In the Tiburon Peninsula, a lack of moderate earthquakes until January 2010 prevented a proper understanding of the seismotectonics of the Enriquillo-Plantain Garden Fault (EPGF); however, the morphotectonics along with the historical and instrumentally recorded earthquakes indicated that the fault system is active and potentially hazardous (Figure 1).

Two recent major earthquakes, the 12-01-2010 (Mw 7.1), hereafter the 2010 earthquake, and the 14-08-2021 (Mw 7.2), hereafter the 2021 earthquake, provide rare glimpses into the system, although at the expense of thousands of lives. It is therefore crucial that we extract all the information we can from these earthquakes, which may help in saving lives during future significant earthquakes in the peninsula. 

The report contains the results of our slip inversion of the two earthquakes as well as our inferences. We estimated the slip distribution by using the Kikuchi and Kanamori slip inversion method (Kikuchi et al., 1993; Kikuchi and Kanamori, 1991; Kikuchi and Kanamori, 1982). Seismic waveforms were obtained from the IRIS database (Incorporated Research Institutions for Seismology, 2021).

1. Slip Distribution of the 12 January 2010 Earthquake

We used 171 P and 93 SH waveforms from 171 stations in the epicentral distances of 24° to 96° in the slip inversion process. We explored a wide range of plausible parameter values until reaching an acceptable solution (Figure 2). Using the strike and dip angles from the Global CMT (GCMT) catalog (i.e., 250° strike and 71° dip), we obtained an average slip angle of 22°, which is consistent with the value reported in the GCMT solution (Dziewonski et al., 1981; Ekström et al., 2012). Additionally, we tested different strike and dip angles; however, the GCMT values turned out to be the best despite the deviation of the strike from EPGF’s trend in the southwest (Figures 3 and 4). Our moment magnitude, 7.06 (~7.1), is slightly larger than the magnitude reported by GCMT, 7.03 (~7.0).

In our experiment, we expect to reduce subjective preferences significantly by using multiple waveforms with equal weights. The downside of this approach is the uneven distribution of seismic stations, which may affect the slip distribution. We used the CRUST1 model (Laske et al., 2013) to determine the crustal structure for the ruptured area and to calculate the average shear modulus, which affects the slip value. According to the slip distribution, the maximum value is more than 3.5 m, and the slip area spans about 50 km in length and about 20 km in width. In conjunction with the rupture initiation point, the slip distribution implies a bilateral rupture.

Figure 1. The historical (pre-1900) earthquakes in Haiti and the western part of the Dominican Republic (National Geophysical Data Center, 2021). Most of the events have no reported magnitude; therefore, we assigned a rough magnitude of 7.0. The yellow labels show the occurrence years. The colored circles show the instrumentally recorded events from 1900 to 11-01-2010, the day before the first major earthquake in this study (United States Geological Survey, 2021). Events that occurred after 1900 are color-coded according to the date of the event, as shown on the left color bar.
Figure 2.  The observed (black) body waveforms are used to derive the slip distribution, shown in Figure 3, by employing the slip inversion method of Kikuchi and Kanamori (1993). The red curves are the synthetic waveforms. The source time function of the event, shown in the upper-right graph, indicates a rupture duration of more than 20 s. The information on the left of each panel includes waveform type (P or SH), amplitude in micrometer, station code, azimuth, and epicentral distance. The number to the right of each panel corresponds to the station number on the map in the middle.
Figure 2. Continued …
Figure 2. Continued …
Figure 3. The slip distribution of the 12-01-2010 Haiti earthquake. The red star at point (0,0) shows the initiation point. The yellow tags correspond with their counterparts in Figure 11. The ground surface coincides with the +10 alongside the up-dip axis.
Figure 4.  Stars represent historical earthquakes (pre-1900), while filled circles show earthquakes from the 1900s until 12-01-2010. The color of circles represents the date of the event. The Enriquillo-Plantin Garden fault system (EPGF) passes through the whole Tiburon Peninsula. LF stands for the Lamentin Fault. The white curves show the slip contours, 0.5 m spacing, for the 12-01-2010 and the 14-08-2021 earthquakes.
Figure 5. The colored circles show the earthquakes from 12-01-2010 to the day before the 14-08-2021 earthquake. The color represents the date of the events as indicated in the left color bar. Most of the events are the aftershocks of the 2010 earthquake, with magnitudes between 3 and 6. EPGF stands for the Enriquillo-Plantin Garden fault while LF represents the Lamentin fault.

