The Kalaotoa Fault: A Newly Identified Fault that Generated the Mw 7.3 Flores Sea Earthquake
Item Type Article
Authors Supendi, Pepen;Rawlinson, Nicholas;Prayitno, Bambang Setiyo;Widiyantoro, Sri;Simanjuntak, Andrean;Palgunadi, Kadek Hendrawan;Kurniawan, Andri;Marliyani, Gayatri Indah;Nugraha, Andri Dian;Daryono, Daryono;Anugrah, Suci Dewi;Fatchurochman, Iman;Gunawan, Mohammad Taufik;Sadly, Muhammad;Adi, Suko Prayitno;Karnawati, Dwikorita;Arimuko, Abraham
Citation Supendi, P., Rawlinson, N., Prayitno, B. S., Widiyantoro, S., Simanjuntak, A., Palgunadi, K. H., Kurniawan, A., Marliyani, G. I., Nugraha, A. D., Daryono, D., Anugrah, S. D., Fatchurochman, I., Gunawan, M. T., Sadly, M., Adi, S. P., Karnawati, D., & Arimuko, A. (2022). The Kalaotoa Fault: A Newly Identified Fault that Generated the Mw 7.3 Flores Sea Earthquake. The Seismic Record, 2(3), 176–185. https://doi.org/10.1785/0320220015 Eprint version Publisher's Version/PDF
DOI 10.1785/0320220015
Publisher Seismological Society of America (SSA)
Journal The Seismic Record
Rights Archived with thanks to The Seismic Record under a Creative Commons license, details at: http://creativecommons.org/
licenses/by/2.0/
Download date 2023-12-29 18:32:14
Item License http://creativecommons.org/licenses/by/2.0/
Link to Item http://hdl.handle.net/10754/680058
The Kalaotoa Fault: A Newly Identified Fault that Generated the M w 7.3 Flores Sea Earthquake
Pepen Supendi*1,2 , Nicholas Rawlinson1 , Bambang Setiyo Prayitno2, Sri Widiyantoro3,4 ,
Andrean Simanjuntak2 , Kadek Hendrawan Palgunadi5 , Andri Kurniawan6, Gayatri Indah Marliyani7 , Andri Dian Nugraha3 , Daryono Daryono2, Suci Dewi Anugrah2, Iman Fatchurochman2, Mohammad Taufik Gunawan2, Muhammad Sadly2, Suko Prayitno Adi2, Dwikorita Karnawati2, and Abraham Arimuko2
Abstract
Cite this article asSupendi, P., N. Rawlinson, B. S. Prayitno, S. Widiyantoro, A. Simanjuntak, K. H. Palgunadi, A. Kurniawan, G. I. Marliyani, A. D. Nugraha, D. Daryono, et al. (2022).
The Kalaotoa Fault: A Newly Identified Fault that Generated theMw7.3 Flores Sea Earthquake,The Seismic Record.2(3), 176–185, doi:10.1785/0320220015.
Supplemental Material We reveal the existence of a previously unknown fault that generated theMw7.3 Flores Sea
earthquake, which occurred on 14 December 2021, approximately 100 km to the north of Flores Island, in one of the most complex tectonic settings in Indonesia. We use a double- difference method to relocate the hypocenters of the mainshock and aftershocks, deter- mine focal mechanisms using waveform inversion, and then analyze stress changes to estimate the fault type and stress transfer. Our relocated hypocenters show that this earth- quake sequence ruptured on at least three segments: the source mechanism of the main- shock exhibits dextral strike-slip motion (strike N72°W and dip 78° NE) on a west–east- trending fault that we call the Kalaotoa fault, whereas rupture of the other two segments located to the west and east of the mainshock (striking west-northwest and southeast, respectively) may have been triggered by this earthquake. The Coulomb stress change imparted by the rupture of these segments on nearby faults is investigated, with a focus on regions that experience a stress increase with few associated aftershocks. Of particular interest are stress increases on the central back-arc thrust just north of Flores and the north– south-striking Selayar fault in the northwest of our study region, both of which may be at increased risk of failure as a result of this unusual earthquake sequence.
