Seismic hazard assessment for greater North China from historical intensity observations

Engineering Geology 164 (2013) 117–130

Contents lists available at ScienceDirect

Engineering Geology journal homepage: www.elsevier.com/locate/enggeo

Seismic hazard assessment for greater North China from historical intensity observations Jingwei Liu a,⁎, Zhenming Wang b, Furen Xie a, Yuejun Lv a a b

Institute of Crustal Dynamics, China Earthquake Administration, Beijing, China Kentucky Geological Survey, University of Kentucky, Lexington, KY, USA

a r t i c l e

i n f o

Article history: Received 11 January 2013 Received in revised form 7 May 2013 Accepted 7 July 2013 Available online 13 July 2013 Keywords: Chinese intensity Seismic hazard Seismic risk Deterministic seismic hazard analysis (DSHA) Probabilistic seismic hazard analysis (PSHA) Seismic design

a b s t r a c t Seismic hazards in greater North China were estimated from 500 years of intensity observations. Historical intensity observations were collected, the completeness of the earthquake catalog was tested, and aftershocks were deleted. The intensity data were digitized and placed in a geographic information system (GIS). Finally, the digitized intensity data were analyzed to determine the frequency–intensity relationship (i.e., seismic hazard curve). Seismic risks were also estimated, assuming a Poisson distribution for earthquake occurrence in time. The results show that greater North China faces significant seismic hazards and risks. The results also show that the current design peak ground acceleration (PGA) for greater North China might not be adequate, particularly for the Beijing, Tianjin, and Tangshan areas. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Greater North China, including the cities of Beijing and Tianjin and the provinces of Hebei, Shandong, Shanxi and Jiangsu (Figure 1), has a long history of earthquakes. Large earthquakes affecting greater North China include the 1556 Huaxian (Ms 8 14 ), 1668 Tancheng (Ms 8 12 ), 1679 Sanhe-Pinggu (Ms8), and 1976 Tangshan (Mw7.8) events (Figure 1). These large earthquakes caused heavy casualties and huge economic losses. For example, the 1556 Huaxian earthquake (Ms8 14 ) killed more than 830,000 people (the highest casualty in the world history) and the 1976 Tangshan earthquake (Mw7.8) leveled all of Tangshan City, killed more than 240,000 people, and caused huge economic losses. Greater North China is also the political, economic, and cultural center of the country. Even a moderate earthquake could cause significant economic losses in the area. For example, the 1983 Heze earthquake (Ms5.9) in Shandong Province caused losses of more than 500 million Chinese yuan (National Bureau of Statistics of China, Civil Affairs Bureau, 1995), and the 1999 Yanggao earthquake (Ms5.6) in Shanxi Province caused losses of about 250 million Chinese yuan (Group of Disasters Assessment of the Earthquake in Datong Yanggao, 2000). With rapid economic development and growing population in major cities, the area is facing significant seismic risk. Recent earthquakes, including the 2008 Wenchuan, China, 2009 L'Aquila, Italy, 2010 Haiti, 2011 Christchurch, New Zealand, and 2011

⁎ Corresponding author. Tel.: +86 13488731603. E-mail address: [email protected] (J. Liu). 0013-7952/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.enggeo.2013.07.002

Japan events, showed that seismic hazards and risks were significantly underestimated by probabilistic approaches, probabilistic seismic hazard analysis in particular (Geller, 2011; Stein et al., 2011, 2012; Kossobokov and Nekrasova, 2012; Peresan and Panza, 2012; Wang et al., 2012). As shown in Fig. 2, the predicted intensities (b VII) in the Wenchuan and Yushu areas were much less than the observed intensities (NIX) in the epicentral areas. These underestimations might lead to under preparation and disaster. The world was shocked when a prison sentence of 6 years was handed down by an Italian court to six scientists, including seismologists, for their inadequate seismic risk assessment and poor communication before the 2009 L'Aquila earthquake. The inadequate seismic risk assessment and poor communication confused the public and left them unprepared for the quake. Thus, not only it is important for earth scientists to conduct scientifically defensible seismic hazard and risk assessments, but also to communicate the assessments clearly and understandably to the public. In this paper, we will explore some key issues in current seismic hazard and risk assessment and present an alternative approach for seismic hazard and risk assessment by utilizing historical intensity observations from greater North China. 2. Seismic hazard and risk The basic concepts of seismic hazard and risk must be discussed in order to better assess and communicate seismic hazard and risk. Although seismic hazard and risk have often been used interchangeably, they are fundamentally different (Reiter, 1990; McGuire, 2004; Wang, 2006, 2007, 2009a, 2011; Wang and Cobb, 2012). Seismic hazard is

118

J. Liu et al. / Engineering Geology 164 (2013) 117–130

Fig. 1. Locations of large historical earthquakes in greater North China. Province names are in italics.

Fig. 2. Design intensity in the Wenchuan and Yushu areas (People's Republic of China National Standard, PRCNS, 2001).

J. Liu et al. / Engineering Geology 164 (2013) 117–130

“the potential for dangerous, earthquake-related natural phenomena such as ground shaking, fault rupture, or soil liquefaction” (Reiter, 1990, p. 3). Seismic risk is “the probability of occurrence of these consequences (i.e., adverse consequences to society such as the destruction of buildings or the loss of life that could result from seismic hazards)” (Reiter, 1990, p. 3). In other words, seismic hazard describes the natural phenomenon or property of an earthquake, whereas seismic risk describes the probability of loss or damage that could be caused by a seismic hazard (Wang, 2006, 2007, 2009a, 2011; Wang and Cobb, 2012). The differences between seismic hazard and risk are illustrated in Fig. 3, which shows that the Wenchuan earthquake and its aftershocks triggered massive rockfall hazards. The driver and pedestrians shown in Fig. 3, who were vulnerable to rockfall hazards, were taking a risk, a chance (probability) of being struck by a rockfall during the period that the car and pedestrians passed through the road section. This example demonstrates that seismic risk is a probable outcome (or consequence) from interaction between a seismic hazard and exposure (pedestrian, car, and driver, who are vulnerable to the seismic hazard). Thus, seismic risk can be qualitatively expressed as Seismic Risk ¼ Seismic Hazard Θ Exposure:

ð1Þ

As shown in Eq. (1), high seismic hazard does not necessary mean high seismic risk, and vice versa. For example, as shown in Fig. 3, there would be no risk (i.e., no probability that the car or pedestrians could be hit by a rockfall) if the driver or pedestrians were not exposed (i.e., the car and pedestrians avoid the road section — no exposure). In general, it may not be possible to mitigate or reduce seismic hazards, as they are naturally occurring phenomena. For example, as shown in Fig. 3, rockfall hazards are difficult to mitigate or reduce along the road section because of the steep slope. Some seismic hazards, such as liquefaction and landslides may be mitigated geotechnically. This

