Seismic hazard and total risk analysis for cascade dams situated in geologically sensitive regions

doi:10.21203/rs.3.rs-1985090/v1

A modified version of a seismic risk assessment method is presented that generalises original risk analytical processes, constructs interrelationships among hazard variables in pre-assign weight-based formats, and establishes priorities to address potentially vulnerable dams. The concerned parameter values contained in each component are labelled based on adjusted grading standards, respecting the regional difference in hydraulic structure attributes and safety acceptability ranges. The cascade dams in Dadu River, China, are selected to test the applicability of the proposed method. We systematically investigate potential seismic hazards to demonstrate the implications of earthquake on dam safety. The results indicate that the adapted parameter ranges correspond with the physical aspects of the dams and potential social-economic consequences in geologically sensitive regions. Correspondingly, the hazard recognition levels accommodate their local near-source earthquake conditions. Moreover, the investigation guides local emergency services with warnings of potential dam damages and ensuing floods when encountering critical earthquake shaking.

seismic hazard assessment

total risk analysis

seismic response parameters

grading standards

dam safety management

Over the past years, the Western Sichuan plateau has experienced high levels of seismic activities, including the universally known Wenchuan Ms 8.0 Earthquake in 2008, which damaged numerous dams in the province and resulted in considerable economic losses worth $128 billion (Wu et al., 2012; Peng et al., 2014). Since then, extensive research attention has been focused on adaptation response to seismic implications from disaster risk management insights, particularly for identified hydropower bases in Sichuan Province, which is a seismic-prone region. Several studies have been conducted to investigate complex topographical and geological conditions, potentially seismic disastrous consequences, stability of hydraulic structures, and unpredictable economic crises (Vanzi et al., 2015; Oyguc et al., 2018; Jiang et al., 2022; Irinyemi et al., 2022). Although these studies primarily focused on analysing the nature of seismic ground motions and regional tectonic activities, only few of them (Hasan et al., 2007 and 2010; Srivastava et al., 2010; Hariri-Ardebili et al., 2018) have investigated seismic performance risk for dams or their facilities. This study contributes to this literature.

Seismic performance risk analysis provides the necessary information to dam safety officials and improves practical decision support related to seismic adaptation. Achieving seismic classification through the identification and determination of various risk parameters that define the security assurance level of dams is the first step in a seismic hazard analysis process (ICOLD, 2014). Many deterministic approaches, such as expert scoring and total risk analysis, have been potential candidates for representing the influence of seismic correlation under various conditions. However, these approaches have limitations (Moratto et al., 2007; Wang et al., 2014; Hariri-Ardebili et al., 2018). For example, scoring procedures are primarily dependent on the interpretation of seismic response characteristics. Moreover, experts cannot always explain their analysis of complex real-world safety assessment issues. Although total risk analysis utilises scientific investigations to measure risk-related variables, there are no uniform standards for its related decision criteria. Moreover, in geologically sensitive regions, it can lead to suspicious and misleading results, i.e. overestimate/underestimate risk values and risk-recognition levels.

Therefore, this method cannot be applied directly to analyse seismic hazards, although one possible strategy to benefit from these approaches may be to regionalise seismic response parameters according to the available damage information. In deterministic procedures, the standardised measure of the influencing factors in each component of the total risk analysis is improved. Particularly, parameter ranges that may represent geologically actives, seismic response characteristics, and dam structural stability in seismic risk ranking are modified based on the best knowledge and/or fundamental specifications of dam safety. Notably, qualitative components do not necessarily separate (seismic) hazards from (structural) vulnerability, as it is compatible with the interactions between the reliability of parameter and catchment characteristics. Subsequently, seismic risk calculation and analysis can provide valuable safety feedback to engineering practice.

Herein, we propose a modified seismic risk analysis method and apply this method to build awareness of potential risk (associated with seismic adaptation) related to cascade dams in geologically active regions. The remainder of this paper is structured to address two objectives: Section 2 introduces an overview of the study region (basin characteristics) and the related data set used for the seismic hazard analyses. Section 3 provides a background on the parameterised approach for total risk analysis (Section 3.1) and elaborates on the employed standardised measure of influencing factors and their differences in parameter ranges between the specified area and other locations (Section 3.2). Section 4 describes potential seismic hazards of different dam sites and applicable total risks under different seismic conditions. Section 5 compares seismic risks on two adjacent dam sites and characterizes possible consequences of dam failure patterns. Section 6 presents the conclusions.

2.1 Site Description

The Dadu River, located in southwestern China and stretching between 99°–104°E longitude and 28°–34N° latitude, is one of the 13 hydropower bases in China, with a feasible installed capacity of 23,480 MW (see Fig. 1). It originates from the Bayanhar Mountain seated in Qinghai province and flows approximately 1,115 km southeastward through nine autonomous regions and cities before debouching into the Min River. More specifically, it runs along complex geo-environment regions, characterized by the amalgamation of crustal blocks (e.g. Bayan Har fold belt and Chuan-Dian block) along different ages of orogenic belts during the Middle Pleistocene. Interactions among these blocks result in prominent tectonic deformation that causes several significant strike-slip faults, among which the most prominent one is the west Aba–sinistral Xianshuihe–Longmenshan–Anninghe fault system. According to the Earthquake in Sichuan Province Chi, the fault zone has experienced significant earthquakes with at least magnitudes of 4.5 or greater in historical times along the southern segment. These earthquakes include the Ms 7.9 Luhuo, Ms 7.5 Miaoxian, Ms 7.0 Lushan, and the universally known Ms 8.0 Wenchuan earthquakes, which formed a rupture length of approximately 330 km as outlined by the aftershocks. Undoubtedly, such undesirable seismicity poses significant threats to the safety of hydropower projects. As such, potential seismic risk assessment for strengthening local emergency services has become indispensable. Therefore, in this study, the cascade reservoir system in the Dadu River was investigated.

