Seismic Hazard Analysis for Archaeological Structures — A Case Study for EL Sakakini Palace Cairo, Egypt

The modern architectural heritage of Egypt is rich, and extensively variable. It covers all kinds of monumental structures from palaces, public buildings, residential and industrial buildings, to bridges, springs, gardens and any other modern structure, which falls within the definition of a monument and belongs to the Egyptian cultural heritage. We present herein a comprhen‐ sive geophysical survey and seismic hazard assesment for the rehabilitation and strengthening of Habib Sakakini’s Palace in Cairo, which is considered one of the most significant architec‐ tural heritage sites in Egypt. The palace located on an ancient water pond at the eastern side of Egyptian gulf close to Sultan Bebris Al-Bondoqdary mosque, a place also called “Prince Qraja al-Turkumany pond”. That pond had been filled down by Habib Sakakini at 1892 to construct his famous palace in 1897.


Introduction
The modern architectural heritage of Egypt is rich, and extensively variable.It covers all kinds of monumental structures from palaces, public buildings, residential and industrial buildings, to bridges, springs, gardens and any other modern structure, which falls within the definition of a monument and belongs to the Egyptian cultural heritage.We present herein a comprhensive geophysical survey and seismic hazard assesment for the rehabilitation and strengthening of Habib Sakakini's Palace in Cairo, which is considered one of the most significant architectural heritage sites in Egypt.The palace located on an ancient water pond at the eastern side of Egyptian gulf close to Sultan Bebris Al-Bondoqdary mosque, a place also called "Prince Qraja al-Turkumany pond".That pond had been filled down by Habib Sakakini at 1892 to construct his famous palace in 1897.
Various survey campaigns have been performed comprising geotechnical and geophysical field and laboratory tests, aiming to define the physical, mechanical and dynamic properties of the building and the soil materials of the site where the palace is founded.All these results together with the seismic hazard analysis will be used for the seismic analysis of the palace response in the framework of the rehabilitation and strengthening works foreseen in a second stage.We present herein the most important results of the field campaign and the definition of the design input motion.
The seismic hazard analysis for El Sakakini Palace has been performed based on historical earthquakes, and maximum intensity.PGA with 10% probability of exceedance in 50 and 100 years is found equal to 0.15g and 0.19g respectively.P-wave and S-wave seismic refraction indicated a rather low velocity soil above the seismic bedrock found at depths higher than 20m.Ambient noise measurements have been used to determine the natural vibration frequency of soil and structure of El-Sakakini Palace.The fundamental frequency of El-Sakakini palace is 3.0Hz very close to the fundamental frequency of the underlying soil, which makes the resonance effect highly prominent.Some floors are considered dangerous since it show several resonance peaks and high amplification factors (4 th and 5 th floors) these floors are made of wood so, warnings to decision makers are given for the importance of such valuable structures.
The seismic design and risk assessment of El Sakakini palace is performed in two steps.In the first one we perform all necessary geotechnical and geophysical investigation together with seismic surveys and seismic hazard analysis in order to evaluate the foundation soil properties, the fundamental frequency of the site and the structure, and to determine the design input motion according to Egyptian regulations.The second phase comprises the detailed analysis of the palace and the design of the necessary remediation measures.IN the present pare we present the results of the first phase.

Seismic hazard 2.1. Historical seismicity
Egypt possesses a rich earthquake catalogue that goes back to the ancient Egyptian times.Some earthquakes are reported almost 4000 years ago.Figure 1 shows the most important historical events affecting ElSakakini palace.We can see that the Faiyum area as well as the Gulf of Suez is the most important earthquake zones affecting the place.

Maximum intensity
Historical seismicity and maximum reported intensity is a good preliminary index of the expected severity of a damaging earthquake.Available isoseismal maps in the time period 2200 B.C. to 1995 were digitized and re-contoured to determine the maximum intensity affecting the place.This was done using a cells value of equal area 0.1 lat.× 0.1 long.Figure 2 present the produced IMM intensity showing that a maximum IMM of VII is good design value.