Introduction
On 14 December 2021, anMw7.3 earthquake occurred beneath the Flores Sea (Fig. 1). According to Badan Meteorologi, Klimatologi, dan Geofisika (BMKG), otherwise known as the Indonesian Agency for Meteorology, Climatology, and Geophysics, the mainshock occurred at 03:20:23 UTC and was located at 7.59° S, 122.24° E, and 10 km depth (green star in Fig.1). BMKG used both regional and teleseismic data with the maximum azimuthal gap of 14° to locate this event. From about one minute following the mainshock, a sequence of after- shocks was recorded that continued until at least 20 March 2022 (Fig. S1, available in the supplemental material to this article), the date on which the most recent data used in this study was downloaded. BMKG issued tsunami warnings after the earth- quake, although it was later confirmed to be a strike-slip mecha- nism. Subsequently, tide gauge data from Flores Island recorded
a 7 cm high tsunami. Based on the BMKG shakemap, the earth- quake produced ground motion measured at between III and VI
1. Department of Earth Sciences—Bullard Labs, University of Cambridge, Cambridge, United Kingdom, https://orcid.org/0000-0002-9784-9865(PS); https://orcid.org/0000-0002-6977- 291X(NR); 2. Agency for Meteorology, Climatology, and Geophysics, Jakarta, Indonesia, https://
orcid.org/0000-0003-0623-0037(AS); https://orcid.org/0000-0002-8408-1236(AA); 3. Global Geophysics Research Group, Faculty of Mining and Petroleum Engineering, Institut Teknologi Bandung, Bandung, Indonesia, https://orcid.org/0000-0002-8941-7173(SW); https://
orcid.org/0000-0002-4844-8723(ADN); 4. Faculty of Engineering, Maranatha Christian University, Bandung, Indonesia; 5. Physical Science and Engineering, King Abdullah University of Science and Technology, Thuwal, Saudi Arabia, https://orcid.org/0000-0002-3550-7341(KHP);
6. Geophysical Engineering Study Program, Faculty of Mining and Petroleum Engineering, Institut Teknologi Bandung, Bandung, Indonesia; 7. Geological Engineering Department, Gadjah Mada University, Yogyakarta, Indonesia, https://orcid.org/0000-0003-1356-9645(GIM)
*Corresponding author: [email protected]
© 2022. The Authors. This is an open access article distributed under the terms of the CC-BY license, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
on the modified Mercalli intensity scale in Flores Island caused severe damage to more than 346 houses and injured at least 11 people.
Flores Island is located in the Sunda–Banda arc transition at the junction between the eastern limit of the Australian plate subduction margin, defined by the Sunda arc, and the western section of the collision zone between the Australian plate and the Banda arc (Hall, 2002;Audley-Charles, 2004). In the past 50 yr, there have been three earthquakes that were followed by
Figure 1.Map of the study area. The green star depicts the 14 December 2021 (Mw 7.3) mainshock and its focal mechanism from Badan Meteorologi, Klimatologi, dan Geofisika (BMKG). The historical earth- quakes that generated tsunamis in the region are denoted by yellow stars, the locations of which are sourced fromSupendiet al.(2020), whereas focal mechanisms are obtained from the Global Centroid Moment Tensor catalog. Inverted red triangles denote the BMKG seismic stations that provided the data used in this study. Major crustal faults (red lines) are sourced fromIrsyamet al.(2017). The study area (blue box) in the context of the broader Indonesian region is shown by the inset map.
a tsunami (yellow stars in Fig. 1), that is, Mw 8.3 in 1977, Mw 7.5 in 1992, and Mw 6.5 in 1995 (Supendi et al., 2020).