119

example also demonstrates that mitigation measures for hazards may differ from measures for seismic risk reduction; but the seismic risk can always be reduced by either mitigating the seismic hazard (e.g., building barriers or other measures), reducing the exposure (e.g., limiting traffic or pedestrians), or both. Seismic risk is the more important factor to consider in policy development. Quantitative assessments of seismic hazard and risk are also necessary, particularly for engineering design and other considerations. 2.1. Seismic risk assessment As shown in Eq. (1), seismic risk assessment is complicated and somewhat subjective, because it depends on the desired measurement of consequence (e.g., level of damage, amount of economic loss, or fatalities) and how the hazard and exposure interact in time and space (Wang, 2006, 2007, 2009a, 2011; Wang and Cobb, 2012). In order to estimate seismic risk, a model has to be assumed or introduced to describe how the hazard and exposure interact in time. Several models, such as the Poisson, Empirical, Brownian Passage Time, and Time-Predictable, have been used to describe earthquake occurrence in time. The most commonly used model in engineering risk estimation is the Poisson model (Cornell, 1968; Milne and Davenport, 1969). If earthquake occurrence in time follows a Poisson distribution, then seismic risk, expressed in terms of a probability of ground motion exceeding a specified level at least once for a given exposure during its lifetime t, can be estimated by P T ¼ 1−e

−τt

ð2Þ

where τ is the average recurrence interval (i.e., return period) of ground motion exceeding the specified level y. Eq. (2) describes a quantitative relationship between seismic hazard (i.e., ground motion exceeding a

Fig. 3. The differences between seismic hazard and risk. Seismic hazard: earthquake-triggered rockfall. Seismic risk: the probability of being struck by a rockfall during the period that the car or pedestrians pass through the road section.

120

J. Liu et al. / Engineering Geology 164 (2013) 117–130 Table 1 The β value for each seismic zone.

Fig. 4. The distribution of earthquakes (Ms ≥ 4 34) in time since 1484.

specified level y with an average recurrence interval τ) and seismic risk (i.e., a probability PT that the ground motion could be exceeded during the exposure lifetime t), assuming that earthquake occurrence in time

Seismic zones

β value

Downstream of Yangtse River–Yellow Sea Tanlu North China Plain Fenwei Ordos Northeast China Yinchuan–Hetao Liupanshan–Qilianshan

0.71 0.56 0.59 0.55 0.43 0.64 0.41 0.44

follows a Poisson distribution. The seismic risk estimate will be different if earthquake occurrence follows other distributions. Eq. (2) has been widely used for risk calculation in earthquake engineering (Cornell, 1968; Milne and Davenport, 1969; McGuire, 2004; Luco et al., 2007), hydraulic engineering (Gupta, 1989), and wind

Fig. 5. (a) Zones with a constant β value and (b) seismo-tectonic zones and the distribution of active Holocene faults (People's Republic of China National Standard, PRCNS, 2001). The seismo-tectonic zone is defined as a geographic region that has similar geologic structures and seismic activity.

J. Liu et al. / Engineering Geology 164 (2013) 117–130

engineering (Sachs, 1978). For example, for a given exposure time of 50 years and ground motion hazard of 0.3 g PGA with a return period of 500 years, Eq. (2) results in a probability of 10%. For a given exposure time of 1 year and flood level of 20 ft with a return period of 100 years (i.e., 100-year flood), Eq. (2) results in a probability of 1%. Thus, under two preconditions, ground motion occurrence follows a Poisson distribution and exposure time is 50 years, ground motions with return periods of 500, 1000, and 2500 years (i.e., seismic hazards)

121

are equivalent to ground motions with 10, 5, and 2 percent probabilities of exceedance in 50 years (i.e., seismic risks). Eq. (2) is derived only from the interaction between hazard and exposure in time at a site, without consideration of physical interactions. In other words, the equation can only determine the probability that an exposure could experience a certain level of hazard, without consideration of its vulnerability (i.e., inability to withstand the effects of a seismic hazard) or the related level of damage or economic

Fig. 6. Frequency–intensity curves for cells which have different numbers of intensity observations with β value of seismic zone.

122

J. Liu et al. / Engineering Geology 164 (2013) 117–130

is a mathematical formulation derived in the 1970s by engineers from a probability analysis on the statistical relationships of earthquake magnitudes, locations, and ground-motion attenuation (McGuire, 2008). Although PSHA has advanced greatly since its inception, all of the advances depend, at their core, on the early formulation by Cornell (1968) (McGuire, 2008). Recent studies (e.g., Wang and Zhou, 2007; Wang, 2009a,b, 2011, 2012; Wang and Cobb, 2012) have shown that the original formulation of PSHA (Cornell, 1968) is scientifically flawed because some of the assumptions that PSHA is based on are not valid in earthquake science and because it contains a mathematical error: equating a dimensionless quantity (the annual probability of exceedance − the probability of exceedance in 1 year) to a dimensional quantity (the annual frequency of exceedance or rate with the unit of per year [1/year]) (Wang, 2011, 2012; Wang and Cobb, 2012). This mathematical error leads to the so-called ergodic assumption —“PSHA treats that spatial uncertainty of ground motions as an uncertainty over time at a single point” (Anderson and Brune, 1999, p. 19) and the confusion between probability and frequency (per year) in modern PSHA: the annual probability of exceedance (i.e., probability of exceedance in 1 year) has been used as the annual frequency of exceedance (1/year) (McGuire, 2004, 2008; Hanks et al., 2012; Musson, 2012a, b). In other words, 0.01 (i.e., 1% − dimensionless quantity) and 0.01 per year (i.e., 1% per year − dimensional quantity) are treated the same in modern PSHA (Hanks et al., 2012; Musson, 2012b). As shown by Musson (2012b), probability is a proportion of years:

loss. The physical interaction between seismic hazard and exposure is complicated and can be determined from a fragility analysis. For example, for certain buildings, there is a relationship between ground motion and damage level, expressed as a fragility curve (Kircher et al., 1997). The damage level can also be related to a level of economic loss or fatalities. Thus, seismic risk, in terms of the probability PD that a level of damage to the exposure could be caused by a seismic hazard over an exposure time, can be estimated from
 −t P D ¼ P T  P V ¼ 1−e τ P V

