BROADBAND SIMULATION FOR A HYPOTHETICAL M W 7.1 EARTHQUAKE ON THE ENRIQUILLO FAULT IN HAITI

Transcription

1 Eleventh U.S. National Conference on Earthquake Engineering Integrating Science, Engineering & Policy June 25-29, 2018 Los Angeles, California BROADBAND SIMULATION FOR A HYPOTHETICAL M W 7.1 EARTHQUAKE ON THE ENRIQUILLO FAULT IN HAITI R. Douilly 1, G. P. Mavroeidis 2 and. E. Calais 3 ABSTRACT In this study, we investigate ground shaking in the vicinity of Port-au-Prince if a hypothetical rupture similar to the 2010 Haiti earthquake occurred on the central segment of the Enriquillo Fault. We use a finite element method and assumptions on regional tectonic stress to simulate the low-frequency ground motion components using dynamic rupture propagation for a 52-km-long segment. We consider eight scenarios by varying parameters such as hypocenter location, initial shear stress and fault dip. The high-frequency ground motion components are simulated using the specific barrier model in the context of the stochastic modeling approach. The broadband ground motion synthetics are subsequently obtained by combining the low-frequency components from the dynamic rupture simulation with the high-frequency components from the stochastic simulation using matched filtering at a crossover frequency of 1 Hz. Results show that rupture on a vertical Enriquillo Fault generates larger horizontal permanent displacements in Léogâne and Port-au-Prince than rupture on a south-dipping Enriquillo Fault. The mean horizontal peak ground acceleration, computed at several sites of interest throughout Port-au- Prince, has a value of ~0.45g, whereas the maximum horizontal peak ground acceleration in Port-au-Prince is ~0.60g. Even though we only consider a limited number of rupture scenarios, our results suggest more intense ground shaking for the city of Port-au-Prince than during the already very damaging 2010 Haiti earthquake. 1 Postodoctoral Researcher, Dept. of Earth Science, University of California Riverside, Riverside, CA ( roby.douilly@ucr.edu) 2 Assistant Professor, Dept. of Civil and Environmental Engineering and Earth Science, University of Notre Dame, Notre Dame, IN Professor and Department Head, Dept. of Geosciences, PSL Research University, École Normale Supérieure, Paris, France

2 Eleventh U.S. National Conference on Earthquake Engineering Integrating Science, Engineering & Policy June 25-29, 2018 Los Angeles, California Broadband Simulation for a Hypothetical M w 7.1 Earthquake on the Enriquillo Fault in Haiti R. Douilly 1, G. P. Mavroeidis 2 and. E. Calais 3 ABSTRACT In this study, we investigate ground shaking in the vicinity of Port-au-Prince if a hypothetical rupture similar to the 2010 Haiti earthquake occurred on the central segment of the Enriquillo Fault. We use a finite element method and assumptions on regional tectonic stress to simulate the low-frequency ground motion components using dynamic rupture propagation for a 52-km-long segment. We consider eight scenarios by varying parameters such as hypocenter location, initial shear stress and fault dip. The high-frequency ground motion components are simulated using the specific barrier model in the context of the stochastic modeling approach. The broadband ground motion synthetics are subsequently obtained by combining the low-frequency components from the dynamic rupture simulation with the high-frequency components from the stochastic simulation using matched filtering at a crossover frequency of 1 Hz. Results show that rupture on a vertical Enriquillo Fault generates larger horizontal permanent displacements in Léogâne and Port-au-Prince than rupture on a south-dipping Enriquillo Fault. The mean horizontal peak ground acceleration, computed at several sites of interest throughout Port-au-Prince, has a value of ~0.45g, whereas the maximum horizontal peak ground acceleration in Port-au-Prince is ~0.60g. Even though we only consider a limited number of rupture scenarios, our results suggest more intense ground shaking for the city of Portau-Prince than during the already very damaging 2010 Haiti earthquake. Introduction The earthquake history of the Enriquillo Plantain Garden Fault (EPGF) in southern Haiti, which last ruptured 250 years ago, combined with geodetic measurements of how fast the fault is building up strain energy, suggests that this fault is currently capable of unleashing a M w 7.2 earthquake. This would be devastating to the nearby densely populated Port-au-Prince region if the entire energy was released in a single event (Manaker et al., 2008). This is the very same region devastated by the 2010 M w 7.0 earthquake that killed more than 250,000 people. Originally thought to have occurred on the EPGF, the 2010 earthquake was shown to have occurred on the nearby Léogâne Fault (Calais et al., 2010; Hayes et al., 2010; Mercier de Lépinay et al., 2011; Meng et al., 2012; Symithe et al., 2013; Douilly et al., 2013, 2015). Despite the rupture being located on a segment adjacent to the main plate boundary fault, the 2010 Haiti earthquake influenced the stress state on the EPGF where significant stress increase was inferred mostly on the top portion of this segment (Symithe et al., 2013; Douilly et al., 2015), suggesting that the EPGF was pushed closer to failure. Thus, the anticipation that the people of Port-au- Prince are now safe for a while before another deadly earthquake strikes is premature. The EPGF, which did not rupture in 2010, could potentially rupture in the near future. To understand the hazard this region is facing, we investigate the ground shaking level in

