Modelling of seismicity in southern pakistan using GIS techniques

This study is aimed to elucidate individual fault response and peculiar earthquake characteristics in complex seismotectonic environment of southern Pakistan (Balochistan, Sindh and frontal offshore areas). The southern Pakistan is a seismic-mélange wherein earthquake prone sources are diversified and closely associated with active plate margins. The spatial patterns of seismicity are significant in reckoning seismic potential of active fault lineaments in intraplate cohorts of Arabian, Eurasian and Indian plates. The tectonic earthquakes during the period of 325 B.C to February 2020 are compiled in a catalog. This updated catalog has been thoroughly processed by standard procedures necessary for magnitude coherency, main shocks declustering, magnitude completeness etc. Since, the faulty blocks are extending beyond international borders, a wide region is outlined for cataloging and thorough screening of fault lines. The fault lineaments are digitized after geo-referencing of structural and tectonic maps. A rigorous effort is made to review literature to gather allied data sets of plate kinematics, GPS constraints across faulty blocks and slip-sense during major events. Individual fault lines are examined carefully to discriminate seismicity yield by considering their geometry and influence of seismotectonic settings. The seismicity characteristics of fault lines are used to ascribe ‘seismicity index’ that helped to rank potential faults from I (least active) to V (highly active). Kernel density maps are prepared to allure vulnerable pockets of the fault lineaments for specific magnitude classes. Seismicity Index and Kernel density models are significant to analyze spatial changes in potential magnitude strength which may be pertinent for urban planners and developers to design earthquake resistant buildings in study area.


Introduction
The seismologists are making efforts to study seismicity patterns in diversified seismotectonic zones of southern Pakistan (Ali and Khan 2015;Aslam and Naseer 2020). The southern Pakistan is a seismic-mélange wherein earthquake prone sources are closely associated with active plate margins and faulted cohorts in intraplate regions of Arabian, Eurasian and Indian plates. The faulted blocks provide anisotropic medium of geological layers across their planes and enable the phenomenon to trap elastic energy in creeping blocks or drifting plate margins (Barnhart et al. 2014). In seismically active regions, the accumulating stresses weaken the rock strength to undergone plausible brittle deformation, thus, engender tectonic earthquakes (DeMets et al. 1990; Kopp et al. 2000). However, the strength of tectonic earthquakes appear randomly in time and space coordinates (Gutenberg and Richter 1956;Shi and Bolt 1982).
The southern Pakistan demonstrates a configuration of active margins of Arabian, Eurasian and Indian plates. Primarily, active lithospheric plates are adding elastic fuel to seismotectonic sources and channelizing elastic stresses through fault splays to the intraplate regions (Rani et al. 2011;Khan et al. 2020).These active seismotectonic features are primitive source of Neotectonic deformation and seismic events over yester years in study area (Pararas-Carayannis 2006;Frohling and Szeliga 2016). The subducting Arabian plate underneath the Afghan block is forming a wide Makran subduction zone (~1000 km, East-West oriented) with deformational front parallel to the Balochistan-Iran coast (Mokhtari et al. 2019). Wherein thrust or reverse faults in Makran accretionary prism (sub-aerial & submarine) are providing evidence of crustal shortening cloned with deformation in overlying sedimentary cover (Smith et al. 2013;Mokhtari et al. 2019). The north-south trending Suleiman -Kirthar ranges (~1250 km) formed by continental-continental collision between Indian and Eurasian plates (Kazmi and Jan 1997). The Suleiman and Kirthar ranges host several thrust fault components actively involved in intraplate seismicity since Eocene Regard et al. 2010). This collisional margin converted into transform fault movement at Chaman-Ornach-Nal fault system (Khan and Khan 2017). The transform boundary enrooting from land to offshore areas through Murray ridge (Siddiqui and Jadoon 2013;Frohling and Szeliga 2016). The Murray ridge (NE-SW oriented,1 860 km long) extends further south in offshore Arabian Sea separating the Indian plate to its east from Arabian plate (Farhoudi and Karig 1977). The spreading of Indian plate from Murray ridge and existence of Sonnee fault are adding complexity to seismotectonic of coastal faults in southern Pakistan (Kopp et al. 2000;Khan et al. 2020). Consequential to rotation of Indian plate block along its western margin cause deformation of Paleo-Miocene strata exposed at lower Sindh (Karachi arc) and Balochistan coastal belt (Schelling 1999;Nabi et al. 2019). Nabi et al. (2019) revealed seismotectonic potential of Runn-of-Kutch wherein shallow depth seismicity associated with compressional stresses in Sindh region. However, the seismotectonic stresses are being accommodated in faulting preserved under sedimentary basinfill deposits of Indus basin and Karachi arc system (Bilham et al. 2007;Sarwar and Alizai 2013;Nabi et al. 2020).