Each designed dam was located in physiographically representative regions (Fig. 1). Particularly, the Xiaerxia-, Dawei-, Bala- and Busigou dams are concentrated on the Bayan Har fold belt, where is an elongated tectonic domain and bounded by the northeast segments of the Xianshuihe fault, Aba sub-block, and the convergence zone of south Gan- and north Chuan. The Shuangjiangkou dam is located near the junction of the Aba sub-block and Longriba fault, which is an active fault zone but has been ignored in the past [Xu et al., 2008]. The Danba-, Houziyan-, Changheba-, Luding- and Yingliangbao dams are situated in the convergent hinge of the Xianshuihe fault and Longmenshan thrust. This hinge is the most significant strike-slip component along these two faults and may influence the tectonic activity of the entire Sichuan Basin. The Dagangshan-, Longtoushi- and Pubugou dams are surrounded by the Longmenshan thrust, Anninghe fault, and Chengdu Plain, which is the active segment that historical and present-day earthquakes often strike. However, these dams are still unconstructed except for the Pubugou reservoirs, they are appropriate for our study to improve the understanding of seismic risk perceptions, owing to the corresponding safety feedbacks could support the analysis of potential consequences during earthquakes.

2.2 Available data

We refer to data of various types collected from different sources. The following list briefly summarizes the dataset used for the seismic hazard analyses in this study. The quality of the datasets is firmly controlled before their release, and the relevant missing historical earthquake is compensated by selecting only the maximum earthquake magnitude.

- Earthquake catalogues: Historical and 21th century instrumentally recorded earthquakes originate from the digital library of China earthquake administration (http://data.earthquake.cn/). The dataset contains seismic hazard maps for mainland China and site-specific seismic hazard parameters, such as seismic sources, magnitudes, hypocentral distances, depths, and quake-affected areas, which represent the most accurate earthquake information covering China.

- Peak ground acceleration (PGA): PGA is retrieved from the Global Earthquake Hazard Distribution Map, which depicts the geographic distribution of the PGA with a 10-% probability of being exceeded in 50 years (http://www.ldeo.columbia.edu/chrr/research/hotspots/coredata.html) [Dilley et al., 2005]. The ranges of PGA values are classified into deciles, 10 classes of approximately equal numbers of grid cells (0.04°×0.04° or 4.5 km at the equator). Correspondingly, the attenuation equations for PGA are determined by previous studies [Lei et al., 2007; Ji et al., 2021].

- Socio-economic characteristics: This refers to the number of inhabitants and economic development in the study area. The Sichuan Statistical Bureau (SSB: http://tjj.sc.gov.cn/sctjj) provides information on the resident population and its variations in each census section. The statistical office of prefecture-level cities provides economical values for different assets, considering their regional economic structures and different classes of residential and industrial buildings and urban public utilities.

- Construction conditions for Dams: Basic engineering documents with specific information, including geological features and tectonic environments in or the vicinity of each reservoir area, were obtained from previous studies (Dadu River Hydropower Development CO., LTD, 2003). The detailed parameters of the reservoirs are presented in Table 1. The possible dam-break flood process for several representative reservoirs is obtained from informed research results [China Institute of Water Resources and Hydropower Research, 2017].

Table 1

Characteristic parameters of selected dams
Dams	Dam height (m)	Built year	Dam type	Purpose	Installed capacity (10³Kw)	Total reservoir capacity (10⁹m³)	Maximum design earthquake
Xiaerxia	223.0	u/c	RF	E + FC	540	2.800	Ⅶ
Bala	142.0	u/c	RF	E	700	0.127	Ⅷ
Dawei	107.0	u/c	RF	E	270	0.140	Ⅷ
Pusigou	133.0	u/c	RF	E	360	0.246	Ⅷ
Shuangjiangkou	270.0	u/c	RF	E + FC	1800	2.900	Ⅷ
Danba	162.0	u/c	RF	E	1300	0.684	Ⅷ
Houziyan	223.5	u/c	RF	E	1700	0.662	Ⅷ
Changheba	238.0	u/c	RF	E	2200	0.985	Ⅷ
Luding	74.0	2000	RF	E + FC	800	0.330	Ⅸ
Yingliangbao	15.0	u/c	RF	E	1300	0.021	Ⅸ
Dagangshan	210.0	u/c	CDAD	E	2600	0.742	Ⅸ
Longtoushi	72.5	2008	RF	E	640	0.115	Ⅷ
Pubugou	186.0	2010	RF	E + FC	3600	5.390	Ⅷ

“E” and “FC” represent energy and flood control, respectively; “u/c” represents the mentioned dams under construction; “RF” and “CDAD” are designed as rockfill- and concrete double-curvature arch dams, respectively.

In this section, we modify a seismic risk ranking method (Bureau et al. 2003) to analyse the anti-seismic capability and improve public awareness of potential seismic hazards. A more detailed description is presented throughout the following subsections.

