Earthquake Risk Assessment in Seismically Active Areas of Qinghai Province Based on Geographic Big Data

Abstract

Earthquakes can cause serious damage to buildings, roads and other infrastructure. The large amount of dust and particulate matter generated when these structures collapse and are damaged can quickly enter the air, leading to a decline in air quality. At the same time, earthquakes may cause secondary disasters such as fires and landslides, which will also produce large amounts of soot and particulate matter, which will have a negative impact on air quality. Therefore, earthquake disaster risk assessment studies are carried out to identify potentially hazardous areas and facilities in advance in order to reduce the air pollution problems that may be caused by earthquakes. Existing research on earthquake disaster risk assessment mainly evaluates earthquake risk from the perspective of geology or seismology, but there are few studies based on multidisciplinary assessment that integrates geology, seismology, engineering and social sciences into socioeconomic factors. To this end, based on remote sensing and GIS technology, this paper takes Qinghai Province, a seismically active area, as the research area, and integrates land use data, natural environment data, social environment data and seismic parameter zoning data to construct a comprehensive assessment model for earthquake disaster vulnerability and risk. The results showed that there were 5 very high-risk areas, 7 high-risk areas, 10 medium-risk areas, 11 low-risk areas and 12 very low-risk areas in Qinghai Province. The high-risk areas are mainly distributed in the central and western parts of Qinghai Province, where the earthquake breeding environment is sufficient, the scale of active faults is huge and the adaptability of the carrier is low. The results of an earthquake disaster risk assessment can provide a reference for the government to formulate environmental protection policies. According to the assessment results, the government can formulate targeted measures to strengthen air pollution control and improve air quality.

1. Introduction

Earthquakes are one of the three main natural disasters in the world [1,2,3]; in recent years, earthquakes and major disasters have occurred frequently across the world, causing countless losses of life and property [4,5,6,7,8,9]. The world’s highest magnitude earthquake was recorded in northern Chile about 3800 years ago, with a magnitude of 9.5 and severe damage to the coastline [10]. Major earthquake events such as the Tangshan earthquake in China caused heavy casualties and economic losses [11,12,13].

At present, people cannot accurately predict the occurrence of short-term earthquakes, but can only focus on prevention [14,15]. Therefore, an earthquake risk assessment should be carried out before the earthquake to minimize the various disaster losses caused by the earthquake. At the same time, carrying out earthquake disaster risk assessment research is not only the basis for national and regional medium- and long-term planning and development, but also an important decision-making basis for earthquake emergencies.

At first, some scholars focused on the traditional model of seismic hazard assessment [16,17,18], such as the probabilistic seismic hazard analysis (PSHA) model, which, as a traditional means of seismic hazard assessment has been used by researchers for nearly 50 years [19,20,21]. Rahman et al. [22] conducted a probabilistic seismic hazard analysis for Bangladesh using background seismicity, crustal faults and subduction zone hypocenter models. However, the use of probabilistic seismic hazard assessment models creates uncertainties in magnitude, location and earthquake recurrence rate [23]. The main contribution of the deterministic seismic hazard analysis (DSHA) method lies in the interpretation of the complexity of seismotectonic zones [24]. Kolathayar et al. [25] used the deterministic assessment method of seismic activity data to assess and analyze the seismic hazard in India. The DSHA model involves some scenario assumptions which may create uncertainty [26]. Subsequently, machine learning techniques became effective and prominent in the field of seismology [27,28]. Commonly used earthquake disaster prediction methods include support vector machine (SVM), artificial neural network (ANN) and convolutional neural network (CNN) [29,30,31]. For example, Chanda et al. [32] used machine learning methods to estimate the magnitude and location of earthquakes. Jena et al. [33] used ANN and analytical hierarchy process (AHP) comprehensive models to study the possible risks of earthquakes to urban populations, and successfully applied the model to the Banda Aceh region of Indonesia. Nowadays, remote sensing and GIS technology are applied in the study of earthquake disaster risk assessment. For example, Omar et al. [34] combined remote sensing and GIS technology to conduct earthquake disaster assessment and analysis in Southwestern Sinai. Sahar et al. [35] used remote sensing images and GIS data to automatically extract buildings and prepare data for urban earthquake risk assessment. Existing studies have shown that the AHP method is one of the commonly used techniques for earthquake disaster and its risk assessment [36,37,38]. For example, Malakar et al. [39] combined the AHP, entropy and ANN comprehensive methods to assess the seismic risk in the Himalayas. Fariza et al. [40] used the AHP method to study the earthquake disaster risk zoning of East Java, Indonesia.

