Analysis on the spatial differentiation characteristics of poverty risk caused by disaster under the stress of geological disasters: a case study of Sichuan Province

Research on the poverty risk caused by geological disasters in disaster-prone areas is a useful exploration to coordinate social economic development with disaster prevention and reduction, and is of great significance to the regional sustainable development. Based on statistical data and spatial data, this paper takes Sichuan Province as the typical research area. Remote sensing and geographic information technology are used to study the poverty risk caused by disasters based on the quantitative evaluation of geological disasters risk and regional development level. The spatial differentiation characteristics of poverty risk caused by disasters are explored on the 1 km × 1 km grid scale. The results indicate that (1) the overall risk of geological disasters in Sichuan Province is relatively high, with high and relatively high risk areas accounting for more than 40% and low and relatively low risk areas accounting for less than 30%. The risks in Mountain and Ravine Areas are significantly higher than other areas. (2) The regional development level in Sichuan Province is relatively high, but with significant spatial differences. The development level of high-altitude areas and remote mountainous areas is quite different from that of the Chengdu Plain in the middle Sichuan Province. The uneven development in the east, middle, and west is a prominent problem. (3) The poverty risk caused by disasters is high, and the spatial pattern presents a characteristic of “high in the west and low in the east” with high positive spatial correlation. High-High Cluster Areas are mainly distributed in western and southwestern Sichuan. Low-Low Outlier Areas are mainly distributed in Chengdu Plain and Hilly Areas of Sichuan Basin. High-Low Outlier and Low–High Outlier Areas occupy a relatively small percentage with scattered distribution. This paper provides some theoretical support for policy formulation and management of coordinated development of regional socioeconomic and ecological environment.