3.1 Seismic risk ranking for large dams

As a preliminary indication of seismic hazard to dams, Bureau et al. (2003) introduced total risk analysis (TRF) to portray potential risk via the summation of seismic performance factors, which highlight the establishment of discriminant functions and are extensively employed by engineers for assigning priorities of critical dams. The performance formulation is represented as

$$TRF=\left( {RRF+DHF} \right) \times PDF$$

1

,

where$RRF$denotes the dam structural components of potential risk and depends mostly on the storage capacity, height, and age of the dam, (i.e. $CRF$, $HRF$and $ARF$, respectively), as given by the expression $RRF=CRF+HRF+ARF$; $DHF$represents the potential downstream consequences and can be expressed as a function of two different factors, $DHF=ERF+DRI$, where $ERF$ denotes downstream evacuation requirements and depends on the exposure of the population and $DRI$represents the downstream damage index based on the values of the private, commercial, industrial, or government properties in the potential flood path; $PDF$ denotes the “damageability” of the dams and is given by $PDF=2.5 \times PDI$, where $PDI$ represents the predicted damage index and is calculated from the geographical relationships developed by Bureau and Ballentine (2002). Notably, the defined $PDF$can identify only the potentially most critical facilities for each postulated earthquake scenario, rather than predicting the failure or non-failure of any specific dam for a given earthquake severity index $ESI$, which denotes the expected ground motion at the dam site for the scenario earthquake considered and is mostly dependent on $PGA$and the Richter or moment magnitude of the causative event, respectively, i.e. $ESI=PGA \times {(M - 4.5)^3}$. The corresponding standardized values for the above-mentioned parameters could be achieved from Bureau (2003), listed in Tables 2 and 3.

3.2 Grading standards of influencing factors

We observed that the standardized values of referring factors, which are set based on the American directionality characteristics, were not consistent with the designed dam structural and socio-economic conditions of the earthquake-prone regions in China. For example, the thresholds of the storage capacity and height for an “extreme” level were set as 6170×10⁴ m³ and 24.4 m, respectively, which are lower than the standard volume and height of a large reservoir (i.e. 1 billion m³ and 70 m) as defined by the Technical Specifications for Construction of Rolled Earth-Rock Dams. This difference would inevitably cause the synthetic risk value higher than its actual unfavourable influence. In this study, we adjusted these predominant parameters to accommodate the classification of weighting factors in the concerned regions (see Table 2). Specifically, the dam storage capacity and height factors (i.e. $CRF$and $HRF$) were redefined according to the corresponding thresholds of 1 billion, 0.1 billion, and 10 million m³ and 70, 30, and 15 m, respectively. These values were chosen because of their ability to readily represent an appropriate relationship between hydrotechnical properties and earthquake resistance in China derived from different dam classifications as per the Technical Specifications for Construction of Rolled Earth-Rock Dams.

For the downstream risk, we selected resident population density and the regional gross domestic product to support the definition of damage magnitude (i.e. $ERF$and $DRI$) as they highlight the relevance of risk perception with realistic society vulnerability features. Furthermore, an attenuation relationship is adopted to calculate the PGA acting on the dam sites owing to the unavailability of site-specific strong ground records, which relate the near-source attenuation of peak horizontal acceleration with earthquake intensities, i.e. surface wave magnitude and site-to-source distance (unit: km). Notably, this functional relationship assumes that the seismic history data contains sufficient information about the magnitudes of the previous earthquakes of the selected zone, and the deficiencies of the relevant earthquake catalogues are compensated by selecting only the maximum earthquake magnitude for any particular year [Huo et al., 1992] formulated as

$$Log{S_a}\left( T \right)={c_1}+{c_2}M+{c_3}{M^2}+{c_4}lg\left[ {R+{c_5}exp\left( {{c_6}M} \right)} \right]$$

2

,

where S_a(T) indicates the acceleration response spectrum (i.e. PGA, unit: m/s²); c₁, c₂, c₃, c₄, c_5, and c₆ are regression coefficients. For the study region, these values can be obtained from Lei et al. (2010b), i.e., the regression coefficients for the entire Sichuan province are stated as $-$1.8244, 1.5408, $-$0.0845, $-$1.6392, 0.8691, and 0.3844, respectively, with a corresponding random error of 0.232.

Table 2

Definition of dam structure influence- and downstream risk factors
Risk factors	unit	Contribution to total risk factor (weighting points)
Risk factors	unit	Extreme	high	moderate	low
Bureau’s definitions
For structure influence
Reservoir storage capacity (CRF)	10⁴m³	> 6170 (6)	6170–123 (4)	123–12.3 (2)	< 12.3 (0)
Dam height (HRF)	m	> 24.4 (6)	24.4–12.2 (4)	12.2–6.1 (2)	< 6.1 (0)
For downstream risk
Evacuation requirements (ERF)	person	> 1000 (12)	1000–100 (8)	100–1 (4)	None (1)
Downstream damage risk (DRI)	None	High (12)	Moderate (8)	low (4)	None (1)
Improved definitions for the selected dam sites
For structure influence
Reservoir storage capacity (CRF)	10⁹m³	> 1 (6)	1–0.1 (4)	0.1–0.01 (2)	< 0.01 (0)
Dam height (HRF)	m	> 70 (6)	70–30 (4)	30–15 (2)	< 15 (0)
For downstream risk
Evacuation requirements (ERF)	person	> 500 (12)	500–100 (8)	100–10 (4)	< 10 (1)
Downstream damage risk (DRI)	10⁹RMB	> 10 (12)	10–4 (8)	4–1 (4)	< 1 (1)