The literature review suggest that plate kinematics augment in development of elastic stresses and foster elastic strain in mechanical faulting either located to seismically sensitive plate margins or active faulty blocks (Page et al. 1979;Byrne et al. 1992; Barnhart et al. 2014;Heidarzadeh and Satake 2015). The regional plate assemblages, geodetic constraints and earthquake focal mechanism suggested that the ongoing oblique extension at Murray ridge with a rate of~3.1 ± 0.7 to 3.7 ± 0.7 mm/yr at southern part, however, to its northern part the Owen Fracture Zone exhibit the dextral strike -slip motion at a rate of 2-4 mm/yr (Minshull et al. 2015). An integrated study of Altamimi et al. (2012) divulged GPS measurements (from 20 stations located in southern Iran, southwestern Pakistan and Oman) coupled with hypocentral depths expound the Makran subduction zone accumulating strain at a rate of between 17:1 þ4:1 À3:5 mm/yr (Frohling and Szeliga 2016). The Chaman-Ornach-Nal fault demonstrate the western margin of Indian plate with the Eurasian plate on terrestrial part of southern Pakistan. At this margin, the accumulating inter-seismic strain was computed using sensitive GPS and InSAR data and Szeliga et al. (2012) concluded that variation in creeping block velocities, for instance, across Chaman fault the convergence rate was 14.1 to 19.5 mm/ year while the avg. compression velocity near southern node of Ornach-Nal fault (nearby shore) was 15.1 mm/yr. Though, the dearth of fixed GPS stations, advanced earthquake instrumentation, broad coverage of geophysical imaging, integration of allied datasets, detailed geophysical and rheological findings about brittle lithosphere and beyond it preclude critical analysis to look-down into dynamic interior of southern Pakistan (Szeliga et al. 2012;Minshull et al. 2015). However, geodetic constraints provided some controls to understand the development of stresses in the faulty blocks of southern Pakistan (Altamimi et al. 2012;Frohling and Szeliga 2016).
USGS catalog of earthquake provides substantial information of more than thousand tectonic earthquakes ordinated form inter-and intraplate regions of southern Pakistan over a century i.e. 1902 to 2015 (Ali and Khan 2015). Thereby, seismicity at a rate of 10-15 per month fill the spatial gap between far-offset epicenters of major events in southern Pakistan. However, recent dataset presented in this study divulge an increase in seismicity populated in Sindh, Balochistan and frontal offshore environments. The epicenters of these seismic events are found closely located to the vulnerable parts of known active faults of Balochistan and Sindh regions. The study area is enriched with fault lineaments widely-spread over spatial coordinates of study area ( Fig. 1) (Qureshi et al. 2001;Bilham et al. 2007;Niamatullah and Imran 2009;Regard et al. 2010;Szeliga et al. 2012;Martin and Kakar 2012). These fault lineaments apparently have a strong association with diversified seismotectonic characteristics of the region. The deep analysis of the earthquake data was suggested to reveal seismogenic characteristics of the faults and to examine their potential seismic hazard in southern Pakistan (Szeliga et al. 2009;Smith et al. 2013).
Hence, it is necessary to study updated seismicity yield of individual faults accompanied in robust seismotectonic environment. Therefore, objectives of present study are: (i) to  (Nabi et al. 2019;Waseem et al. 2019;Smith et al. 2013;Ul-Hadi et al. 2013;Szeliga et al. 2012;Martin and Kakar 2012;Regard et al. 2010;Niamatullah and Imran 2009;Bilham et al. 2007;Kukowski et al. 2000;Qureshi et al. 2001) examine the seismicity yield associated with individual faults; (ii) to identify characteristic vulnerability of fault lineaments (spatial clustering of peculiar magnitude levels) in southern Pakistan.