3 southern Haiti if a hypothetical seismic event, similar in magnitude to the 2010 M w 7.0 Haiti earthquake, occurred on the EPGF segment adjacent to the Léogâne Fault and close to the capital city of Port-au-Prince. To achieve this objective, we use a hybrid method that combines dynamic rupture simulations at low frequencies with stochastic simulations at high frequencies. We consider a range of rupture scenarios on the EPGF segment and generate synthetic ground motion at hypothetical stations. Readers should refer to the recently published article by Douilly et al. (2017) for detailed results pertaining to this study. Methodology We consider a 52-km-long segment of the EPGF close to the heavily populated city of Port-au-Prince based on geological mapping that shows a continuous fault segment. Since the dip of the EPGF segment remains debated, we consider two fault geometries a vertical plane as inferred by Mann et al. (1995) and a plane dipping to the south at 65 as proposed by Prentice et al. (2010). Nevertheless, the strike and length of both faults are consistent with fault trace information published in the literature (Mann et al., 1995). We use the finite element software CUBIT to generate a model that is 150 km long, 200 km wide and 100 km deep and we discretize the model space using tetrahedral elements of 250 m size along the fault surface. We incorporate the resulting mesh into the finite element solver PyLith (Aagaard et al., 2013; Douilly et al., 2015) to carry out the dynamic rupture simulations. In order to simulate the low-frequency dynamic rupture propagation, a priori information such as initial stress conditions and frictional parameters are needed. In this work, we consider a maximum horizontal stress orientation of N50 E, consistent with the transpressional stress regime as indicated by previous geodetic studies (Calais et al., 2010, Calais et al., 2016), to estimate the initial state of stress on the EPGF. We also adopt a slip-weakening friction law with a static friction coefficient of 0.6 and a dynamic friction coefficient of To initiate the rupture, we generate a circular crack of 2.5 km radius centered at the hypocenter location where we impose the shear stress to be 5% greater than the failure stress (Day, 1982; Madariaga et al., 1998). In addition, we consider heterogeneity in the initial shear stress with two scenarios. In scenario A, the fault contains two patches of 6.5 km radius over which shear stress is 15% greater than initial shear stress on the fault. In scenario B, we increase shear stress by 15% in the upper 5 km of the fault in order to mimic shear stress increase found along the upper part of the EPGF segment adjacent to the Léogâne Fault resulting from the 2010 Haiti earthquake (Symithe et al., 2013; Douilly et al., 2015). Due to this heterogeneity, regions of higher shear stress are subject to an increase in stress drop to 7.7 MPa for a vertical fault and 7.0 MPa for a southdipping fault, which is within the observed earthquake stress drop range. To generate the high-frequency components of the synthetic ground motion we used the Specific Barrier Model (SBM) (Papageorgiou and Aki, 1983a, 1983b; Halldorsson and Papageorgiou, 2005) and the stochastic modeling approach (Boore, 1983; Shinozuka, 1988). In the SBM, which is particular case of a composite seismic source model, the fault is discretized into non-overlapping circular sub-events (shear cracks) of equal diameter (also known as barrier interval) that cover a rectangular fault. As the rupture front propagates along the fault, the subevents are assumed to break randomly and independently from each other and a local stress drop occurs on each of the sub-events. The rupture starts at the center of each sub-event and spreads outward at a constant velocity until the rupture is arrested by the barriers. This modeling approach generates the high-frequency components of the synthetic ground motion. To obtain the broadband synthetics we combine the low-frequency ground motions from the dynamic