Table 3

Risk evaluation factors of aging dams
Dam age/ Year	< 1900	1900–1925	1925–1950	1950–1975	1975–2000	> 2000
ARF	6	5	4	3	2	1

4.1 Preliminary indication of seismic hazard of dams

Possible seismic sources were identified based on the historical records of earthquake hazards and the investigated active tectonics, as shown in Figs. 1 and 2a. Notably, the spatial distribution of seismic hazards exhibits a stronger response to their surrounding active fault features, which can confidently be viewed as “background” seismic sources. Many historical earthquakes with magnitude greater than 6 have occurred within less than 100 km from large reservoirs, illustrating that “potential influences” on dam sites are concentrated in the upper and middle reaches of the Dadu River as their general reliance on the complex environment. Correspondingly, the distribution of the PGA curves for the entire basin is presented in Fig. 2b. Remarkably, the structures here are under the influence of potential seismic destruction with high vulnerability, although the map displayed the basin profiles from 1978 to 2016. This is because, within a wide range, the PGA values lie between Ⅴ to Ⅸ, which are equivalent to their values from 0.10 to 0.60 g. Such undesirable values are consistent with the general investigations i.e. high levels of seismicity occur at the middle reaches.

Table 4 shows the overall classification of the dam sites. This classification provides a preliminary indication for seismic hazards based on the International Commission on Large Dam (ICOLD) approach (2014). As shown, these sites are subjected to relatively high earthquake loads with magnitudes of Ⅶ, Ⅷ, or Ⅸ. Moreover, critical zones surround these subject dams distributed only within 40 km because active faults near the sites have negative implications on the stability of their hydraulic structures. Correspondingly, four dams (i.e. the Xiaerxia-, Dawei-, Pusigou- and Shuangjiangkou dams) exhibit smaller PGA values (i.e. less than or equal to 0.10 g). These four dams are classified into hazard class II with moderate hazard, which is consistent with the anticipated level influence of a near-source zone and the inherent design requirement. For the Bala-, Danba-, and Pubugou dams, the mean PGA values are within a wide range (0.15–0.20 g). These dams correspond to hazard class II with moderate hazard, indicating that these detrimental values exhibit a relatively weak influence on the mentioned seismic activity. The remaining six dams exhibit higher PGA values (i.e. more than or equal to 0.25 g). Particularly, the Luding- and Yingliangbao dams are subjected to PGA of 0.35 and 0.40 g, respectively, with active faults closer than 10 km from the site, representing the highest damage or failure risk under earthquake loading, and they are classified into hazard class Ⅳ with an extreme-risk rating. Compared with the Luding- and Yingliangbao dams, the Houziyan-, Changheba-, Dagangshan- and Longtoushi dams have high-risk ratings under a similar distribution as those of the critical zones because they are located relatively far away from the heart of the epicentre region.

Table 4

Seismic hazard class and rating for selected dam sites
Dams	Seismic intensity at dam site	PGA (g)	Epicentral distance (km)	Hazard class	Hazard rating
Xiaerxia	Ⅶ	0.10	32	I	Low
Bala	Ⅶ	0.15	32	II	Moderate
Dawei	Ⅶ	0.10	< 40	I	Low
Pusigou	Ⅶ	0.10	< 40	I	Low
Shuangjiangkou	Ⅷ	0.10	32	I	Low
Danba	Ⅵ	0.15	< 40	II	Moderate
Houziyan	Ⅶ	0.20	< 40	III	Moderate
Changheba	Ⅷ	0.25	< 40	III	High
Luding	Ⅷ	0.35	< 40	Ⅳ	Extreme
Yingliangbao	Ⅸ	0.40	< 40	Ⅳ	Extreme
Dagangshan	Ⅷ	0.30	25	III	High
Longtoushi	Ⅷ	0.25	< 40	III	High
Pubugou	Ⅶ	0.15	< 40	II	Moderate
PGA values at each dam site are determined from seismic zonation map for mainland China (see Fig. 2b).

4.2 Potential seismic hazard under recorded maximum historic earthquake

Table 5 describes the results of potential seismic risk analysis for each selected dam according to the recorded maximum historic earthquake near the dam sites (M_site). Notably, the obtained risk rating for each dam accommodates their local near-source earthquake conditions, i.e. the calculated TRF values of most of the dams are identified approximately as moderate-level (i.e. the TRF lies between 25 and 125). However, the Shuangjiangkou- and Luding dams have high-level risk ratings with TRF values great than 125, illustrating that both dams show potential damage or failure risk when subjected to the loading of a relatively higher magnitude earthquake. Moreover, the seismic influence indexes for the Luding- and Yingliangbao dams are relatively greater (i.e. PGA ranges from 0.350 to 0.420g), implying a multiple-hazard threat to the safety of the dam systems and their downstream areas because the recorded maximum historic earthquake magnitudes (M_site = 7.5) for both dam sites have shorter hypo-central distances (i.e. 9.36 and 4.89 km, respectively). For weight point characteristics, the defined risk factors exhibit maximum values for the mentioned key water-control dam sites, demonstrating negative implications on the safety of high dams or large reservoir structures and the functionality of their downstream socio-economic activities. Moreover, for the Houziyan-, Changheba-, and Luding dams, the social parameters (i.e. ERF and DRI) present relatively greater values i.e. 12 (extreme) and 8 (high), respectively, showing the natural attribute of the vulnerability of their downstream society, especially at such densely developed areas.

