Muksin Jurnal15
Muksin Jurnal15
Muksin Jurnal15
○
E
ABSTRACT
On 6 December 2016, an M w 6.5 earthquake occurred in highlighting the need for better understanding of active fault-
Pidie Jaya, Aceh, about 30 km to the north of the Sumatran ing and seismic hazard in this region.
fault (SF) that killed more than 100 people and destroyed
∼3000 buildings. Mainshock focal mechanism inversions
using regional Agency for Meteorology, Climatology, and
Geophysics of Indonesia (BMKG) broadband data and tele- Electronic Supplement: Damage photos, figures of relative main-
seismic waveform data all indicate a strike-slip event with a shock relocation, depth resolution of focal mechanism solution
centroid depth of 11–15 km. The observed macrointensity of the mainshock, focal mechanism solution of the mainshock
data show that most of the damaged buildings are distributed using a cut-and-paste (CAP) method, teleseismic finite-fault
along the coast, approximately perpendicular to the ruptured inversion of the mainshock for the fault plane with a strike of
fault strike instead of parallel with it. The strong shaking and 148°, and the double-difference hypocenter relocation of the
damage sites are primarily located on the coastal sedimentary aftershocks using hypoDD.
soils, highlighting the importance of site conditions in deter-
mining risk. We used one-month data recorded by nine INTRODUCTION
temporal broadband stations to locate aftershocks with grid-
search and double-difference algorithms, thereby resolving a On the early morning of Wednesday, 7 December 2016,
linear trend of seismicity aligned in a northeast–southwest 05:03:33 local time (6 December 22:03:33 UTC), an M w 6.5
direction. The refined aftershock locations indicate a left- earthquake rocked the northern coastal area of north Aceh in
lateral rupture that is in agreement with the preliminary the Pidie Jaya region (hereafter, Pidie Jaya earthquake). The
finite-fault slip inversion as well as geomorphic signatures Agency for Meteorology, Climatology, and Geophysics of In-
of local geological structure. Using a well-located M L 4.2 donesia (BMKG) reported an epicenter location of 5.29° S and
aftershock for path calibration, we relocated the mainshock 96.22° E, with a shallow depth of 15 km. The Pidie Jaya earth-
epicenter with regional P-wave arrivals. The refined epicenter quake occurred at the moment before dawn when many people
falls within the cloud of the well-located aftershocks, whereas were still asleep and many others were in the mosques for
locations from the global and regional catalogs are located morning prayer. The earthquake caused heavy damage within
10–20 km away. Aftershock focal mechanisms determined a radius of ∼35 km of the epicenter. In total, 101 people were
by the first motion reveal similar solutions as the mainshock. killed (mainly caused by buildings collapsing), over 800 people
This earthquake sequence ruptured a previously unidentified were injured, thousands were displaced, and thousands of
fault that is either located at the west of the fault that pro- buildings collapsed or were severely damaged. The estimated
duced the 1967 M w 6.1 earthquake sequence or is actually at economic loss was about $139 million U.S. according to the
the same fault. The Pidie Jaya earthquake and other off-SF National Disaster Management Authority of Indonesia. This
events suggest strong distributed crustal deformation in Aceh, earthquake was well recorded by the national strong-motion
Table 1
Hypocenter Locations of the Mainshock from Three Institutions
Origin Time (UTC)
Number (yyyy/mm/dd hh:mm:ss) Latitude (°) Longitude (°) Magnitude (M w ) Depth (km) Source
1 2016/12/06 22:03:35 5.290 96.220 6.5 15 BMKG
2 2016/12/06 22:03:32 5.281 96.108 6.5 8.2 USGS
3 2016/12/06 22:03:33 5.320 96.070 6.5 10 GFZ
BMKG, Agency for Meteorology, Climatology, and Geophysics of Indonesia; USGS, U.S. Geological Survey.
▴ Figure 3. Postevent survey photographs of closely adjacent sites showing drastically different responses of buildings to similar
ground-shaking intensity. (a) Site 31: complete collapse of a building, Kec. Meurah Dua at 5.233° N, 96.268 ° E; and (b) site 32: minimal
damage of an office at the same subdistrict at 5.235° N, 96.274° E.
intensities derived from interviews have larger uncertainties, a few damage reports from inland areas to the south of the main-
because they are subjective; on the other hand, inspections of shock epicenter. The direction of the building collapse or move-
buildings resulted in more accurate intensity estimations. Photo- ment agreed well with a left-lateral strike-slip focal mechanism
graphs of damage at these sites (numbered in Fig. 2) can be (e.g., sites 10, 11, 27, and 40 in Ⓔ Fig. S1), consistent with the
found in Ⓔ Figure S1 (available in the electronic supplement fault geometry defined by the relocated aftershocks. These pho-
to this article), with representative damage shown in Figure 3. tos provide independent evidence for the ruptured fault plane,
The number of fatalities and damaged buildings are shown in preferring the one with a strike of 055°. The aftershock distri-
Table 3. The maximum inferred intensity is V in SIG-BMKG or bution also implicates the fault plane with a strike of 055°, rather
IX in MMI, corresponding with total building collapse (e.g., than the orthogonal plane with a strike of ∼149°, which is more
Fig. 3). It is likely that the complete building collapse was caused parallel to the coast and to the SF.