Existing research on earthquake disaster risk assessment mainly assesses earthquake risk from the perspective of geology or seismology, but ignores the multidisciplinary assessment that combines geology, seismology, engineering and social sciences, and does not consider socioeconomic factors such as regional population density, economic level and infrastructure development. These factors directly affect the degree of impact of the earthquake on human society and the ability to recover after the disaster. In the seismic risk assessment model, based on the sensitivity analysis, the weight and influence degree of different social and economic factors in the seismic risk assessment are evaluated. However, a multidisciplinary approach combined with socioeconomic factors can help to fully understand the potential risk and vulnerability of earthquake disasters, and then determine the earthquake-affected areas and formulate targeted disaster prevention and mitigation strategies. In addition, Qinghai Province is located in the Qinghai–Tibet Plateau seismic belt (the largest seismic belt in China), where the geological structure is complex, the scale of active faults is huge, the earthquake breeding environment is sufficient and the region shows the largest and strongest seismic activity. Therefore, in order to quickly and effectively carry out emergency rescue in the event of an earthquake and minimize casualties and property losses, it is particularly important to carry out earthquake risk assessment research in Qinghai Province.

Therefore, this paper takes Qinghai Province, a seismotectonic active area, as the research area, and based on remote sensing and GIS technology, integrates land use data, terrain and other natural environment data to construct an exposure model of earthquake-stricken areas, providing important data support for regional-scale earthquake disaster risk assessment. At the same time, social environmental data are introduced to construct the calculation factors of seismic carrier sensitivity and adaptability, and the risk grade and spatial pattern distribution of earthquake disaster vulnerability are obtained quantitatively, so as to accurately evaluate the earthquake risk. The research results can provide national services for earthquake disaster risk prevention and control and earthquake science and technology, and have important practical significance in urban planning, earthquake risk management, reducing economic losses and reducing casualties.

2. Materials

2.1. Study Area

Qinghai Province is located in the inland northwest of China (31°39′~39°19′ N, 89°35′~103°04′ E), with a total area of about 722,300 km², most of which is located at high altitudes, and the average altitude of the province is more than 3 km [41]. The climate of Qinghai Province is characterized by being dry and windy, cold and hypoxic, and with strong solar radiation. The geographical location of the study area is shown in Figure 1.

Qinghai Province is located in the northeast of the Qinghai–Tibet tectonic block, with strong tectonic movements and huge active faults. Qinghai is mainly dominated by the famous East Kunlun active fault belt, the Altun Mountain active fault belt and the Qilian Mountain active fault belt, which is the main place for stress release and adjustment during crustal deformation. All kinds of secondary active fault belts are spread throughout almost all of Qinghai Province, which have a strong influence on the tectonic movement of Qinghai Province and the surrounding provinces, forming seismic active belts. Yushu, Qinghai Province, is located at the junction of the Asian plate and the Tibet plate. On 14 April 2010, a 7.1-magnitude earthquake occurred in Yushu, Qinghai Province, with a maximum intensity of 9 degrees. The focal depth was 33 km, the affected area was about 35,800 square kilometers, including 4000 square kilometers in the worst-hit areas and 1000 square kilometers in the worst-hit areas, and 270 people are still missing, with nearly 250,000 people having been affected and direct economic losses reaching more than 61 billion yuan [42]. With the severe global earthquake situation, the uncertainty of earthquake disaster risk increases. In order to better grasp the earthquake situation and earthquake disaster risk situation of Qinghai Province, it is necessary to carry out earthquake disaster risk assessment research in Qinghai Province.

2.2. Data Source

The dataset used in this study includes four types, namely land use data, natural environment data, social environment data and peak ground acceleration data. (1) Land use data include full-element land cover data and the single-element water coefficient covering the study area. They are mainly used to calculate the exposure value in the subsequent earthquake vulnerability assessment model. Among them, the spatial resolution of the full-element land cover data European Space Agency (ESA) world cover (https://zenodo.org/record/5571936, accessed on 9 February 2023) is 10 m, and the dataset contains 10 first-class land cover types. The single-element water system data come from the National Basic Geographic Information Center (https://www.webmap.cn/commres.do?method=dataDownload, accessed on 13 February 2023). (2) Natural environment data include digital elevation data, slope data, fault zone data and earthquake point data. They are mainly used to calculate the exposure value in the subsequent earthquake vulnerability assessment model. Among them, the digital elevation data come from the USGS Earth Resources Observation and Science (EROS) Center, with a spatial resolution of 30 m, and the corresponding slope data are calculated based on the digital elevation data. The fault zone and earthquake point data come from the National Earthquake Science Data Center (https://data.earthquake.cn/, accessed on 10 February 2023). (3) The social environment data come from China’s 7th (2020) census data and the 2020 statistical yearbook issued by the statistics bureaus of counties in Qinghai Province, which provide the most detailed and accurate current population, GDP and other related data. They are mainly used to calculate the sensitivity and adaptability values in the subsequent earthquake vulnerability assessment model. (4) The peak acceleration data of the earthquake motion are derived from the earthquake motion parameter values in the “China Earthquake Parameter Zoning Map” (GB18306-2015) (https://data.earthquake.cn/, accessed on 12 February 2023). They are mainly used for subsequent earthquake hazard assessment. Figure 2 shows the land use data and natural environment data used in this paper.