Probabilistic hazard assessment
An improved earthquake catalogue for Egypt and surrounding areas affecting El Sakakini Palace has been prepared for the purposed of this study partially based on recent work of Gamal and Noufal, 2006.The catalogue is using the following sources: • For the period 2200 B.C to1900: Maamoun,1979

Probabilistic hazard assessment
An improved earthquake catalogue for Egypt and surrounding areas affecting El Sakakini Palace has been prepared for the purposed of this study partially based on recent work of Gamal and Noufal, 2006.The catalogue is using the following sources: -For the period 2200 B.C to1900: Maamoun ,1979  The horizontal peak ground acceleration over the bedrock of El Sakakini area was estimated using Mcguire program 1993.37 seismic source zones were used to determine the horizontal PGA over the bedrock (Figure 3), while PGA attenuation formula of Joyner and Boore, 1981 was used because of its good fitting to real earthquake data in Egypt.A complete analysis for the input parameters to estimate the PGA values over the bedrock can be found in Gamal and Noufal, 2006.

EL Sakakini Palace
( ) ( ) The probabilistic analysis provided the following results: The peak horizontal acceleration in gals with 10 % probability of exceedance over 50 years is 144cm/sec 2 (or 0.147g) For 10% probability in 100 years the estimated PGA for rock conditions is 186 (cm/sec 2 ) (or 0.19g) (Figures 3 and 4).These values are quite high and considering the local amplification they may affect seriously the seismic design and stability of El.Sakakini Palace.

Geotechnical investigation
Core drilling is among the routine methods for subsurface exploration.Most commonly, NXsize core drill is used, representing a hole diameter of 76 mm (3") and a core diameter of 54 mm (2 1/8").The drilling often has multiplier purposes, of which the following are in most cases the most important: Verification of the geological interpretation.Detailed engineering geological description of rock strata.To obtain more information on rock type boundaries and degree of weathering.
To supplement information on orientation and character of weakness zones.To provide samples for laboratory analyses.Hydro geological and geophysical testing.Input data for engineering classification of rock masses.
The geotechnical investigation, six geotechnical boreholes with Standard Penetration Test (SPT) measurements have been carried out in the archaeological site included the drilling of three geotechnical boreholes with integral sampling to a depth 20 meters, one borehole to depth 15 meters and two boreholes to depth 10 meters at six locations in the site.The geotechnical data also indicated the ground water level at the archaeological site.We did all the boreholes inside the site with hand boring machine.
The results of laboratory tests which have been carried out on the extracted soil samples from the boreholes, which include specific gravity (Gs), water content (Wn), saturated unit weight (γsat), unsaturated unit weight (γunsat), Atterberg limits and uniaxial compressive strength (UCS), in addition to the ground water table (GWT), are shown in the figures (7a,7b).
The shear wave profile obtained by using ReMi compared very well to geotechnical boreholes and geophysical survey data.In addition, the shear wave profile obtained by using ReMi Performed much better than commonly used surface shear-wave velocity measurements.
Geotechnical boreholes (1) through (3) indicated that: Filling of Fill (silty clay and limestone fragments, calc, dark brown) From ground surface 0.00m to 3.50m depth.Sand Fill (silty clay, medium, traces of limestone& red brick fragments, calc, dark brown) From 3.50m to 5.00m depth.Silty clay, stiff, calc, dark brown From 5.00m to 6.50m depth.Clayey silt, traces of fine sand & mica, yellowish dark brown From 6.50m to 8.50m depth.Silty sand, fine, traces of clay & mica.Dark brown.From 8.50m to 11.00m depth.Sand, fine, some silt, traces of mica, yellowish dark brown.From 11.00m to 14.00m depth.Sand, fine to medium, traces of silt& mica, tracesof fine to medium gravel, traces of marine shells, yellowish dark brown.From 14.00m to 16.00m depth.Sand, fine, traces of silt & mica, yellowish dark brown.From 16.00m to 18.00m depth.Sand & Gravel, medium sand, graded gravel, traces of silt, yellow darkbrown.From 18.00m to 20.00m depth.End of drilling at 20.00m.
Geotechnical boreholes (4) through (6) indicated that: Fill (silt, clay and fragments of limestone and crushed brick, from ground surface 0.00m to 4m depth.Fill (silty cal with medium pottery and brick fragments, calc dark brown) from 4 to 5 m depth.Brown stiff silty clay and traces of limestone gravels, from 5.00m to 7.50m depth.silt, traces of brown fine sand & traces of clay from 12.00m to 14.00m depth.Dark brown clay silt with traces of fine sand.from 14.00m to 15.00m depth.Engineering Seismology, Geotechnical and Structural Earthquake Engineering