4 rupture simulations with the high-frequency ground motions from the stochastic simulations using matched filtering at a crossover frequency of 1 Hz (Mavroeidis and Scotti, 2013). Results A cumulative moment release of M w ~7.1 is observed for our simulations and is consistent with historical earthquakes and fault segmentation in southern Haiti. Results for the horizontal permanent displacements from the low-frequency simulations show that stations located north of the ruptured segment experience larger horizontal permanent displacements for a rupture on a vertical fault than for a rupture on a south-dipping fault. In addition, the horizontal permanent displacements for scenario B is typically greater than for scenario A due to the fact that the slip patches for scenario B are predominantly concentrated near the surface and hence generate larger horizontal permanent displacements close to the fault. Furthermore, we observe larger vertical permanent displacements for the south-dipping plane than for the vertical plane for both scenarios A and B. Also, in contrast to horizontal permanent displacements, stations located directly south of the ruptured segment experience larger vertical permanent displacements than stations located north of the ruptured segment for both scenarios A and B. Therefore, we infer that a vertical (resp. south-dipping) fault plane generates larger horizontal (resp. vertical) permanent displacements at stations located north (resp. south) of the fault. Similar to the low-frequency simulations presented above, our broadband simulations show that stations located north of the ruptured fault segment experience higher Peak Ground Acceleration (PGA) for a vertical fault than for a south-dipping fault due to the shorter distance of these stations to the vertical fault plane. We compare our synthetic PGA values with the GMPE proposed by Boore and Atkinson (2008) and overall our results are in good agreement with the GMPE curves. In addition, we observe that stations in Port-au-Prince have a mean horizontal PGA of ~0.45g (resp. ~0.35g) for a rupture on a vertical (resp. south-dipping) fault, which is about times greater than the estimated PGA from the 2010 Haiti earthquake on the Léogâne Fault, where the maximum PGA reached values up to ~0.6g (Douilly et al., 2017). Conclusions Our broadband ground motion simulations show that the mean peak ground acceleration in Portau-Prince is ~0.45g, twice as much as the estimated peak ground acceleration during the 2010 Haiti earthquake. We emphasize that our simulations do not take into account lithological or topographic site effects, which tend to amplify ground shaking. Therefore our simulations should be considered as the lower bound of the ground motion variability spectrum. The fact that the intensity of ground motion experienced in Port-au-Prince during the 2010 earthquake could be surpassed during the rupture of the fault segment modeled in this study should serve as a warning for Haitian engineers and the population in general.