Introduction
Looking back on the history of world development, the development of human society has always been accompanied by natural disasters, diseases, and epidemics, resulting in a large number of casualties and heavy economic losses . China is a mountainous country with mountain areas accounting for 69.4% of the total land, the unique energy gradient in the mountain area usually induces geological disasters such as debris flow, landslide, and collapse, causing serious loss of people's lives and property accompany by restricting regional development (Cui 2014). Due to geographical constraints, the economic development of mountain areas is lagging. All the 14 contiguous povertystricken areas in China before 2020 are in mountain areas, where low-income people are concentrated. The fragile ecological environment, frequent disasters, and backward social as well as economic development have led to 20% of Chinas poor farmers being impoverished caused by disasters, which is ranking second among the factors causing poverty.
Recently, more and more scholars begin to pay attention to the relationship between poverty and geological disaster. From the perspective of the relationships among economic poverty, disaster risk, and environmental degradation, economic poverty is the external driving force of the vicious cycle of "frequent disaster-ecological degradation-poverty intensification" (Ding et al. 2013;Andrew and Mikhail 2017). The asset stock of low-income families in disasterprone areas is not enough and being vulnerable, which not only makes their assets vulnerable to lose, but also seriously worsens their income sources in the future, and then affects their post-disaster reconstruction and recovery capacity, thus making them fall into the persistent poverty trap (Bidisha et al. 2021;Sakai et al. 2017;Carter and Barrett 2006).
At the beginning, scholars were more concerned with the probability of the occurrence of geological disasters, focusing on the prevention and loss of geological disasters and assessment of vulnerability Dominey-Howes 2002;Niu et al. 2012), and the methods adopted mainly include analytic hierarchy process (Zheng et al. 2021), neural network method (Aditian et al. 2018), informative method , and random forest model (Zhang et al. 2017). As the research further develops, scholars should not only pay attention to the possibility of geological disasters, but also pay attention to the losses induced by them (Fedeski and Gwilliam 2007;Malheiro 2006). Therefore, it has become a hot topic to study the impact of geological disasters on human socioeconomic development from the perspective of vulnerability. Generally speaking, the higher the vulnerability of geological disasters in a certain area, the greater the potential loss it will caused (Li et al. 2015). The reasons leading to spatial differences of geological disaster vulnerability are not only related to the frequency of geological disasters, but also related to local residents' risk perception of disasters and their disaster prevention capacity (Khan et al. 2020). Studies on poverty or regional development are mainly carried out from singledimensional and multi-dimensional perspective (Jin et al. 2020;Ding et al. 2013;Jin et al. 2020;Liu et al. 2020). The field of single-dimensional poverty mainly focuses on income level, such as the median and average of disposable income of residents (Wang and Sun 2021), while the multidimensional poverty perspective mostly includes economic, social, and natural dimensions, involving economy, medical care, resources, education, development, and other indicators (Schleicher et al. 2018;Jin et al. 2020;Liu et al. 2020). Scholars mostly use polyhedron method (Jin et al. 2020), spatial econometric model (Zhang and Yang 2019), and A-F method (Wen et al. 2018) to identify and measure multidimensional poverty and then propose measures to promote regional development based on the evaluation results (Liu et al. 2016. Geological disaster and poverty have a geographical spatial attribute (Bathrellos et al. 2012;Besagni and Borgarello 2019;Iparraguirre 2012). In recent years, scholars have made positive progress on the coupling among geographic patterns of poverty, poverty, and geographic environmental factors (Berberich 2019;Elwood et al. 2017;Jessie et al. 2016;Zhou et al. 2021;. Bird and Shepherd (2003) pointed that the spatial poverty trap is the area with a small stock of "geographic capital" and with high incidence of poverty. The economic structure of agriculture in povertystricken areas is relatively homogeneous, and the livelihood of farmer is extremely dependent on natural resources and ecosystem services. That is, the geographical distribution of environmental vulnerability and poverty is highly coupled (Barbier 2010). Ding et al. (2013) constructed an evaluation index system of disaster and poverty based on PSR model, quantitatively studied the coupling relationship between them, and believed that disaster and poverty tend to form a vicious cycle (Ding et al. 2013). Relevant studies have proved that geological disasters have a comprehensive impact on residents' lives, housing, income, and assets and play a leading role in poverty vulnerability (Xu et al. 2017. The problem of poverty caused by disasters has gradually attracted human attention. Scholars believed that the disaster factor in poverty structure system is objective and cannot be separated out (Shang and Shao 2018). In addition, they also believed that farmers whose main source of household income is agricultural business are mostly affected by natural disasters and are more likely to experience poverty and return to poverty caused by disasters (Zhang et al. 2020).
At present, most scholars focus on the natural attributes of geological disasters, but lacking researches on the social attributes between disasters and regional development, especially the quantitative evaluation from the geospatial perspective. As the number of geological disasters, which threatened people and threatened property rank the top disasters in China, Sichuan Province is one of the provinces with the most serious geological disasters in China and the largest relatively poor population in China, with a prominent contradiction of unbalanced and insufficient regional development (Qiu and Zeng 2017;Yong et al. 2020). Therefore, this paper is taking Sichuan Province as a research interest and constructs a geological disaster-induced poverty risk evaluation model based on the relationship between disasters and regional development. The paper also quantitatively studies the poverty risk caused by disasters and its geospatial pattern in Sichuan Province, thus providing a theoretical basis for coordinated management between administration of geological disasters and relative poverty.

Study area
Sichuan Province, located in Southwest China and in the upper reaches of the Yangtze River, lies between 92°21′ and 108°12′ E and 26°03′ ~ 34°19′ N, with a length of more than 1075 km from east to west and a width of more than 900 km from north to south (Fig. 1). The geographical environment of Sichuan Province is complex, geological structure is changeable, the neotectonics is active, the geological disasters are frequent, and the spatial distribution types and development characteristics are different. The total area of geological disaster-prone areas accounts for 97% of the total land area of Sichuan Province, which is one of the provinces that has the most severe geological disasters in China (Xu 2006). Sichuan Province has identified 35,000 potential geological disaster sites in 2020. At the same time, Sichuan has four contiguous poverty-stricken areas before 2020, namely, the Qinling-Bashan Mountains, the Wumeng Mountains, Liangshan Yi Ethnic Region, and the Tibetan-related areas in Sichuan, along with 88 poverty-stricken counties. By the end of 2013, the rural poverty population was 6.25 million, of which more than 60% were located in 88 poverty-stricken counties, making it one of the impoverished counties and provinces with the most impoverished population in China (Qiu and Zeng 2017).