Table 5

Estimation of seismic hazard for different dam sites based on recorded maximum historic earthquake (M_site)
Dams	Site parameters			Influence indexes			Weighting points for risk factors					TRF	DR
Dams	M_site	R	PGA	ESI	PDI	PDF	CRF	HRF	ARF	ERF	DRI	TRF	DR
Xiaerxia	5.4	37.42	0.024	0.100	0.997	2.493	6	6	1	12	12	92.223	Ⅱ
Bala	5.4	26.62	0.038	0.100	0.997	2.493	4	6	1	2	4	42.373	Ⅱ
Dawei	5.4	31.22	0.031	0.100	0.997	2.493	4	6	1	2	4	42.373	Ⅱ
Pusigou	6.0	39.71	0.042	0.141	1.010	2.525	4	6	1	2	2	37.875	Ⅱ
SJK	6.0	33.19	0.051	0.172	1.382	3.455	6	6	1	12	12	127.835	Ⅲ
Danba	6.0	36.70	0.045	0.151	1.359	3.398	4	4	1	4	8	71.348	Ⅱ
Houziyan	6.5	31.09	0.086	0.100	0.997	2.493	4	6	1	12	8	77.268	Ⅱ
Changheba	6.5	28.24	0.096	0.100	0.997	2.493	4	6	1	12	12	87.238	Ⅱ
Luding	7.5	9.36	0.350	9.442	1.711	4.278	6	6	1	12	8	141.158	Ⅲ
Yingliangbao	7.5	4.89	0.420	11.341	1.747	4.368	4	4	1	4	8	91.718	Ⅱ
Dagangshan	6.2	40.98	0.047	0.100	0.997	2.493	4	6	1	8	8	67.298	Ⅱ
Longtoushi	6.0	8.52	0.205	0.100	0.997	2.493	4	6	1	8	12	77.268	Ⅱ
Pubugou	6.0	49.23	0.030	0.101	0.998	2.495	6	6	1	12	12	92.315	Ⅱ
“SKJ” represents Shuangjiangkou reservoir; “DR” represents Dam risk rating; ‘R’ indicates site-to-source distance with a unit of “kilometre”.

4.3 Potential seismic hazard under sympathetic faults distribution

Table 6 illustrates the detailed analysis results for each selected dam under the consideration of a regional seismological background and sympathetic fault distribution. Here, the PDF values for the dam sites with respect to the wide PGA range (0.01 and 0.9 g) are relatively high (i.e. the upper bound of the PDF values changes from 3.088 to 4.808). Additionally, the synthetic risk classification is closely correlated to the “damageability” characteristics of the selected dams, i.e. the hazard recognition levels for the Shuangjiangkou-, Changheba-, Luding-, Longtoushi-, and Pubugou dams were identified into class II or III (a moderate and high-risk rating). Notably, the intervals of the PGA values for the different dams presented the variation of a source-to-site distance to more precisely display the dam structural response. Subsequently, the upper and lower bounds were obtained according to the minimum and maximum distances between the dam sites and the earthquake source on the disaster-affected faults (see Fig. 1). Interestingly, the upper bound of the PGA at the Shuangjiangkou dam reached 0.9 g owing to this critical area and was extremely close to the active faults (only 0.64 km), which could cause damaging displacement of the structure. Additionally, the weighting points for each risk factor exhibited the same values as that of the seismic hazard analysis at the dam site presented in Table 5, demonstrating negative implications on the safety of the high dams or large reservoir structures and the functionality of their downstream socio-economic activities. However, these weighting points could be improved if the assessment schemes were analysed by employing out-of-date criteria, which inevitably presented unacceptably overestimated TRF values and recognition risk levels of these structures, although strengthening community resilience to seismic-related disaster prevention and protection capabilities is more profitable. Such features also imply that these high dams and large reservoirs are in challenging circumstances associated with the actual cumulative effect of multiple hazards. Therefore, the government should focus on and provide practical support to promote more effective precautionary measures.

Table 6

Qualitative estimation of the potential seismic hazard for different dams based on the spatial distribution of the disaster-affected faults
Dams	PGA	PDF	Weighting points for risk factors					TRF	DR
Dams	PGA	PDF	CRF	HRF	ARF	DRI	ERF	TRF	DR
Xiaerxia	0.01 ~ 0.04	2.471 ~ 3.228	6	6	1	12	12	91.43 ~ 119.44	Ⅱ
Bala	0.01 ~ 0.06	2.471 ~ 3.176	4	6	1	2	4	42.01 ~ 53.99	Ⅱ
Dawei	0.01 ~ 0.06	2.471 ~ 3.088	4	6	1	2	4	42.01 ~ 52.50	Ⅱ
Pusigou	0.01 ~ 0.10	2.471 ~ 3.550	4	6	1	2	2	37.07 ~ 53.25	Ⅱ
Shuangjiangkou	0.01 ~ 0.90	2.485 ~ 4.808	6	6	1	12	12	91.95 ~ 177.9	Ⅱ~Ⅲ
Danba	0.01 ~ 0.15	2.478 ~ 3.340	4	4	1	4	8	52.04 ~ 70.14	Ⅱ
Houziyan	0.01 ~ 0.25	2.480 ~ 3.496	4	6	1	12	8	76.88 ~ 108.38	Ⅱ
Changheba	0.01 ~ 0.30	2.488 ~ 3.734	4	6	1	12	12	87.08 ~ 130.69	Ⅱ~Ⅲ
Luding	0.01 ~ 0.35	2.515 ~ 4.488	6	6	1	12	8	83.00 ~ 148.10	Ⅱ~Ⅲ
Yingliangbao	0.01 ~ 0.45	2.515 ~ 4.663	4	4	1	4	8	52.82 ~ 97.92	Ⅱ
Dagangshan	0.03 ~ 0.14	2.510 ~ 3.348	4	6	1	8	8	67.77 ~ 90.40	Ⅱ
Longtoushi	0.04 ~ 0.27	2.488 ~ 3.515	4	6	1	8	12	77.13 ~ 108.97	Ⅱ~Ⅲ
Pubugou	0.05 ~ 0.12	2.493 ~ 3.451	6	6	1	12	12	92.24 ~ 127.69	Ⅱ~Ⅲ

