote tide gauges. Slip is tightly bounded between 25 and 40 km depth, the up‐dip limit of slip in the earthquake is resolved to be well‐inland of the shelf break, and the rupture extent along strike is well‐constrained. The coseismic slip increased Coulomb stress on the shallow plate boundary extending to the trench, but the frictional behavior of the megathrust below the continental slope remains uncertain.

Application of an Asymptotic Solution of the Problem of Linear Wave Propagation on Water to the Approximation of Tsunami Mareograms of 2011 Obtained at Two DART Stations

Application of an Asymptotic Solution of the Problem of Linear Wave Propagation on Water to the Approximation of Tsunami Mareograms of 2011 Obtained at Two DART Stations

A Comparative Analysis of Coastal and Open-Ocean Records of the Great Chilean Tsunamis of 2010, 2014 and 2015 off the Coast of Mexico

The three great earthquakes off the coast of Chile on 27 February 2010 (Maule, Mw 8.8), 1 April 2014 (Iquique, Mw 8.2) and 16 September 2015 (Illapel, Mw 8.3) generated major transoceanic tsunamis that spread throughout the Pacific Ocean and were measured by numerous coastal tide gauges and open-ocean DART stations. Statistical and spectral analyses of the tsunami waves from the events recorded on the Pacific coast of Mexico enabled us to estimate parameters of the waves along the coast and to compare statistical features of the events. We also identified three coastal “hot spots” (sites having maximum tsunami risk): Puerto Angel, Puerto Madero and Manzanillo. Based on the joint spectral analyses of the tsunamis and background noise, we have developed a method for using coastal observations to determine the underlying spectrum of tsunami waves in the deep ocean. The “reconstructed” open-ocean tsunami spectra are in close agreement with the actual tsunami spectra evaluated from direct analysis of the DART records offshore of Mexico. We have further used the spectral estimates to parameterize the energy of the three Chilean tsunamis based on the total open-ocean tsunami energy and frequency content of the individual events.

Comparison of solutions of a problem of Cauchy–Poisson type under discontinuous and smooth initial values with the marigrams of Tsunami 2011 obtained from DART stations

The problem of Cauchy–Poisson type is considered in the framework of potential tsunami model with impulse cylindical compactly supported source discontinuous with respect to the radial horizontal coordinate under the assumption that the depth of the liquid is constant. An analytical solution of the problem for a special choice of values of the parameters of the source is given. The connection the solution thus obtained with nonstandard characteristics introduced by Maslov is discussed. The graphs of the time history of the free surface elevation are compared with the marigrams, received for the 2011 tsunami from the DART stations 21418, 21413, and 51407, and also with graphs calculated for the well-known model with a “simple” source.

Joint modeling of teleseismic and tsunami wave observations to constrain the 16 September 2015 Illapel, Chile, Mw 8.3 earthquake rupture process

AbstractThe 16 September 2015 Illapel, Chile, Mw 8.3 earthquake ruptured ~170 km along the plate boundary megathrust fault from 30.0°S to 31.6°S. A patch of offshore slip of up to 10 m extended to near the trench, and a patch of ~3 m slip occurred downdip below the coast. Aftershocks fringe the large‐slip zone, extending along the coast from 29.5°S to 32.5°S between the 1922 and 1971/1985 ruptures. The coseismic slip distribution is determined by iterative modeling of teleseismic body waves as well as tsunami signals recorded at three regional DART stations and tide gauges immediately north and south of the rupture. The tsunami observations tightly delimit the rupture length, suppressing bilateral southward extension of slip found in unconstrained teleseismic‐wave inversions. The spatially concentrated rupture area, with a stress drop of ~3.2 MPa, is validated by modeling DART and tide gauge observations in Hawaii, which also prove sensitive to the along‐strike length of the rupture.

Application of array backprojection to tsunami prediction and early warning

AbstractTeleseismic and static geodetic data have weak constraints on the offshore slip while tsunami data are limited by their availability, so predictions of tsunami waves in the near‐field remain challenging. In this study, we develop a near‐field tsunami prediction approach based on seismic array backprojections (BP). In this approach, the rupture area is first estimated by enclosing the BP radiators. Then slip models with uniform slip are constructed based on statistical scaling relations between rupture area and seismic moment to predict the near‐field tsunami waveforms. The method is applied to the 2011 Tohoku, 2014 Iquique, and 2015 Illapel tsunami events, and the model predictions are compared with tsunami recordings at 57 tidal gauges and nine DART stations. Results show that the average error of arrival time and amplitude nearshore is approximately −15 to +5 min and 0.5 m, respectively, which are sufficiently small for tsunami warning purposes.