3. Methods

Based on geographic big data and combined with GIS technology, this paper conducts a comprehensive assessment of earthquake disaster risk in Qinghai Province. The overall technical process is shown in Figure 3. Firstly, the earthquake hazard assessment results are obtained based on the peak acceleration of ground motion. Secondly, the earthquake vulnerability assessment results are obtained based on land use data, natural environment data and social environment statistics. Finally, earthquake risk is evaluated based on earthquake hazard and earthquake vulnerability.

3.1. Hazard Assessment Method

Earthquake hazard is the possibility of the frequency and magnitude of earthquakes in a certain research area in the future. The magnitude of earthquakes in the same area is different, and the seismic hazard is also different. In this study, the earthquake risk evaluation basis is based on the earthquake motion parameter values in the “Zoning Map of Earthquake Motion Parameters in China”. The specific steps are as follows: Firstly, the peak acceleration of the earthquake in the county area is obtained. The peak acceleration of ground motion is an important index to measure the strength of earthquake action. The “Zoning Map of Earthquake Motion Parameters in China” gives the peak acceleration of ground motions under the basic acceleration response spectrum characteristic period value of Class II site conditions in the township governments of various provinces and cities above the county level in China. Secondly, obtain the county-level intensity. The “Zoning Map of Earthquake Motion Parameters in China” gives a comparison table of peak acceleration and intensity of ground motion for Class II sites according to the peak acceleration of earthquake motion (

Table 1. Comparison table of peak ground acceleration, seismic intensity and seismic hazard level.

PGA/g	$0.02\le {\mathit{a}}_{\mathit{m}\mathit{a}\mathit{x}\mathit{I}\mathit{I}}<0.04$	$0.04\le {\mathit{a}}_{\mathit{m}\mathit{a}\mathit{x}\mathit{I}\mathit{I}}<0.09$	$0.09\le {\mathit{a}}_{\mathit{m}\mathit{a}\mathit{x}\mathit{I}\mathit{I}}<0.19$	$0.19\le {\mathit{a}}_{\mathit{m}\mathit{a}\mathit{x}\mathit{I}\mathit{I}}<0.38$	$0.38\le {\mathit{a}}_{\mathit{m}\mathit{a}\mathit{x}\mathit{I}\mathit{I}}<0.75$
Intensity	V	VI	VII	VIII	IX
level	1	2	3	4	5
explain	Very low	Low	Moderate	High	Very high

). Finally, the acquired intensity grades of each county are used as indicators to measure the seismic hazard of the region.

3.2. Vulnerability Assessment Method

Vulnerability refers to the natural, social, economic and environmental factors that a disaster-bearing body shows in the face of potential disaster risk, such as physical exposure, inherent sensitivity to external shock and human ability to resist risks associated with a disaster-bearing body [43,44]. Earthquake vulnerability assessment is a complex problem, including the population, economy, environment and so on [45]. Therefore, it is very important to select an advanced and complete vulnerability equation to accurately evaluate the carrier vulnerability. Based on existing studies [46,47], Formula (1) of the vulnerability assessment model was adopted in this study to carry out vulnerability assessment and analysis of the seismic carrier.

$$V_i=\frac{E_i\times S_i}{A_i}$$

(1)

where $V_i$ is the vulnerability index of county region $i$; $E_i$ is the exposure of county $i$; $S_i$ is the sensitivity of county $i$; and $A_i$ is the adaptability of county $i$. An increase in exposure and sensitivity leads to an increase in vulnerability, while an increase in adaptability leads to a decrease in vulnerability. These components are conceptualized and quantified in a detailed analysis below.

Exposure (E) refers to the number or value of the carrier within the range of risk factors [48]. For the risk of earthquake disaster, the greater the exposure, the greater the risk [49]. According to the disaster risk assessment guidelines issued by the State Oceanic Administration of China (http://www.soa.gov.cn/zwgk/zcgh/ybjz/201601/t20160115_49734.html, accessed on 20 February 2023) and the land use data classification system obtained, this paper reclassifies land use types into 7 categories, namely built-up, cropland, vegetation, water, bareland, snow/ice and wetland. Different land use types correspond to one or several types of hazard-bearing bodies. For example, the hazard-bearing bodies mainly included in cropland are crops, and the hazard-bearing bodies mainly included in built-up are buildings and population. In addition, the hazard-bearing body is located in the natural environment, and its exposure will change with the environment. Therefore, this paper adopts a method to determine the exposure value of a disaster-bearing body considering the influence of the natural environment. This method evaluates the influence of the natural environment on the exposure of a disaster-bearing body from five aspects: elevation, slope, distance from water system, distance from fault zone and distance from earthquake point. Under the same earthquake damage condition, the higher the elevation, the steeper the slope, and the closer it is to the water system, fault zone and earthquake point of an area, the higher its exposure degree, and vice versa.

