The majority of tectonics along the northwestern margin of Pakistan owes to head on collision of the Eurasian with Indian plate. This oblique translation and rotation between these continental plates lead to the formation of the Chaman fault accommodating left-lateral shear. The eastern side of Indian plate comprises a series of fold and thrust belts of Sulaiman & Kirthar Ranges which are a southwestern extension of the Himalayan mountain belt (Bender and Raza, 1995; Kazmi and Jan, 1997). The Kirthar fold-thrust belt (KFTB) bounds the Sulaiman fold thrust belt (SFTB) in the south and these fold-thrust belts are separated by a tight hairpin bend of Quetta Syntaxis, Katawaz basin, containing semi-rigid mass is sandwiched between SFTB and CF and was responsible for the lobate shape of Sulaiman Range (Chandra, 1979). The significant neo-tectonic structure within this region includes Sulaiman & Kirthar Ranges, Chaman Fault Zone and Makran Range (Frohling and Szeliga, 2016).
During Eocene time, the Indian and Eurasian plates accreted sediments collectively along the western side of Pakistan and are inferred to represent a Pakistan-Iran Makran Microplate (P-IMM). The deformation and deposition of accreted sediments along the southern side of (P-IMM) subduction lead to the formation of Makran Accretionary prism/wedge and marks the offshore boundary as ‘Oman Trench’ (Burg et al., 2011). These accretionary wedge sediments scraped off the Arabian plate during the Tertiary period and stretches from Iran to Pakistan and off the coast (White, 1982; Harms et al., 1984). The termination extent of an accretionary prism is 450km in the north against the two mountain ranges of Chagai and Raskoh signifying the magmatic arc of the Makran subduction regime. The Chagai magmatic arc having Quaternary-Tertiary magnetism extends northward into Afghanistan (Shareq et al., 1977). The same arc is terminated by Chaman Fault Zone (CFZ) and Harirud Fault Zone (HFZ) from east to west and is convex towards the southward side (Farah et al., 1982).
The geodynamics of the Chagai-Makran region has a unique tectonic regime. In the south, the region is bounded by the subduction zone of the Cretaceous age. This subduction zone itself is considered to be a rare surviving feature when compared to Zagros subduction in Iran and the Indus suture zone in the Himalayas which have been eliminated by major continental collisions. Makran accretionary prism lies along the southern region and is considered to be one of the biggest accretionary prisms (Fig. 3). This accretionary prism is the hanging wall of Makran subduction zone and occupies a majority of the southern Makran region. The same accretionary wedge forms during the subduction of the Arabian plate’s oceanic portion against the overriding Eurasian plate’s Afghan Micro-continent. A transform boundary connecting the Chaman fault, Ornach-Nal fault and Murray Ridge with Owen Fracture Zone in the south and Heart fault in the north. The major structural high within the region is Murray Ridge. This onshore ridge considers equivalent to the offshore Kirthar Fold Belt in Pakistan (Clift et al., 2002; Edwards et al., 2000). The Murray Ridge is bounded by the offshore Indus Basin which serves as a major plate boundary between the Arabian and Eurasian Plates (Edwards et al., 2000).
Seismic Data Compilation and Methodology
Local seismic data from the Centre for Earthquake Studies (CES) and Pakistan Metrological Department (PMD) were used to locate earthquakes within the Khuzdar region. These seismic networks consist of 28 seismic stations, operating in the upper, central and southern parts of Pakistan i.e. Kohat-Potwar plateau, Punjab plains and big cities including Peshawar, Islamabad, Karachi, Lahore and Quetta. These seismic networks have been functional since 1976 and data is transmitted through the satellite network. The density of stations in non-uniform but threshold magnitude varies between central and southern parts of Pakistan as 1.0 and 2.7 respectively. The majority of these seismic stations are strong-motion and broadband seismometers that are operating in continuous mode and are provided data with a 24-bit digitizer.
SEISAN software package is used to determine hypocentral parameters i.e. depth, origin time and location using HYPOCENTER code (Lienert, 1991; Lienert et al., 1986; Lienert and Havskov, 1995; Ottemöller et al., 2016). Usually, the stations with good azimuthal coverage (> 180°) yield better results in focal depth and focal mechanism. Therefore, in the current analysis, only those stations are considered to have a good signal-to-noise ratio (SNR) as well as improved azimuthal coverage. The azimuthal coverage of the Khuzdar earthquake is above 150°. Similarly, having lower RMS (Root Mean Square) values for the location of the event on different seismic stations reflects a small residual between predicted and observed arrival times (Lienert and Havskov, 1995). The same recorded event has 0.8 sec residual time (RMS) based on local seismic data.
To have consistent results of source inversion and focal mechanism solution (FMS), stations having high noise, amplitude or time problems were removed. For evaluating moment tensor inversion, we used only nine out of 28 seismic stations (Fig. 4). Seismic waveform data were pre-processed using Seismic Analysis Code (SAC) software. Near subsurface effects were removed by filtering source data between 0.07 and 0.1 Hz (Fukuyama and Dreger, 1998). Seismic instrument response, mean and trend were also removed from the data. Previously available FMS within 100km of the Khuzdar event are taken from the Harvard Centroid Moment Tensor (CMT) catalog. The majority of these events were crustal depth (~ 30km) and moderate to large magnitude earthquakes (4.3–7.7). These focal mechanism solutions were further used to determine regional stresses using the Win-Tensor software suite.