Earthquakes in the region are a consequence of subduction, collision, back-arc thrust faults, and terrestrial faults (Irsyam et al., 2017). Since 9–10 Ma, the Australian continent has been actively colliding with the western section of the Banda arc (Keep and Haig, 2010) beneath the islands of Timor, Sumba, and Flores (Hall, 1997). Based on seismicity data from 1963 through to 2021 (Fig. S2a), there have not been any shallow earthquakes (depth <100 km) with a magnitude Mw ≥ 7.0 in the region, and very few moderate earthquakes (Mw ≥ 5.0) at depths <100 km (see Figure S2b). Because of the struc- tural complexity of the western Banda arc and lack of data, not all faults have been properly mapped, including the one that ruptured to produce the 14 December 2021Mw7.3 earthquake.
To identify the fault responsible for the event, we relocate the mainshock and aftershocks of the earthquake sequence to determine more precise hypocenter locations compared to standard BMKG locations, undertake focal mechanism analy- sis, and investigate changes in fault stress. The goal is to char- acterize the fault in terms of its type, its kinematics, and the static stress changes caused by theMw 7.3 main event.
Data and Method
We use earthquake arrival times from the BMKG catalog between 14 December 2021 and 20 March 2022, which were extracted from continuous data recorded by the broadband seismic stations shown in Figure 1 (red inverted triangles).
A total of 1456 events with a magnitude range 1.9–7.3 were identified from 19,151P-wave and 5239S-wave picks obtained from the mainshock and aftershocks (Fig. S1). BMKG uses LocSAT (Bratt and Nagy, 1991)—a linearized inversion rou- tine that is part of the SeisComP3 program (Hanka et al., 2010) to help construct their earthquake catalog from the arrival time picks they make. As part of this process, hypocen- ter locations were determined using the IASP91 reference velocity model (Kennett and Engdahl, 1991).
Earthquake relocation
The double-difference method (Waldhauser and Ellsworth, 2000), implemented via the HypoDD program (Waldhauser, 2001), is used to relocate the mainshock and associated aftershocks based on data recorded by nearby stations (see Fig. 1). This is achieved by minimizing the residuals of the differences between theoretical and observed travel times for nearby earthquake pairs recorded at each station, under
the assumption that nearby sources have almost identical ray paths and travel times if the receiver is distant.
Therefore, event pairs are identified by locating nearby earth- quakes with similarP andStravel times (based on locations, origin times, and arrival time picks from the BMKG catalog).
The hypocentral separation maximum was set to 200 km, the maximum number of neighbors per event was set to 100, and the minimum number of links needed to define neighbors was set to 10. We imposed the maximum distance between cluster centroid and station of 200 km. These input parameter choices were made after extensive testing to deter- mine the combination of values that achieved the best results, noting that the final distribution of relocated hypocenters does not change significantly across a reasonable range of input values. We used a 1D velocity model based on CRUST 1.0 (Laskeet al., 2013) in the region of interest, combined with AK135 (Kennett et al., 1995), which provides velocities in the uppermost mantle.
Similar to the study bySupendiet al.(2022), we assess loca- tion uncertainty through application of a statistical resampling scheme based on the“bootstrap”method (Billings, 1994) for all relocated events. All residuals had Gaussian noise (standard deviation of 0.1 s) added to simulate realistic levels of noise in the test dataset. Using this dataset, we relocated all events and calculated their perturbation with respect to the initial double-difference location. This entire process was repeated 200 times, thus allowing for the estimation of 95% confidence ellipsoids for each event.
Focal mechanisms
We determine focal mechanisms of the mainshock and two aftershocks using Kiwi Tools(Heimann, 2011). We used wave- form traces from the broadband stations of the BMKG seismic network with epicentral distances up to 700 km. The inversion was achieved by fitting the amplitude spectra of the unrotated components (vertical, north–south, and east–west) with a dif- ferent band-pass filter at each iteration. We used the frequency range 0.015–0.05 for the mainshock and 0.02–0.06 Hz for the aftershocks. We used the low-frequency filter to stabilize the inversion process and reduce dependence on short-scale length crustal heterogeneities (Cescaet al., 2010), which are otherwise not considered in the modeling. The reference velocity model used in the generation of synthetic seismograms is IASP91 (Kennett and Engdahl, 1991). The misfit (M) between observed and synthetic ground motion is calculated using an L2 norm, defined as:
EQ-TARGET;temp:intralink-;df1;47;261
M
P
i usi−uoi2 P
i uoi2 ; 1
in whichusi anduoi are synthetic and observed displacements, respectively, for station componenti.
