S51C-01
An Extension to Short Distances of Real-Time Estimators of Seismic Sources
Working from a data set of seismograms from 61 large events recorded at epicentral distances between 5 and 35 degrees, we assess the validity at short distances of the teleseismic (35 to 80 deg.) method of Newman and Okal (1998) using the ratio of estimated energy to moment (theta) to characterize a seismic source in real time. An evaluation of short-distance data from 17 events common to this study and that of Newman and Okal shows good agreement in calculated theta values. Deficient values of theta in this study are observed for the December 2004 N. Sumatra (theta -6.82), June 1994 S. of Java (-6.70), September 1992 Nicaragua (-6.44), July 2006 S. of Java (-6.04), and November 2006 Kuril Island (-6.00) events; all of these have been previously recognized as having unusual sources. Other notable events with slightly less-deficient theta values are June 2001 Peru (-5.81) and February 1996 Peru (-5.67). For this data set we also present an extension of the discriminant tau method of Reymond et al. (AGU, 2006) based on evaluation of the envelope of 2-4 Hz bandpass filtered P arrival windows. Preliminary results show that both theta and tau discriminants may be used to improve real-time identification of seismic sources with the potential to generate anomalously large tsunamis.
S51C-02
Tsunami Runup in the Middle Kuril Islands from the Great Earthquake of 15 Nov 2006
Two expeditions to the middle Kuril Islands [IMGG FED RAS, NSF Kurils Biocomplexity Project] in the summer of 2007 yielded tsunami runup and inundation measurements from the 15 Nov 2006 Mw 8.3 subduction-zone earthquake, and possibly from the 13 Jan 2007 Mw 8.1 earthquake seaward of the subduction zone. Both earthquakes produced measurable tsunamis in the far field, the 13 Jan tsunami significantly smaller; the 15 Nov tsunami did some damage in the harbor of Crescent City, CA. Ours are the first near-source measurements because no one lives in the middle Kurils. Moreover, because KBP visited many of the same sites in summer of 2006, we have numerous before-and-after comparisons, including quantified erosion. We measured 120 profiles and made more than 300 runup measurements. We found dramatic tsunami effects of erosion and deposition, with widespread runup of 8-12 m, up to about 20 m, between and including Simushir and Matua islands. In most cases, we measured runup with a transit and surveying rod, producing a topographic profile from sea level to the slope above runup indicators; in some cases, we used a hand level and tape. Runup/inundation criteria were generally subhorizontal lines of floatable debris, typically wood, plastic, glass floats, and styrofoam. Single occurrences, e.g., of a plastic bottle were not considered adequate. Corroborative evidence, not used independently, included limits of consistently oriented stems of tall grasses and flowers, limit of sand and gravel deposits above turf and dead vegetation, and elevation of fresh erosion of turf from slopes landward of the beach plain. Currently we are compiling, correcting and vetting our measurements, which will be submitted to online databases. Topographic profiles obviously had an effect on the data, with short, steep profiles generating high runup and short inundation; most beach-ridge profiles had longer inundation and shorter runup. However, at Ainu Bay on Matua Island, we found as much as 18-20 m of runup at the landward limit of inundation 350-400 m inland. Our survey did not include islands north of Matua; the next major island, Shiashkotan, is almost 100 km away, so it is hard to predict effects there. Some preliminary results [m runup height at landward limit, m inundation distance]: Matua, Pacific coast [6-17, 50-100]; Matua, South Bay [6-8, 100-220]; Matua, Ainu Bay [12-20, 70-400]; Rasshua, Okhotsk coast, [5-10; 50-65]; Ryponkicha, Pacific coast [10-11, 45-55]; Ketoi south coast [6-9; 45-65]; NE Simushir [7-19; 50-150]; SE Simushir [5-7; 80-140].
S51C-03
Deep Ocean Tsunami Detection: Third Generation DART
Deep Ocean Assessment and Reporting of Tsunamis (DART) buoys have demonstrated their value in providing the appropriate data to forecast tsunamis in real time. The NOAA experimental tsunami forecast system has used real-time DART data with impressive accuracies for the Tonga Island, May 2006, Kurile Island, November 2006, Kurile Island, January 2007, Solomon Island April, 2007, and Peru, August 2007 tsunamis. Recent developments in DART technology hold promise for reduced costs to operate and maintain the current array of 35 DART buoys. Details of the new technology and costs savings will be presented.
S51C-04
Assessing tsunami hazard to the U.S. East Coast using relationships between submarine landslides and earthquakes
Our knowledge of the size distribution and recurrence interval of earthquakes is better than that for submarine landslides. Because approximately 90% of landslide-generated tsunamis worldwide are associated with earthquakes, it is useful to try and predict the size and recurrence of submarine landslides from those of earthquakes. Several approaches for estimating the potential of earthquake-induced landslide tsunamis along the U.S. Atlantic continental margin are investigated and compared: calculating catastrophic slope failure conditions on this margin due to horizontal acceleration by earthquakes, using compilations of maximum observed distance to liquefaction [Ambraseys, 1988], and using compilations for the maximum area affected by earthquake-induced subaerial landslides [Keefer, 1984; Rodriguez et al., 1999]. The calculated slope stability predicts similar or slightly smaller triggering distances and failure areas than for subaerial observations. The results from all 3 approaches compare well with the slope failure observations of the M=7.2, 1929 Grand Banks earthquake. In terms of maximum triggering distance, these approaches suggest that an M=7.5 earthquake (the largest expected earthquake along the U.S. east coast) must be within 100 km of the continental slope to induce a catastrophic slope failure, a location which almost everywhere is offshore. The observed rate of seismicity offshore the U.S. Atlantic coast is very low with the exception of New England, where some micro-seismicity is observed. An extrapolation of annual strain rates from the Canadian Atlantic continental margin suggests that the New England margin may experience the equivalent of a magnitude 7 earthquake on average every 600-3000 years. In terms of minimum triggering magnitude, only earthquakes with magnitude larger than 5.5- 6 could probably cause a sufficiently large submarine failure to generate a devastating tsunami under specific conditions: if the epicenter is optimally located within the vulnerable area, if much of the area indeed fails, and if the slide moves rapidly.