Short-term forecast of tsunami occurred on April 1, 2014 on the Kuril Islands coast

The results of forecasting the Chilean tsunami in the Pacific Ocean on April 1,2014are presented. For the first time in Russia the forecast was prepared in the near real-time mode 9.5-10.5 hours before the tsunami attacked the Russian coast. A good agreement was obtained with the tsunami forms registered by DART stations along the US West Coast and the Aleutian and Kuril Islands. The information about the expected tsunami meets the requirements to the short-term tsunami forecast formulated by UNESCO Intergovernmental Oceanographic Commission.

Observations and Numerical Modeling of the 2012 Haida Gwaii Tsunami off the Coast of British Columbia

A major (M w 7.7) earthquake occurred on October 28, 2012 along the Queen Charlotte Fault Zone off the west coast of Haida Gwaii (formerly the Queen Charlotte Islands). The earthquake was the second strongest instrumentally recorded earthquake in Canadian history and generated the largest local tsunami ever recorded on the coast of British Columbia. A field survey on the Pacific side of Haida Gwaii revealed maximum runup heights of up to 7.6 m at sites sheltered from storm waves and 13 m in a small inlet that is less sheltered from storms (Leonard and Bednarski 2014). The tsunami was recorded by tide gauges along the coast of British Columbia, by open-ocean bottom pressure sensors of the NEPTUNE facility at Ocean Networks Canada’s cabled observatory located seaward of southwestern Vancouver Island, and by several DART stations located in the northeast Pacific. The tsunami observations, in combination with rigorous numerical modeling, enabled us to determine the physical properties of this event and to correct the location of the tsunami source with respect to the initial geophysical estimates. The initial model results were used to specify sites of particular interest for post-tsunami field surveys on the coast of Moresby Island (Haida Gwaii), while field survey observations (Leonard and Bednarski 2014) were used, in turn, to verify the numerical simulations based on the corrected source region.

Tsunami surges around the Hawaiian Islands from the 1 April 2014 North Chile Mw 8.1 earthquake

AbstractThe 1 April 2014 Iquique Mw 8.1 earthquake ruptured a segment of the megathrust fault offshore of northern Chile and generated a moderate‐size tsunami across the Pacific. Tide gauges in Hawaii recorded over 1 m of wave height despite the long distance from the source and position away from the main radiated energy lobe. Inversion of global teleseismic body waves combined with forward modeling of the tsunami at four near‐field DART stations arrives iteratively at a self‐consistent finite‐fault model with very compact dimensions. The slip distribution produces a NNE‐SSW trending seafloor uplift patch that enhances the tsunami directionality in the WNW, resulting in good matches to observed DART and tide gauge records around the Hawaiian Islands. The relatively large waves at selected locations in Hawaii can be attributed to a combination of the spatial slip distribution and the resulting short‐period waves that triggered localized resonance over the insular shelves. This event highlights the importance of characterizing detailed slip distributions in analysis or forecasting of tsunamis even for a compact source.

Characteristics of the 2011 Great Tohoku Tsunami on the Russian Far East Coast: Deep-Water and Coastal Observations