Data sources and processing
The data used in this paper mainly include spatial data, socioeconomic data, and demographic data as follows: ( All the data were entered into GIS geodatabase after preprocessing. Furthermore, the projection and coordinate system of all the data were transformed into UTM84N and WGS-84 projection, by GIS software.

Geological disaster risk assessment
Deterministic c oefficient method The amount of information can objectively reflect the contribution of evaluation factors to the risk of geological disasters under different classification standards. The greater the index is, the higher the risk of geological disasters will be. Combined with the content of information, this paper uses the certainty factor (CF) to determine the weight. CF was proposed by Shortliffe and Buchanan (1987), which is usually used for sensitivity analysis of different factors, and gradually used for calculating the weight of each factor (Liang et al. 2019). The method is as following: In this formula, PP a is the ratio of the number of geological disasters in factor a to the area of factor a; PP s is the ratio of the number of geological disasters in the study area to the area of the study area. The variation range of CF is [− 1,1]; a positive value indicates a high certainty of geological disaster occurrence, which is more likely to occur; a negative value represents a lower certainty of geological disaster occurrence, which is less likely to occur; when the calculation result is close to 0, it means that the factor cannot determine whether the geological disaster is likely to occur or not in this classification.
The weight ω i is calculated as follows: In this formula, CF (i, max) is the maximum value to determine the coefficient in each classification of factor i, and CF (i, min) is the minimum value to determine the coefficient in each classification of factor i.

Risk assessment method
The disaster index is an indicator reflecting the scale of geological disasters in each evaluation unit, and the calculation formula is as follows (Liang et al. 2019): In this formula, V i is the evaluation unit. V i is the comprehensive risk index of geological disasters of the evaluation unit. ω j is the weight value of the evaluation index j of the evaluation unit i. y j is the normalized value of the evaluation index j of the evaluation unit i.
Geological disaster risk assessment index Combining with relevant research (Liang et al. 2019;Luo et al. 2020) and field investigation, this paper comprehensively works on the considerations of topography, land cover, lithology, meteorology, and hydrology. Seven indicators are selected as the risk assessment factors of geological disasters. The deterministic coefficient method was used to calculate the weights of each indictor, and the calculation process is shown in formulas (1) and (2). The evaluation index system is shown in Table 1.

Regional development level evaluation method
Comprehensive development index method The comprehensive development index (CDI) is used to reflect the comprehensive development level in the study area. The larger the index value, the richer the region will be and vice versa. The CDI could be calculated as following : In this formula, CDI is the comprehensive development index; F ij is the indicator value after standardized processing, ω ij represents the index weight; ω i represents the dimension weight; n is the number of dimensions; and m is the number of indicators corresponding to a certain dimension.
The evaluation indicators of CDI include both positive and negative values. Therefore, the data standardization in this paper adopted the extreme difference standard method, and the specific formulas are as follows: In those formulas, Y ij is the indicator value after standardized processing; X ij is the original data of the evaluation index j of county i in Sichuan Province. X max and X min are the maximum and minimum values of the evaluation index j, respectively. The calculation method of index weight can be divided into subjective method and objective method. In order to reduce subjective biases in expert judgments, incomplete data, and objective biases caused by data quality, the analytic hierarchy process (AHP) and entropy method are used to calculate weights, respectively (Ni et al. 2009;Chen et al. 2009). Then, the subjective and objective weights are added to get the comprehensive weight.
CDI evaluation system According to specific research needs, the relationships between research results and socioeconomic development level in Sichuan Province are considered (Alkire and Foster 2011;Schleicher et al. 2018;Jin et al. 2020;Liu et al. 2020). And we measured CDI from four dimensions (resource, economic, income, education, and medical care) and 18 indicators. The weights of each indicator were calculated by using AHP and entropy method. The evaluation indicator and weights are shown in Table 2.

Evaluation method of the poverty risk caused by geological disaster
According to the general risk assessment formula, the poverty risk caused by geological disasters consists of geological disaster risk and the regional development level (Tian and Zhang 2016). The formula is as follows: In this formula, R m is the risk index of poverty caused by geological disasters. The higher the index, the higher the poverty risk caused by disasters is. V i is the risk of geological disasters. CDI j is a regional comprehensive development index.
The evaluation process of poverty risk caused by geological disasters is shown in Fig. 2.