The aforementioned analysis highlights deterministic safety evaluations from an overall seismic performance perspective with strong consideration given to regional geological features and historic earthquake information. For a comprehensive understanding of earthquake effects on the dam safety threat, seismic risk probabilities and potential consequences of dam failure pattern are objectively discussed in the following sections.

5.1 Probability of seismic risk on two adjacent dam sites

The probability of seismic risk on two adjacent dam sites is summarized using a histogram, as shown in Fig. 3 (based on Lei et al., 2010a in Appendix A). Notably, only a single bar was shown at the Xiaerxia dam site because no data was available for its upstream segment risk. Moreover, the risk probabilities of different river segments and each dam site exhibited similar alteration characteristics, i.e., the fluctuations of these two series were consistent. However, only the corresponding fluctuation degrees varied with locations. This agreed with realistic arrangements of seismic risk influence features. Specifically, the river segment risk between each adjacent pair of dams featured relatively high probability values, particularly between the Dagangshan- and Longtoushi dams (i.e. the value reached 2.29×10^− 4). Conversely, the ultrahigh river segment risk probability corresponded to a higher-level hazard, demonstrating that more attention should be given to possible hydraulic-structure destruction when dams encounter massive earthquakes, as they would inevitably result in significant damage to the downstream areas. Importantly, a noticeable polarity (with a value quite close to 0) was presented for the estimated segment risk performance between the Shuangjiangkou- and Danba dams, which represented a stronger aseismic capacity because their relatively higher seismic for quantification parameters posed relative smaller exceedance probabilities. Moreover, the seismic risk probability series on each dam site were relatively high, i.e., most of them were generally beyond their corresponding design level (i.e. 2×10^− 4). However, the values at the Xiaerxia-, Shuangjiangkou-, and Pubugou dams were steady within acceptable safety levels (i.e. 2×10^− 4, 1.25×10^− 4, and 1.5×10^− 4, respectively), which could function as “risk diversification” and solve the contradiction of dam safety requirements and potential seismic implications.

5.2 Potential consequences of dam failure pattern

Herein, we qualitatively characterise possible consequences of dam failure patterns, rather than a specific physical process. Therefore, a comparison between the potential dam-break flood volume and the available detention zone by Zhang et al. (2016) was performed, with details in Appendix B. Table 7 illustrates the overall evaluation results according to the assumption that a reservoir encounters an unexpected dam breach (under 2/3 of the corresponding reservoir storage capacity). Notably, we ignored three dams, Bala, Dawei, and Pusigou, in high seismic sensitivity zones. These three dams are insignificant because their total reservoir capacities are considerably smaller than those of the upstream Xiaerxia and downstream Shuangjiangkou. Table 7 shows that a series of dams are respectively described as “O-topping” when the Shuangjiangkou- and Changheba dam failures occur, suggesting that the downstream reservoir cannot maintain its security for the control of unexpected floods and the occurrence of overtopping events. This can subsequently result in considerable disasters in human lives and properties downstream. However, such features agree relatively well with their pre-concerted design purpose, particularly for the Shuangjiangkou- and Pubugou dams, which could be considered as a ‘buffer’ for eliminating or alleviating the secondary flood-hazard tension induced by earthquakes. Notably, the security states of the Houziyan- and its surrounding dam were expressed in two different situations by considering their similar storage capacities. These situations included ‘C-damage’ at the Changheba dam and a chain of ‘O-topping’ for the selected dams until risk termination at the Pubugou dam. By comparing those features for the Xiaerxia- and Shuangjiangkou dam groups, the limited water diversion capacities of the Houziyan-, Changheba- and their surrounding dams would pose a minor water release and relatively fewer adverse consequences. This implies that the relative contribution of failure mechanisms to social vulnerability in their downstream areas is also relatively weaker. However, the potential destructiveness of uncontrolled extreme water release cannot be ignored.

Table 7

Qualitative estimation of the possible dam risk pattern for different dams, when the earthquake-induced dam break event happens
Dams	Xiaerxia	Shuangjiangkou	Danba	Houziyan	Changheba	Luding	Liangyingbao	Dagangshan	Longtoushi	Pubugou
Xiaerxia	N/A	C-damage
Shuangjiangkou		N/A	O-topping	O-topping	O-topping	O-topping	O-topping	O-topping	O-topping	C-damage
Danba			N/A	O-topping	Ok
Houziyan				N/A	C-damage
Changheba					N/A	O-topping	O-topping	O-topping	O-topping	Ok
Luding						N/A	O-topping	Ok
Liangyingbao							N/A	Ok
Dagangshan								N/A	O-topping	Ok
Longtoushi									N/A	Ok
Pubugou										N/A

‘N/A’ represents the dam break events happening under 2/3 of the corresponding reservoir storage capacity; ‘O-topping’ represents the case wherein the reservoir cannot maintain its security for the unexpected floods control and occurrence of overtopping events, which inevitably result in the destruction of hydraulic structure and also cause serious chain reaction; “C-damage” represents case wherein the reservoir can capture and manage the unexpected floods although it can inevitably influence the downstream flood levee system; ‘Ok’ represents the case wherein the reservoir can effectively utilize its available detention volume to maximize the safety of downstream areas, equivalent to ‘risk termination’.