Additional to waveform data and source mechanism, the previous seismicity around 50km and 100km of the Khuzdar region were also analyzed to understand spatio-temporal patterns. First, threshold magnitude for twelve years period (between 2010 and 2022) was calculated using the frequency magnitude distribution of events. Threshold magnitude is the minimum magnitude above which the catalog is complete. It is calculated using the maximum curvature method of Wiemer and Wyss (1997).
For finite region and time earthquake frequency magnitude distribution decay as a power law called Gutenberg-Richter (GR) law (Gutenberg and Richter, 1942). This is one of the most documented and general empirical relations used for earthquake studies. If N(M) is a number of earthquakes of magnitude M above Mc, then frequency magnitude distribution (FMD) can be expressed as:
$$N\left(M\right)={10}^{a-bM}$$
where a and b are productivity and relative abundance of large to small events respectively (Gutenberg and Richter, 1942; Wiemer and Wyss, 1997). These parameters describe different characteristics of an earthquake, such as if an earthquake catalog is complete for a certain time and magnitude then a can be used for earthquake recurrence time and activity rate. This is a key input parameter for seismic hazard studies. The scaling parameter b (b-value) slope of FMD is extensively used to elaborate different aspects of earthquakes (Schorlemmer and Wyss, 2005; Farrell and Smith, 2009; Gulia and Wiemer, 2010; Scholz, 2015; Khan and Mittnik, 2018; Kulhanek, Persson and Nuannin, 2018). Although for a long time window its value is close to unity (b ≈ 1) but significant variations have been observed for a shorter time window, in the range of 0.5 to 2.5 (El-Isa and Eaton, 2014). Typically FMD is affected by several factors and finding the mechanisms which are responsible for these variations is still debatable. However many studies attributed these variations to changes in stress conditions (Scholz, 2015), acoustic emissions of rocks (Scholz, 1968; Meredith, Main and Jones, 1990; Amitrano, 2003; W. Goebel et al., 2013; Martínez-Garzón et al., 2014), with depth (Wiemer and Wyss, 1997; Zúñiga and Wiemer, 1999; Gerstenberger, Wiemer and Giardini, 2001; Spada et al., 2013 and Scholz, 2015), with the tectonic regime (Schorlemmer, Wiemer and Wyss, 2005), loading rate (Amelung and King, 1997), and structural or compositional properties (Wyss, Sobolev and Clippard, 2004). Other reasons for b-value perturbations may be the degree of material heterogeneity or crack density (Mogi, 1963; Sobiesiak et al., 2007) or the change of thermal gradients (Warren and Latham, 1970).
Focal Mechanism Solution
The focal mechanism solution (FMS) is a powerful tool to investigate seismo-tectonic deformations and present-day stress patterns along tectonic regimes through seismological observations (Palano et al., 2012; Ousadou et al., 2014). These studies are important for the characterization of faulting styles and the rupture process. The seismic waveform that was recorded at a certain observatory has a linear relationship between source-time function, Green’s function and instrument response (Aki and Richards, 1980; Jost and Herrmann, 1989). The summation of many delta functions is supposed to be a source-time function when the instrument response is known of each station. Similarly, the amplitude and source-time function also depend upon the size of an earthquake (Stein and Wysession, 2002). Green’s function on the other hand depends upon the velocity model used and depicts elastic or inelastic seismic wave propagation within the subsurface. The elastic response represents wave propagation, reflection and refraction through the boundaries; whereas, its attenuation is inelastic in nature along the medium through which it propagates (Stein and Wysession, 2002). To compute Green’s function (synthetic ground motions), Kohketsu (1985) discrete wavenumber technique was used which was applicable for horizontal layered structures (Kohketsu, 1985).
FMS strongly relies upon the velocity models being used and the anisotropy of earth subsurface structures (Vavrycˇuk, 2005). To have minimum artifacts and reliable solution of inversion results, an appropriate velocity model should be selected (Kissling and Kradolfer, 1995). Two velocity structures are available for Pakistan. The regional scale (3-D) velocity model was proposed by Johnson and Vincent (2002) and developed for India and Pakistan (Johnson and Vincent, 2002). The local scale (1-D) velocity model was proposed by Soomro and Iqbal (2021) and it is validated for the regions of central and upper Indus Basin, Pakistan (Soomro et al., 2021). They used 486 events for earthquake location and suggested an average residual redunction in RMS was less than 0.9s. We used Soomro and Iqbal (2021) local velocity model, because they used the same local seismic network as that for the Khuzdar event.
To investigate the FMS of the Khuzdar earthquake, we used the moment tensor technique of Yuji Yagi and Nishimura's (2011) for source mechanism of the mainshock. It resolved the depth of the event along with fault parameters i.e. strike, rake and dip (Yagi and Fukahata, 2008). The source geometry was characterized by combining six components of the moment tensor. The model parameters, such as depth and nodal planes were taken on basis of best fit, representing the lowest variance reduction between observed and synthetic waveforms. The waveform having the best match between observed and synthetic depicts low variance. Similarly, the variance is sensitive to the amplitude and time shift of both observed and synthetic seismograms. To have maximum cross-correlation and minimum variance value, a small shifting of the seismic signal was performed. During the inversion practice, all seismograms must be assigned with uniform weights.