5 References Aagaard, B., Knepley, M., and Williams, C. (2013). A domain decomposition approach to implementing fault slip in finite-element models of quasi-static and dynamic crustal deformation. J. Geophys. Res. Solid Earth 118, Boore, D.M. (1983). Stochastic simulation of high-frequency ground motions based on seismological models of the radiated spectra. Bull. Seismol. Soc. Am. 73, Boore, D. M., and G. M. Atkinson (2008). Ground-motion prediction equations for the average horizontal component of PGA, PGV, and 5%-damped PSA at spectral periods between 0.01 and 10.0 s, Earthq. Spectra 24, no. 1, Calais, E., Freed, A., Mattioli, G., Amelung, F., Jónsson, S., Jansma, P., Hong, S.-H., Dixon, T., Prépetit, C., and Momplaisir, R. (2010). Transpressional rupture of an unmapped fault during the 2010 Haiti earthquake. Nat. Geosci. 3, Calais, É., Symithe, S., Mercier de Lépinay, B., and Prépetit, C. (2016). Plate boundary segmentation in the northeastern Caribbean from geodetic measurements and Neogene geological observations. Comptes Rendus Geosci. 348, Day, S.M. (1982). Three-dimensional simulation of spontaneous rupture: the effect of nonuniform prestress. Bull. Seismol. Soc. Am. 72, Douilly, R., Haase, J.S., Ellsworth, W.L., Bouin, M.-P., Calais, E., Symithe, S.J., Armbruster, J.G., de Lépinay, B.M., Deschamps, A., Mildor, S.-L., et al. (2013). Crustal Structure and Fault Geometry of the 2010 Haiti Earthquake from Temporary Seismometer Deployments. Bull. Seismol. Soc. Am. 103, Douilly, R., Aochi, H., Calais, E., and Freed, A. (2015). Three-dimensional dynamic rupture simulations across interacting faults: The Mw7. 0, 2010, Haiti earthquake. J. Geophys. Res. Solid Earth 120, Douilly, R., G. P. Mavroeidis, and E. Calais (2017). Simulation of broadband strong ground motion for a hypothetical Mw 7.1 earthquake on the Enriquillo Fault in Haiti, Geophys. J. Int., 211, Halldorsson, B., and Papageorgiou, A.S. (2005). Calibration of the specific barrier model to earthquakes of different tectonic regions. Bull. Seismol. Soc. Am. 95, Hayes, G., Briggs, R., Sladen, A., Fielding, E., Prentice, C., Hudnut, K., Mann, P., Taylor, F., Crone, A., Gold, R., et al. (2010). Complex rupture during the 12 January 2010 Haiti earthquake. Nat. Geosci. 3, Madariaga, R., Olsen, K., and Archuleta, R. (1998). Modeling dynamic rupture in a 3D earthquake fault model. Bull. Seismol. Soc. Am. 88, Manaker, D.M., Calais, E., Freed, A., Ali, S., Przybylski, P., Mattioli, G., Jansma, P., Prépetit, C., and De Chabalier, J. (2008). Interseismic plate coupling and strain partitioning in the northeastern Caribbean. Geophys. J. Int. 174, Mann, P., Taylor, F.W., Edwards, R.L., and Ku, T.-L. (1995). Actively evolving microplate formation by oblique collision and sideways motion along strike-slip faults: An example from the northeastern Caribbean plate margin. Tectonophysics 246, Mavroeidis, G.P., and Scotti, C.M. (2013). Finite-Fault Simulation of Broadband Strong Ground Motion from the 2010 Mw 7.0 Haiti Earthquake. Bull. Seismol. Soc. Am. 103, Meng, L., Ampuero, J.-P., Sladen, A., and Rendon, H. (2012). High-resolution backprojection at regional distance: Application to the Haiti M7. 0 earthquake and comparisons with finite source studies. J. Geophys. Res. Solid Earth 117,

6 Mercier de Lépinay, B., Deschamps, A., Klingelhoefer, F., Mazabraud, Y., Delouis, B., Clouard, V., Hello, Y., Crozon, J., Marcaillou, B., Graindorge, D., et al. (2011). The 2010 Haiti earthquake: A complex fault pattern constrained by seismologic and tectonic observations. Geophys. Res. Lett. 38, L22305, doi: /2011gl Papageorgiou, A.S., and Aki, K. (1983a). A specific barrier model for the quantitative description of inhomogeneous faulting and the prediction of strong ground motion. I. Description of the model. Bull. Seismol. Soc. Am. 73, Papageorgiou, A.S., and Aki, K. (1983b). A specific barrier model for the quantitative description of inhomogeneous faulting and the prediction of strong ground motion. Part II. Applications of the model. Bull. Seismol. Soc. Am. 73, Prentice, C., Mann, P., Crone, A., Gold, R., Hudnut, K., Briggs, R., Koehler, R., and Jean, P. (2010). Seismic hazard of the Enriquillo-Plantain Garden fault in Haiti inferred from palaeoseismology. Nat. Geosci. 3, Shinozuka, M. (1988). State-of-the-art report: engineering modeling of ground motion, in Proceedings of the 9th World Conference on Earthquake Engineering (9WCEE), Tokyo, Japan, 2 9 August Symithe, S.J., Calais, E., Haase, J.S., Freed, A.M., and Douilly, R. (2013). Coseismic Slip Distribution of the 2010 M 7.0 Haiti Earthquake and Resulting Stress Changes on Regional Faults. Bull. Seismol. Soc. Am. 103,