Coulomb stress
By constraining fault geometry and slip using the aftershock distribution and focal mechanisms computed earlier, we determined the Coulomb static stress change (King et al., 1994) caused by the coseismic slip through application of Coulomb 3.3 (Toda et al., 2011). The subsurface ruptures (length = 98.5 km, width = 22.5 km, and slip = 1.5 m for cluster
1; length = 38.9 km, width = 13.2 km, and slip = 0.8 m for cluster 2; and length = 53.1 km, width = 15.7 km, and slip
= 0.9 m for cluster 3) were defined using the empirical equation from Wells and Coppersmith (Wells and Coppersmith, 1994).
We calculated static stress changes assuming an elastic half- space, as implemented in Okada (1992), for the three strike- slip source faults (average of fault parameters for each cluster are shown in Table1) at 12.2 km depth, in which we assume a Poisson ratio of 0.25 and a coefficient of friction of 0.4 (Song et al., 2018). Four separate scenarios were tested, each corre- sponding to a different receiver fault on which future ruptures may occur. These include the central back-arc thrust near Flores, segments 1 and 2, and the Selayar fault near the southern tip of Sulawesi.
Table 1
Focal Mechanism Solution Data for Earthquakes That Occur in Each of the Three Clusters and Correspond to the Fault Segments That Are Interpreted
Date
(yyyy/mm/dd)
Time (UTC) (hh:mm:ss.s)
Latitude (°)
Longitude (°)
Depth
(km) Mw
Strike (°)
Dip (°)
Rake
(°) Note
2021/12/14 03:20:22.5 −7.5551 122.248 12.19 7.3 288 78 169 Segment 1
2021/12/14 03:23:30.1 −7.5204 122.2765 9.91 5.7 285 78 169
2021/12/14 03:38:13.5 −7.6137 121.5439 11.24 5.5 263 70 168
2021/12/14 03:47:01.5 −7.498 121.8055 17.61 5 275 79 170
2021/12/14 03:47:03.3 −7.538 121.6917 12.74 5.3 278 72 179
2021/12/14 04:16:12.6 −7.5994 121.6205 19.69 4.5 270 71 140
2021/12/14 04:34:02.9 −7.5336 121.6346 9.54 4.8 278 65 167
2021/12/14 05:46:41.3 −7.5071 121.7708 10.79 5 283 76 177
Average 278 74 167
2021/12/14 03:25:30.9 −7.6619 122.2914 12.2 5.8 319 62 163 Segment 2
2021/12/14 03:41:55.6 −7.7817 122.4038 9.91 5.5 301 80 172
2021/12/14 22:17:05.8 −7.6537 122.3442 9.02 4.7 335 70 160
Average 318 71 165
2021/12/15 03:52:34.5 −7.2538 121.2138 10.73 4.5 300 82 172 Segment 3
2021/12/15 09:14:30.1 −7.3775 121.2445 9.14 4.2 281 79 168
2021/12/15 22:24:01.6 −7.2622 121.1407 9.03 4.6 279 81 155
2021/12/18 02:10:42.5 −7.2326 121.1348 12.14 4.5 283 77 166
2021/12/21 11:14:55.0 −7.3238 121.2896 9.5 4.3 277 72 175
Average 284 78 167
Results and Discussion
We relocated the mainshock and a total of 1403 aftershocks from the 2021 Flores Sea earthquake from 14 December 2021 to 20 March 2022 (Fig.2); the remaining 53 events (less than 4% of the total) were relocated above the surface and, therefore, discarded. This reduction in the total number of events is typical, and likely due to data noise and use of an approximate reference model conspiring to introduce errors into the relocation. The BMKG catalog, from which initial locations and arrival time’s picks are sourced, typically uses fixed depths for shallow events, which is likely why this incon- sistency was not picked up on earlier. The difference between the initial locations (from BMKG) and relocated aftershocks
can be compared in cross section and map view (Fig. S3).