The source region of the catastrophic Tohoku tsunami of 11 March 2011 was not far from the Russian Far East coast. The Sakhalin Tsunami Warning Center at Yuzhno-Sakhalinsk issued an Alert for threatened coasts of the Kuril Islands and Kamchatka; the observed tsunami heights were up to 2–2.5 m along the Pacific coast of the Kuril Islands. The tsunami was clearly recorded by a number of coastal tide gauges and by deep-ocean bottom pressure stations located in the vicinity of the Kuril Islands, as well as by the Russian open-ocean DART station 21401, located eastward from the South Kuril Islands. The data from other DART stations, in particular those located near Japan and the Aleutian Islands, were used for comparison. The records from these instruments were used to estimate the major characteristics of the tsunami waves, including arrival times, maximum heights, durations of the signal and main wavelength periods, as well as for comparison with the results of numerical modelling. In contrast to deep-sea stations where the first waves were the highest, at Russian coastal sites the highest waves occurred several hours after the arrival of the first tsunami wave. Further analysis indicated significant differences in spectral characteristics of tsunami waves propagating eastward toward North America and those directed in the northeast direction towards the Russian Far East. The main peaks of the eastward propagating tsunami waves were relatively high-frequency (periods ranging from 6–8 to 15–20 min), while those propagating in the northeast direction were mainly low-frequency (ranging from 25–40 to 60–80 min). Additionally, pronounced spectral peaks with similar long periods were found in the “middle-field” records at Hanasaki (on the northeastern coast of Hokkaido Island), at Yuzhno-Kurilsk (Kunashir Island) and at Malokurilsk (Shikotan Island). At remote stations, the resonant periods associated with local topographic features were predominant in the spectra.

The open ocean energy decay of three recent trans‐Pacific tsunamis

The 2009 Samoa (Mw 8.1), 2010 Chile (8.8), and 2011 Tohoku (9.0) earthquakes generated destructive tsunamis recorded by a large number of DART stations in the Pacific Ocean. High‐resolution (15 s) DART records yield mean energy decay times for these events of 17.3, 24.7, and 24.6 h, respectively. We attribute these differences to the frequency content of the tsunamis. Specifically, the Samoa tsunami was a “high‐frequency” event with periods of 2–30 min whereas the Chile and Tohoku tsunamis were “broad‐band” events with periods of 2–180 min. Differences in frequency content are linked to differences in the source parameters: Samoa was a relatively small deep‐water earthquake while Chile and Tohoku were extensive shallow‐water earthquakes. Frequency‐dependent analysis of the Chile and Tohoku tsunamis indicates that shorter period waves attenuate much faster than longer‐period waves (decay times range from 15 h for 2–6 min waves to 29 h for 60–180 min waves).

The 2010 Chilean Tsunami Off the West Coast of Canada and the Northwest Coast of the United States

The major (M w = 8.8) Chilean earthquake of 27 February 2010 generated a trans-oceanic tsunami that was observed throughout the Pacific Ocean. Waves associated with this event had features similar to those of the 1960 tsunami generated in the same region by the Great (M w = 9.5) 1960 Chilean Earthquake. Both tsunamis were clearly observed on the coast of British Columbia. The 1960 tsunami was measured by 17 analog pen-and-paper tide gauges, while the 2010 tsunami was measured by 11 modern digital coastal tide gauges, four NEPTUNE-Canada bottom pressure recorders located offshore from southern Vancouver Island, and two nearby open-ocean DART stations. The 2010 records were augmented by data from seven NOAA tide gauges on the coast of Washington State. This study examines the principal characteristics of the waves from the 2010 event (height, period, duration, and arrival and travel times) and compares these properties for the west coast of Canada with corresponding properties of the 1960 tsunami. Results show that the 2010 waves were approximately 3.5 times smaller than the 1960 waves and reached the British Columbia coast 1 h earlier. The maximum 2010 wave heights were observed at Port Alberni (98.4 cm) and Winter Harbour (68.3 cm); the observed periods ranged from 12 min at Port Hardy to 110–120 min at Prince Rupert and Port Alberni and 150 min at Bamfield. The open-ocean records had maximum wave heights of 6–11 cm and typical periods of 7 and 15 min. Coastal and open-ocean tsunami records revealed persistent oscillations that “rang” for 3–4 days. Tsunami energy occupied a broad band of periods from 3 to 300 min. Estimation of the inverse celerity vectors from cross-correlation analysis of the deep-sea tsunami records shows that the tsunami waves underwent refraction as they approached the coast of Vancouver Island with the direction of the incoming waves changing from an initial direction of 340° True to a direction of 15° True for the second train of waves that arrived 7 h later after possible reflection from the Marquesas and Hawaiian islands.