Exploratory spatial analysis method
Global Moran's I Global Moran's I can describe the spatial characteristics of geographical environment elements and detect the spatial autocorrelation of spatial elements in the whole study area to identify the spatial distribution pattern of the study area on the whole (Fu et al. 2017). Global Moran's I is used to describe the spatial differentiation characteristics of poverty risk caused by disasters in the study area (Ma et al. 2021). The formula is as follows: In this formula, w ij is the spatial weight; x is the mean value of the attribute; x i and x j are the attribute values of the elements i and j, respectively; n is the number of units, and the correlation is considered to be significant when |Z|> 1.96 (Meng et al. 2005).

Local indicators of spatial association
In order to make up for the defect that Moran's I cannot point out the spatial distribution of the risk clustering of poverty caused by disasters or abnormal occurrences, the spatial pattern of poverty caused by disasters risk in Sichuan Province was analyzed using local indicators of spatial association (LISA) (Fu et al. 2017). Local autocorrelation reveals the local clustering characteristics of spatial unit attributes by analyzing the difference degree and significance level between spatial units and surrounding units (Ma et al. 2021). The formula is as follows:

Analysis of geological disaster risk
Formulas 1-3 were used to calculate the geological disaster index by GIS software. Using natural breakpoint method, the study area was classified into five grades: relatively low risk, low risk, moderate risk, high risk, and relatively high risk (Fig. 3).
The spatial distributions of geological disaster are shown in Fig. 3. The relatively high-risk areas are mainly distributed in Western Sichuan Alpine Valley, Mountain, and Ravine region of Southwest Sichuan and Mountain Areas around Sichuan Basin. High-risk areas are less distributed in Chengdu Plain and Hilly Areas of Sichuan Basin, while the other 4 areas have a higher high-risk distribution proportion. Fig. 2 The evaluation process of poverty risk caused by geological disasters Fig. 3 Classification results of geological disaster risk Similar to high-risk areas, the moderate risk areas are less distributed in Chengdu Plain, and the other 5 areas account for a relatively high moderate risk proportion. The relatively low-risk areas are concentrated in Hilly Areas of Sichuan Basin, while the risks in other 5 areas are relatively not significant. The low-risk areas are mainly distributed in Chengdu plain and Hilly Areas of Sichuan Basin, but are only distributed sporadically in other 4 areas.
The statistics results of numbers of geological disaster risk areas of various grades are as follows. The area of relatively high risk area is 73,574.81 km 2 and accounts for 15.14% of total areas, distributed in the Mountain and Ravine Area of West Sichuan and Mountain Areas around Sichuan Basin in flakes. The area of high-risk area is 137,037.55 km 2 and accounts for 28.20%, distributed in flakes and groups around Western Sichuan Plateau as well as relatively high-risk area. The area of moderate risk area is 130,263.67 km 2 and accounts for 26.80%, distributed throughout the study area with the pattern of strips and groups. The area of relatively low-risk area is 99,829.32 km 2 and accounts for 20.54%, distributed in strips in West Sichuan and South Sichuan distributed in groups in East Sichuan in groups. The area of low-risk area is 45,294.65 km 2 and accounts for 9.32%, distributed in flakes in the Chengdu Plain, distributed in groups in the plateaus of Southeast Sichuan and northwest Sichuan, distributed in strips in the riverside along South Sichuan. The risk of geological disasters in Sichuan Province is relatively high in the whole area and widely distributed, and the distribution is closely related to the topography and landforms, thus presenting significant geographic spatial differences.