Table 2. Exposure indicators and weights.

Indicator	Judgment Criterion	Exposure Value	Weight
Land use	Built-up	1.0	0.024
	Cropland	0.8
	Vegetation	0.6
	Water	0.4
	Bareland	0.4
	Snow/Ice	0.2
	Wetland	0.2
Elevation (km)	<2	0.2	0.132
	2–3	0.4
	3–4	0.6
	4–5	0.8
	>5	1.0
Slope (°)	<2	0.2	0.225
	2–6	0.4
	6–15	0.6
	15–25	0.8
	>25	1.0
Distance to fault zone (km)	<1.0	1.0	0.110
	1.0–3.0	0.8
	3.0–5.0	0.6
	5.0–10.0	0.4
	>10.0	0.2
Distance to earthquake point (km)	<10	1.0	0.458
	10–20	0.8
	20–30	0.6
	30–40	0.4
	>40	0.2
Distance to water (km)	<0.5	1.0	0.051
	0.5–1.0	0.8
	1.0–2.0	0.6
	2.0–5.0	0.4
	>5.0	0.2

lists land use types, five natural environment indicators considering the impact of the natural environment and their corresponding exposure values. In this study, the classification and weight distribution of each index were determined by References [47,50,51,52] and expert experience. Figure 4 shows the spatial distribution of exposure of each index.

Sensitivity (S) refers to the degree to which a disaster recipient is affected by a natural disaster event, generally an inherent attribute of the disaster recipient [53]. For example, children and the elderly are among the most vulnerable groups in society and may find it more difficult to take appropriate measures to protect themselves in the face of sudden disasters. Women are less functional than men, and they are more likely to be affected during disasters. The sensitivity of this study is shown by the total population, the proportion of the female population, the proportion of the population under 14 years old and the proportion of the population over 65 years old.

Adaptation (A) refers to the capacity of the system, in terms of behavior, resources and technology, to respond to adverse impacts such as climate change or disasters [54]. In the study of Liu et al. [47], general public budget expenditure and GDP were selected. The more developed the city is, the more resources it has to prevent and resist disasters directly or indirectly. The per capita disposable income of urban and rural residents represents the individual’s economic situation, which means having available resources to absorb, reduce and recover from losses. Adequate medical personnel and infrastructure can also help improve regional resilience and mitigate the immediate impact of disasters. Therefore, the adaptability of this study is presented by the expenditure of the general public budget, GDP, per capita disposable income of urban residents and rural residents, and the number of medical personnel and health institutions.

According to the results of the exposure assessment considering the natural environment, the land use type area within the range of 0.2~1.0 was counted by taking the county as the unit, and the land use area corresponding to different exposure values was used as the exposure index of the city. In order to avoid the redundancy of indicators, exposure values were divided into four levels at equal intervals. Finally, 14 indicators were used to evaluate the vulnerability of each county to earthquake disasters.

Table 3. Indicators for assessing vulnerability index to earthquake disaster.

Vulnerability Dimension	No.	Indicator	Impact to Vulnerability
Exposure	1	Areas of land use with the exposure value of 0.2–0.4	+ 1
	2	Areas of land use with the exposure value of 0.4–0.6	+
	3	Areas of land use with the exposure value of 0.6–0.8	+
	4	Areas of land use with the exposure value of 0.8–1.0	+
	5	Total population	+
	6	Percentage of females	+
Sensitivity	7	Percentage of population aged 14 and under	+
	8	Percentage of population aged 65 and above	+
Adaptability	9	General public budget expenditure	- 1
	10	GDP	-
	11	Urban disposable income per capita	-
	12	Rural disposable income per capita	-
	13	Number of hospital medical staff	-
	14	Number of medical institutions	-

¹ “+” indicates the indicator tends to increase vulnerability; “-” indicates the indicator tends to decrease vulnerability.

shows the detailed information of each indicator system.