by a combination of poor structural design, strong shaking, and We therefore assume a relocated hypocenter of 96.225°
site effect. At several sites, neighbored structures to collapsed E/5.235° N/15 km, M w of 6.5, and a strike-slip focal mechanism
buildings exhibited only light damage, indicating very different of the mainshock and predict the distribution of peak ground
building design and quality. For instance, sites 31 (Fig. 3a) and velocity (PGV) and peak ground acceleration (PGA) based on
the ground-motion prediction equation (GMPE) in Newmark
32 (Fig. 3b) are directly adjacent, but the degrees of damage are
and Hall (1982) and Zhao et al. (2006). We then convert the
substantially different. The most severe building damage was
PGV into intensity, with V S30 (the average shear-wave velocity
found near the relocated mainshock epicenter (see the details
down to 30 m) taken into account (Fig. 2; Wald et al., 2005).
of relocation in the Seismicity Relocation and Aftershock Focal
The conversion of ground motion to intensity is based on the
Mechanism Inversion section). The damage zone spreads in a empirical formula of the ground-motion-to-intensity conversion
nearly east–west direction, parallel to the coastline, with only equation as described by Worden et al. (2012). It shows that the
maximum predicted intensity is VIII in MMI at the relocated
Table 3 epicenter location. This predicted intensity is slightly smaller
Casualties and Building Damages Produced by the than that implied by the observed damage (see Fig. 2 and Ⓔ
Earthquake (BNPB) Fig. S1; e.g., indexes 18, 38, and 40), probably due to poor build-
ing quality. The predicted intensity map also shows an elongated
Number Description Total distribution of high intensity along the coast. The east–west-ori-
1 Deaths 101 ented damage zone in the residential area was likely strongly re-
2 Injured 857 lated to site effects rather than to the finite rupture process,
3 House damaged because the soft soil (very low V S30 ) zone is distributed along
Heavy 2992 the coast instead of along the strike of the fault (Fig. 4). Further
Moderate 94 inland to the south, the population is relatively low, the area is
Light 8582 hilly, and the soil is harder and thinner than in the coastal area.
Thus, the damage reported in this area is much less.
BNPB, National Disaster Management Authority of The thickness of sediment in the coastal region is ∼4:0 km
Indonesia. according to CRUST1.0 (Laske et al., 2013). However,
▴ Figure 5. HVSR at Pijay-Net stations, solid and dashed lines showing the average HVSR and its 95% confidence level, respectively. The
variable frequency response patterns and different peak magnitudes of HVSR illustrate the importance of local site conditions for shallow
amplification of seismic waves. Site classifications based on HVSR (see table 2 in Zhao et al., 2006) are shown on each subfigure below
the station names.
▴ Figure 8. Aftershock distribution and focal mechanism solution of the 2016 Pidie Jaya earthquake after relocation using grid-search
and hypoDD methods. Numbers on the focal mechanism solution are the event ID in the Ⓔ electronic supplement. Green squares indicate
the three events with their waveforms recorded at Station A22 (gray triangle) as shown in Figure 9. Blue circles indicate the sequence of
an historically significant M 6.1 earthquake in 1967.
hypoDD (Waldhauser and Ellsworth, 2000; Waldhauser, (shown as blue circles). The surface projection of the relocated
2001). The relocation results in map view and vertical profiles seismicity shows a clear lineation that has a fault-length dimen-
are shown in Figure 8. We also plotted USGS locations for the sion of ∼20 km and strikes toward an azimuth of ∼N042°,
historical 1967 M 6.1 earthquake sequence on the same figure agreeing well with the strike of one of the fault-plane solutions
▴ Figure 10. The teleseismic finite-fault inversion result of the mainshock. (a) The teleseismic displacement waveform fits where the
actual data are shown in black and the synthetic in red. The station names are indicated at the beginning of each waveform pair along
with the azimuth (upper) and epicenter distance (lower) in degree. The peak amplitude in the data in micrometers is displayed at the end
of each waveform. (b) The depth profile of the finite-fault slip model in which the contours indicate the rupture time and the arrows
indicate the slip direction. (c) The moment rate function of the finite-fault model.
plitude on the tangential components, which is typical for a FINITE-FAULT SLIP MODEL
strike-slip focal mechanism. As shown, the centroid depth
of the earthquake is ∼11 km, which is in agreement with the Although the magnitude of the Pidie Jaya earthquake is rela-
finite-slip distribution derived from teleseismic inversion, as tively small (M w 6.5) compared with other damaging events,
shown in the Finite-Fault Slip Model section. (The relocated we attempted a teleseismic finite-fault inversion to resolve the
aftershocks and focal mechanism can be found in the Ⓔ elec- fault plane of the earthquake. We downloaded broadband
tronic supplement.) teleseismic P waves recorded by the Global Seismic Network