ð3Þ

where PV is the exposure's vulnerability to damage (i.e., probability of damage versus the level of ground motion y). As shown in Eq. (3), reducing vulnerability PV through strengthening the built environment will reduce risk. This also demonstrates that better engineering design for buildings and other structures is an effective way to reduce seismic risk. Thus, detailed seismic risk assessment is complicated and beyond earth science, and seismic hazard assessment is a key component of seismic risk assessment. 2.2. Seismic hazard assessment The main goal of seismic hazard assessment is to quantify ground motion and its return period (i.e., recurrence interval) at a site or in a region using scientific information obtained from instrumental, historical, and geologic observations made by earth scientists (geologists, geophysicists, and seismologists). There are two general approaches for seismic hazard assessment: empirical and theoretical. The empirical method utilizes observational data directly to develop seismic hazard curves for the site(s) of interest (Milne and Davenport, 1969; Bozkurt et al., 2007; Xie et al., 2011). The key limitation of the empirical method is that observations are scarce in most regions of the world. Two theoretical approaches are widely used for seismic hazard assessment: probabilistic seismic hazard analysis (PSHA) (Cornell, 1968; McGuire, 2004, 2008) and deterministic seismic hazard analysis (DSHA) (Krinitzsky, 1995, 2002). PSHA and DSHA use the same seismologic and geologic information, but define and calculate seismic hazard fundamentally differently. PSHA was intended to estimate the probability that ground motion will exceed a given level from all earthquakes, whereas DSHA estimates the ground motion from one earthquake or a set of scenario earthquakes. In other words, PSHA emphasizes probability and depends on statistical models of earthquakes, whereas DSHA emphasizes ground motion and depends on physical models of earthquakes.

P ¼ Y e =Y t

ð4Þ

where Ye is the number of years with exceedances and Yt is the total number of years. The probability defined in Eq. (4) is dimensionless — the number of years over the total number of years, and its inverse is still dimensionless. However, Musson (2012a, p. 717) stated, “The inverse of the probability is a number of years.” Thus, PSHA has become an erroneous mathematical (computer) model without an earth-science basis, and its results are artifacts. PSHA analysts have become experts in probability theory, instead of experts in earth science. This can be seen in recent publications by Scherbaum and Kuehn (2011), Hanks et al. (2012), and Musson (2012a, 2012b). Therefore, the use of PSHA in seismic hazard and risk assessments is problematic. PSHA could either underestimate ground motion hazards, such as the 2008 Wenchuan, China, 2009 L'Aquila, Italy, 2010 Haiti, and 2011 Christchurch, New Zealand, earthquakes, or overestimate ground motion hazards, such as the extreme ground motion of 11 g PGA and 13 m/s PGV for the Yucca Mountain nuclear waste repository site (Hanks, 2011) and the higher design ground motion for the New Madrid region of the central United States (Wang and Cobb, 2012). These over- and underestimations can have dire consequences for society.

2.2.1. PSHA As pointed out by Hanks (1997, p. 369), “PSHA is a creature of the engineering sciences, not the Earth sciences, and most of its top practitioners come from engineering backgrounds.” This is because PSHA Table 2 Return period of different intensities for major cities in the study area. Cell no.

66 345 645 816 2720 3427 5132 6066 7876 8162 8582 9574 10802

City

Hangzhou Shanghai Hefei Nanjing Xi'an Zhengzhou Linfen Jinan Taiyuan Shijiazhuang Yinchuan Tianjin Beijing

Longitude

120.169°E 121.459°E 117.276°E 118.769°E 108.944°E 113.641°E 111.515°E 117.001°E 112.565°E 114.495°E 106.274°E 117.194°E 116.368°E

Latitude

30.294°N 31.230°N 31.865°N 32.053°N 34.267°N 34.758°N 36.078°N 36.650°N 37.874°N 38.033°N 38.467°N 39.126°N 39.931°N

Number of observations

Return period (year) I = VII (VAR.)

I = VII (VAR.)

I = IX (VAR.)

33 39 34 46 36 38 51 56 55 58 44 56 67

1029 762 277 501 99 340 91 194 144 163 100 107 112

5231 3873 1003 2547 350 1322 319 756 507 634 274 416 437

N10,000 N10,000 3634 (1887) N10,000 1229 (625) 5144 (1770) 1123 (398) 2941 (1203) 1781 (450) 2465 (238) 751 (127) 1619 (188) 1700 (550)

(564) (338) (54) (112) (34) (111) (45) (44) (37) (50) (54) (39) (41)

(1375) (1612) (224) (1780) (128) (233) (77) (283) (222) (138) (57) (122) (123)

J. Liu et al. / Engineering Geology 164 (2013) 117–130

Fig. 7. Return periods of I ≥ VII (a), VII (b), and IX (c) in the study area.

123

124

J. Liu et al. / Engineering Geology 164 (2013) 117–130

Fig. 8. Intensity distribution for return periods of 100 (a) and 500 years (b).

2.2.2. DSHA, neo-DSHA, or scenario hazard analysis DSHA estimates ground motion from a single earthquake or several scenario earthquakes (i.e., maximum magnitude, maximum probable, or maximum credible earthquakes) that have maximum impact at a site. Ground motion hazard derived from DSHA has a clear physical and statistical meaning. Recent efforts in DSHA have focused on computer simulation for ground-motion hazard quantification (Panza et al., 2001; Irikura and Miyake, 2011; Zuccolo et al., 2011; Peresan and Panza, 2012) — called Neo-DSHA (Panza et al., 2001) or scenario hazard analysis (Wang and Cobb, 2012; Wang et al., 2012). Neo-DSHA or

scenario hazard analysis has several advantages, such as (1) it results in ground motion with clearly understood physical and statistical meaning, (2) it is easy to understand, and (3) it utilizes ground-motion simulation. One of the drawbacks in traditional DSHA (Krinitzsky, 1995, 2002) is that “frequency of occurrence is not explicitly taken into account” (Reiter, 1990, p. 225). In other words, the temporal characteristic of earthquakes (i.e., recurrence interval or frequency and its associated uncertainty) is not addressed in traditional DSHA. The temporal characteristic of earthquakes and resulting ground motions at a site is an integral part of seismic hazard and must be considered in engineering

Table 3 Relationship between intensity and peak ground acceleration (People's Republic of China National Standard, PRCNS, 2001). Peak ground acceleration (g)

b0.05

0.05

0.10–0.15

0.15–0.20

0.20–0.30

0.30–0.40

≥0.40

Chinese intensity

bVI

VI

VII

VII

VIII

VIII

≥IX

J. Liu et al. / Engineering Geology 164 (2013) 117–130

125

Fig. 9. Equivalent PGA distribution (a) and design PGA (b) (People's Republic of China National Standard, PRCNS, 2001) in the study area.