An approximate method of short-term tsunami forecast and the hindcasting of some recent events

Abstract. The paper presents a method for a short-term tsunami forecast based on sea level data from remote sites. This method is based on Green's function for the wave equation possessing the fundamental property of symmetry. This property is well known in acoustics and seismology as the reciprocity principle. Some applications of this principle on tsunami research are considered in the current study. Simple relationships and estimated transfer functions enabled us to simulate tsunami waveforms for any selected oceanic point based only on the source location and sea level data from a remote reference site. The important advantage of this method is that it is irrespective of the actual source mechanism (seismic, submarine landslide or other phenomena). The method was successfully applied to hindcast several recent tsunamis observed in the Northwest Pacific. The locations of the earthquake epicenters and the tsunami records from one of the NOAA DART sites were used as inputs for the modelling, while tsunami observations at other DART sites were used to verify the model. Tsunami waveforms for the 2006, 2007 and 2009 earthquake events near Simushir Island were simulated and found to be in good agreement with the observations. The correlation coefficients between the predicted and observed tsunami waveforms were from 0.50 to 0.85. Thus, the proposed method can be effectively used to simulate tsunami waveforms for the entire ocean and also for both regional and local tsunami warning services, assuming that they have access to the real-time sea level data from DART stations.

Effective resolution and drift of Paroscientific pressure sensors derived from long‐term seafloor measurements

Starting with the installation of the NOAA tsunami early warning system DART in 1983, the Paroscientific Digiquarz® Broadband Depth Sensor has been deployed for long‐term ocean bottom pressure measurements in numerous marine investigations. The sensor turned out to be very reliable with a good signal repeatability within a broad measuring range and high resolution. However, up to now its long‐term drift, noise level, and effective resolution under in situ conditions at the seafloor are only known from a few published studies and for a few sensors. In this study we analyze 118 long‐term seafloor pressure time series (longer than 2 months and up to a maximum time period of 9 years) to investigate effective resolution and long‐term drift under in situ conditions. The data are from DART, IFM‐GEOMAR, LOLEM, and CORK stations. The noise level of the data and therefore the mean sensor resolution are lower than an upper threshold of 20.7 Pa (corresponding to 2.07 mm equivalent water level) for stations installed directly on the seafloor. The noise level of CORK sensors is much larger, with a median value of 127 Pa. Long‐term sensor drift is assumed to be composed of an initial exponential part and a subsequent linear drift. In situ mean drift is −0.88 ± 0.73 kPa/a, determined from all available data with a slight increase of drift with deployment depth. For the first time, we are able to quantify effective pressure resolution and drift of the widely used Paroscientific pressure gauges under in situ conditions. Our results provide important constrains for the interpretation of seafloor pressure records. However, in situ self‐calibration procedures for the Paroscientific sensors are still needed in order to benefit from high effective resolution and long‐term stability and also to increase accuracy especially for the monitoring of geodynamic processes at the seafloor.

Near-source observations and modeling of the Kuril Islands tsunamis of 15 November 2006 and 13 January 2007

Abstract. Two major earthquakes near the Central Kuril Islands (Mw=8.3 on 15 November 2006 and Mw=8.1 on 13 January 2007) generated trans-oceanic tsunamis recorded over the entire Pacific Ocean. The strongest oscillations, exceeding several meters, occurred near the source region of the Kuril Islands. Tide gauge records for both tsunamis have been thoroughly examined and numerical models of the events have been constructed. The models of the 2006 and 2007 events include two important advancements in the simulation of seismically generated tsunamis: (a) the use of the finite failure source models by Ji (2006, 2007) which provide more detailed information than conventional models on spatial displacements in the source areas and which avoid uncertainties in source extent; and (b) the use of the three-dimensional Laplace equation to reconstruct the initial tsunami sea surface elevation (avoiding the usual shallow-water approximation). The close agreement of our simulated results with the observed tsunami waveforms at the open-ocean DART stations support the validity of this approach. Observational and model findings reveal that energy fluxes of the tsunami waves from the source areas were mainly directed southeastward toward the Hawaiian Islands, with relatively little energy propagation into the Sea of Okhotsk. A marked feature of both tsunamis was their high-frequency content, with typical wave periods ranging from 2–3 to 15–20 min. Despite certain similarities, the two tsunamis were essentially different and had opposite polarity: the leading wave of the November 2006 trans-oceanic tsunami was positive, while that for the January 2007 trans-oceanic tsunami was negative. Numerical modeling of both tsunamis indicates that, due to differences in their seismic source properties, the 2006 tsunami was more wide-spread but less focused than the 2007 tsunami.