In this study, we systematically investigated the seismic performance risk of the cascade dams in the Dadu River. A modified version of a seismic risk ranking method was described that suggested dam safety management priorities through synthetic risk, which enabled simultaneous consideration of the physical conditions of dam structures and potential social-economic consequences. In the proposed method, predominant risk parameters were adjusted to appropriately describe the weight proportion of hazard variable impacts, thereby increasing the chances of achieving a reliable total risk and facilitating hazard perception in engineering practices. This was because the specified and valued parameters of the original ICOLD approach were lower than their corresponding Chinese directionality characteristics in earthquake-prone regions. Then, we demonstrated the application of the proposed approach by analysing the influence of potential seismic sources in fragile geological regions and evaluating seismic hazards under two representative situations, i.e. the recorded maximum historic earthquake and spatial distribution of sympathetic faults. We also discussed the potential consequences of dam failure patterns.

Overall, the proposed parametric method could efficiently investigate potential seismic hazards in the concerned regions, and the corresponding hazard recognition levels for the different dams were consistent with their surrounding seismic source features. The key results are as follows: Relatively high-level risk ratings occurred at the Shuangjiangkou-, Changheba-, Luding-, Longtoushi-, and Pubugou dams, with high PGA values and weighting points, which indicate the significant negative effects that could occur on their downstream areas when encountering critical tremors due to earthquakes. The rest of the dams exhibited relatively low dam structural and social parameters and were identified as moderate-level seismic risk (correspond to a relatively weak influence). Nonetheless, considerable attention should be given to the stabilities of their hydraulic structures because this could inevitably cause a secondary flood-hazard tension and pose an additional risk to the cascade reservoirs. In addition, the pre-defined parameters accommodated the risk problem encountered by different reservoirs near the seismic sources and thus provided effective damage magnitudes for avoiding unexpected overestimation.

Additionally, earthquake-induced dam damage was represented via seismic risk probabilities and possible dam failure patterns. The results indicated a strong consistency between the hazard recognition levels and seismic risk probability series on each dam site. Moreover, their corresponding probabilities were generally beyond a defined acceptable value, which corresponded to a relatively high hazard level. Further, we demonstrated the key role of secondary flood-hazard tension, especially when the Xiaerxia-, Shuangjiangkou-, Houziyan- and Changheba dam failures occurred. Moreover, significant changes in reservoir release may cause damage to the downstream channels and turbines. Although we demonstrated a rough analysis process for qualitatively characterizing the possible dam failure patterns, their delineation was consistent with the actual flood risk perceptions of reservoirs. The ability to associate potential safety problems with reservoir operational performance may be an area for future study. Future research will explore a more systematic analysis for dealing with prominent features of threatening scenarios and integrating them with reservoir-protection goals from a precautionary perspective.

CRediT authorship contribution statement

Rui Zhang: Methodology, Writing - Original draft, Writing - review & editing, Funding acquisition.

Bende Wang: Conceptualization, Formal analysis, Data curation, Supervision, Funding acquisition.

Declaration of Competing Interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data Availability Statement

Users can access the data used in this study by contacting the corresponding author and data archiving is underway though GitHub.

Acknowledgements

This work was supported by the Second Tibetan Plateau Scientific Expedition and Research program [grant number 2019QZKK0903], and the National Key Research and Development Program of China [grant number 2018YFC0407705]. The first author was supported by a fellowship from the Chinese government for her visit to Texas A & M University. The authors are also grateful to the Institute of Mountain Hazards and Environment and Dalian University of Technology for providing the related resources. The remarks provided by the editor and associate editor are also greatly appreciated.