design and other policy considerations. As pointed out by Wang et al. (2004), a scenario earthquake can always be associated with a recurrence interval and its uncertainty. As shown by Peresan et al. (2013),

the recurrence interval (i.e., the return period) for ground motion exceeding a given level at a site is related to the recurrence intervals of earthquakes that have impacts on the site and can be determined

Table 4 Average occurrence intervals of characteristic earthquakes on some active faults in the study area. Fault

Seismo-structural zones

Historical earthquake

Characteristic Average magnitude occurrence interval

References

Tangshan Fault New Xiadian Fault Tanlu Fault (Juxian–Tancheng section) Haicheng–Jinzhou Fault Huoshan Mountain Front Fault

North part of North China Plain North Part of North China Plain West Shandong East Liaoning Middle and East Shanxi

7 12–8 8 8 12 7–7 12 8

1500–7500 6000–7000 3000–4000 5500–6000 1500–2000

Wang (1984); Liu et al. (1997) Xiang et al. (1988); Ran et al. (1997) Lin and Gao (1987); Cao et al. (1994) Han et al. (1991) Su et al. (2003)

Northern Xiqinling Faults Huashan Mountain Front Fault Luoshan Mountain Eastern Foot Fault Wutaishan Mountain Northern Foot Fault

Weihe Weihe Tianjingshan–Liupanshan Shanxi–Hebei–Inner Mongolia

Tangshan Ms7.8 in 1976 Sanhe-Pinggu Ms8 in 1679 Tancheng Ms8 12 in 1668 Haicheng Ms7.3 in 1975 Zhaocheng–Hongdong Ms8 in 1303 Tianshui Ms7 12 in 734 Huaxian Ms8 in 1556 Zhongning Ms7 14 in 1561 Yuanping–Daixian Ms7 12 in 512

7 12 8 7 14 7 12

−1000 2000–2500 2000–3000 −2500

Cao et al. (2003) Li (1992) Min et al. (1992, 1993, 1996, 2000) Chen and Huang (1992); Liu et al. (1991)

126

J. Liu et al. / Engineering Geology 164 (2013) 117–130

J. Liu et al. / Engineering Geology 164 (2013) 117–130

accordingly. In other words, DSHA can also provide the frequency or recurrence interval (i.e., return period) of ground motion. Thus, DSHA is more appropriate for seismic hazard assessment. This has been demonstrated by the satisfactory performance of DSHA maps for the design ground motions for buildings (Building Seismic Safety Council, BSSC, 2009) and bridges (Mualchin, 2011) in California and the design ground motions in Japan (Kuramoto, 2006). 2.2.3. Empirical seismic hazard analysis Similar to the empirical flood hazard analysis in hydraulic engineering, an empirical seismic hazard assessment has been used to derive seismic hazard in terms of ground motion or intensity versus its occurrence frequency (per year) at a site of interest (Milne and Davenport, 1969; Bozkurt et al., 2007; Xie et al., 2011). The limitation of the empirical method is that it requires a long history of observations. Thus, the empirical method can only be applied to limited areas that have sufficient observations. We applied empirical analysis to derive seismic hazard in terms of intensity versus occurrence frequency (per year) in greater North China, utilizing about 500 years of historical observations in this paper.

Similar to the Gutenberg–Richter relationship, the hazard curve, or relationship between intensity and mean occurrence frequency at the center of each individual cell, follows: logð f Þ ¼ α−β  I

Log ðR0 Þ ¼ 0:48  M–1:57

vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 0X 1 u u 2 u ð − logf Þ logf i c B C R¼u 2 A t1−@X
logf i − logf c X

ð5Þ

where R0 is the spatial radius from the main shock for foreshocks and aftershocks in kilometers, and M is magnitude. Considering that the precise location precision of earthquakes is often unknown, the minimum value of R0 was set at 5 km. Records of 652 independent events, including 366 historical earthquakes from 1484 to 1911 (China Earthquake Administration, CEA, 1995) and 286 earthquakes from 1912 to 2008 (China Earthquake Administration, CEA, 1999), were obtained. The intensity scale used in this study is the Chinese intensity scale, with 12 grades, I through XII. Some of the intensity maps only showed felt range and had no intensity values. The intensity attenuation relationship of Wang et al. (2000) was used to calculate intensity distribution, Ia and Ib from the major axis and minor axis, respectively: Ia ¼ 5:019 þ 1:446M−4:136 logðRa þ 24Þ σ ¼ 0:517

ð6Þ

Ib ¼ 2:240 þ 1:446M−3:070 logðRb þ 9Þ σ ¼ 0:517

ð7Þ

where M is magnitude, Ra and Rb are radius from major and minor axes in kilometers, respectively, and σ is the standard deviation. Based on the coverage area of intensity data, geological characteristics, and population density, the study area was divided into 0.1° × 0.1° cells in order to depict the characteristics of spatial distribution of seismicity in the study area. The total number of cells obtained was 15,562. The intensity maps were digitized using ArcGIS under the WGS1984 coordinate system at the centers of 0.1° × 0.1° cells.

ð8Þ

where f is the frequency for intensity exceeding I, and α and β are constants determined by least-squares fitting (Milne and Davenport, 1969; Bozkurt et al., 2007; Xie et al., 2011). The range of β for all cells is quite large, from 0.184 to 1.477, respectively. In order to constrain the large variation of β values, a constant value was applied to all cells in a zone with similar seismo-tectonic characteristics (Figure 5). Table 1 shows the β value for each zone. In order to test how the frequency–intensity curve satisfies observations in individual cells, the regression coefficient and least-squares residual (∏) for the frequency–intensity curve are used. They follow:

3. Seismic hazard assessment in greater North China China has long and rich historical records on earthquakes. According to Huang et al. (1994a, b), in North China, except for Inner Mongolia and the Yellow Sea, the earthquake catalog is complete for MS ≥4 34 since 1484. From the earthquake catalogs (Gu, 1983; Xie et al., 1983–1987, 1989; China Earthquake Administration, CEA, CEA, 1995, CEA), 856 earthquakes and their intensity observations (maps) were obtained, including MS ≥4 34 earthquakes since 1484 and all MS ≥7 earthquakes since the beginning of recorded history. Fig. 4 shows the distribution of earthquakes in time. Foreshocks and aftershocks were eliminated from the database using the G–C model (Chen et al., 1998):