Following HypoDD relocation, the magnitudes of the travel- time residuals are clearly reduced (see Fig. S4), which is sug- gestive of more robust hypocenters. A significant proportion of events in the BMKG catalog are fixed at 10 km depth, but these are now distributed in depth (Fig. S3) following relative relo- cation. Relative errors in location based on the bootstrap analy- sis method are shown in Figure3. Mean horizontal and vertical
Figure 2.Relocated epicenters of theMw7.3 mainshock (green star) and aftershocks. Colored dots depict the time of occurrence of the after- shocks (measured in days) relative to the mainshock.
location errors are typically less than 3 km (represented by the longest radius of the ellipsoid), and the corresponding maximum mislocations are less than 7 km (Fig.3and Fig. S5). However, due to limited station coverage to the east–northeast of the earth- quakes we study (Fig.1), horizontal location uncertainty is larger in the east–west direction compared to the north–south direc- tion, although depth uncertainty is still greater.
The distribution of the relocated aftershocks exhibit an interesting pattern, with at least three clusters (Fig.2) appear- ing to be associated with the Mw 7.3 earthquake, which we interpret as a sign of slip along what we call the Kalaotoa fault system. Cluster 1 is caused by slip on the∼100 km long east– west main fault that also accommodates the Mw 7.3 earth- quake, which appears to trigger the∼40 km long southeast-ori- ented segment in the east (cluster 2) and∼50 km long west- northwest-oriented segment in the west (cluster 3). The width
of these three segments, based on visual analysis of the after- shock distribution, is∼20 km. Such faults appear to be quite rare in Indonesia, but can be found in other regions, for exam- ple, theMw6.1 Ludian earthquake, China (Niuet al., 2019), the 2016 Mw 7.1 Kufeastmamoto earthquake, Japan (Lin and Chiba, 2017), and on the 2019 Ridgecrest earthquake sequence, California (Fialko and Jin, 2021). In addition, there is another cluster to the west of cluster 3 that we interpret to occur on an extension of the Selayar fault (Fig. 4), which is a normal fault (Irsyamet al., 2017). Furthermore, another cluster to the south of cluster 2 may be associated with the central back-arc thrust
Figure 3.(a) Map view of relative relocation 95% confidence ellipsoids for each of the earthquakes in the sequence, (b) longitude slice, and (c) latitude slice.
fault that was triggered by the Mw 7.3 event (Fig. 4). In this study, it is not easy to differentiate between what could be regarded as an aftershock or a triggered event, which is why these terms are used somewhat interchangeably. However, given that the mainshock occurs on the fault segment that defines cluster 1, it would be reasonable to define events in the other clusters as triggered events, particularly as finite fault predictions from teleseismic data estimate the rupture length of theMw7.3 event to be less than 100 km.
The misfit between the observed and synthetic traces following inversion for the mainshock and aftershock focal mechanisms has a value of ∼0.45, indicating a satisfactory fit according to the criterion ofCescaet al.(2010). We produce double-couple parameters that correspond to a right-lateral fault, based on the aftershock distribution for the earthquake, with an average strike of N82°W and dip 74° SW (segment 1), an average strike of N42°W and dip 71° SW (segment 2), and
an average strike of N76°W and dip 78° SW (segment 3), as shown in Figure 4and Table 1. Interestingly, both the after- shock distribution and focal mechanisms suggest that segment 3 is offset to the north of segment 1 by∼10–15 km; this“gap”
appears to be accompanied by a decrease in aftershock activity and is worthy of further investigation. Uncertainties associated with the focal mechanisms determined for the mainshock, and two representative aftershocks for segment 2 and segment 3 are shown in Figures S9–S11, and indicate that they are well con- strained by the data.
Figure 4.Focal mechanism solutions and relocated hypocenters associated with theMw 7.3 mainshock and associated aftershocks. The dashed yellow line suggests that theMw7.3 mainshock and aftershock distri- bution are caused by three fault segments. Focal mechanisms depict three segments (segment 1, green; segment 2, red; and segment 3, yellow) and are not color coded by depth.