S51C-05
Tsunami awareness saves Solomon Islanders on 1 April 2007
On April 1, 2007 at 20:39:56 UTC (local time: UTC+11), a magnitude Ms 8.1 earthquake occurred 50 km off the New Georgia Islands in the Solomon Sea generating a locally focused tsunami striking more than 300 coastal communities in the Solomon Islands. A reconnaissance team deployed within one week investigated 65 coastal settlements on 13 remote Islands and measured run-up heights of 12 m, local flow depths of 5 m as well as tectonic uplift up to 3.6 m and subsidence down to -1.5m. This South Pacific archipelago's worst disaster since WWII resulted in 52 confirmed death and 36'000 directly affected – roughly half of these numbers are children. The ground shaking pinned people to the ground and palm trees bounced back and forth with leafs touching the ground. The ancestral heritage "run to high ground after an earthquake" passed on to younger generations by survivors of a smaller 1952 tsunami triggered an immediate spontaneous self evacuation, which dramatically reduced the death toll in the small evacuation window of a few minutes between the end of the ground shaking and the onslaught of the tsunami. The survivors remained traumatized by the tsunami, afraid of the sea and living in evacuation camps on the hills illustrating the importance of community-based education and awareness programs.
S51C-06
Understanding tsunami by landslides as the next challenge for hazard, risk and mitigation: Insight from multi-material hydrocode modeling
Sliding bodies, either traveling inside the water column or impacting into water, generate waves. Depending on the volume, velocity, and material properties of the moving mass, waves of significant height and length are generated. Such waves can be considered tsunami waves, even if their characteristic is profoundly different from "classical" tsunami generated by earthquakes. Most models of landslide-induced tsunami waves focus on the propagation of the generated waves as they are non-linear and require the application of higher-order terms in the governing equations. The actual generation process in these models is often simplified and tuned to two specific scenarios that can be realized in laboratory experiments: (1) the sliding body is represented by a rigid or (2) a viscous body. In real landslides, the rheology of the sliding body is unknown and varies between these two end-member scenarios. We present new modeling results to gain a better understanding of the relevant parameters affecting the characteristic of generated tsunami waves by landslides. The applicability of the multi-material hydrocode iSALE we used in our models is demonstrated by reproducing experiments conducted with rigid and viscous sliding bodies. Our results provide new insight of the slide dynamics, the coupling between the sliding body and the slope and the water column, repsectively. They confirm the known positive effect of increased velocity and increased volume of the sliding body on the initial wave characteristics. In addtion we have investigated how the rheological properties of the sliding body affect the wave characteristic.
S51C-07
Probabilistic Tsunami Hazard Assessment for California from Distant Sources
California has a substantial tsunami hazard from distant sources. Tsunamis are infrequent, and while there have been substantial advances in emergency preparedness, few -if any- communities have a quantitative understanding of the risk in terms of return periods or probabilities of exceedance. Other than the early work of Houston and Garcia for flood insurance studies and the pioneering 2006 FEMA report of Geist et al for Seaside Oregon, there have been few quantitative studies on probabilistic hazard assessment. To quantitatively evaluate the farfield tsunami threat, time dependent and time independent methods are used to compute probabilities of exceedance for wave heights offshore. Following NOAA's practice we use Green's function style sources which are then superposed to construct composite fault motions and then calculate the resultant evolution to the region of interest. We will provide results with probabilities of exceedance of specific heights, for ports and harbors in California. Northern California has the highest risk with shorter periods with offshore wave heights >1m every 4-10 years and >4m every 100-500 years - the range of return periods reflects different calculation methods. For reference, the wave height offshore Crescent City during the 15 November 2007 Kuril islands tsunami was ~1m. In Southern California, >1m offshore waves are expected every 22-50 years, and >1.5m waves every 100 to 150 years. Depending on the local coastal topography, these values may imply substantial inundation. http://www.usc.edu/dept/tsunamis
S51C-08
Dynamic Models of Earthquakes and Tsunamis
The abundance of inhabited cities lining the coasts of earthquake-prone regions serves as a motivation to investigate the effects of fault dynamics on tsunami generation. In particular, time dependent tsunami sources are important when investigating tsunami run-up in the near-field. While typical tsunami codes use a simple dislocation model with constant slip as a source, we link a fully dynamic earthquake model to tsunami generation. We construct a finite element model of a fault system based on the splay fault geometry inferred to exist in the Nankai subduction zone off of Japan (Cummins & Kaneda, 2000), and link the results to a finite- difference depth-averaged tsunami propagation code. We find that a simple regional stress field in the splay geometry is improbable since it is nearly impossible to produce a self-consistent model of an earthquake under such a condition. We also find that whether the earthquake ruptures the upper splay in addition to or instead of the main subduction fault depends upon the presence or absence of barriers along the main fault plane. The tsunamis resulting from these two different rupture patterns differ strongly, emphasizing the effect of fault geometry on tsunami generation and near-source run-up.