127

П¼

ð logf i − logf c Þ n

ð9Þ

2

ð10Þ

where n is the number of the intensity data, R is the regression coefficient, ∏ is the least-squares residual, and fi and fc are observed and predicted frequencies, respectively. The results show that in most cells, the frequency–intensity curve can satisfy observations very well (R N 0.900). Fig. 6 shows the data points, intensity–frequency curve, obtained α, regression coefficient, and least-squares residual with β value of different seismic zones for the Beijing, Tianjin, Tangshan, Shijiazhuang, Taiyuan, and Shenyang cells. We can estimate the mean occurrence frequency (f) or mean return period (1/f) (i.e., the reciprocal of the mean occurrence frequency) for a given intensity at each cell using Eq. (8). We can also estimate the intensity for a given frequency or return period at each cell from Eq. (8). Table 2 lists the return periods of different intensities for major cities in the study area. Fig. 7a, b, and c show the return periods with intensity I ≥ VII, VIII, and IX for all cells in the study area, respectively. Fig. 8 shows the intensity distribution for return periods of 100 and 500 years, respectively. Table 2 and Fig. 7 show that the return periods for I ≥ VII and IX are greater than 500 years in the study area. 4. Discussion Intensity scale, as shown in Appendix A, is a comprehensive measurement (index) of human reactions and buildings damaged by ground motion and its induced hazards. In other words, intensity is not a true ground motion measurement, but rather a measure of the consequences resulting from the interactions between humans and buildings and ground motion. Thus, an intensity scale not only depends on the strength of ground motion, but also on human reaction, building strength, and site conditions. Intensity scales are different for different countries, because human reactions and buildings are different. For example, the Chinese intensity scale is different from the Japanese and Italian scales. However, there is a general statistical relationship between intensity and ground motion. For example, there is a relationship between the Chinese seismic intensity scale and peak ground acceleration (PGA) (Table 3) (People's Republic of China National Standard, PRCNS, 2008). Similar statistical relationships between seismic intensity and ground motion can be found in the United States, Italy, and

Fig. 10. Exceedance probabilities of I = VII, VII, and IX in 50 years in the study area. The requirements for reliable cells are that the number of intensity observations in an individual cell is more than 15, the number of different intensity values is equal to or more than III, and the regression coefficient of the intensity–frequency fitting curve is equal to or greater than 0.85.

128

J. Liu et al. / Engineering Geology 164 (2013) 117–130

other countries (Ambraseys, 1973; Wald et al., 1999; Molchan et al., 2002, 2004). Thus, seismic intensity can be converted into ground motion by using the statistical relationship. For example, we can use Table 3 to convert the Chinese intensity scale into PGA. Fig. 9a shows the equivalent PGA with a return period of 500 years for the study area. As discussed earlier, if earthquake occurrences follow a Poisson distribution in time and exposure time for buildings is 50 years, the PGA with a return period of 500 years is equivalent to the PGA with a 10 percent probability of exceedance in 50 years, based on Eq. (2). Fig. 9b is the design PGA zonation map for greater North China with 10 percent exceedance probability in 50 years, which was produced from PSHA (People's Republic of China National Standard, PRCNS, 2001). The PSHA used in People's Republic of China National Standard (PRCNS) (2001) is a traditional PSHA with some modifications in the input parameters. For example, in order to depict the spatial asymmetry, both the seismic statistical region and potential source zone were used as inputs (Han et al., 2002; Lu, 2006; Gao et al., 2008). As shown in Fig. 9, the distribution of high PGA and low PGA values is similar, but the PGA values derived from this study are higher than those derived from PSHA (People's Republic of China National Standard, PRCNS, 2001). The biggest difference is that the PGA results for the North China Plain and Ordos basin are much higher than the current design PGA (People's Republic of China National Standard, PRCNS, 2001). This suggests that the current design PGA map of China (People's Republic of China National Standard, PRCNS, 2001) might not be adequate for seismic design in the study area, particularly in the Beijing, Tianjin, and Tangshan areas. As shown in Table 2, the mean return periods for intensities greater than VIII are larger than 500 years that is longer than the observation period (i.e., 500 years). In other words, the mean return periods were extrapolated from the 500 years of observation. The calculated mean return periods for intensity greater than VIII are consistent with the recurrence intervals of the characteristic earthquakes determined from geological, geodetic, and other studies. The North China Block is controlled by normal strike-slip or normal faults. As shown in Fig. 5, the study area covers 21 seismo-tectonic subzones, and seven of them have the potential for a great earthquake with magnitude equal to or greater than Ms8. These active seismo-tectonic zones are distributed in the north China Plain and around the Ordos Block, which have a shorter return period of strong intensity (I ≥ IX). A comparison of the resulting return periods with different intensities (Figure 7) and distribution of seismo-tectonic zones (Figure 5) shows that the areas with higher frequency of intensity (I ≥ VIII, IX) are consistent with the areas that have higher potential for large earthquakes with magnitude equal to or greater than Ms7.5, whereas other areas that have a lower frequency of intensity I ≥ VIII and IX, such as Yangtse River Delta, have lower potential for strong earthquakes (≥Ms7.5). Table 4 lists

the recurrence intervals of characteristic earthquakes on some active faults in the study area. In addition, the average recurrence interval of Ms8 (the Yinchuan–Pingluo earthquake in 1739) in the Yinchuan subzone was estimated to be about 2300 to 3000 years (Deng and Liao, 1996). We can estimate seismic risk in terms of probability of an intensity exceeding a specified level (I) for an exposure with a life of 50 years at each cell using Eqs. (2) and (8). The exceedance probabilities for intensity I = VII, VIII, and IX in 50 years are shown in Fig. 10. As shown in Fig. 10, the middle part of the study area has high seismic risk, especially the Beijing–Tianjin–Tangshan area, Shanxi Province, the central section of Shanxi Province, and the joint area between Shanxi Province and Ningxia municipality. Table 5 lists the 50-year exceedance probabilities of I = VII, VIII, and IX for some major cities in greater North China.

5. Conclusions Seismic hazards in greater North China were estimated from historical intensity observations since 1484 in terms of intensity versus annual frequency (hazard curve). The results show that the mean return periods for intensity I ≥ VII are less than 400 years in most parts of greater North China. Tangshan and its vicinity is the area that has the highest frequency of intensity, I ≥ VII. The return period of I ≥ VIII is about 437 years for Beijing, 416 years for Tianjin, 350 years for Xi'an, 507 years for Taiyuan, and 274 years for Yinchuan. The areas with higher intensities are consistent with the areas with higher seismicity. The results show that greater North China is facing high seismic hazards. Seismic risks, in terms of the probability of an intensity exceeding a given value in 50 years, in greater North China were also estimated, assuming a Poisson distribution for earthquake occurrence in time and an exposure lifetime of 50 years. The results show that greater North China is facing significant seismic risk. The resulting risk estimates also show that the current design PGA map for greater North China (People's Republic of China National Standard, PRCNS, 2001) might not be adequate, particularly in the Beijing–Tianjin–Tangshan area.

