Abstract: More than 300 earthquakes have been reported by the Colombian Seismic Network (RSNC) in a seismic zone in southwestern Colombia since 2012 ranging from Ml 0.8 up to Ml 5.3, with at least 3 M4.5+ earthquakes in a span of 1 month in 2016. Some of these events have been widely felt in major cities in Colombia and are well recorded by the local network. We study the temporal and spatial evolution of these earthquakes, and in particular focus in the short seismic sequence in 2016. Earthquakes show migration towards the SW, along an apparent fault trace and show clear Omori-type aftershock behavior. Using an empirical Green’s function approach, we study the changes in source parameters of the seismic sequence and try to identify if ruptures show any preferential directivity. Temporal or spatial variations of these source parameters may be indicative of changes in the medium before and after the mainshock.
Poster:
poster_AA
Abstract: Mexico City’s building code considers rules to include the bidirectional building response in the design process. The so-called α combination rules account for the impact of 100% and α times the ground motion acting in orthogonal directions. Thus, it allows estimating the forces a structure will resist. Here, based on fully 3D ground motion simulations including the soil-structure interaction, we present a preliminary analysis of the uncertainty in the α parameter in soft soil deposits in Mexico City. The study is performed for a typical structure located in the Lake zone coupled with a realistic velocity model of the basin and a detailed model of the building. The seismic wave propagation and the building response is performed using the Finite Element Method for eight scenarios, mostly subduction earthquakes. We observe that differences in the optimal α parameter estimated per simulation are not adequate to establish a preferred value.
Poster:
Poster_59_JoaquinSanchez
Abstract: With the exception of isolated and largely near-shore deployments of ocean-bottom seismometers (OBSs), most seismic instrumentation is located on land, although two-thirds of Earth’s surface is covered with oceans. Large earthquakes are generally confined to subduction zones or other plate boundaries, leading to an uneven distribution of seismic sources. This heterogeneity, coupled with the land-based limitations for most for the Earth’s interior, leads to significant unsampled parts of the Earth. Our work is motivated by the planning of a Joint Task Force to develop concepts and applications for Science Monitoring and Reliable Telecommunication (SMART) cables. Over a million kilometers of submarine telecommunication cables currently exist, which are unavailable to the scientific community for acquisition of geophysical data. If these cables are gradually replaced by SMART cables with oceanographic and seismic sensors at roughly 75 km intervals, one significant benefit to our science will be the near-ubiquitous extent of seismic receivers across the oceans, affording an unprecedented opportunity for both monitoring and modeling. In previous work we presented ray tracing through a 1D reference model to predict improvements to ray coverage afforded by sensors on SMART cables, compared to existing land-based seismic network coverage. Here we extend that modeling, tracing P and S rays through the SALSA3D global tomographic model. We compare results of this exercise to those for the iasp91 model with, and without, the SMART cable sensors.
Poster:
SSA_2018_poster_Rowe_final
Abstract: We determine empirical relationships between instrumental peak ground motions and observed intensities for three recent Chilean megathrust (interplate subduction) earthquakes: the November 14, 2007 Mw 7.7 Tocopilla earthquake, and two great earthquakes, the February 27, 2010 Mw 8.8 Maule earthquake and the April 1, 2014 Mw 8.2 Iquique earthquake. Most great (M > 8) earthquakes, like these, occur within subduction zones. Yet, few ground motion to intensity conversion equations (GMICEs) exist for subduction earthquakes.We pair instrumental peak ground motions, both acceleration (PGA) and velocity (PGV), with intensities derived from on-site surveys of earthquake damage and volunteered felt reports. We fit a linear predictive equation between the geometric mean of the maximum PGA or PGV of the two horizontal components and intensity using an ordinary least squares regression. We use a weighting scheme to express the uncertainty of the intensity value assigned in each pairing based on the proximity of the instrument (station) to the nearest intensity observation. We find the best fitting predictive equations (GMICEs) for MMI from PGA, in cm/s/s, or PGV, in cm/s, are: PGA MMI = 2.10 log PGA + 0.26 II ≤ MMI ≤ V MMI = 6.78 log PGA – 10.54 V < MMI ≤ IX PGV MMI = 3.72 log PGV + 1.10 II ≤ MMI ≤ V MMI = 5.05 log PGV – 0.55 V < MMI ≤ IX We find existing GMICEs consistently overpredict intensity for these events, and existing inverse GMICEs underpredict the measured peak ground motions. We evaluate the effectiveness of existing GMICEs to reduce the misfit between predicted and observed intensities. We show that predictions for Chilean subduction events are improved when the corrections to global GMICEs are applied to peak ground motions instead of intensities. We conclude that the distinctive characteristics of Chilean megathrust earthquakes warrant the development of separate GMICEs for specific subduction zone regions.