Geophysical campaign 4.1. P-wave refraction
A total of 10 seismic profiles are conducted at El Sakakini palace area (Figure 8).All profiles are carried out using 12 receivers, P-type geophones with 5m intervals and 2 shots.The forward and reverse shots were carried at a distance of 1 m at both ends.The seismic shots layouts are described in Table 1.The conducted profiles are interpreted using time-term inversion method; an example of the conducted profiles and corresponding geoseismic model is shown in Figure 9. Table 2 summarizes the measured Vs values and the corresponding soil thicknesses.The soil stratification is not uniform and horizontal, as it should be expected for a filled area.However it is possible to distinguish the following three main layers: the soil layering can be summarized in the following (table 2).
Soil A-Fill (<300 m/s): A surface highly heterogeneous material (mainly man-made fill) with an average thickness of 10 m and an average velocity Vs lower than 300m/s.It is composed of very loose and low strength sediments such as silt, clay and limestone fragments.It is not found in all locations.
Soil B-Clayey soil (400-600 m/s): Below the surface layer (soil A) there is a clayey or silty clay layer with an average thickness of 10 m meters and Vs velocity 400-600 m/s.

Refraction-microtremor (ReMi method)
We have used the ReMi (refraction microtremors) method to determine the S-wave seismic velocity with depth.The method is based on two fundamental ideas.The first is that common seismic-refraction recording equipment, set out in a way almost identical to shallow P-wave refraction surveys, can effectively record surface waves at frequencies as low as 2 Hz (even lower if low frequency phones are used).The second idea is that a simple, two-dimensional slowness-frequency (P-f) transform of a microtremors record can separate Rayleigh waves from other seismic arrivals, and allow recognition of true phase velocity against apparent velocities.Two essential factors that allow exploration equipment to record surface-wave velocity dispersion, with a minimum of field effort, are the use of a single geophone sensor at each channel, rather than a geophone "group array", and the use of a linear spread of 12 or more geophone sensor channels.Single geophones are the most commonly available type, and are typically used for refraction rather than reflection surveying.There are certain advantages of ReMi method: it requires only standard refraction equipment, widely available, there is no need for a triggering source of energy and it works well in a seismically noisy urban setting.(Louie, 2001, Pullammanappallil et al. 2003).
A 12 channel ES-3000 seismograph was used to measure background 'noise' enhanced at quiet sites by inducing background noise with 14Hz geophones in a straight line spacing 5m Figure 5 shows the map were ReMi measurements were made.Almost all the sites were noisy.In particular big hammer used to break some rocks generated noisy background at El Sakakini Palace.30files of 30sec records (unfiltered) of 'noise' were collected at each site.Five profiles were taken inside the Palace (Figure 8).To ReMi-5 (Figure 9).

5-FREQUENCY CHARACTERSITICS OF THE SOIL AND THE BUILDING USING MICROTREMRS
Microtremors are omnipresent low amplitude oscillations (1-10 microns) that arise predominantly from oceanic, atmospheric, and urban or anthropogenic actions and disturbances.The implicit assumption of early studies was that microtremors spectra are flat and broadband before they enter the region of interest (soil or building).When microtremors enter preferable body it changes and resonate depending on the nature of the material, shape, and any other characteristics of this body.
It may be considered to compose of any of seismic wave types.We have two main types of microtremors, Local ambient noise coming from urban actions and disturbances and long period microtremors originated from distances (e.g.oceanic disturbances).There is still a debateongoing on the characteristics of the ambient noise that should be used for site characterization and ground response.While some are using only the longer period microtremors originated from farther distances (e.g.