After the establishment of the vulnerability assessment index system, it is necessary to determine the weight of each index. In order to improve the reliability of earthquake disaster vulnerability assessment results, the entropy weight method [55] was adopted in this paper to evaluate earthquake vulnerability. Information entropy is a measure to measure the amount of information contained in a system. If the variation degree of data value of an index is greater, the amount of information contained by the index will be greater, the information entropy of the index will be smaller, and the corresponding weight will be higher, and vice versa. The entropy weight method is an evaluation method that comprehensively measures the information provided by each index and determines the index weight based on information entropy. The specific steps are as follows: (1) standardize the original data matrix; (2) calculate the proportion of each evaluated unit under each index; (3) calculate the entropy and difference coefficient of each index; and (4) calculate the weight of each index and the vulnerability index value of the evaluated unit.

Table 4. Classification of vulnerability index.

Vulnerability Value	Vulnerability Level	Vulnerability Interpretation
(0.000, 0.035]	1	Very low
(0.035, 0.068]	2	Low
(0.068, 0.169]	3	Moderate
(0.169, 0.458]	4	High
(0.458, 1.000]	5	Very high

shows the classification principles of the vulnerability index values.

3.3. Risk Assessment

The specific process of earthquake disaster risk assessment in this study is shown in Figure 5. The risk matrix method is a kind of risk ranking tool; it can identify risks, and is a combination of qualitative and quantitative methods. The traditional risk level uses the risk and probability model to classify the two factors that determine the risk size, namely the probability and the degree of disaster loss, thus forming the risk matrix. In this study, earthquake risk is jointly determined by earthquake risk and vulnerability. These two variables are input into the improved risk matrix, with earthquake risk as the vertical coordinate and comprehensive vulnerability as the horizontal coordinate. With reference to the research results of Nyimbili et al. [56], earthquake risk is divided into five levels (

Table 5. Classification of earthquake disaster risk levels.

	Vulnerability	1 Level	2 Level	3 Level	4 Level	5 Level
Risk
1 level		1 level	1 level	1 level	2 level	3 level
2 level		1 level	1 level	2 level	3 level	4 level
3 level		1 level	2 level	3 level	4 level	5 level
4 level		2 level	3 level	4 level	5 level	5 level
5 level		3 level	4 level	5 level	5 level	5 level

), which are: level 1 (very low risk), level 2 (low risk), level 3 (moderate risk), level 4 (high risk) and level 5 (very high risk).

4. Results

4.1. Hazard

Figure 6 shows the earthquake hazard map calculated based on the ground motion parameter values, which directly expresses the degree of spatial variation in the earthquake hazard in the study area. The earthquake intensity in Qinghai Province of China is high. The earthquake high-risk areas include Mangya, Qilian, Maqin and Dari, which are mainly distributed in the north, northeast and southeast of Qinghai Province, while the other areas belong to the earthquake risk areas. The results indirectly indicate that Qinghai Province is one of the main areas of earthquake activity in China, with a wide range of earthquake activity. This is mainly due to the special geographical position of Qinghai Province, for example, Qinghai Province is located at the intersection of the Indian Ocean plate and the Eurasian plate, located in the Mediterranean–Himalayan seismic belt, as well as the special geological structure of Qinghai Province, seismic fault zones and terrain changes and other factors, which together lead to its high earthquake risk. Expert analysis indicates that, in the next few years, the earthquake safety situation in Qinghai Province is still not optimistic, with a severe and complex earthquake safety situation, leading to the province’s earthquake prevention and disaster reduction work putting forward new requirements and new challenges. The earthquake hazard map produced in this paper provides the spatial pattern distribution of the possible damage degree caused by earthquakes and provides understandable information. At the same time, such maps are the basic maps used by administrative agencies for earthquake prevention and disaster reduction services.

4.2. Vulnerability

Earthquake vulnerability in Qinghai Province, China, was evaluated based on land use data, social environment statistics and natural environment data, and the results were divided into five levels: very low vulnerability, low vulnerability, moderate vulnerability, high vulnerability and very high vulnerability (Figure 7). As can be seen from Figure 7, there are four areas with very high earthquake vulnerability in Qinghai Province, namely Geermu, Dulan, Zaduo and Yushu, which are mainly distributed in the central and western parts of the study area. In these highly vulnerable areas, the development of earthquake prevention and relief, the construction of emergency shelters, the reserve of emergency materials and the number of medical personnel are relatively backward, leading to a high vulnerability value. Mangya, Zhiduo, Qumalai, Nangqian, Delingha, Maduo, Xinghai and Menyuanhuizu are all high vulnerability areas, mainly distributed in the central and western parts of the study area. There are many areas with very low earthquake vulnerability in Qinghai Province, mainly including Wulan, Gangcha, Haiyan, Huangyuan, Guide, Guinan, Pingan, Chengxi, Tongren, Xunhuasalazu, Tongde and Henanmengguzu, which are concentrated in the eastern part of the study area. These areas with very low earthquake vulnerability have obvious advantages in economic development, social security, and development of earthquake prevention and disaster reduction, and their development is relatively balanced, so they show low earthquake vulnerability. There are many distribution areas of high and very high earthquake vulnerability in Qinghai Province. The reasons include its special geographical location and geological structure, geomorphological factors, economic development and population density, earthquake resistance of buildings and disaster response ability. Therefore, in order to reduce the loss and impact of earthquakes, it is necessary to strengthen the construction of earthquake monitoring and early warning systems, improve the earthquake resistance of buildings and strengthen the disaster response ability. The earthquake vulnerability map made in this study has important practical significance for the national and local governments to formulate earthquake prevention and disaster reduction policies and reduce the loss and injury from earthquake disasters.