Analysis of regional development
Considering the county-level administrative divisions as the evaluation unit, the comprehensive development index (CDI) of each region was calculated respectively by GIS software according to formulas 4-6. Using natural breakpoint method, the development level index of 183 districts and counties was divided into 5 grades (Fig. 4), which are high, relatively high, medium, relatively low, and low, respectively.
As shown in Fig. 4, the development of every district and county in Sichuan province varies greatly, and the contradiction of unbalanced and insufficient development is more is prominent. From the perspective of spatial pattern, the areas with low and relatively low comprehensive development level are mainly distributed in the Western Sichuan Plateau Area, Mountain and Ravine Area of Western Sichuan, and Mountain and Ravine region of Southwest Sichuan. The areas with medium comprehensive development level are mainly distributed in Hilly Areas of Sichuan Basin and Mountain Areas around Sichuan Basin, while the other areas are scattered. The areas with relatively high development level are mainly distributed in Hilly Areas of Sichuan Basin and Chengdu Plain area; areas with high level development level are relatively small, mostly distributed in Chengdu Plain Area.
The normal distribution characteristics are significant considering the proportions of various development grades. The high development level counties accounts for 6.56% of total counties in number, distributed in the main urban area of Chengdu in groups and scattered in Fucheng District of Mianyang City, Cuiping District of Yibin City, East District of Panzhihua City, etc. The relatively high development level counties accounts for 25.14% in number, distributed in Chengdu and its surrounding areas in flakes, and it is obviously influenced by the radiation of Chengdu City. The number with medium level counties accounts for 33.33%, mainly located in cities and prefectures such as Guangyuan City, Bazhong City, and Nanchong City in the northeast and Leshan City on the southern edge of the Chengdu Plain. The number with relatively low level counties accounts for 24.04%, the number with low level counties accounts for 10.93%. They are distributed in the Western Sichuan Plateau and Mountain and Ravine Area of Southwest Sichuan, which is related to the long-term low economic level of these areas. In general, with the Chengdu Plain being in the center of Sichan Province, the level of comprehensive development in Sichuan Province is lower as it goes outwards of Chengdu Plain. The level of comprehensive development in mountainous areas and high-altitude areas is obviously low, which is basically consistent with the spatial distribution of 88 counties out of poverty, indicating that these areas will remain the work focus of agriculture, rural areas, and peasantry in the future.

Overall characteristics of poverty risk caused by disasters
Based on the evaluation of geological disaster risk and regional development, the poverty risk caused by disasters in the study area was calculated. The natural breakpoint method was used to divide the risk of poverty into 5 grades: high, relatively high, medium, relatively low, and low (Fig. 5). The overall spatial pattern of poverty risk in the study area is high in the west and low in the east. Due to the lag of comprehensive development, higher social and economic vulnerability, and weak ability to resist geological disasters, families' livelihood in Western Sichuan Alpine Valley as well as Western Sichuan Alpine is more vulnerable to geological disasters than that in the central and eastern regions of Sichuan Province. While poverty risk caused by disasters is higher. Instead, cities such as ChengDu and Mianyang in the central part, Suining and Nanchong in the east, Yibin and Luzhou in the southeast, and Panzhihua in the south of Sichuan province all have relatively low risk of poverty caused by disasters, which are mainly attributed to their high social and economic level or low risk of regional geological disasters.
The statistical results of the areas of poverty risk grades caused by disaster are presented in Fig. 6. The area of high risk area was 73,574.81 km 2 , and the area of relatively high risk area was the largest, reaching 137,037.55 km 2 . The area of medium risk area is close to that of high risk area, accounting for 130,263.67 km 2 . The area of low risk area is 99,829.32 km 2 . The area of relatively low risk area is accounting for the smallest which is 45,294.65 km 2 . These results indicate a high overall poverty risk caused by disasters in Sichuan Province.