Acknowledgments We gratefully thank the financial support from the Institute of Crustal Dynamics, China Earthquake Administration (No. ZDJ2012-13, Spatial distribution of b value and its relationship with source stress). We thank Meg Smath of the Kentucky Geological Survey for editorial help. We also thank two anonymous reviewers for their valuable comments and suggestions that improve this manuscript greatly.

Table 5 Seismic risk for major cities in the study area. Cell no.

City

Longitude

Latitude

Number of observations

50 year probability of I ≥ VII (%)

50 year probability of I ≥ VIII (%)

50 year probability of I ≥ IX (%)

66 345 645 816 2720 3427 5132 6066 7876 8162 8582 9574 10802

Hangzhou Shanghai Hefei Nanjing Xi'an Zhengzhou Linfen Jinan Taiyuan Shijiazhuang Yinchuan Tianjin Beijing

120.169°E 121.459°E 117.276°E 118.769°E 108.944°E 113.641°E 111.515°E 117.001°E 112.565°E 114.495°E 106.274°E 117.194°E 116.368°E

30.294°N 31.230°N 31.865°N 32.053°N 34.267°N 34.758°N 36.078°N 36.650°N 37.874°N 38.033°N 38.467°N 39.126°N 39.931°N

33 39 34 46 36 38 51 56 55 58 44 56 67

5 6 17 9 40 14 42 23 29 26 39 37 36

1 1 5 2 13 4 14 4 9 8 17 11 11

0.2 0.3 1 0.4 4 1 4 2 3 2 6 3 3

J. Liu et al. / Engineering Geology 164 (2013) 117–130

129

Appendix A. The intensity scale used in this study is the Chinese seismic intensity scale (2008). According to this scale, intensity is divided into 12 grades, I through XII. Following is an abbreviated description of the 12 levels of the Chinese seismic intensity scale

Intensity

Person feeling

Building damage

Other damage

I II

Not felt Felt only by one or two persons at rest indoors Felt by a few persons at rest indoors Felt indoors by many, outdoors by few. A few persons awakened Felt indoors by nearly everyone, outdoors by many. Many awakened.

Windows, doors make slight sound Windows, doors make sound

Suspender jiggles Suspender wiggles, vessels make sound Suspender strongly wiggles. Unstable objects overturn.

0.31 (0.22–0.44)

Furniture and objects move, riverside and soft soil crack, one or two brick chimneys slightly cracked

0.63 (0.45–0.89)

Objects fall down, riverside collapses, soft soil heavily cracked, many brick chimneys moderately damaged.

1.25 (0.90–1.77)

Dry and hard soil cracks, most brick chimneys heavily damaged.

2.50 (1.78–3.53)

Dry and hard soil heavily cracked, bedrock cracked, landslides appear, many brick chimneys destroyed.

5.00 (3.54–7.07)

XII

All destroyed.

Landslip and earthquake faults appear, arch bridge on bedrock destroyed, nearly every brick chimney destroyed Earthquake faults take long rupture, and masses of landslip Ground, mountains and rivers acutely changed.

10.00 (7.08–14.14)

XI

Doors, windows, roofs vibrate. A few walls slightly cracked. A few chimneys on the roof may be damaged. a A: A few moderately damaged, many slightly damaged or almost well a B: One or two moderately damaged, a few slightly damaged, almost well a C: One or two slightly damaged, nearly every house well A: A few destroyed or heavily damaged, many moderately or slightly damaged. B: A few moderately damaged, many slightly damaged or well. C: A few moderately or slightly damaged, many well. A: A few destroyed, many heavily or moderately damaged. B: One or two destroyed, a few heavily damaged, many moderately or slightly damaged. C: A few heavily or moderately damaged, many slightly damaged. A: Many heavily damaged or destroyed. B: A few destroyed, many heavily or moderately damaged. C: A few destroyed or heavily damaged, many moderately or slightly damaged. A: Nearly every house destroyed. B: Almost destroyed. C: Many destroyed or heavily damaged. A, B, C: Nearly every house destroyed.

III IV V

VI

Many persons cannot stably stand, a few frightened and escape to outdoor.

VII

Nearly everyone frightened and escape to outdoors. Felt by riders on bikes and drivers in cars.

VII

Many persons cannot make a move

IX

Moving persons tumble.

X

Riders tumble, unstable persons threw up

PGA (m/s2)

a Buildings are divided into three types: A: Old houses built with wood, soil, stone and brick. B: Unfortified monolayer or multilayer brick house. C: Fortified monolayer or multilayer brick houses according to intensity VII.

References Ambraseys, N.N., 1973. Earth Sciences in Archaeology and History. Eng. Antiquity Publications, Gloucester. Anderson, J.G., Brune, J.N., 1999. Probabilistic seismic hazard analysis without the ergodic assumption. Seismological Research Letters 70 (1), 19–28. Bozkurt, S.B., Stein, R.S., Toda, S., 2007. Forecasting probabilistic seismic shaking for greater Tokyo from 400 years of intensity observations. Earthquake Spectra 23, 525–546. Building Seismic Safety Council (BSSC), 2009. NEHRP Recommended Provisions for Seismic Regulations for New Buildings, Federal Emergency Management Agency (FEMA) Report P-750, 2009 ed. , p. 372. Cao, H., Li, J., Cui, Z., Zhao, Q., 1994. Characteristic slip behavior of the Holocene fault in the central section of the Tanlu fault zone and the characteristic earthquakes. Inland Earthquake 8 (4), 297–304 (in Chinese). Cao, J., Liu, B., Wen, X., 2003. Determination of the average recurrence intervals of characteristic earthquakes and estimate of earthquake risk on northern Xiqinling faults. Journal Seismology Research 4, 372–381 (in Chinese). Chen, G., Huang, Z., 1992. Basic features of prehistorical earthquake during the period of Holocene in Shanxi province. Journal of Seismology 1, 10–18 (in Chinese). Chen, L., Liu, J., Chen, Y., Chen, L., 1998. Aftershocks deletion in seismicity analysis. Chinese Journal of Geophysical 41 (S), 244–252 (in Chinese). China Earthquake Administration (CEA), 1995. Historical Strong Earthquake Catalog of China (2300 B.C.–1911 A.D.). Earthquake Publishing House, Beijing (in Chinese). China Earthquake Administration (CEA), 1999. Recent Earthquake Catalog of China (1912–1990, MS ≥ 4.7). Chinese Science and Technology Press, Beijing (in Chinese). Cornell, C.A., 1968. Engineering seismic risk analysis. Bulletin of Seismological Society of America 5 (5), 1583–1606. Deng, Q., Liao, Y., 1996. Paleoseismology along the range-front fault of Helan Mountains, north central China. Journal Geological Research 101 (B3), 5873–5893.