Dataset and methods
The complimentary tasks of this study are (i) compilation and processing of earthquake data (ii) gathering of allied data sets (fault lineaments, GPS constraints etc.) (iii) utilization of data sets im mapping and estimation of seismicity index. These holistic approaches helped to encode the seismological conductivity of faults through presentation of GIS-based map layouts.
i) Compilation of Earthquake dataset.
The primary data of earthquake events is accessed from open source earthquake databases such as US Geological Survey -NEIC and Pakistan Metrological Department-National Seismic Monitoring Network. The seismological datasets from both the sources are self-reviewed and necessarily annotated for cataloging. The earthquake catalog shall require uniformity of parameters in compiled catalog. The magnitudes of those events (4031 events, spreading over 7 ºN to 32º N and 55º E to 71º E) are provided in different magnitude scales. Thereby, all magnitudes scales are translated into unified moment magnitude (Mw) which is considered most reliable magnitude scale for cataloging (Scordilis 2005;Al-Ahmadi et al. 2014;Khan et al. 2018).
The historic events during 325 B.C. to 1900 are obtained from the catalog of Khan et al. 2018. The strength of these events are calibrated in Mw. There are 37 such events before the instrumental age of earthquake recording i.e. pre-1900. The historic events have been merged with the primary data inventory of this study to compile a 'final catalog' containing 4459 events unified to Mw. This unified catalog has been methodically processed to filter deviations of noisy data i.e. foreshocks, aftershocks and spatially outliers. The declustering exercise is performed using Reasenberg method which found 135 clusters of foreshocks and aftershocks populated with 975 events. The epicenters out-lying the spatial boundary of study area are 341, these anomalies are removed from the catalog. Thereby, the events are reduced to 3275 in the 'final catalog' occurred during the period of 325 B.C. to 2020 February. The estimated magnitude of completeness (Mc) for final catalog is Mw 5.3 (at max. curvature solution), however, the threshold magnitude of catalog is Mw 3.7. The focal depth of the catalog varies from 0 to 121 km.
An inventory of geological fault lines in southern Pakistan has been developed after digitization of geo-referenced structural maps from previous studies (Qureshi et al. 2001;Kukowski et al. 2000;Bilham et al. 2007;Niamatullah and Imran 2009;Regard et al. 2010;Szeliga et al. 2012;Martin and Kakar 2012;Hadi et al. 2013;Nabi et al. 2019;Waseem et al. 2019). The fault lineaments are stored as vector layer in ArcGIS format shown in Fig. 1. These fault lineaments are providing the base map grid to reference the seismicity characteristics to delineate the seismicity profiles at high spatial resolution in southern Pakistan.
The land and sea surface topography of southern Pakistan has been archived from the global terrain model for ocean and land i.e. GEBCO (www.gebco.net/data_and_products/ gridded_bathymetry_data/gebco_2019). The global gridded bathymetric dataset is presenting topography at 15 arcsecond intervals (accessed on 15 February 2020). The digital boundaries of transform and subduction plates are accessed in ArcMap format from the global project of Coffin et al. (1997). The parameters of fault plane are quested from the Global Centroid Moment Tensor (gCMT) database for major earthquakes (M w >5.0) occurred in the southern Pakistan. The CMT data base provided fault plane parameters along with moment magnitude of events (Dziewonski et al. 1981;Ekström et al. 2012). Thereby, the converted Mw in this study have also been validated for events Mw > 5.0 from gCMT catalog.
iii) Utilization of dataset The earthquake data set comprises source parameters such as epicentral location (latitude and longitude), moment magnitude, focal depth, time (hours, min), date of occurrence (year, month, day) provide basic information of earthquake events. Since there were 3278 events in the initial earthquake catalog, therefor, a matrix of 3278 × 9 is prepared for estimation of Mc and b-values by using Zmap application developed by Wiemer (2001). These seismicity variables (in the final catalog) are utilized for seismicity mapping and modeling of earthquake potential associated with each fault lineaments.