4.3. Risk

This section provides a comprehensive assessment of earthquake risk regionalization in Qinghai Province, China, based on earthquake risk and earthquake vulnerability (Figure 8). The earthquake risk zoning map of China’s Qinghai Province is divided into five categories: very low risk, low risk, moderate risk, high risk and very high risk. The results showed that there were 5 very high-risk areas, 7 high-risk areas, 10 moderate-risk areas, 11 low-risk areas and 12 very low-risk areas. The spatial pattern of earthquake risk regionalization in Qinghai Province is obvious. The very high-risk areas and high-risk areas are mainly distributed in the central and western part of Qinghai Province, while the very low-, low- and moderate-risk areas are mainly distributed in the eastern part of Qinghai Province. In the central and western part of Qinghai Province, the economic development level is low, the medical conditions and the evacuation facilities are not perfect, and the carrier’s adaptability is low. As a result, the vulnerability of these regions to earthquake disasters is high. If an earthquake occurs, it will cause huge loss of personnel and property, and the risk of earthquake is relatively high. Therefore, it is suggested that Qinghai Province should reduce regional vulnerability according to the actual situation of counties and districts, focus on governance for regions with higher risk levels and improve regional carrying capacity, so as to realize the symbiosis between society and earthquake disasters. Compared with the whole province, the eastern part of Qinghai province has better medical conditions, more evenly distributed shelters and better facilities. The professional knowledge and skills of earthquake prevention and disaster reduction personnel are relatively strong, and the vulnerability to earthquake disasters is relatively low, resulting in a lower risk level of earthquake disasters. The research results are of great significance for safeguarding people’s lives and property, promoting social and economic development, enhancing public awareness and supporting government decision making.

5. Discussion

5.1. Validation

The earthquake risk zoning map provides important supporting data to guide earthquake prevention and disaster reduction, and has important practical significance for reducing economic losses and casualties [57]. According to the China Earthquake Networks Center (CENC) (http://www.ceic.ac.cn/history, accessed on 10 April 2023) and the National Earthquake science Data center (https://data.earthquake.cn/, accessed on 12 April 2023) access to China’s Qinghai province in 2009–2022 earthquake directory (Figure 9), the accuracy of the resulting risk zoning map has been verified. Figure 9 shows that there has been frequent earthquake activity in Qinghai Province from 2009 to 2022, and earthquakes are mainly concentrated in the medium-, high- and very high-risk areas of Qinghai Province. Among them, earthquakes with magnitudes greater than 6 were mainly distributed in the Zaduo, Yushu, Maduo, Menyuanhuizu and Delingha regions of Qinghai Province. Therefore, we find that the earthquake risk zoning map of Qinghai Province is consistent with historical earthquakes. The results of this study can provide important reference for the investigation and management of earthquake disaster risks in Qinghai Province.