Bureau and G, Ballentine, GD. A comprehensive seismic vulnerability and loss assessment of the state of South Carolina using HAZUS—part vi. Dam inventory and vulnerability assessment methodology. In: 7th US national conference on earthquake engineering, Boston, MA, 21–25 July 2002, pp.1943–1953. Oakland, CA: Earthquake Engineering Research Institute.
Bureau, G.J. Dams and appurtenant facilities. In Dams and Appurtenant Facilities in Earthquake Engineering Handbook; Chenh, W.F., Scawthorn, C., Eds.; CRS Press: Boca Raton, FL, USA, 2003; pp. 26.1–26.47.
China Institute of Water Resources and Hydropower Research, 2017. A Final Report on the Research of Risk Assessment and Prevention in Cascade Reservoirs. Tech. Rep. National Natural Science Foundation of China (in Chinese).
Dadu River Hydropower Development CO., LTD., 2003. A generalized hydropower bases: Reservoirs’ planning & design statements in Dadu River, Tech. Rep. Ministry of Water Resources, China (in Chinese).
Dilley, M., R.S. Chen, U. Deichmann, et al. 2005. Natural Disaster Hotspots: A Global Risk Analysis v1.0. Washington D.C.: The World Bank. https://doi.org/10.7916/D8NC68P8.
Hasan T., İsmail Z., Ahmet O. et al. 2007. Seismic hazard and total risk analyses for large dams in Euphrates basin, Turkey. Eng. Geol., 89 (1-2), 155-170. https://doi.org/10.1016/j.enggeo.2006.10.003
Hasan T., Seyrek E., 2010. Total risk analyses for large dams in Kizilirmak basin. Turkey Nat Hazards Earth Syst Sci 10(5):979–987. https://doi.org/10.5194/nhess-10-979-2010.
Hariri-Ardebili MA, Nuss LK, 2018. Seismic risk prioritization of a large portfolio of dams: Revisited. Adv Mech Eng 10(9):1–20. https://doi.org/10.1177/1687814018802531.
ICOLD, 2014. Selecting seismic parameters for large dams, guidelines, bulletin 148, Committee on Seismic Aspects of Dam Design, International Commission on Large Dams (ICOLD), Paris
Irinyemi, S.A., Lombardi, D., Ahmad, S.M., 2022. Seismic risk analysis for large dams in West Coast basin, southern Ghana. J Seismol 26, 101–116. https://doi.org/10.1007/s10950-021-10045-w.
Ji, K., Wen, R., Ren, Y. et al., 2021. Disaggregation of probabilistic seismic hazard and construction of conditional spectrum for China. Bull Earthquake Eng 19, 5769–5789. https://doi.org/10.1007/s10518-021-01200-2.
Jiang Y., Zhang R., Wang B., 2022. Scenario-based approach for emergency operational response: Implications for reservoir management decisions. Int. J. Disaster Risk Reduct, 10, 2212-4209. https://doi.org/10.1016/j.ijdrr.2022.103192.
Lei J., Gao M., Wu J., Kang C., 2010a. A bi-site seismic hazard analysis method and its application. Acta Seismologica Sinica, 32 (3): 310-319. https://dx.doi.org/10.3969/j.issn.0253-3782.2010.03.006.
Lei J., Gao M., Lu H., Kang C., 2010b. Effects of seismic belt division scheme on seismicity parameters. Acta Seismologica Sinica, 32 (4): 457-465. https://dx.doi.org/10.3969/j.issn.0253-3782.2010.04.008.
L. Moratto, B. Orlecka-Sikora, G. Costa, et al., 2007. A deterministic seismic hazard analysis for shallow earthquakes in Greece, Tectonophysics, 442 (1–4), 66-82. https://doi.org/10.1016/j.tecto.2007.05.004.
Oyguc, R., Toros, C., & Abdelnaby, A. E., 2018. Seismic behavior of irregular reinforced-concrete structures under multiple earthquake excitations. Soil Dyn. Earthq. Eng, 104, 15-32. https://doi.org/10.1016/j.soildyn.2017.10.002.
Peng, M., Zhang, L.M., Chang, D.S., Shi, Z.M., 2014. Engineering risk mitigation measures for the landslide dams induced by the 2008 Wenchuan earthquake. Eng. Geol. 180, 68–84. https://doi.org/10.1016/j.enggeo.2014.03.016.
Srivastava, A., Babu, G.L.S., 2010. Total risk rating and stability analysis of embankment dams in the Kachchh Region, Gujarat, India. Eng. Geol., 115(1-2), 68–79. https://doi.org/10.1016/j.enggeo.2010.06.011
Vanzi, I., Marano, G.-C., Monti, G., & Nuti, C., 2015. A synthetic formulation for the Italian seismic hazard and code implications for the seismic risk. Soil Dyn. Earthq. Eng, 77, 111-122. https://doi.org/10.1016/j.soildyn.2015.05.001.
Wu, J., Li, N., Hallegatte, S. et al 2012. Regional indirect economic impact evaluation of the 2008 Wenchuan Earthquake. Environ Earth Sci 65, 161-172.https://doi.org/10.1007/s12665-011-1078-9.
Wang J., Wu Y., 2014. A new seismic hazard analysis using FOSM algorithms, Soil Dyn. Earthq. Eng, 67, 251-256. https://doi.org/10.1016/j.soildyn.2014.10.006.
Water Resources Ministry of China, 2020. Technical specifications for construction of rolled earth-rock dams, Beijing. http://zwgk.mwr.gov.cn/jsp/yishenqing/appladd/biaozhunfile/flash/pdf/20210316145921154.pdf.
Xu X., Wen X., Chen G., Yu G., 2008. Discovery of the Longriba Fault Zone in Eastern Bayan Har Block, China and its tectonic implication. Sci China Ser D-Earth Sci, 51 (9), 1209-1223. https://doi.org/10.1007/s11430-008-0097-1.
Zhang R., 2016. Analysis of the overtopping failure probability in cascade reservoirs and its emergency response strategies. PhD Thesis, Dalian University of Technology, China (in Chinese)

No competing interests reported.

Supplementary.docx

Seismic hazard and total risk analysis for cascade dams situated in geologically sensitive regions

Status:

Journal Publication

Version 1

Abstract

Figures

1 Introduction

2 Study Area And Datasets

2.1 Site Description

2.2 Available data

3 Methodology

3.1 Seismic risk ranking for large dams

3.2 Grading standards of influencing factors

4 Results

4.1 Preliminary indication of seismic hazard of dams

4.2 Potential seismic hazard under recorded maximum historic earthquake

4.3 Potential seismic hazard under sympathetic faults distribution

5 Discussion

5.1 Probability of seismic risk on two adjacent dam sites

5.2 Potential consequences of dam failure pattern

6 Conclusion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1