The capabilities of ArcMap tools have been explored for data analysis, clipping and mapping of desired earthquake data. Since, spatial coordinates of epicenter of each event served as primary key of earthquake dataset. The clipping tools of ArcMap are utilized to select the events within the buffer around a fault line and tectonic lineaments. The clipping tool improve the efficiency in getting data points in response of spatial queries. The polygon geometry generated the numbers of events (Ni) lie within the buffer routing the fault. However, critical selection is made to reduce the overlapping of events in close distanced fault lines. Further, the nature of fault planes or its slip in major events is also considered intact. These events are exported to examine earthquake characteristics of the corresponding fault line fault line and potential tectonic margin. These spatial sub-sets of final catalog are vital in defining seismicity index to retrieve probabilistic seismicity hazard.
Geoprocessing tool of ArcMap employed the spinel interpolation algorithm to compute interpolates from points using a minimum curvature spline technique (Scott and Janikas 2010). Further, the capabilities of geographical information system (GIS) are valuable for geospatial modelling of earthquake data by utilizing spatial statistical analyses tools. Kernel density algorithm (KDA) is a nonparametric spatial interpolation method applied to analyze first-order properties of spatial data distribution i.e. earthquake events by computing the event's density i.e. magnitude per unit area of ground (Kagan and Jackson 1994;Woo 1996;Botev et al. 2010). KDA considers the point data of earthquake and anxious fault segment produced specific magnitude strengths while calculating the Kernel density (KD). The general expression of KD in two-dimensional space is given by: where, KD (s) is the density at location s, r is the search radius or bandwidth of the KDA, n is the number of sampling points, and k is the weight of a point i at distance dis to location s. The surface value is the peak measured at the location of point and reduces with increasing distance from the point, reaching zero at the search radius distance from the point. The variable k is usually modeled as a kernel factor of the ratio between d is and r. GIS-based procedure of KDA articulates Kernel density of epicenters by computing magnitude per unit area of ground. Since the epicenters of this catalog were spreading over teleseismic distances (> 10º), the 'Geodesic method' is considered instead of the 'Planar method' to maintain correct distance estimation across the latitudinal arcs. The Geodesic method considers the curvature of the spheroid, thus measure precise area-based density of epicenters. The KD models are projected at datum of WGS 1984 spheroid (semi-major axis: 6378137.0; semi-minor axis: 6356752.314245179 and inverse flattening factor: 298.257223563). The KD models for each magnitude class of earthquake events were presented as a raster surface graded in pixel values (KD value) encoded with specific color spectrum. The thematic surface models of KD are valuable in determination of fault pockets under seismic influence, explore spatial changes in seismicity strengths and investigate characteristics of spatial patterns in various source zones (Ramanna and Dodagoudar 2012;Stock and Smith 2002).

Results and Discussion
The outlook of regional seismicity manifests that lithospheric mosaic is enriched with seismicity wealth (Fig. 2). It is observed that the earthquakes are widely distributed in terrestrial and offshore dimensions outlined within southern Pakistan. Analytical analysis of spatial dispersion of seismicity helped in defining seismicity patterns engendering from active margins of Indian, Eurasian and Arabian plates and their associated faults in intraplate regions (Fig. 2). It can be comprehended that the epicenters of earthquake are wellpopulated on respective platforms e.g. Eurasian plate~55%, Indian plate~40% and Arabian plate~5%. General distribution of earthquakes suggest that nnorthern part of study area (onshore region) has more earthquake epicenters than its southern part (offshore). Least seismicity distributed in offshore areas resting over Indian and Arabian plates. Probably, the lithospheric mass of Arabian plate (under northern Arabian sea) is more stable except its nauseating margin at Murray ridge and under-thrusting shoulder at Makran subduction zone. It is assumed that a relatively 'free-way' subduction is ongoing at 4 mm/year along the Makran subduction front, but probably locked at 33 km (Smith et al. 2013).
The depth distribution suggest that the lower part of Indian plate generated few events having focal depth > 70 km, particularly located in southern Sindh and frontal offshore of Indian Ocean. It is speculated that the Indian plate is carrying more seismicity load, drifting to the north with collisional margin at its western margin along which greater number of deep events were populated (Fig. 2). The high density of the shallow depth events indicating degree of deformation within sedimentary cover resting over Indian substrate and Eurasian platform. It is apparently illustrated that the intraplate regions of Eurasian and Indian plates are relatively more seismic prone than Arabian plate region in study area.