Gao, M., Xiao, H., Yan, W., Yu, Y., Chen, K., Chen, X., 2008. The importance of seismic zonation in moderate earthquake area and the recent technical progress. Technical Earthquake Disaster Prevention 3 (1), 1–7 (in Chinese). Geller, R.J., 2011. Shake-up time for Japanese seismology. Nature 472, 407–409. Group of Disasters Assessment of the Earthquake in Datong Yanggao, 2000. Losses assessment of seismic disasters of the Ms5.6 earthquake in Datong-Yanggao, Shanxi of 1999. Earthquake Research Shanxi 1, 7–15 (in Chinese). Gu, X., 1983. Earthquake Catalog of China (1831 B.C.–1969 A.D.). Science Press, Beijing (in Chinese). Gupta, R.S., 1989. Hydrology and Hydraulic Systems. Prentice Hall, Englewood Cliffs, New Jersey 739. Han, D., Lei, Q., Liu, J., Shi, S., 1991. An ancient earthquake discovered in Haicheng Earthquake region. Seismological Research Northeast China 4, 44–48 (in Chinese). Han, Z., Wu, L., Yu, G., Ye, Y., 2002. The high gradients of bouguer gravity anomalies and seismotectonics of moderate earthquakes in Jianghuai region, China. Seismological Research China 18 (3), 230–238 (in Chinese). Hanks, T.C., 1997. Imperfect science: uncertainty, diversity, and experts. Eos, Transactions of the American Geophysical Union 78, 369–377. Hanks, T.C., 2011. Extreme ground motion. Seismological Research Letters 82, 305–306. Hanks, T.C., Beroza, G.C., Toda, S., 2012. Have recent earthquakes exposed flaws in or misunderstandings of probabilistic seismic hazard analysis? Seismological Research Letters 83, 759–764. Huang, W., Li, W., Cao, X., 1994a. Study of the completeness of the continental earthquake catalog of China—an example for the North China. Seismology 16 (3), 273–280 (in Chinese). Huang, W., Li, W., Cao, X., 1994b. Study of the completeness of the continental earthquake catalog of China—the beginning year distribution of complete earthquake observations in subzones. Seismology 16 (4), 423–432 (in Chinese). Irikura, K., Miyake, H., 2011. Recipe for predicting strong ground motion from crustal earthquake scenarios. Pure and Applied Geophysics 168, 85–104.

130

J. Liu et al. / Engineering Geology 164 (2013) 117–130

Kircher, C.A., Nassar, A.A., Kustu, O., Holmes, W.T., 1997. Development of building damage functions for earthquake loss estimation. Earthquake Spectra 13 (4), 663–682. Kossobokov, V.G., Nekrasova, A.K., 2012. Global seismic hazard assessment program maps are erroneous. Seismic Instruction 48, 162–170. Krinitzsky, E.L., 1995. Deterministic versus probabilistic seismic hazard analysis for critical structures. Engineering Geology 40, 1–7. Krinitzsky, E.L., 2002. How to obtain earthquake ground motions for engineering design. Engineering Geology 65, 1–16. Kuramoto, H., 2006. Seismic design codes for buildings in Japan. Journal Disaster Research 1, 341–356. Li, Y., 1992. Research on Ground Fissures in Xi'an Region and Active Faults in Weihe Basin. Earthquake Publishing Press, Beijing (in Chinese). Lin, W., Gao, W., 1987. The recurrence intervals of large earthquake in the Yishu fault zone. Earthquake Research China 3 (3), 34–39 (in Chinese). Liu, G., Yu, S., Zhang, S., Dou, S., Xu, Y., Fan, J., 1991. The northern active fault zone of Wutaishan Mountain in Shanxi province. Studies of Active Faults 1, 128–129 (in Chinese). Liu, J., Song, H., Wu, Y., Liu, G., 1997. Tangshan earthquake fault kinemics and recurrence interval. Seismology 19, 566–573 (in Chinese). Lu, S., 2006. Publicizing Textbook (GB17741-2005) for ‘Evaluation of seismic safety on engineering sites’ (in Chinese). China Standard Publishing House, Beijing 68. Luco, N., Ellingwood, B.R., Hamburger, R.O., Hooper, J.D., Kimball, J.K., Kircher, C.A., 2007. Risk-targeted versus current seismic design maps for the conterminous United States. Structural Engineers Association of Kentucky 2007 Convention Proceedings. Fair Oaks, California, pp. 163–171. McGuire, R.K., 2004. Seismic hazard and risk analysis. Earthquake Engineering Research Institute Report MNO-10, p. 240. McGuire, R.K., 2008. Probabilistic seismic hazard analysis: early history. Earthquake Engineers & Structural Dynamics 37, 329–338. Milne, W.G., Davenport, A.G., 1969. Distribution of earthquake risk in Canada. Bulletin of Seismological Society of America 59, 729–754. Min, W., Chai, Z., Wang, P., Yang, P., 1992. Preliminary Study on the Holocene active fault features at the eastern Piedmont of the Luoshan Mountain. Earthquake Research China 8 (4), 49–54 (in Chinese). Min, W., Chai, Z., Wang, P., Yang, P., 1993. The study on the paleoearthquakes on the eastern Piedmont fault of the Luoshan Mountains in Holocene. Plating Earthquake Research 5 (4), 97–107 (in Chinese). Min, W., Wang, Z., Jiao, D., Deng, Q., 1996. The eastern Piedmont fault of the Luoshan Mountains, Yantongshan Mountain fault and Ms#earthquake of 1561. Studies of Active Faults 5, 181–188 (in Chinese). Min, W., Zhang, P., Deng, Q., 2000. Preliminary study on recurrence of regional paleoearthquakes. Seismology 22 (2), 163–170 (in Chinese). Molchan, G., Kronrod, T., Panza, G.F., 2002. Shape analysis of isoseismals based on empirical and synthetic data. Pure and Applied Geophysics 159, 1229–1251. Molchan, G.M., Kronrod, T.L., Panza, G.F., 2004. Shape of empirical and synthetic isoseismals: comparison for Italian M b =6 earthquakes. Pure and Applied Geophysics 161, 1725–1747. Mualchin, L., 2011. History of modern earthquake hazard mapping and assessment in California using a deterministic or scenario approach. Pure and Applied Geophysics 168, 383–407. Musson, R.M.W., 2012a. PSHA validated by quasi observation means. Seismological Research Letters 83 (1), 130–134. Musson, R.M.W., 2012b. On the nature of logic tree in probabilistic seismic hazard assessment. Earthquake Spectra 28 (3), 1291–1296. National Bureau of Statistics of China, Civil Affairs Bureau, 1995. Report of the Damage Caused by Disaster in China. Chinese Statistics Press, Beijing (in Chinese). Panza, G.F., Romanelli, F., Vaccari, F., 2001. Seismic wave propagation in laterally heterogeneous anelastic media: theory and applications to seismic zonation. Advances in Geophysics 43, 1–95. People's Republic of China National Standard (PRCNS), 2001. Seismic Ground Motion Parameter Zonation Map of China, GB 18306-2001. China Zhijian Publishing House, Beijing (in Chinese). People's Republic of China National Standard (PRCNS), 2008. The Chinese Seismic Intensity Scale, GB/T 17742-2008. China Zhijian Publishing House, Beijing (in Chinese). Peresan, A., Panza, G.F., 2012. Improving earthquake hazard assessments in Italy: an alternative to “Texas sharpshooting”. Eos 93, 538.