Insight into the possible locations of future earthquakes (Stein and Lisowski, 1983; King et al., 1994) can be gained by comparing the Coulomb stress change and aftershock dis- tribution. The stress changes caused by theMw7.3 mainshock (Fig. 5) reveal several interesting patterns. In the case of Figure 5a, it seems clear that there is a stress increase at the endpoints of the source faults, which is unsurprising because that would be, in which we might expect future failure on the same fault surface to occur; Figure 5b is fairly similar to Figure5a, as one might expect; Figure5cis interesting, because it shows that there is a stress increase on the back-arc thrust fault, in which we see aftershocks, which may reflect rupture on that fault. However, there are regions of stress increase on either side of these aftershocks, which may suggest that failure is now more likely in these regions of the back-arc thrust fault.
Similarly, Figure5dis interesting because there is a clear zone
of stress increase at the southern end of the Selayar fault, but not a lot of aftershocks, except on its western segment. This suggests that risk of failure on this fault has increased. A more detailed analysis of the implications of relatively high static stress transfer to the northwest and southeast with respect to the mainshock is urgently needed to investigate possible future fault rupture scenarios.
Figure 5.Modeled Coulomb stress change for four different receiver faults (a) strike-slip fault corresponding to segment 1 that ruptured to produce theMw7.3 mainshock; (b) strike-slip fault corresponding to segment 2;
(c) thrust fault with strike equal to in which segment 2 intersects the central back-arc thrust; (d) roughly north–south normal fault corre- sponding to the Selayar fault. The red and blue colors denote positive and negative Coulomb stress changes, respectively. The green star and white circles identify the locations of the mainshock and aftershocks, respec- tively. The green lines denote the three source faults.
The complex fault system in this region accommodates the varying orientations of the ongoing plate convergence in the eastern Sunda–Banda arc, which forms a curvilinear feature bounded to the south by the northern Australian continental margin and to the east by the Arafura Sea margin (Norvick, 1979; Hinschberger et al., 2005; see inset map in Fig. 1).
Further complexity is added by the presence of the transition from subduction to arc-continent collision to the south of Flores, which involves a slab tear and underthrusting fore-arc lithosphere (Zhang and Miller, 2021). It is, therefore, unsurpris- ing that a variety of fault types across a range of orientations are present to accommodate these changes. Identification and under- standing of these fault systems is a major challenge, but essential if we are to understand seismic hazard in southeast Indonesia.
Conclusions
TheMw 7.3 Flores Sea earthquake and associated aftershocks were caused by slip along the newly identified dextral strike- slip Kalaotoa fault system, which appears to be composed of three segments, with lengths of∼100,∼50, and∼40 km. The segment to the west is offset to the north of the central segment by∼10–15 km. Our Coulomb stress change predictions based on the three source segments were undertaken for a number of realistic scenarios that include strike-slip, normal, and thrust faults, all of which occur in the study region as a result of its complex tectonic setting. We find that there are a lack of aftershocks in regions of high Coulomb stress change that may eventually yield a large earthquake on adjoining fault sys- tems, including the central back-arc thrust adjacent to Flores and the Selayar fault near Sulawesi. A more detailed study of the implications of relatively high static stress transfer to these two areas with respect to the main shock is urgently needed to properly assess the region’s seismic hazard potential.
Data and Resources
The earthquake dataset exploited in this study is sourced from Badan Meteorologi, Klimatologi, dan Geofisika (BMKG). All figures in the main article and supplemental material were gen- erated using The Generic Mapping Tools (Wessel and Smith, 1998). Topography and bathymetry data were obtained from the Digital Elevation Model Nasional (https://tanahair .indonesia.go.id/demnas/#/demnas) and Batimetri Nasional (https://tanahair.indonesia.go.id/demnas/#/batnas). The earth- quake relocation catalog data are available at https://doi.org/
10.5281/zenodo.6592435. The supplemental material includes Figures S1–S11. All websites were last accessed in March 2022.
Declaration of Competing Interests
The authors acknowledge that there are no conflicts of interest recorded.
