Evaluating Seismic Hazard Map for Speci ed Seismic Source Fault Using GIS-based 3D Deterministic Model: a Case Study of Low Seismicity Region in Egypt


 Although Egypt may be considered as an area of low seismicity, it has experienced the recurring destructive effects of earthquakes throughout its history. The damage distribution is significantly affected by fault geometry and the local site effects. Thus, an appropriate assessment of seismic hazard became a major challenge for efficient mitigation of the seismic risk. This study develops the Geographical Information System (GIS) based three-dimensional (3D) deterministic model to evaluate the seismic hazard map for specified seismic source fault for Egypt. The geometry of a fault plane is generated by a fault-modeling algorithm in which multiple 3D plane sets are interpreted by fault trace, geology, and topography. For estimating the local site effects, the study area is modeled by a 3D grid-mesh system and the ground motion propagation is estimated at each grid by considering the spatial analysis of average shear-wave velocity and a soil susceptibility map. The developed model is applied to simulate the seismic hazard maps in particular the October 12, 1992, Dahshour earthquake that caused huge disasters. The simulated seismic hazard maps of the October 12, 1992, Dahshour earthquake are validated at the observation of an isoseismal map and the evidence of the locations that caused serious damages. Furthermore, an evaluation of the expected Dahshour earthquake with magnitude 6.5 and October 11, 1999, Beni Suef earthquake is conducted for seismic risk mitigation study. By utilizing the developed model, our results encourage the GIS approach for seismic hazard analysis where 3D models can lead to a more accurate assessment.

It is essential to compute a deterministic seismic hazard map in three-dimensional (3D) 127 model parameters and to provide automation, intellection, and visualization in the 128 scenario-based simulations for Egypt. Moreover, the spatial quantification of the various 129 ground effects at a specific site is an important analysis of the ground motion propagation 130 for a study area. The recent development of Geographical Information Systems (GIS) 131 comprises a technology designed to support integrative modeling and to conduct 132 interactive spatial analysis for understanding various 3D processes (Fortheringham and 133 Wegener 2000). In the present study, a deterministic calculation model using GIS that 134 realistically computes the seismic hazard map for specified seismic source fault is 135 established to obtain a reliable evaluation of seismic risk for Egypt. A fault modeling 136 algorithm in GIS is developed for generating a 3D fault geometry interpreted by fault 137 trace, topography, and geology. A grid-mesh system in GIS is utilized as a 3D model 138 parameter to estimate the ground motion propagation in the study area by considering the 139 spatial analysis of average shear-wave velocity and the soil susceptibility map. For a 140 preliminary assessment, the ground motion attenuation relationships by Si and 141 Midorikawa (1999) are adopted to establish the 3D seismic calculation modeling 142 algorithm using GIS functions for evaluating several scenario earthquakes in the study 143 area. The developed model is applied to evaluate the seismic hazard maps for the 144 Dahshour earthquake and Beni Suef earthquake based on a scenario study. 145 146 2. Seismic Source Faults in Egypt 147 2.1 Historical earthquake sites 148 The study area (Fig. 1) is located in the northern part of Egypt as a part of the Nile 149 river valley, which covers southwestern Cairo including the El-Faiyum, Girza, and Beni 150 Suef areas. According to Badawy andHorvath (1999a, 1999b), most of the earthquake 151 events mainly occur in the northern part of Egypt and are related to the plate boundaries. 152 The relative motion of the Sinai sub-plate concerning the Suez rift and to the Aqaba-Dead 153 Sea rift characterizes the major source of seismic activities in the northern part of Egypt. 154 A relatively low rate of historical earthquake activity also continues on-trend to the 155 southwest Beni Suef area. Fig. 2 represents the study area remarked by dotted lines and 156 selected historical earthquakes from 1990 to 2018 beneath southwestern Cairo that has 157 been the important site of several earthquake events, in particular, a magnitude 5.9 158 Dahshour earthquake struck about 40km southwestern Cairo on October 12, 1992, caused 159 tremendous damage. 160

Geological and tectonic setting 161
Geology of the study area has been investigated by a variety of authors (e.g. Hume 162 1911;Said 1962;Said and Martin 1964