In addition, we compare the research results with the risk assessment results of strong earthquakes that may occur in mainland China during 2021–2030 conducted by Shao et al. [58]. Based on the data on the strong earthquake recurrence model, the cumulative probability of each target fault in the next 10 years is given by the recurrence period and elapsed time of each fault, which are adopted from relevant studies such as seismological geology, geodesy and historical earthquake records. Figure 10 shows the dangerous areas of strong earthquakes covering Qinghai Province. We find that there are eight dangerous areas for strong earthquakes in Qinghai province predicted by Shao et al. They are the middle part of the Altun fault zone, the eastern segment of the Altun fault zone, the west-middle section of the Qilian fault zone, the east-middle section of the Qilian fault zone, the West Datan–East Datan segment of the East Kunlun fault zone, the Feng volcanic Yushu River section, the eastern section of the Kunlun fault zone–northern section of the Longriba fault zone and the middle and western section of the western Qinling Mountains. By comparison, it is found that the strong earthquake danger areas distributed in the central and western regions belong to the high and very high earthquake danger areas obtained in this paper, while the eastern strong earthquake danger areas basically belong to the moderate earthquake danger areas obtained in this paper. When Shao et al. comprehensively analyzed the risk of strong earthquakes that may occur in mainland China in the next 10 years, they took active tectonic blocks as the basis and considered the boundary and interior of active tectonic blocks. The fault activity states were estimated via the methods with clear physical significance, such as the fault segments with seismic gap, motion strongly locked, sparse small–moderate earthquakes and apparent Coulomb stress increase. Based on these scientific strategies and combined with the prediction of various methods, the risk areas of strong earthquakes that may occur in mainland China during 2021–2030 are finally predicted. The reliability of the research results largely depends on fault detection and geophysical observation. In this paper, based on remote sensing and GIS technology, a comprehensive assessment model of vulnerability and risk of earthquake disasters is constructed by integrating land use data, natural environment data, social environment data and seismic parameter zoning data. However, Shao et al.’s study is based on the strong earthquake recurrence model, and the cumulative probability of each target fault in the next 10 years is given by the recurrence period and elapsed time of each fault, which are adopted from relevant studies such as seismological geology, geodesy and historical earthquake records. The evaluation models of the two different methods can predict the earthquake risk to a certain extent, but the model in this paper is more suitable for a large-scale region, and Shao et al.’s model is more suitable for a small-scale region.

5.2. Analysis of Influencing Factors of Earthquake Risk Assessment

The areas with high and very high earthquake risk in Qinghai Province are mainly located in the central and western parts of Qinghai Province, accounting for 68.08% of the total study area. These areas with high earthquake risk also show high earthquake vulnerability. The Tibetan Plateau is the result of the Indian plate colliding with the Asian plate in the north, and Qinghai Province is located in the northeast of the Tibetan tectonic block, where the earthquake breeding environment is sufficient and the scale of active faults is huge. The influencing factors of high and very high-risk earthquake areas in Qinghai Province mainly include the following:

(1)

Research shows that the vast majority of earthquakes on Earth occur on active faults [59]. In Figure 11, the Altun fault belt in the north of Qinghai Province and the East Kunlun fault belt in the middle of Qinghai Province are its main trunk, and are the main places for stress release and adjustment during crustal deformation. The paleoseismic research since late Quaternary shows that the paleoearthquakes on the Altun active fault zone are frequent and strong, and there are at least ten paleoseismic surface rupture deformation belts along the Altun active fault zone, each of which has had several strong earthquakes [60]. According to the official determination of the China Earthquake Networks, a 5.8-magnitude earthquake occurred in Mangya City, Qinghai Province, on 16 June 2021, and the epicenter of the earthquake was located on the fault along the southern margin of the Altun Mountains and the northwestern margin of the Qaidam Basin.

The East Kunlun fault zone in the central part of Qinghai Province is the most frequent fault zone with earthquake activity above magnitude 7 and the longest fault zone in Qinghai Province. According to historical earthquake data, there have been four earthquakes of magnitude 7 or higher on the East Kunlun fault zone since 1900.

The Ganzi–Yushu–Feng volcanic fault in southwest Qinghai Province is a sinistral strike-slip fault. Due to the dynamic process caused by the Indian plate pushing northward towards the Qinghai–Tibet Plateau, this fault is one of the faults with a large crustal active scale and frequent earthquake activity in the southwest region of Mainland China at present, which has triggered a series of large earthquakes. For example, on 14 April 2010, a 7.1-magnitude earthquake occurred in Yushu City, Qinghai Province. The earthquake occurred in the Ganzi–Yushu–Feng volcanic fault zone, belonging to the boundary between the Qiangtang Block and Bayan Khara block. According to the historical geological data, the average left-lateral slip rate of the fault is 8~14 mm/a, and the vertical slip rate is about 0.6~1.2 mm/a. Due to the strong plate activity, due to the constant accumulation of strain energy, when the block boundary reaches or exceeds the threshold degree, it will be released in the form of an earthquake, causing serious damage to the surface.

The Lenglongling fault zone in eastern Qinghai Province, which is distributed along the ridge of Lenglongling, is a sinistral strike-slip structural zone in the Qilian–Haiyuan tectonic belt. It is composed of several faults of different lengths in a left-order oblique sequence. Under the long-term compression of active faults, the Lenglongling fault zone accumulates energy and releases stress, which leads to the occurrence of moderate and strong earthquakes. For example, the Shandan earthquake of magnitude 7.2 in 1954, and the Menyuan earthquakes of magnitude 6.4 in 1986, magnitude 6.4 in 2016 and magnitude 6.9 in 2022 all occurred along this fault.