Frequency charactersitics of the soil and the building using microtremrs
Microtremors are omnipresent low amplitude oscillations (1-10 microns) that arise predominantly from oceanic, atmospheric, and urban or anthropogenic actions and disturbances.The implicit assumption of early studies was that microtremors spectra are flat and broadband before they enter the region of interest (soil or building).When microtremors enter preferable body it changes and resonate depending on the nature of the material, shape, and any other characteristics of this body.
It may be considered to compose of any of seismic wave types.We have two main types of microtremors, Local ambient noise coming from urban actions and disturbances and long period microtremors originated from distances (e.g.oceanic disturbances).There is still a debateongoing on the characteristics of the ambient noise that should be used for site characterization and ground response.While some are using only the longer period microtremors originated from farther distances (e.g.

Instrumentation and data acquisition
A high dynamic range Seismograph (Geometrics ES-3000 see Figure 13) mobile station with triaxial force balance accelerometer (3 channels), orthogonally oriented was used.The station was used with 4Hz sensors to record the horizontal components in longitudinal and transverse directions in addition to the vertical components.For the data acquisition and processing we followed the following steps: • Recording 10-min of ambient noise data using a mobile station moving among variable soil stations or El Sakakini building floors/ • Zero correction to the total 10-minnoise at time domain • Subdivision of each 10-minsignal into fifteen 1-min sub-windows, Each of these series was tapered with a 3-sec hanning taper and converted to the frequency domain using a Fast Fourier transform, • Smoothing the amplitude spectrum by convolution with 0.2-Hz boxcar window, • Site response spectrum for a given soil site (or certain floor) is given by dividing the average spectrum of this site over the spectrum of the reference site.The reference site is choose carefully in the site as deepest and calmest station in the basement floor with least soil response (usually we choose a certain basement floor location with least soil response to be used as reference site).
• Smoothing the final response curves by running average filter for better viewing.A complete description of the methodology can be found in Gamal and Ghoneim, (2004).

Ground response
Figure 14 shows the locations of microtremors stations used to determine the ground response at EL Sakakini Palace area.The predominant frequency of the ground at EL Sakakini Palace is about 3 Hz (see Figure 15 & Table 1), a value almost identical to the theoretical estimation according to Kennett and Kerry (1979) (Figure 16 & Table 4).The amplification factor is about 2, which is relatively low.The fundamental natural frequency of vibration is always the most important frequency that insert the maximum earthquake vibration energy into structure.However when we find other mode of vibrations with big amplification factors we consider this as a warning that this structure may suffer from vibration.This could be very good warning for its unstable performance during vibration.Engineering Seismology, Geotechnical and Structural Earthquake Engineering            Table 5.Natural frequencies of vibration of El Sakakini Palace.

CONCLUSIONS
ElSakakini Palace is an important monument in Egypt.We presented the main results of the seismic hazard analysis and the geophysical campaign to estimate the main characteristics of the ground response and the structure.Based on the available maximum intensity maps for historical earthquakes (>2200BC) the maximum Mercalli Intensity expected at ElSakakini Palace site is VII.
The peak horizontal acceleration at the seismic rock basement found at -35m approximately, and for 10% probability of exceedance in 50 years is 144 cm/sec 2 (0.147g), while for 100 years is 186 (cm/sec 2 ) (0.19 g).We determined the average soil profile using different geophysical campaigns.It is found that the upper layer has an average shear wave velocity lower than 300m/s and a thickness of 5 to 10meters.It is a man made fill material in rather loose conditions.Below there is a clayey material with average Vs velocity equal to 400-600m/s.At -35m in average we found saturated compacted sand and gravels with Vs velocity exceeding 700m/s.It is considered as the seismic bedrock for the foreseen detailed site-specific analysis of the ground response.Table 5. Natural frequencies of vibration of El Sakakini Palace.