2. Slip Distribution of the 14 August 2021 Earthquake

In order to calculate the slip distribution of the 2021 earthquake, we used a similar procedure as applied to the 2010 earthquake. 100 P and 106 SH waveforms from 127 seismic stations in the epicentral range of 24° to 96° were used in the slip inversion (Figure 6). The preferred values for the strike, 266°, and dip, 64°, angles come from the GCMT solution. We determined the best GCMT-derived values after many trials. Nevertheless, our derived average slip angle, 38°, differs from the value reported in the GCMT solution, 43°.

Using the source time function, shown in the upper-right panel of Figure 6, it is derived that the rupture was about 48 s long. There is an early prime segment that corresponds to the dominant slip area, and a late minor segment that corresponds to the two small slip patches (Figure 7 and Figure 8). The principle slip area is about 60 km in length and 24 km in width, with a maximum slip exceeding 3 m. Slip distribution indicates a westward unilateral rupture.

Figure 6The observed (black) body waveforms are used to drive the slip distribution, Figure 7, by employing the slip inversion method of Kikuchi and Kanamori (1993). The red curves are the synthetic waveforms. The source time function of the event indicates a rupture duration of about 48 s. The information on the left of each panel includes waveform type (P or SH), amplitude in micrometer, station code, azimuth, and epicentral distance. The number to the right of each panel corresponds with the station number on the map in the middle.
Figure 6. Continued …
Figure 6. Continued …
Figure 7. The slip distribution of the 14-08-2021 Haiti earthquake. The red star at (0,0) shows the nucleation point. The contour interval is 0.5 m. The yellow tags correspond with their counterparts in Figure 11.
Figure 8. The colored circles show the seismicity for 14- to 27-08-2021 in Tiburon Peninsula. The color of circles shows the date of events. The white curves show the slip contours for the 2010 and 2021 earthquakes. EPGF stands for the Enriquillo-Plantin Garden fault while LF represents the Lamentin fault.

3. Discussion and Results

The slip distributions of both 2010 and 2021 earthquakes indicate a transpressive tectonic regime (Figures 9 and 10); however, the dominant mechanism of aftershocks of the 2010 earthquake is the pure reverse. The substantial reverse component in such systems contribute to the mountain building and changes of the topography alongside the causative system (Brueckner et al., 2009; Fossen and Tikoff, 1993; Frehner, 2016; Rastogi, 2004; Searle et al., 1998; Spotila et al., 2007; Tikoff and Peterson, 1998). Intuitively, the variations in topography along the system explain the changes in the degree of coupling and the strength of the corresponding fault segments. It is possible to examine the above notion by looking at the slip distribution of these two events.

Figure 11 shows the elevation profile alongside the Enriquillo-Plantin Garden fault system. The southwestern part of the 2010 slip deviates from the EPGF trend. However, the principle slip for the 2010 earthquake is located in its northeast, toward point L, and decreases by moving toward point K. Point L matches with the peak of the topography. This observation indicates that the topography between L and M corresponds with either unreleased or previously released seismic energy. At point M, the east-west EPGF connects to the SE-trending Lamentin Fault (LF).  The known seismic activity at this area includes two events in 1751 and one in 1784 (Figure 4). These three historical events with the period of 33 years have likely been on both EPGF and LF. As such, it is unlikely that Port-au-Prince is a safer place as a result of seismic energy release in 2010.

Figure 9. All the available CMT solutions from the GCMT database that fall within the Tiburon Peninsula. EPGF stands for the Enriquillo-Plantin Garden fault while LF represents the Lamentin fault.