About 60% of earthquake epicenters neatly follow the trends of plate margins and fault lineaments in offshore or onshore regions of southern Pakistan. There are more major tectonic events (stared in Fig. 2) originated from Indian plate at its intraplate region and its western margin, however, some of such events are loacted on intraplate region of Eurasian plate and its margins with Arabian plate (to its north at Makran subduction zone). Major events having 7.0 < Mw > 8.1 are triggered from both the latitudinal plate margins of Arabian plate with the Indian plate at Carlsberg ridge and with Eurasian plate at Makran subduction zone. Further, it is highlighted that the earthquake density is more pronounced Fig. 2 The seismicity distribution of the catalog compiled in this study. The regional plate margins are labeled in legends. The tectonic lineaments and fault lines are displayed in grey. The city names are: G = Gwadar P = Pasni O = Ormara K = Karachi B j =Bhuj Np = Nagar Parkar Hy = Hyderabad S = Sukkur. The cell size at which the output raster will becreated at Carlsberg ridge and Makran subduction zone than Murray ridge in offshore environments. Comparatively, the longitudinal plate margin at Murray ridge produced relatively moderate to low magnitude of events during the study period. However, few seismic events are dispersed with the widening of Murray ridge at its northern and middle parts. The epicenteral clustering is found concentrated at the northern part of Murray ridge, probably, reflecting the complexity due to inter-junction of transform faults (ONF, fulted Murray ridge, Sonnee fault etc.) in Makran accretionary zone (sub marine) at frontal Balochistan coast (Regard et al. 2010).
It is also significant to investigate the density of interplate events. The seismicity parallel to the western margin of Indian plate is observed relatively denser, probably, indicating the width of deformed margin characterized with transpressional stresses therein. Khan et al. (2020) studied comparative seismicity of major strike slip faults forming western boundary of Indian plate and observed that transform faults of Chaman fault system are accommodating differential stresses in shallow depths. This study may help to delineate the crustal locking with reference to confined seismicity to transform boundary i.e. Chaman-Ornach Nal fault system. The moderate magnitude earthquake epicenters alleviate at eastern side along the transform Ornach-Nal fault (probably coupling with Murray ridge). The seismicity is mounting at western side of Makran as well due to pairing of subduction thrust with Minab transform fault (Fig. 2). The seismicity distribution reveals amassed clusters engendered from massive deformation along several thrust faults associated with curl of Suleiman lobe [north-eastern part of Balochistan / northern part of western margin of Indian plate]. The small to medium sized towns (Quetta, Sibbi, Fort Munro, Zindapir, Uch, etc.) settled in these Tertiary rocks which were rocked several times by strong amplitudes of major to moderate magnitude events (Szeliga et al. 2012;Barnhart et al. 2014).
Despite seismotectonic margins of Indian, Eurasian, and Arabian plates, the intraplate regions of lithospheric mosaic also share paramount contribution (remaining 30% of events in the catalog, having shallow focal depth (0-10 km)) to total seismicity yield of the region. These events are concentrated within or near distance to the potential faults of the study area. Figure 2 helps to figure out and concentrate on specific intraplate regions of southern Pakistan. The intraplate seismicity found off the tectonic lineaments fuelled by trust faults in Makran onshore following the trends of subduction zone (parallel to the coast) while the strike -slip faults are accommodating the compressional stresses (due to convergence) at western margin of Indian plate, undergoing push of Arabian plate along the Makran subduction zone and oblique divergence at Murray ridge. Since, the intraplate region receive the elastic stresses from conjugate seismotectonic environments, it is idealized that the local creeping of the blocks has integral role in clustering of intraplate earthquakes and their pattern definitions.
The spatial distribution of earthquakes are found relatively rich in several intraplate cohorts on onshore region. The intraplate seismicity scatters in eastern part of study area (desert and plains of Sindh), Indus delta and Sindh coastal belt. The intraplate earthquakes originating from Sindh plains are critical to investigate due to absence of known surface faulting (Nabi et al. 2020). However, mapping of faulted outcrops in Nagar Parker, substantial information of seismic sources or active faults (in Runn of Kutch, Allah Band, northeast of Karachi, Bhambore, etc.) are obtained with the help of previous seismicity records. It is evident from the seismicity located to the eastern part of study area that numerous active faults are preserved under sedimentary basin-fill deposits of Indus basin (Sindh plains) and the Runn-of-Kutch zone which are presenting the degree of crustal vulnerability of Indian plate at various depths. This study expound spatial association between tectonic earthquakes and individual faults in southern Pakistan. The fault lines are exposed to the onshore region located to the north of active Makran subduction zone (Balochistan) and within lower Indus basin (Sindh). The geophysical prospecting reported mechanical faulting since Jurassic . It is speculated that the stresses build at interplate margins, are feeding the seismic stresses promulgated to the intraplate regions of Indian and Eurasian plates.