Acknowledgments
Badan Meteorologi, Klimatologi, dan Geofisika (BMKG) is thanked for providing data from their earthquake archive for use in this study. The University of Cambridge funded this research through a Herchel Smith Research Fellowship awarded to Pepen Supendi Konsorsium Gempabumi dan Tsunami. BMKG is also acknowledged for their support.
References
Audley-Charles, M. G. (2004). Ocean trench blocked and obliterated by Banda forearc collision with Australian proximal continental slope, Tectonophysics 389, nos. 1/2, 65–79, doi: 10.1016/
j.tecto.2004.07.048.
Billings, S. D. (1994). Simulated annealing for earthquake location, Geophys. J. Int.118,no. 3, 680–692, doi:10.1111/j.1365-246X.1994.
tb03993.x.
Bratt, S. R., and W. Nagy (1991). The LocSAT Program, Science Applications International Corporation (SAIC), San Diego, California.
Cesca, S., S. Heimann, K. Stammler, and T. Dahm (2010). Automated procedure for point and kinematic source inversion at regional dis- tances,J. Geophys. Res.115,no. B6, doi:10.1029/2009JB006450.
Fialko, Y., and Z. Jin (2021). Simple shear origin of the cross-faults ruptured in the 2019 Ridgecrest earthquake sequence, Nature Geosci.14,no. 7, 513–518, doi:10.1038/s41561-021-00758-5.
Hall, R. (1997). Cenozoic plate tectonic reconstructions of SE Asia, Geol. Soc. Lond. Spec. Publ. 126, no. 1, 11–23, doi: 10.1144/
GSL.SP.1997.126.01.03.
Hall, R. (2002). Cenozoic geological and plate tectonic evolution of SE Asia and the SW Pacific: computer-based reconstructions, model and animations, J. Asian Earth Sci. 20, no. 4, 353–431, doi:10.1016/S1367-9120(01)00069-4.
Hanka, W., J. Saul, B. Weber, J. Becker, P. Harjadi, Fauzi, and GITEWS Seismology Group (2010). Real-time earthquake moni- toring for tsunami warning in the Indian Ocean and beyond,Nat.
Hazards Earth Syst. Sci.10,no. 12, 2611–2622, doi:10.5194/nhess- 10-2611-2010.
Heimann, S. (2011). A robust method to estimate kinematic earth- quake source parameters,Ph.D Thesis, University of Hamburg.
Hinschberger, F., J.-A. Malod, J.-P. Réhault, M. Villeneuve, J.-Y. Royer, and S. Burhanuddin (2005). Late Cenozoic geodynamic evolution of eastern Indonesia, Tectonophysics 404, nos. 1/2, 91–118, doi:
10.1016/j.tecto.2005.05.005.
Irsyam, M., S. Widiyantoro, D. H. Natawidjaya, I. Meilano, A.
Rudyanto, S. Hidayati, W. Triyoso, N. R. Hanifa, D. Djarwadi, L.
Faizal, et al. (2017). Peta sumber dan bahaya gempa Indonesia tahun 2017, Pusat Penelitian dan Pengembangan Perumahan dan
Permukiman, Kementerian Pekerjaan Umum dan Perumahan Rakyat (in Indonesian).
Keep, M., and D. W. Haig (2010). Deformation and exhumation in Timor: Distinct stages of a young orogeny,Tectonophysics 483, no. 1, 93–111, doi:10.1016/j.tecto.2009.11.018.
Kennett, B. L. N., and E. R. Engdahl (1991). Travel times for global earthquake location and phase identification,Geophys. J. Int.105, no. 2, 429–465, doi:10.1111/j.1365-246X.1991.tb06724.x.
Kennett, B. L. N., E. R. Engdahl, and R. Buland (1995). Constraints on seismic velocities in the Earth from travel times,Geophys. J. Int.
122,no. 1, 108–124, doi:10.1111/j.1365-246X.1995.tb03540.x.
King, G. C. P., R. S. Stein, and J. Lin (1994). Static stress changes and the triggering of earthquakes,Bull. Seismol. Soc. Am.84, no. 3, 935–953.