The slip distribution for the 2021 earthquake shows a main large slip area and two minor patches at its western end (Figures 7 and 8). These small patches were persistent in almost all solutions; therefore, they are not artifacts, although their slip direction seems inconsistent. Two sub-patches make up the main patch: a shallow strong patch and a relatively deeper but weak patch (Figure 7). According to Figure 11, the strong patch correlates with the peak between F and G, while the infirm patch correlates with the lowland part between E and F. The minor slip patches locate between points B and C, corresponding with another peak on the profile. The area between B and C is slightly moved and has not ruptured totally. The blank span between C and E, with a length of 30 km, corresponds with a dominant peak on the profile, which indicates that the span should be a robust barrier. The slip distribution affirms this notion, as no dislocation took place in the span (Figure 7).  The rupture history in the upper-right panel of Figure 6 reaffirms this conclusion. It consists of a prime part, 3 to 24 s, and a minor later part, 36 to 47 s. They correspond with the major and the minor patches on the slip distribution (Figure 7). The prime part of the source time function displays an abrupt increase at 3 s and a sharp decline in 22-24 s, which indicates that the rupture process suddenly halted at point E. However, about 12 s later, the area between C and B dislocated slightly. The abrupt stopping of the rupture is also evident from the sharp decline of the slip values at 36 km along strike in Figure 7. It is evident from all these indications that a strong barrier exists, hereinafter called Macaya.

The absence of slip along the Macaya barrier does not arise from a recent rupture of the segment. The largest aftershock of the event on 15-08-2021 with magnitude 5.7 ruptured a small part of the western end of the Macaya barrier, indicating it is in a mature stage of the earthquake cycle (Figure 10).

To estimate the magnitude of this likely earthquake, we assume a correlation between the seismic moment of earthquakes and the corresponding area under the elevation profile. This is a valid assumption as the seismic moment corresponds with the fault area of the ruptured segment. To examine the above suggestion, we considered the area under the elevation profile for 2021 as a benchmark and computed the magnitude of the 2010 earthquake. We obtained a value of 7.21 for the 2010 earthquake, while its actual magnitude is 7.06. Therefore, a tolerance for the assumption is about 0.15 units.

The minimum expected rupture span that will include the Macaya barrier is the range A to E; however, it can further extend in the ocean from the west side. By such an assumption, the minimum expected earthquake should have a magnitude of about 7.6±0.15. 

With a similar comparison based on the area under a curve, the immediate segment to the east of the 2021 earthquake, segment G to I, may rupture in a single earthquake with a magnitude of 7.2

Figure 10. The GCMT solutions for the 14-08-2021 earthquake and two of its large aftershocks (Dziewonski et al., 1981; Ekström et al., 2012). EPGF stands for the Enriquillo-Plantin Garden fault while LF represents the Lamentin fault.
Figure 11. The elevation profile along part of the Enriquillo-Plantin Garden fault system. The similar yellow tags (A to M) in the map and the profile show the corresponding points.

4. What should be done?

Possibly the rational solution lies in strengthening the buildings and ensuring that building codes are followed. That is important and necessary, but it is a lengthy process and is dependent on the economic situation. The authorities should take appropriate action on that front, if at all possible. This does not relieve seismologists of the responsibility of taking further action. We now know the location of the Macaya barrier, so researchers should follow and add every observation to the jigsaw puzzle of likely scenarios. Despite being challenging, the task is not insurmountable. As a result of the current study, the area west of the Macaya boundary, near point C, is considered a probable rupture initiation area. The down-dip bound of such an extensive barrier is also an appropriate location for rupture initiation. As a result, it is vital to establish geodetic and seismic networks with online analysis capabilities around the Macaya barrier. Daily data from publicly available sources may be collected and analyzed to determine the total electron content (TEC); (Galvan et al., 2011; J. Y. Liu et al., 2001, 2009, 2011; Jann Yenq Liu et al., 2006; Shah et al., 2020; Xu et al., 2011).

Observations from the above tasks alone do not suffice to predict the onset of a rupture; however, the correlation between variations in the geophysical properties of the barrier area and abnormal observations by nearby residents may help. “Abnormal observation” is a general term and should not be confined to known earthquake precursors only. It is therefore imperative that local residents and seismologists work together. Meanwhile, an analysis of the vital information on precursors is essential before the looming earthquake. 

An earthquake is the culmination of a long process. Such a process can be accompanied by ample time. Large shallow earthquakes are accompanied by robust physical and chemical processes. As the rupturing time approaches, these processes accelerate both in intensity and occurrence.

The residents of Haicheng are largely responsible for the successful prediction of the 04-02-1975 earthquake (Mw 7.3; Qidong et al., 1981; Scholz, 1977; Wang et al., 2006). Nowadays, mobile phones facilitate this type of cooperation. A system should be in place for automatically analyzing the messages of residents. It is necessary to maintain a database of the contact information of the residents.