Peresan, A., Magrin, A., Nekrasova, A., Kossobokov, V.G., Panza, G.F., 2013. Earthquake recurrence and seismic hazard assessment: a comparative analysis over the Italian territory. Proceedings of the 9th International Conference on Earthquake Resistant Engineering Structures, 8–10 July, 2013. Coruña, Spain. Ran, Y., Deng, Q., Yang, X., Zhang, W., Li, R., Xiang, H., 1997. The 1679 Sanhe–Pinggu earthquake fault and recurrence interval. Seismology and Geology 19, 193–201 (in Chinese). Reiter, L., 1990. Earthquake Hazard Analysis: Issues and Insights. Columbia University Press, New York. Sachs, P., 1978. Wind Forces in Engineering, Second ed. Pergamon Press Inc., Elmsford, New York 400. Scherbaum, F., Kuehn, N.M., 2011. Logic tree branch weights and probabilities: summing up to one is not enough. Earthquake Spectra 27 (4), 1237–1251. Stein, S., Geller, R.J., Liu, M., 2011. Bad assumptions or bad luck: why earthquake hazard maps need objective testing. Seismological Research Letters 82, 623–626. Stein, S., Geller, R.J., Liu, M., 2012. Why earthquake hazard maps often fail and what to do about it. Tectonophysics 562–563, 1–25. Su, Z., Yuan, Z., Zhao, J., 2003. A review on studies concerned with the 1303 Hongtong, Shanxi, earthquake of M8. Earthquake Research Shanxi 3, 4–9 (in Chinese). Wald, D.J., Quitoriano, V., Heaton, T.H., Kanamori, H., 1999. Relationships between peak ground acceleration, peak ground velocity, and modified Mercalli intensity in California. Earthquake Spectra 15 (3), 557–564. Wang, T., 1984. Recurrence interval of the Tangshan earthquake. Seismology and Geology 6, 77–83 (in Chinese). Wang, Z., 2006. Understanding seismic hazard and risk assessments: an example in the New Madrid Seismic Zone of the central United States. Proceedings of the 8th National Conference on Earthquake Engineering, April 18–22, San Francisco, Calif., paper 416. Wang, Z., 2007. Seismic hazard and risk assessment in the intraplate environment: the New Madrid Seismic Zone of the central United States. In: Stein, S., Mazzotti, S. (Eds.), Continental Intraplate Earthquakes: Science, Hazard, and Policy Issues. Geol. Soc. Am., Special Paper, 425, pp. 363–373. Wang, Z., 2009a. Seismic hazard vs. seismic risk. Seismological Research Letters 80 (5), 673–674. Wang, Z., 2009b. Comment on “Sigma: Issues, Insights, and Challenges” by Fleur O. Strasser, Norman A. Abrahamson, and Julian J. Bommer. Seismological Research Letters 80, 491–493. Wang, Z., 2011. Seismic hazard assessment: issues and alternatives. Pure and Applied Geophysics 168, 11–25. Wang, Z., 2012. Comment on “PSHA validated by quasi observational means” by R.M.W. Musson. Seismological Research Letters 83, 714–716. Wang, Z., Cobb, J.C., 2012. A critique of probabilistic versus deterministic seismic hazard analysis with special reference to the New Madrid seismic zone. Geol. Soc. Am., Special Paper, pp. 493–513. Wang, Z., Zhou, M., 2007. Comment on “Why do modern probabilistic seismic-hazard analyses often lead to increased hazard estimates?” by Julian J. Bommer and Norman A. Abrahamson. Bulletin of Seismological Society of America 97, 2212–2214. Wang, S., Yu, Y., Gao, A., Yan, X., 2000. Development of attenuation relations for ground motion in China. Earthquake Research China 2, 99–106 (in Chinese). Wang, Z., Woolery, E.W., Shi, B., Kiefer, J.D., 2004. Reply to comment on “Communicating with uncertainty: a critical issue with probabilistic seismic hazard analysis” by C.H. Cramer. Eos, Transactions of the American Geophysical Union 85, 283–286. Wang, Z., Butler, D., Woolery, E.W., Wang, L., 2012. Seismic hazard assessment for the Tianshui Urban Area, Gansu Province, China. International Journal Geophysics 10. Xiang, H., Fang, Z., Xu, J., Li, R., Jia, S., Hao, S., Wang, J., Zhang, W., 1988. Tectonic background and recurrence interval of the Sanhe–Pinggu earthquake. Seismology and Geology 10, 15–28 (in Chinese). Xie, Y., Cai, M., Wang, H., 1983–1987. The Corpus of Chinese Historical Earthquake Observations (I–V). Chinese Science and Technology Press, Beijing (in Chinese). Xie, Y., Cheng, D., Ding, X., 1989. The Uniform Magnitude Catalog of China Earthquakes (1900–1980, Ms ≥ 4.7). Earthquake Publishing House, Beijing (in Chinese). Xie, F., Wang, Z., Liu, J., 2011. Seismic Hazard and Risk Assessments for Beijing–Tianjin– Tangshan, China, area. Pure and Applied Geophysics 168, 731–738. Zuccolo, E., Vaccari, A., Peresan, A., Panza, G.F., 2011. Neo-deterministic and probabilistic seismic hazard assessments: a comparison over the Italian territory. Pure and Applied Geophysics 168, 69–83.