Laske, G., G. Masters, Z. Ma, and M. Pasyanos (2013). Update on CRUST1.0—A 1-degree Global Model of Earth’s Crust, EGU General Assembly 2013, 7–12 April, 2013, Vienna, Austria.
Lin, A., and T. Chiba (2017). Coseismic conjugate faulting structures produced by the 2016 M w 7.1 Kumamoto earthquake, Japan,J.
Struct. Geol.99,20–30, doi:10.1016/j.jsg.2017.05.003.
Niu, Y., S. Wang, W. Zhu, Q. Zhang, Z. Lu, C. Zhao, and W. Qu (2019). The 2014 Mw 6.1 Ludian earthquake: The application of RADARSAT-2 SAR interferometry and GPS for this conjugated ruptured event, Remote Sens. 12, no. 1, 99, doi: 10.3390/
rs12010099.
Norvick, M. S. (1979). The tectonic history of the Banda arcs, eastern Indonesia: A review,J. Geol. Soc.136,no. 5, 519–526, doi:10.1144/
gsjgs.136.5.0519.
Okada, Y. (1992). Internal deformation due to shear and tensile faults in a half-space,Bull. Seismol. Soc. Am.82,no. 2, 1018–1040, doi:
10.1785/BSSA0820021018.
Song, G., C. Wang, and B. Wang (2018).Structural Health Monitoring (SHM) of Civil Structures, MDPI AG, Illustrated edition.
Stein, R. S., and M. Lisowski (1983). The 1979 homestead valley earthquake sequence, California: Control of aftershocks and post- seismic deformation, J. Geophys. Res.88, no. B8, doi: 10.1029/
JB088iB08p06477.
Supendi, P., A. D. Nugraha, S. Widiyantoro, C. I. Abdullah, N.
Rawlinson, P. R. Cummins, C. W. Harris, N. Roosmawati, and M. S. Miller (2020). Fate of forearc lithosphere at arc-continent collision zones: Evidence from local earthquake tomography of the Sunda-Banda arc transition, Indonesia, Geophys. Res. Lett.
47,no. 6, doi:10.1029/2019GL086472.
Supendi, P., S. Widiyantoro, N. Rawlinson, A. Wibowo, P. Priyobudi, K. H. Palgunadi, A. D. Nugraha, I. Imran, G. I. Marliyani, D.
Daryono,et al.(2022). Analysis of the 2021 Semangko Bay earth- quake sequence in southern Sumatra, Indonesia, using broadband seismic network data,Seismol. Res. Lett.doi:10.1785/0220210304.
Toda, S., R. S. Stein, V. Sevilgen, and J. Lin (2011). Coulomb 3.3 Graphic- rich deformation and stress-change software for earthquake, tectonic, and volcano research and teaching-user guide, Rept. 2011–1060, Reston, VA, Open-File Rept., doi:10.3133/ofr20111060.
Waldhauser, F. (2001). hypoDD—A program to compute double- difference hypocenter locations, Open-File Rept. 2001–113, doi:
10.3133/ofr01113.
Waldhauser, F., and W. L. Ellsworth (2000). A double-difference earthquake location algorithm: Method and application to the northern Hayward fault, California,Bull. Seismol. Soc. Am. 90, 1353–1368, doi:10.1785/0120000006.
Wells, D. L., and K. J. Coppersmith (1994). New empirical relation- ships among magnitude, rupture length, rupture width, rupture area, and surface displacement,Bull. Seismol. Soc. Am.84,no. 4, 974–1002.
Wessel, P., and W. H. F. Smith (1998). New, improved version of generic mapping tools released, Eos Trans. AGU 79, no. 47, 579–579, doi:10.1029/98EO00426.
Zhang, P., and M. S. Miller (2021). Seismic imaging of the subducted Australian continental margin beneath Timor and the Banda arc collision zone, Geophys. Res. Lett. 48, no. 4, doi: 10.1029/
2020GL089632.
Manuscript received 15 April 2022 Published online 1 August 2022