5. Concluding Remarks

  1. A rupture of the Macaya barrier can lead to an earthquake with a magnitude of 7.6±0.15 or greater.
  2. A dense geodetic and seismic network with online analysis capability is necessary in order to cover the Macaya Barrier.
  3. The occurrence of the 2010 earthquake does not rule out the occurrence of another significant earthquake around Port-au-Prince.
  4. Rupture of the immediate segment to the east of the 2021 earthquake, the distance between points G and I in Figure 11, may generate a 7.2 earthquake.
  5. Daily analysis of the Total Electron Content data (TEC) is necessary.
  6. In order to get reports on abnormal observations, residents should be closely involved. There is a need for automatic analysis of the messages.
  7. It is essential to have a database of the contact information of the residents that can operate efficiently.

Acknowledgments

The figures are created using GMT (Wessel et al., 2019). SAC software is used widely (P. Goldstein and Snoke, 2005; Peter Goldstein et al., 2003). We appreciate the developers for sharing their codes. The seismic waveforms are downloaded from the Incorporated Research Institutions for Seismology (IRIS, http://www.iris.edu/mda). We appreciate Hadi Ghofrani for peer-reviewing this report.

References

Brueckner, H. K., Avé Lallemant, H. G., Sisson, V. B., Harlow, G. E., Hemming, S. R., Martens, U., et al. (2009). Metamorphic reworking of a high pressure-low temperature mélange along the Motagua fault, Guatemala: A record of Neocomian and Maastrichtian transpressional tectonics. Earth and Planetary Science Letters284(1–2). https://doi.org/10.1016/j.epsl.2009.04.032

Dziewonski, A. M., Chou, T. A., and Woodhouse, J. H. (1981). Determination of earthquake source parameters from waveform data for studies of global and regional seismicity. Journal of Geophysical Research86(B4). https://doi.org/10.1029/JB086iB04p02825

Ekström, G., Nettles, M., and Dziewoński, A. M. (2012). The global CMT project 2004-2010: Centroid-moment tensors for 13,017 earthquakes. Physics of the Earth and Planetary Interiors200201. https://doi.org/10.1016/j.pepi.2012.04.002

Fossen, H., and Tikoff, B. (1993). The deformation matrix for simultaneous simple shearing, pure shearing and volume change, and its application to transpression-transtension tectonics. Journal of Structural Geology15(3–5). https://doi.org/10.1016/0191-8141(93)90137-Y

Frehner, M. (2016). 3D fold growth in transpression. Tectonophysics693. https://doi.org/10.1016/j.tecto.2016.01.002

Galvan, D. A., Komjathy, A., Hickey, M. P., and Mannucci, A. J. (2011). The 2009 Samoa and 2010 Chile tsunamis as observed in the ionosphere using GPS total electron content. Journal of Geophysical Research: Space Physics116(6). https://doi.org/10.1029/2010JA016204

Goldstein, P., and Snoke, J. A. (2005). SAC Availability for the IRIS Community. DMS Electronic Newsletter7(1).

Goldstein, Peter, Dodge, D., Firpo, M., and Minner, L. (2003). 85.5 SAC2000: Signal processing and analysis tools for seismologists and engineers. In International Geophysics (Vol. 81). https://doi.org/10.1016/S0074-6142(03)80284-X

Incorporated Research Institutions for Seismology. (2021). Wilber 3: Select Event. Retrieved August 16, 2021, from https://ds.iris.edu/wilber3/find_event

Kikuchi, M., and Kanamori, H. (1991). Inversion of complex body waves – III. Bulletin – Seismological Society of America81(6).

Kikuchi, M., Kanamori, H., and Satake, K. (1993). Source complexity of the 1988 Armenian earthquake: evidence for a slow after-slip event. Journal of Geophysical Research98(B9). https://doi.org/10.1029/93jb01568

Kikuchi, Masayuki, and Kanamori, H. (1982). Inversion of complex body waves. Bulletin of the Seismological Society of America72(2), 491–506. https://doi.org/10.1785/BSSA0720020491

Laske, G., Masters, G., Ma, Z., and Pasyanos, M. (2013). Update on CRUST1.0—A 1-degree global model of Earth’s crust. In EGU General Assembly 2013 (Vol. 15).