Spatial distribution characteristics of poverty risk caused by disasters
The evaluation results of poverty caused by disasters in the study area were gridded with a grid size of 1 km × 1 km, and the Moran's I at the grid scale is calculated. The Moran's I of poverty caused by disasters in Sichuan Province is 0.767, the value of p < 0.001, indicating that there is a significant positive spatial correlation between disaster and poverty risk in Sichuan Province at the grid scale of 1 km × 1 km. The disaster risk caused by poverty is spatially dependent and clustered. That is, the risk of poverty caused by disasters has regional characteristics.
The LISA is used to study the local clustering characteristics of spatial unit attributes of poverty risk caused by disasters in Sichuan Province, and the spatial clustering is divided into four types: High-High Cluster, research grid, and neighboring grids are at high risk of poverty caused by disasters; High-Low Outlier, the risk of poverty caused by disasters in the research grid is high, while the risk of poverty caused by disasters in the neighboring grid is low; Low-High Outlier, the risk of poverty caused by disasters in the research grid is low, and the risk of poverty caused by disasters in the surrounding neighborhood grid is relatively high; Low-Low Cluster, the risk of poverty caused by disaster is low in research grid and neighboring grid. The local spatial autocorrelation of poverty caused by disaster is shown in Fig. 7.
The results show that the spatial clustering characteristics of poverty risk caused by disasters in Sichuan Province are different, showing a clustering characteristic of "high in the west and low in the east." High-High and Low-Low Cluster is obvious, while the High-Low and Low-High Outlier is not significant. The details are as follows: (1) High-High Cluster Area: total area of 142,418.80 km 2 .
The type is mainly distributed in most of the districts and counties of Ganzi Prefecture, Wenchuan County, Li County and Mao County in Aba Prefecture, and Liangshan Prefecture in western Sichuan. These areas account for 29.30% of the province's land area. Located in the southeastern edge of the Qinghai-Tibetan Plateau, the above three areas, with the main topography of high Mountains, Ravines, and Plateaus, are all contiguous poverty-stricken areas in China until 2020. At the same time, geological disasters occur frequently in this area, and the damage intensity is high. In recent years, major natural disasters impact frequently in these areas, such as Wenchuan earthquake, Xinmo Village high-position landslidein Diexi Town, Wenchuan torrent and debris flow, and Danba County torrent and debris flow. The interweaving of development lagging and the geological disasters is the most important characteristic in High-High Cluster Area.
(2) Low-Low Cluster Area: With a total area of 163,890.440 km 2 , these areas are mainly distributed in the Chengdu Plain represented by Chengdu; also distributed in Hilly Areas of Sichuan Basin represented by Neijiang, as well as a few places of Panzhihua in southern Sichuan, it totally accounts for 33.72% of the province's land area. The Low-Low Cluster Area is characterized by plains and shallow hills with high regional economic level. These areas have fewer geological disasters accompanied by strong regional disaster capacity and low disasters-caused poverty risk.
(3) The areas of High-Low Outlier and Low-High Outlier are 152.63 km 2 and 798.17 km 2 respectively, accounting for 0.03% and 0.16% of the province's land area, respectively. Those two area types are mainly scattered in the borders between High-High and Low-Low Cluster Areas.

Discussion
In this study, the geological disaster risk evaluation system was constructed, and the geological disaster risk in the study area was evaluated combined with the determined coefficient method. Overall, the risk of geological disasters in Sichuan province is high, and the spatial differentiation is significant. The reasons for the spatial differentiation are related to topography, geological conditions, and seismic activities (Wen et al. 2021;Xu 2006 (Xu 2006;Wen et al. 2021;Xiong et al. 2017). Although the frequency of landslide disasters is high in Hilly Areas of Sichuan Basin, , there are relatively fewer disasters such as debris flow, mountain torrents, and collapses in these areas; the regional landslide disasters are also in relatively small scale. Accompanying with strong disaster prevention and bearing capacity, the risk of geological disasters is relatively low Ding et al. 2013;Liu et al. 2018). Therefore, corresponding prevention and control measures should be taken according to the grade of geological disasters to ensure the safety of people's property. Based on the multi-dimensional poverty perspective, AHP and entropy method were used to calculate the CDI of the study area, and the results showed that the comprehensive development level of Western Sichuan and Southwestern Sichuan was low, the comprehensive development level of Eastern Sichuan was moderate, and the comprehensive development level of Chengdu Plain and its neighboring regions was high, which is similar to the research results of Chen et al (2020). By 2020, Sichuan Province had 88 poverty-stricken counties and four contiguous poverty-stricken areas (Qinling-Bashan Mountain area, Wumeng Mountain area, Liangshan Yi Ethnic Region and Tibetan Plateau area). These areas are mainly distributed in the Western Sichuan Alpine Valley, Mountain and Ravine Area of Southwest Sichuan, and Mountain Areas around Sichuan Basin, which have the largest distribution of relatively poor population in Sichuan Province (Qiu and Zeng 2017). The difference of regional development level in Sichuan Province lasts for a long time (Deng and Bai 2020), which is closely related to disaster vulnerability (Sun et al. 2017). Therefore, it is important to further understand the relationship between disasters and economic differences and collaborate to create a policy mechanism that is conducive to the virtuous cycle of economic development and resources environment.
From the perspective of the spatial pattern of the risk of poverty caused by disasters, the Tibetan Areas in Northwest Sichuan Plateau and the Liangshan Yi Ethnic Region in the Mountain and Ravin Areas of Southwest Sichuan have the highest risk of poverty caused by disasters, followed by Mountain Areas around Sichuan Basin. The risk in Chengdu Plain is the lowest, which is similar to the findings of Sun et al. (2017). The Chengdu Plain has a low risk of geological disasters, high comprehensive development level, and strong carrying capacity, but the population density is relatively large, so it is necessary to focus on population vulnerability indicators (Sun et al. 2017). The Western Sichuan Alpine Valley, Mountain and Ravine Area of Southwest Sichuan, and Mountain Areas around Sichuan Basin are the areas with the highest intensity of disaster damage and the largest number of relatively poor population, and it is necessary to strengthen disaster management and poverty management and rely on community power to cope with disasters to enhance the synergy between disaster reduction and poverty alleviation (Kirby et al. 2019;Xu et al. 2020). Hilly Areas of Sichuan Basin can take advantage of the location of Chengdu-Chongqing economic circle to build an opening pattern at a wide range and at all levels as well as promote the coordinated development of regional social economy and natural environment on the whole.