The spinel interpolation method smoothen the spatial data variations. The GPS velocities of creeping faulty blocks in Makran region has been observed by Altamimi et al. (2012). The GPS velocities are significantly varying within local coordinates of study area such as relatively more active faulty blocks observed in the northeastern part of Balochistan. Wherein, complex geological processes augment accumulation of seismotectonic stresses evident from velocities of creeping blocks, although, high density of earthquake revealed the capacities of rocks to retain elastic stresses along the pockets of Zhob fault, Kohlu Fault and Ziyarat-Harnai fault (Figs. 2 and 3a). The interpolation of velocity components show relative changes across the faulty blocks e.g. transform Chaman fault, Gahzaband fault, Makran subduction zone ( Fig. 3a and b).
The dearth of installed seismometers and accelerometers in Sindh and Balochistan regions preclude the mapping of stresses being accumulating in various faulty blocks. It is challenging to know the fault block is either creeping or locked in active seismotectonic region of study area, though, the interpolation of focal depths and GPS velocities across faulty blocks are helpful to understand anomalies. Contrary to creeping of the blocks, the locking segments are 'more vulnerable' and alarming due to experiencing inter-seismic stage and accumulating stresses rather releasing low magnitude earthquakes. However, it is suggested to establish advance earthquake stations for monitoring and determination of continuous stresses, deformation rates, lithospheric kinematics at high resolution along the regional seismotectonic margins, i.e. oceanic-continent convergence at Makran, ocean-continent oblique transform motion at Murray ridge and continentcontinent collision at western margin of Indian plate. Later, the vulnerable zones such as eastern part of Karachi arc, northeast of Hyderabad region, Runn of Kutch, shall be included for monitoring of GPS velocities across the faulty blocks. A detailed geophysical study is suggested in offshore environment to map the interlocking pockets of Arabian slab and interlinking configuration of northern Murray ridge with transform boundary faults. Fig. 3 Interpolation maps of GPS velocities relative to the stable Eurasian plate as defined by Altamimi et al. (2012). (a) Vertical velocity in the study area (b) Horizontal velocity in the study area. The black square represents the location of GPS station A careful screening of seismic events concurrence to the fault lineaments helped to examine the seismicity profile in study area. The seismicity profile of fault lines comprehend significant information such as active segment of each fault lineament; major and minor magnitude of associated events; shallowest and deepest hypocenter; recurrence period for credible magnitude; spatial distribution of earthquake density (high, low); total number of events; and b-value. Further, the earthquakes catalog of this study is used to estimate b-values by the maximum curvature solution of frequency-magnitude distribution of data points. The statistical procedure of maximum curvature solution considers best-fitting of a master-line at frequency-magnitude-distribution plot of the dataset (Ashadi et al. 2015). Subsequently, Mc and b-values are determined for best-fitted slope at maximum curvature solution. The estimation of the recurrence period for credible magnitude along the key faults was estimated by following equation. The b-value estimation are integral component for Tr estimations (Table 1)  Though, a rigorous effort is made to obtain key information related to platform velocities and geometries/slip behavior of faulty blocks, slip rate in major events from literature review of Bilham et al. (2007) With some limitations, it may be basic but credible information of the faulty zones in study area. The seismicity characteristics such as platform velocities, geometries of faulty blocks, epicenteral patterns, reckoning of potential fault strength, bvalues and recurrence interval etc. are helpful in hazard assessment of an area (Chen et al. 1998;Clévédé et al. 2009). The data set of fault lineaments helped to delineate the seismic conductivity of each key fault line using updated earthquake catalog in this study (Table 1). This database is further used to define seismicity index (SI) for the first time in southern Pakistan. SI ranked the key fault lines with their seismic potential and seismotectonic characteristics. Since recurrence interval depends on the number of events, credible earthquake from fault, Mc and b-value. Thus, Tr has vital role to define weighted function for SI, such as Seismicity Index-I for Tr < 50 years; Seismicity Index-II for 51 < Tr < 100 years; Seismicity Index-III for 101 < Tr < 200 years; Seismicity Index-IV for 201 < Tr < 500 years; SI = Rank-V for Tr > 501 years.