Conclusions
ElSakakini Palace is an important monument in Egypt.We presented the main results of the seismic hazard analysis and the geophysical campaign to estimate the main characteristics of the ground response and the structure.Based on the available maximum intensity maps for historical earthquakes (>2200BC) the maximum Mercalli Intensity expected at ElSakakini Palace site is VII.
The peak horizontal acceleration at the seismic rock basement found at -35m approximately, and for 10% probability of exceedance in 50 years is 144 cm/sec 2 (0.147g), while for 100 years is 186 (cm/sec 2 ) (0.19 g).We determined the average soil profile using different geophysical campaigns.It is found that the upper layer has an average shear wave velocity lower than 300m/s and a thickness of 5 to 10meters.It is a man made fill material in rather loose conditions.Below there is a clayey material with average Vs velocity equal to 400-600m/s.At -35m in average we found saturated compacted sand and gravels with Vs velocity exceeding 700m/s.It is considered as the seismic bedrock for the foreseen detailed site-specific analysis of the ground response.
Based on the ambient noise campaign the fundamental frequency of the ground is of the order of 3.0 to 3.5sec very close to the fundamental frequency of the palace.Resonance phenomena should be expected and considered seriously in the detailed analysis of the structure.There are strong evidences that the upper two stories with wooden floors, which are presenting high amplification factors, are subjected to several damages and degradation of their bearing capacity.

Figure 1 .Figure 2 .
Figure 1.Important and historical earthquakes occurred in and around El Sakakini Palace area in the period 2200 B.C to1995.

Figure 2 .
Figure 2. Maximum intensity zonation map based on the historical seismicity reported in the time period 2200 BC to1995

Figure 4 .
Figure 4. a: Peak Horizontal Acceleration in gals (cm/sec 2 ) for the seismic bedrock with10 % probability of exceedance in 50 years.b: Peak Horizontal Acceleration in gals (cm/sec 2 ) for the seismic bedrock with 10 % probability of exceedance over 100years

Figure 8 .
Figure 8. Location of the P-wave seismic refraction, S-wave refraction and ReMiprofiles conducted at ElSakakini Palace.

Figure 10 -
11 shows an example of the dispersion curves and its P-F image (Remi Spectral ratio of surface waves) for refraction microtremors profile ReMi-1.The estimated average Vs for all profiles are shown in Figure12.

Figure 13 .
Figure 13.High dynamic range ES-3000 Geometrics mobile station and triaxial geophone used 4 Hz to drive soil response of El-Sakakini Palace.

Figure 16 .
Figure 16.Theoretical ground response analysis at EL Sakakini Palace using Kennett at al. (1979) method.

Figure 17 .
Figure 17.Location of stations at the basement of EL-Sakakini Palace.

Figure 21 .
Figure 21.Natural frequency of vibration for basement floor.

12 Figure 21 .
Figure 21.Natural frequency of vibration for basement floor.

Figure 22 .
Figure 22.Natural frequency of vibration for the 1 st floor.

Figure 25 .
Figure 25.Natural frequency of vibration of the 4 th floor (wooden).

2 Figure 24 .
Figure 24.Natural frequency of vibration for the 3 rd floor.

Figure 24 .
Figure 24.Natural frequency of vibration for the 3 rd floor.

Figure 25 .
Figure 25.Natural frequency of vibration of the 4 th floor (wooden).

2 Figure 25 .
Figure 25.Natural frequency of vibration of the 4 th floor (wooden).

Figure 26 .
Figure 26.Natural frequency of vibration of the 5 th floor (wooden).

Figure 26 .
Figure 26.Natural frequency of vibration of the 5 th floor (wooden).

Table 2 .
Below soil B there is a stiff soil layer with various thicknesses.it shows a considerable increase of Vs seismic velocity reaching sometimes values as high as 1300m/s.The soil is composed of compacted stiff saturated sand and gravel with an average Vs velocity equal or higher than 700m/s.It may be considered as the "seismic bedrock" for the local site amplification analyses.P-wave refraction geophysical campaign conducted at El-Sakakini palace area.

Table 3 .
Fundamental frequencies and amplification factors at five locations

Table 4 .
Parameters used for the Kennett and Kerry method(1979)