Conclusions
Based on the quantitative evaluation of geological disaster risk and regional development, the poverty risk caused by disasters and their spatial differentiation characteristics were explored in Sichuan Province as the research interest. The results are as follows: (1) The overall risk of geological disasters in Sichuan Province is relatively high, with high and relatively high risk areas accounting for 43.34% of the total areas, low and relatively low risk areas accounting for 29.86%, and medium risk areas accounting for 26.80%. Due to the topography, geological structure, human activities, and other factors, the risk of geological disasters has some obvious spatial clustering and differentiation. The risk of geological disasters in Mountain and Ravine Area is significantly higher than that in other areas, while the risk of geological disasters in Chengdu Plain and its eastern adjacent areas is significantly low.
(2) In recent years, the medical care, education security, infrastructure construction, and other improvements in Sichuan Province have been improved gradually, the residents' income and the efficiency of resource utilization have been improved constantly. The national economy and society are developing steadily. The percentage of areas in high and relatively high development level is 31.70%, the percentage of medium development level areas accounts for 33.33%, and the percentage of low and relatively low areas is 34.97%. The spatial differences of regional development level are prominent. The overall development of high altitude areas and remote mountain areas in western Sichuan is low on the whole, while the development level of plain areas and hilly areas are high. (3) The area proportion of the poverty risk grades caused by disasters in Sichuan Province is as follows: medium risk area > relatively low risk area > relatively high risk area > low risk area > high risk area. The spatial pattern shows the characteristic of "high in the west and low in the east." There was a significant positive spatial correlation between poverty risk and disasters. High-High Cluster mainly distributed in Garze Tibetan Autonomous Prefecture, Liangshan Yi Autonomous Prefecture, and other areas in western Sichuan and southwestern Sichuan, accounts for 29.30% of the total areas. Low-Low Cluster areas are mainly distributed in Chengdu Plain and Hilly Areas of Sichuan Basin, accounting for 33.72% of the total areas; High-Low and Low-High Outlier Areas are small and scattered randomly. (4) There is a geospatial coupling between regional development level and geological disaster risk in Sichuan Province. It is necessary to coordinate society and economic development with disaster reduction and prevention management, for example, develop ecological agriculture and green industries according to local conditions. It is also important to improve the industrial infrastructure towards the direction of green and sustainable development, thus reducing the dependence on natural resources and enhancing the ability to resist disasters from the regional and personal perspectives; in the end, we can ensure the coordinated development of economy, society, and ecology. and Technology Program (2020YFS0308) and Open Foundation of the Research Center for Human Geography of Tibetan Plateau and Its Eastern Slope (Chengdu University of Technology).

Data availability
The data that support the findings of this study are available upon reasonable request from the authors.

Declarations
Ethics approval and consent to participate Not applicable.

Consent for publication
The author agrees to participate in the publication of the paper.

Competing interests
The authors declare no competing interests.