This analysis helped to rank the potential faults, for instance, Kathiwar, Bhuj, Surjan, Gaj, Kirthar, Mughal Kot, Kingri, Pardan, Harnai, Ziarat, Ghazaband, Sonmiani, Nai Rud, Kulri, Murray Ridge, Ornach Nal fault more sensitive faults categorized in seismicty index rank-I. SI rank-I indicates low values which reflects these faults are more susceptible to tectonic earthquakes with less inter-seismic interval. The faults such as Makran subduction zone, Kalat, Mach, Dalbandin and Malik Salar faults are ranked with highest value of SI. Seismicity index rank-V inferred these fault with longer recurrence for major seismic event. Although, the recurrence interval is necessary for seismic hazard assessment and predicting an earthquake in an area. The projecting earthquake estimations shall précised with fault kinematics in its most immediate surroundings, fault mechanics and differential slip rates along the suspected potential fault (Papazacho et al., 2004). The faulty regions characterized with low b-values are more vulnerable to future earthquakes (Rani et al. 2011).

Spatial analysis of epicenters of magnitudes
As an integral part of this study, the literature review and earthquake data analysis suggest that earthquake prone processes and seismic potential sources are vibrant in study area thus variations in magnitude strengths are noticeable. The magnitude of an earthquake quantifies the amount of elastic strain energy released during tectonic earthquake. It is revealed that changes in magnitude strength are not just the statistical data variations, but these changes are closely articulated with the seismo-geological characteristics such as brittle deformation, focal depth, micro-environment of faulting, platform stability, fault mechanics etc. The catalog of earthquakes for 3275 events is fragmented into several magnitude classes based on Mw (Table 2). Table 2 shows basic statics of the earthquake point data classes such as minimum magnitude of the class, maximum magnitude of the class, standard deviation of the events in specific class. The data of these magnitude classes is used for Kernel density analysis.
KD distribution maps show significant changes in spatial patterns of earthquake epicenters for potential magnitude classes associated with fault lineaments (Fig. 4). The KD algorithm generated continuous raster maps of earthquake density in a nexus of faulty blocks, by computing the locations of earthquake concentration and respective magnitude values (Papazachos et al. 2004). Higher number of events with higher magnitude will give high density values and vice versa. Henceforth, the KD maps illuminate active pockets with ccontrasting buffers of kernel density from background values. Similarly, it elucidate individual fault response in different magnitude classes and subsequently ascertain aseismic segment/spatial locking of the fault.
The KD map for earthquakes having moment magnitude 4.0 to 5.0 shown highest Kernal density at western segment of Carlsberg ridge, northwestern Minab fault, Ghazaband fault, Runn-of-Kucth and Suleiman lobe (Fig. 4a). The KD highlight two anomalous areas i.e. faulted blocks of Suleiman lobe (northeastern Balochistan) and Runn-of-Kutch (southeastern Sindh). The faults in Suleiman lobe are characterized with thrust or reverse faulting penetrated deep beyond the basement (Garcia et al. 2006;Iqbal and Helmcke 2004). However, active offshoots of reverse faults disturbed the Tertiary rocks to host moderate to major magnitude events concentrate at a depth range of 0-35 km. Thereby, KD algorithm reflects a polygon buffering the curvature Suleiman lobe (Fig. 4a).The high values of KD also isolated the Runn-of-Kutuch zone under the influence of shallow ruptured zones in surroundings of Bhuj fault, Banni fault, Kutch Mainland fault etc. However, KD drops immediately while entering towards Indus delta and   KD trends for magnitude class ranging in Mw 5.0 to 6.0, depicting some connecting surfaces spread over southern Pakistan (Fig. 4b). There are four elliptical patches of low KD (purple color) for earthquakes magnitude 5 < Mw > 6 along the Murray ridge segments. High KD polygons identified vulnerable pockets located to the convex of Suleiman lobe and Minab fault (Iran). The progressively decreasing KD contours in the vicinity of Quetta and Khuzdar outsourced from high values of KD peak. The moderate KD values prominent in Pasni, Gwadar, Bhuj, and Nasratabad (Iran) regions. It has been speculated that the kernel density for higher magnitude earthquakes increases in terrestrial part of southern Pakistan, probably, presents northern part as "central-eye" for future earthquake events. The connection or extension of faults such as Hoshab and Panjgur in Iran faults nearby splays of Nasratabad fault need to be investigated with more data. Similarly, the coupling of transform fault i.e. Ornach-Nal fault with the Murray ridge and uncertainty of the triple junction are suggested to investigate.