Poster:
SSA_2018_Poster_Final
Abstract: The island of Hispaniola, of which the Dominican Republic (DR) comprises the eastern two-thirds, is subject to a complex array of tectonic stresses. The southward-dipping North American tectonic plate is subducting obliquely under the Caribbean plate in the north, two significant transform fault systems transect the island from east to west, and the island is believed to have accreted from at least three distinct terranes through transform and convergent motions. Further, the Dominican Republic is the site of dramatic topography, from a broad, flat but uplifted plain in the east to the highest mountains in the Caribbean. These features are largely the result of offsets that were accompanied by earthquakes and the historical record contains numerous descriptions of devastating events. However, in the modern, instrumented era the frequency of large earthquakes has been fairly low, which has made it difficult to characterize of the range of potential earthquake sources. The Centro Nacional de Sismologia (CNS), an organized research unit of the Autonomous University of Santo Domingo, is tasked with maintaining the network, analyzing its data, and compiling a comprehensive earthquake catalog. With the current network, which was expanded in 2013, approximately one thousand earthquakes per year are recorded and located in the DR with magnitudes between 2 and 6. We will present the focal mechanisms of ~146 earthquakes with magnitude 4 or greater recorded by ten or more stations since 2013. (The largest event to have occurred within the borders of the DR since 2013 is a magnitude 5.8.) Mechanisms, determined with first motion polarities and S/P amplitude ratios, for events with focal depths in the shallow to mid-crust share properties that correspond to surface tectonics but a set of intermediate-depth events beneath the eastern DR and the Mona Passage show a greater variety.
Poster:
Poster-Jottin-Eugenio-Jay-Victor-fv
Abstract: We investigate shear-wave velocity structure of the eastern Caribbean lithosphere using continuous seismic records from broadband stations deployed around the study area. We construct cross-correlation functions of ambient noise for two time periods, 2003-2005 and 2011-2013, based on data availability. Phase velocities of fundamental-mode Rayleigh waves are measured from the daily-stacked cross-correlation functions using a phase-matched-filter based method, and they are further projected on a set of 2°x2° grids, as suggested by checkerboard tests. The grid phase velocity dispersion data are inverted for 1D shear velocity profiles. The 3D velocity model indicates that the mean shear velocity in the lithospheric mantle is lower than the global average. Similar results have been reported beneath other large igneous provinces. Meanwhile, the lithosphere-asthenosphere boundary is located at depths between ~50 km and ~75 km, which is thinner than a normal oceanic lithosphere with an age of ~126-150 Ma. We speculate that low velocities in the lithosphere are associated with its composition together with a slightly elevated thermal state. Significant amount melts are likely produced beneath the thickened oceanic crust during the formation of the Caribbean Large Igneous Province (CLIP). Meanwhile the plume head has also been eroding the Caribbean lithosphere. In a later stage after the magmatic event, the melts crystalized and formed pyroxenite, which possesses a low seismic velocity but a high density, resulting in a thin lithosphere with a reduced wavespeed. A thick buoyant crust overlying on a thin denser lithospheric mantle may explain the subduction configuration in the eastern Caribbean.
Poster:
Poster SSA (1)
