Assessment of seismic sources and capable faults through hierarchic tectonic criteria: implications for seismic hazard in the Levant

We present a methodology for mapping faults that constitute a potential hazard to structures, with an emphasis on ground shake hazards and on surface rupture nearby critical facilities such as dams and nuclear power plants. The methodology categorises faults by hierarchic seismo-tectonic criteria, which are designed according to the degree of certainty for recent activity and the accessibility of the information within a given region. First, the instrumental seismicity is statistically processed to obtain the gridded seismicity of the earthquake density and the seismic moment density parameters. Their spatial distribution reveals the zones of the seismic sources, within the examined period. We combine these results with geodetic and pre-instrumental slip rates, historical earthquake data, geological maps and aerial photography to define and categorise faults that are likely to generate significant earthquakes (M ≥ 6.0). Their mapping is fundamental for seismo-tectonic modelling and for probabilistic seismic hazard analyses (PSHAs). In addition, for surface rupture hazard, we create a database and a map of Quaternary capable faults by developing criteria according to the regional stratigraphy and the tectonic configuration. The relationship between seismicity, slip dynamics and fault activity through time is an intrinsic result of our analysis that allows revealing the dynamic of the deformation in the region. The presented methodology expands the ability to differentiate between subgroups for planning or maintenance of different constructions or for research aims, and it can be applied in other regions.


31
The global population growth and the establishment of sensitive facilities, such as 32 nuclear power plants or dams, increase the seismic risk to higher levels and require 33 profound understanding of the seismic hazard (e.g. Marano et al., 2010). Probably the most 34 famous example is the destruction of the Fukushima nuclear power plant by the tsunami 35 caused by the 2011 M w = 9.0 Tohoku-oki earthquake, which has been affecting an 36 extensive region ever since. 37 A basic step in seismic hazard evaluation is defining and characterising faults that 38 constitute a potential hazard. Because earthquakes are stochastic processes that trigger 39 different hazards (such as ground shaking, tsunamis, landslides, liquefaction and surface 40 rupture) and the planning of different infrastructures requires different safety standards, 41 mapping and categorising hazardous faults is generated according to specific requirements. 42 In this paper, we present a methodology for mapping and categorising faults, which can 43 be applied for the evaluation of different seismic hazards. To generate our maps and to 44 classify the faults in them, we combine seismological analysis with geologic and geodetic 45 information. The methodology is implemented for generating regional maps of the "main 46 seismic sources" and of "capable faults". The former are the regional faults that should be 47 considered for ground shaking models and Probabilistic Seismic Hazard Analysis (PSHA), 48 and the latter constitute surface rupture hazards that should be considered for siting 49 facilities with environmental impact, such as dams and nuclear plants, or other vulnerable 50 facilities. We apply hierarchic criteria for categorising faults according to the specific 51 hazard. 52 We demonstrate our methodology for the Israel region, a seismically-active zone 53 mainly affected by the Dead Sea Transform fault system (DST; Fig. 1). First, we determine 54 the main seismic sources in Israel and its vicinity, focusing on faults that are likely to 55 generate significant earthquakes. Subsequently, we present the procedure to determine and 56 map faults that constitute a potential hazard of surface rupture for sensitive facilities. We 57 design the criteria according to the likelihood of surface rupture along specific faults. 58 Despite the limited duration of the instrumental record, it constitutes one of the main 59 direct evidence of fault activity in the current tectonic configuration. Probabilistic analyses 60 a series of asymmetric folds, strike-slip faults, and monoclines (Eyal and Reches, 1983; 91 Sagy et al., 2003). Regional uplift began from the end of the Eocene and the area was 92 intermittently exposed to erosional processes (Picard, 1965). The African-Arabian plate 93 broke along the suture of Gulf of Aden -Red Sea during the Miocene, generating the Suez 94 rift and the DST which separate the Sinai sub-plate from the African and the Arab plates 95 (Fig. 1). The Suez rift, however, has shown relatively minor signs of deformation since the 96 end of the Miocene (Garfunkel and Bartov, 1977;Joffe and Garfunkel, 1987;Steckler et 97 al., 1988). In the easternmost Mediterranean Sea, the deformation concentrates along the 98 convergent Cyprian Arc (Fig. 1), where the Anatolian plate overrides the plates of Africa 99 and Sinai (e.g., Mckenzie, 1970). Mediterranean region (Fig. 1). Its northern section crosses northwest Syria in a N-S 104 orientation; several recent large earthquakes were attributed to this section during the past 105 two millennia (Meghraoui et al., 2003). The middle section of the DST is the Lebanon 106 restraining bend (LRB; Fig. 1), characterised by transpression deformation (Quennell, 107 1959). This section is branched to a few segments that transfer the main component of the 108 strike-slip motion in Lebanon area (Gomez et al., 2003;. The Israel region is located 109 along the southern section of the DST but seismically it is also affected by the activity of 110 the middle part. 111 The southern part of the DST (Fig. 1) is dominated by a sinistral displacement of ~105-112 km over the last ~16-20 million years (Quennell, 1959;Garfunkel, 1981;. It is  (Neev et al., 1976;Sadeh et al., 2012) where the southern part is assumed 131 to be relatively rigid, while northward, normal faults orientated E-W generate S-N 132 extension expressed by graben and horst structures (Ron and Eyal, 1985

140
The database of faults that were active in the recent geological history is mainly based 141 on high-resolution geological maps. As of January 2019, 71 geological map sheets in the 142 scale of 1:50,000 are available for this study, out of the 79 sheets required to cover the 143 whole state of Israel (Fig. A1). The 1:200,000 geological map of Israel (Sneh et al., 1998) 144 is utilised where 1:50,000 data are absent. Included also are faults defined as active or 145 potentially active during the last 13,000 years, for the Israel Standard 413 (building code) 146 "Design provisions for earthquake resistance of structures" (Sagy et al., 2013). In addition, 147 some faults, which have not been mapped (or not updated yet) crossing Quaternary units 148 in the geological maps, are marked here as Quaternary faults based on evidence presented 149 in scientific publications, reports, and theses (see Table A1).