It is significant to notice that Murray ridge shows least KD values for earthquake magnitude ranging between 6.0 and 7.0. The juncture between Murray ridge and Carlsberg ridge shown with higher KD values for high magnitude events (Fig. 4c). The size of higher KD value polygons increases for high magnitude events at four benchmark locations such as Pasni-Gwadar region, Kallat, Suleimanlobe and Minab fault (Iran). High to moderate KD values allure 'active segments' with high yield of events for this class. Least KD for higher magnitude events (purple color) helped to understand 'locked or aseismic zones' along the seismotectonic lineaments. The expending surfaces of lowto-moderate KD values for this class spread over Sindh. Despite, high KD contours for high magnitude events are prominent in Hyderabad, Bhambore, Bhuj regions probably, under the influence of devastating events of Sindh during historic records. It may be hypothesized that the low bouger gravity at high topography of Suleiman range has higher Kernel densities computed by KD algorithm. Although, the southern region of Balochistan and Sindh has low frequency of high magnitude earthquakes in recent times, but the potential faults engendered large magnitude earthquakes in historic records are found in low seismicity index. The results of this study would contribute to mild the curiosity of people living with earthquakes and uncertainty about seismological suspects in southern Pakistan.
There are 19 highest magnitude events having moment magnitude ranging between 7.0 and 8.1. The earthquake data records shown that ample data size (about 50% of these events) occurred in historic times i.e. pre-1900. The KD map illuminates most vulnerable zones for the highest magnitude class. There are closed spheres located to offshore areas i.e. along the Carlsberg ridge, Makran subduction zone and Runnof-Kutch (Fig. 4d). These high KD contours divulge major tsunamigenic events e.g. Makran earthquake of 325 BC, 1945 and 1483 and Carlsberg ridge earthquake of 1954. The estimates for least KD values, probably, helped to recognize 'cut-off' value for potential earthquake along the Murray ridge which remain silent for 6 < Mw > 8. There are four elliptical spheres of higher KD values for highest magnitude event, located to the terrestrial part of southern Pakistan (Sindh and Balochistan). The downward progression of KD levels from peak values encompasses major cities of Balochistan and northern Sindh. The epicenters of of devastating events in historic times cause expansion of peak KD values for higher magnitude events in surroundings of Indus Delta (Bhambore, Bhuj and Hyderabad region).

Conclusions
This study is pertinent to address emerging trends of GIS applications in resolving seismological information in complex geological environments. The Mw-based updated catalog of earthquakes has been utilized to study the spatial patterns of earthquakes and their parameters collinear to the known faults in southern Pakistan. It is challenging to elucidate the seismicity parameters of each fault in active seismotectonic zones but GIS-based mapping tools are helpful to conduct detailed analysis. These characteristics were systematically involved in development of seismicity index, which ranked the potential faults in rank-I. Kernel density of higher magnitude earthquakes increases in terrestrial part of southern Pakistan, revealing the earthquake prone faults in Balochistan and Sindh. Kernel density maps (for Mw > 4.0) allure the spatial patterns of earthquakes with higher magnitudes are notably concentrated in the central, northeastern and southern parts of study area.. Spatial analysis of data sets for seismic events suggest stronger spatial associations between earthquake epicenters and faults oriented N-S, and weaker spatial associations of epicenters with some -unknown faults in the plains of Indus Basin. The spatial distribution of earthquake events along active faults and Kernel Densities shows the northeastern Balochistan more significant with higher number of earthquakes and higher probability of the high-magnitude earthquake events during instrumental age. Yet the active faults are more pronounced and sensitive for developers and policy makers. The seismicity index of potential faults in the region may help in designing tangible measures against seismicity hazard for major cities adjacent to highly indexed fault lines in vibrant southern Pakistan.