Liu, J. Y., Chen, Y. I., Chuo, Y. J., and Tsai, H. F. (2001). Variations of ionospheric total electron content during the Chi-Chi earthquake. Geophysical Research Letters28(7). https://doi.org/10.1029/2000GL012511

Liu, J. Y., Chen, Y. I., Chen, C. H., Liu, C. Y., Chen, C. Y., Nishihashi, M., et al. (2009). Seismoionospheric GPS total electron content anomalies observed. before the 12 May 2008 Mw7.9 Wenchuan earthquake. Journal of Geophysical Research: Space Physics114(4). https://doi.org/10.1029/2008JA013698

Liu, J. Y., Le, H., Chen, Y. I., Chen, C. H., Liu, L., Wan, W., et al. (2011). Observations and simulations of seismoionospheric GPS total electron content anomalies before the 12 January 2010 M7 Haiti earthquake. Journal of Geophysical Research: Space Physics116(4). https://doi.org/10.1029/2010JA015704

Liu, Jann Yenq, Tsai, Y. Ben, Ma, K. F., Chen, Y. I., Tsai, H. F., Lin, C. H., et al. (2006). Ionospheric GPS total electron content (TEC) disturbances triggered by the 26 December 2004 Indian Ocean tsunami. Journal of Geophysical Research: Space Physics111(5). https://doi.org/10.1029/2005JA011200

National Geophysical Data Center (2021). World Data Service (NGDC/WDS): NCEI/WDS Global Significant Earthquake Database. https://doi.org/10.7289/V5TD9V7K

Qidong, D., Pu, J., Jones, L. M., and Molnar, P. (1981). A preliminary analysis of reported changes in ground water and anomalous animal behaviour before the 4 February 1975 Haicheng earthquake. Earthquake Prediction: An International Review. https://doi.org/10.1029/me004p0543

Rastogi, B. K. (2004). Damage due to the M-w 7.7 Kutch, India earthquake of 2001. Tectonophysics390(1-4), 85–103. https://doi.org/10.1016/j.tecto.2004.03.030

Scholz, C. H. (1977). A physical interpretation of the Haicheng earthquake prediction. Nature267(5607). https://doi.org/10.1038/267121a0

Searle, M. P., Weinberg, R. F., and Dunlap, W. J. (1998). Transpressional tectonics along the Karakoram fault zone, northern Ladakh: constraints on Tibetan extrusion. Geological Society Special Publication135. https://doi.org/10.1144/GSL.SP.1998.135.01.20

Shah, M., Ahmed, A., Ehsan, M., Khan, M., Tariq, M. A., Calabia, A., and Rahman, Z. ur. (2020). Total electron content anomalies associated with earthquakes occurred during 1998–2019. Acta Astronautica175. https://doi.org/10.1016/j.actaastro.2020.06.005

Spotila, J. A., Niemi, N., Brady, R., House, M., Buscher, J., and Oskin, M. (2007). Long-term continental deformation associated with transpressive plate motion: The San Andreas fault. Geology35(11). https://doi.org/10.1130/G23816A.1

Tikoff, B., and Peterson, K. (1998). Physical experiments of transpressional folding. Journal of Structural Geology20(6). https://doi.org/10.1016/S0191-8141(98)00004-2

United States Geological Survey (2021). Search Earthquake Catalog. Retrieved from https://earthquake.usgs.gov/earthquakes/search/

Wang, K., Chen, Q. F., Sun, S., and Wang, A. (2006). Predicting the 1975 Haicheng earthquake. Bulletin of the Seismological Society of America96(3). https://doi.org/10.1785/0120050191

Wessel, P., Luis, J. F., Uieda, L., Scharroo, R., Wobbe, F., Smith, W. H. F., and Tian, D. (2019). The Generic Mapping Tools Version 6. Geochemistry, Geophysics, Geosystems20(11). https://doi.org/10.1029/2019GC008515

Xu, T., Chen, Z., Li, C., Wu, J., Hu, Y., and Wu, Z. (2011). GPS total electron content and surface latent heat flux variations before the 11 March 2011 M9.0 Sendai earthquake. Advances in Space Research48(8). https://doi.org/10.1016/j.asr.2011.06.024

Last Updated on August 29, 2022