150
The establishment of Quaternary formation database (Table A2) (Table A3). The latter are limited to the extensions of mapped faults that are within Israel,165 and/or the main DST segments. The criteria for selecting these faults are discussed in 166 section 6.  169 We analyse the spatial distribution of seismic events in order to reveal the regional 170 seismic pattern, which helps to define the main seismic sources and develop an independent 171 criterion for Quaternary active faults. So as to define the seismicity-based criterion, we 172 design seismic criteria that are based on the distribution of two parameters that are, to a 173 large extent, independent: the earthquake kernel density and the seismic moment kernel 174 density. We demonstrate the methodology and then present the results below.

195
In order to assess the applicability of the following seismic processing and analysis, we  In order to quantitatively characterise the regional seismicity and associate the 203 earthquakes with mapped faults we examine two parameters: a) earthquake kernel density 204 and b) seismic moment ( 0 ) kernel density. Both parameters are obtained through the 205 following spatial data processing. A regional scan is carried out in a 0.5-km interval 2D  The earthquake kernel density parameter, , is calculated by counting all the 219 weighted events within a 6-km radius from each grid point, dividing their sum by the 220 sampler area (πr 2 ) and normalising by the duration of the earthquake catalogue: where N is the total number of events within the radius r, d(n) is the distance between an

225
The 0 kernel density parameter,    This contrast is predominant in the Sea of Galilee, which contains high earthquake kernel 281 density ( Fig. 3) but is less significant in the seismic moment kernel density (Fig. 4). DST and its main branches. We now combine these data with geologic, geodetic and paleo-286 seismologic measurements to generate the main seismic sources map, which displays 287 regional faults that demonstrate slip rates inferred as ≥ 0.5 mm/yr during the Holocene.  (Table 2), we estimated the slip rates along these fault zones as 0.5 -1 mm/yr. All the fault segments are located inside (or partly inside) the overlap zone which defined by the two 324 seismological analyses (Fig. 6). for better evaluation of its seismic potential. 349 We additionally note that large earthquakes along the Cyprian Arc ( Fig. 1) can also 350 generate tsunamis that might affect the coastline of Israel (Salamon et al., 2007). This 351 source is not analysed and mapped here, but should be taken into account in regional 352 seismo-tectonic models.  Quaternary period is selected as the time reference for sensitive facilities due to two main 373 reasons: a) we assume that faults that were active during the present regional stress regime 374 (Zoback, 1992) are more likely to activate in the near future. The regional stress state within . We note that "regional stress 377 field" (Zoback, 1992) as a criterion for active faulting is closely related to the "tectonic  and only if they do not match it, they are examined according to the second criterion, and 383 so on. Where geological evidences are absent, we utilise a seismological criterion (Fig. 6), 384 under the assumption that faults associated with seismically active subzones are more 385 likely to have ruptured the surface in the Quaternary compared to others.

386
Finally, because of the limitation of our database, mapped capable faults (Fig. 7) are 387 limited to Israel region, unless their continuations spread to the neighbouring countries.  (Table A2) to rupture the earth's surface at least once since the Quaternary.

395
This criterion is mainly related to zones covered by Quaternary units.    Two regional fault maps are presented; one is relevant for regional ground shaking models 449 (Fig. 5), and the other for surface rupture nearby facilities that are particularly vulnerable 450 to this hazard (Fig. 7). In addition to the approach of classifying faults by the recency of The DST accommodates most of the seismic activity, but also contains zones of very 497 sparse seismicity (Fig. 6). The seismicity distribution maps (Figs. 3, 4)  research is required for better characterisation of this activity and its relationship to the 529 regional tectonics.

530
Finally, relatively long E-W trending faults (SNB) cross the south of Israel and Sinai 531 and some of them are marked as Quaternary faults (Fig. 7, Fig. A4). However, there are no 532 geologic or geodetic indications for any activity along them since the early Pleistocene, 533 and the associated seismic activity mostly concentrates in their junctions with the DST. We 534 therefore assume that these dextral oblique slip faults are inactive in the present regional 535 stress field, and their reactivation may generally decrease with increasing distance from the 536 DST. developing an interdisciplinary regional database and hierarchical seismo-tectonic criteria.

542
With respect to the specific dictated requirements, faults that are potential sources for the success of this selection is demonstrated by the match between the geological-categorised 549 faults and the seismicity criterion (Fig. A4). The union zone defined by these two statistical 550 distributions is efficient in both definition of the main seismic sources (Fig. 6) and in 551 categorising capable faults (Fig. 7).  junctions. In addition, we identify a zone of seismicity that seems to diverge from the main 580 fault zone towards ~NW (EBL in Fig. A5; Fig. 6). Its orientation and a few independent 581 evidences imply that it reflects extension-related activity, accommodated by (subsurface?) 582 fault system that branch off the DST.