(2)

Earthquake risk and vulnerability are the main factors that affect earthquake risk, showing the possibility of earthquake disaster [61,62]. The peak acceleration of ground motion in Qinghai province is relatively high, and the whole province is basically in the earthquake medium–high-intensity area of degree Ⅶ or above, which indicates that the ground and the buildings on the ground are affected and damaged greatly. Therefore, in 2021, the General Office of the People’s Government of Qinghai Province issued the 14th Five-Year Plan for Earthquake Prevention and Disaster Reduction of Qinghai Province. This is of great significance for Qinghai Province to further improve the basic capacity of earthquake prevention and disaster reduction at the grassroot level, promote the modernization of earthquake prevention and disaster reduction undertakings in the new era, and deepen the reform and development of earthquake prevention and disaster reduction undertakings. In order to reduce the earthquake disaster, we can increase the stiffness and strength of the building structure, increase the seismic support, use seismic materials, adopt seismic isolation technology, shock absorption technology, etc., and through reasonable seismic design and reinforcement measures, we can greatly improve the seismic ability of the building and reduce the loss caused by the earthquake.

(3)

The central and southern part of Qinghai Province is the Qingnan Plateau with an average altitude of more than 4000 m, accounting for more than half of the total area of the province. This natural barrier has brought great difficulties for earthquake relief work. Its lofty mountains form the basic skeleton of Qinghai’s landscape and witness the uplifting northward march of the Tibetan Plateau, making Qinghai and its neighbors the most active intracontinental orogenic belt in the world. The geological structure and active fault system within the orogenic belt are very complex, resulting in an earthquake-prone area.

5.3. The Limitation of Remote Sensing and GIS Technology in Earthquake Disaster Risk Assessment

At present, researchers use remote sensing and GIS technology for earthquake hazard, earthquake vulnerability and earthquake risk assessment, and other related studies [35,63,64,65,66]. The study of Manfre et al. [67] emphasized the importance of remote sensing, GIS and GPS data in earthquake risk and natural disaster management. Remote sensing technology, as the main data acquisition means in the field of earth observation, is an important tool to realize the analysis of geospatial products, and can meet the operational requirements of decision support systems for all types of natural disasters [68]. Surface topography image analysis technology based on GIS technology has also been widely applied in natural disaster assessment studies. For example, Nobrega et al. [69] classified surface exposure according to the slope model to determine the sensitivity of erosion and landslide. The existing methods of data acquisition and analysis provide basic information and technical support for earthquake disaster risk assessment. Therefore, the carrier data needed for the research can be acquired quickly, with high precision and low cost, and the regional-scale earthquake disaster risk assessment can be carried out in the study area.

However, there are still some difficulties in real-time data acquisition of remote sensing satellites, mainly reflected in the following aspects: spatial, spectral and temporal resolution, spatial coverage, and two-dimensional and three-dimensional capabilities. In addition, the ability to interpret and extract satellite imagery of specific disasters, such as the number of damaged floors, is a major challenge. Various thematic maps can be generated based on remote sensing and GIS technology to provide data support for earthquake probability analysis. However, at present, a comprehensive GIS-based model has not been established for disaster assessment and analysis in earthquake-prone areas, and it is difficult to implement a suitable method, which is currently a challenge. Of course, it is also necessary to pay attention to the cost of establishing an earthquake disaster risk assessment model, so as to reduce or not cause any economic losses as far as possible, present the most accurate assessment and prepare for future earthquake disasters. In the future, in order to better serve the country’s earthquake prevention and disaster reduction, it is suggested that the International Space Committee provide high-quality remote sensing image data free of charge in earthquake-stricken areas, and increase investment in UAV technology in earthquake-stricken areas where cloud interference is more serious.

6. Conclusions

Based on remote sensing and GIS technology, this paper comprehensively evaluates the earthquake hazard risk in Qinghai Province, China. The results show that there are five risk levels of earthquake disaster in Qinghai Province, which are: very low-risk area, low-risk area, moderate-risk area, high-risk area and very high-risk area. Among them, there are 5 very high-risk areas, 7 high-risk areas, 10 moderate-risk areas, 11 low-risk areas and 12 very low-risk areas. The central and western parts of the study area have a large scale of active faults, high altitude, and imperfect medical conditions and shelter facilities, and these areas have greater vulnerability and risk of earthquake disasters. Compared with the whole province, the eastern part of the study area has better medical conditions and refuge places, and the vulnerability and risk of earthquake disaster are lower.

The accuracy of this study is closely related to the accuracy of the land cover data, natural environment data and social environment data obtained. In the future, land cover data with higher precision and resolution can be introduced to meet the requirements of high-quality data for earthquake disaster research. Although there are some shortcomings in this study, the proposed method can be effectively applied to earthquake disaster risk assessment, and can provide reference for the formulation of earthquake prevention and disaster reduction strategies.