A 3D fault modeling algorithm in GIS 183
The most simplistic method for 3D fault modeling is a simple flat plane. By setting a 184 dip and azimuth the inferred fault can be created. However, if the case involves multiple 185 faults that intersect or consist of multiple parts of a fault plane, manual fault mapping is 186 labor-intensive and time-consuming to determine 3D models (Admasu et al. 2006). In 187 this study, automatic fault mapping is accomplished by using multiple spatial datasets 188 interpreted by fault trace, geology, and topography. 189 In the computational implementation system, the 3D fault modeling is accomplished 211 by the use of GIS geoprocessing tools. Fig. 5a shows the actual stages of the schematic 212 flow model of geoprocessing tools that comprises a 110-step process to provide an 213 automatic computation of 3D faults for speeding up interpretation on large 3D datasets. 214 The developed schematic flow model is represented as a diagram that chains together 215 sequences of processes using the output of one process as the input to another process. 216 Accordingly, an adjustment model for 3D fault planes can be performed by only changing 217 the input parameters without going through a long-stage process. However, the opportunity to capture recorded ground motion data from the earthquakes 228 in the northern part of Egypt was lost primarily because of the lack of strong motion 229 instrumentation, inadequate geographic coverage of instrumentation networks presently 230 installed, and inadequate dynamic range of the seismographic instrumentation in place 231 (Thenhaus et al. 1993). Lacking the ground motion records, there is always the question 232 of whether anomalously located areas of damage are due to deficient construction 233 practices or an actual physical enhancement of ground motion propagation. 234 3.2 A 3D grid-mesh system for site amplification factor 235 It has been recognized by researchers that soft and young sediments covering firm 236 bedrock can amplify seismic ground motions and cause severe damages during an 237 earthquake event. The shear-wave velocity of shallow sediments is very important in 238 ground motion propagation. Anderson et al. (1986) noted that the strata in the top 30 239 meters have a considerable influence on the character of the created seismic ground 240 motions. The average shear-wave velocity from the surface to 30-meter depth (AVS or 241 Vs30) is a well-known parameter for estimating the site amplification factor (ARV). 242 Table 1 shows the definition in terms of AVS and simple geological descriptions by the 243 National Earthquake Hazard Reduction Program (NEHRP). The ARV between surface 244 and firm bedrock has the following relationship with AVS, described as Eqn. 1. 245 (Midorikawa 1994) 246 logARV = 1.83 -0.66 logAVS, for (100 < AVS < 1500) (1) 247 where, 248 ARV: the site amplification factor for peak ground velocity (m/s) 249 AVS: the average shear-wave velocity from surface to 30m depth (m/s) 250 In this study, the AVS dataset from the United States Geological Survey (USGS) is 251 used to obtain the AVS contours interpolated from grid points at a 5-kilometer scale as a 252 proxy for the spatial analysis of the site amplification factor. Since the AVS values should 253 be used at finer scales for the study area, the soil susceptibility map by 1:50,000 scale and 254 the interpolated AVS contour map are utilized further by spatial correlation analysis to 255 estimate the AVS at a 50-meter grid scale. This soil susceptibility map is generally based 256 on the distribution of shallow soil profiles over rock beneath the Nile river valley. In the GIS calculation model, the 3D grid-mesh system (see Fig.5b) is used for seismic 262 hazard simulations that will define the accuracy and resolution of the simulation results, 263 both of which will affect the GIS computation time and level of detail in the results of the 264 seismic hazard map. Therefore, it's important to have a site amplification factor that is 265 linked in the 3D grid-mesh for computational simulation. Grid-mesh involves the ground 266 elevation and a site amplification factor to represent the arrangement and spacing between 267 each grid point. The study area which varies in elevation across different mountain ranges 268 or complex surface geological boundaries may require high meshing density to ensure 269 accuracy, which increases the computation time in the simulation. In this study, a 270 rectangular grid by 50-meter is designed to represent the study area for use in a seismic Seismic intensity (SI) is calculated using the relation between the intensity and the 304 peak ground velocity, described as Eqn. 5. 305 SI = 2.68 + 1.72 logPGV, for (4 ≤ I ≤ 7) (5) 306 where, 307 SI: seismic intensity 308

3D calculation modeling algorithm in GIS 309
There are two spatial datasets of the input parameter for conducting a computational 310 process of the seismic hazard assessment within GIS. The first dataset is 3D polygons as 311 a fault geometry model, and the second dataset is a 3D grid-mesh system as a ground 312 motion propagation model. Other parameters such as earthquake magnitudes, hypocenter 313 depth, fault types, etc. are specified in the 3D polygon dataset. The 3D seismic hazard 314 calculations are performed within GIS. The function of GIS is used as a spatial-temporal 315 database for extracting, calculating, and updating the input data. 316 values for all the point coordinate at the designed 3D grid-mesh system for each 3D 343 polygon, and calculates the seismic hazard based on the earthquake parameters (Fig. 8). validation studies in the region, the 3D input parameters used for the seismic hazard maps 375 are fruitful to be based for reliable disaster risk reduction in the region. The detailed 376 results of the seismic hazard maps are described in two following sections. 377

The Dahshour earthquake 378
The first step of analysis includes the validation of the simulation model of the October 379 12, 1999, Dahshour earthquake at the observation of isoseismal map according to the 380 MMI scales, as currently available data. Fig. 9 shows the location of the designed 381 simulation region for evaluating seismic hazard maps as a study area marked by a red 382