Discussion on the coupling relationship between flood risk and population vulnerability from climate justice

The accelerated urbanization has created an increasingly complex social ecosystem containing the complicated relationship among ecology, politics, economy, and society, which makes the impact of urbanization increasingly unpredictable. Research by van den Berg & Keenan (2019) shows that the unequal development of urbanization is a major cause of climate vulnerability, which makes vulnerable groups more susceptible to the impacts of climate change. In addition, climate change has intensified the complexity of urbanization. It deepens the existing vulnerability and poses greater risks to society, the economy, and the environment (Shokry et al. 2020; Su 2020). Therefore, we must view the impact of climate change on social vulnerability as a complex and persistent hazard. Moreover, the combined influence of climate justice, flood risk, and population vulnerability has rarely been studied (Mahmoud & Gan 2018; Zhou et al. 2019).

Climate change tends to intensify the global hydrological cycle. The most obvious impact is to make the spatial and temporal distribution of precipitation more uneven and bring more extreme rainfall. This further harms the lives and properties of residents and causes more losses (Chang & Su 2021). At the same time, the socioeconomic status of different social groups produced by social differentiation determines the degree to which people are affected by climate change. The research on climate justice showed that climate change, especially the flood hazard it brings, may have a greater impact on vulnerable groups (the elderly, children, and marginalized population) (Holland 2017; Yang et al. 2021). From the perspective of climate justice, the main goal is to mitigate the disadvantaged position of vulnerable groups in the world when disasters occur. However, vulnerable groups often prioritize economic benefits rather than flood risk (Running 2015). They pay more attention to the daily living environment, such as violence, unemployment, and crime instead of the impact of climate change (Lizarralde et al. 2021). Also, wealthier regions and privileged citizens can benefit from better environmental conditions. This also highlights the lack of economic opportunities for vulnerable groups in the environment where they live. However, nature has a good memory. When facing future climate change, vulnerable groups may pay a greater price and may become even poorer due to the impact of climate-related floods.

The combined influence of urbanization and flood risk has had a disproportionate negative impact on the vulnerable groups in the city. To reduce the flood risk of the entire city, it is usually necessary to restrict the construction of flood plains or relocate existing buildings. However, for vulnerable groups, settling in these areas can meet their needs for livelihood and work, and it is an effective way to get rid of poverty. Therefore, flood-hit areas are often the only settlements that vulnerable groups can afford (van den Berg & Keenan 2019). People still know very little about how social differentiation affects flood risk. Some researchers believe that the wealthier people are, the less susceptible they are to climate change, and the most marginalized groups are more easily hit by climate changes such as floods (climate change alters the distribution and intensity of rainfall, often causing greater and more frequent floods). There are also some researchers who point out that the poor are more resistant to floods because they are more mobile and have fewer assets, and thus they suffer fewer losses and can more easily recover from the disasters. However, the rich people may be more vulnerable to flood since their resources are bound to the specific location of the land (Friend & Moench 2013). Nevertheless, the uneven distribution of flood risk and population vulnerability highlights the increasing importance of the research on climate justice. This research analyzed the coupling relationship between flood risk and population vulnerability, so as to reveal the spatial inequality between disasters and vulnerable groups and to promote the understanding of climate justice.

The unequal development of urbanization has made certain groups in the society more vulnerable. For example, many vulnerable groups living in cities do not have access to reliable and affordable basic services (Juhola et al. 2016). Vulnerable groups are unlikely to prioritize long-term floods caused by climate change over economic growth. They also do not support restrictions on economic activities and believe the climate policies that restrict their own development are fundamentally unfair (Parks & Roberts 2006; Norgaard 2012). Relatively wealthy people are more likely to separate environmental issues from economic issues and actively participate in the climate change agenda. For example, wealthier families build flood-prevention walls around their houses to withstand the damage from floods. In addition, high land prices in non-flooded areas have further intensified climate injustice because the rich are able to bear higher housing costs and relocate to safer areas. Therefore, vulnerable groups are forced to relocate to high-risk areas where land prices are low (Mavromatidi et al. 2018). It can be seen that vulnerable groups are often marginalized and are ignored in the evaluation and policy intervention. Thus, a highly adaptable system in urbanization is not necessarily fair, nor is it beneficial to vulnerable groups (Friend & Moench 2013). It is believed that understanding the relationship between flood risk and population vulnerability in urbanization is of great significance, which can promote climate justice.

Flood risk and vulnerable groups are often characterized by their location in urban space (Su et al. 2021). Coupled thinking can provide an analysis framework for the multi-scale complex connection between flood risk and population vulnerability, so that the reasons for the coupled spatial relationship between flood risk and population vulnerability under urbanization can be quantified. Coupling originated in physics, which is the concept of the interaction between two or more systems (Cui et al. 2019). In recent years, this concept has been applied to the urbanization–environment system, such as the study on the coupling relationship among industry, environmental pollution, housing, and population (Cai et al. 2021). Therefore, the coupling coordination degree (CCD) model was used in this study to measure whether there is a coupling relationship between flood risk and population vulnerability in a certain area, so as to further reveal the coordination degree between flood risk and vulnerable groups. To understand why the vulnerable groups are spatially coupled with flood risk, it is necessary to explore the driving factors of the relationship between them. Based on the results of coupling, this research used the Tobit model to analyze the factors of urbanization contributing to this relationship. This study explored the spatial and statistical connections between flood risk and population vulnerability and provides a theoretical basis for further research on climate justice.

Although cities are paying increasing attention to climate justice, there seems to be no acceptable empirical method to test the climate justice between flood risk and population vulnerability. This research may contribute to this field. We try to achieve the following goals by providing a new framework for the coupling between flood risk and population vulnerability: (Ⅰ) to establish a model for the coupling between flood risk and population vulnerability to investigate the main driving factors of climate injustice and (ii) to make the geographic and statistical relationships between flood risk and population vulnerability clear since these relationships have triggered the unbalanced and unfair urban development. In short, this research provides a new understanding of climate justice and injustice, so that the problems the vulnerable groups cannot handle can be discovered and solved.

In this research, the concept of population vulnerability was included into the scope of disaster management. To verify the spatial coupling relationship between flood risk and population vulnerability, the composition of population vulnerability needs to be defined, but vulnerability is relative. It is believed that those with low income are more vulnerable to climate change than wealthy people. Age, health, race, income, and education level will all affect the vulnerability of the population (Mavromatidi et al. 2018; Wilk et al. 2018; Drakes et al. 2021). Factors such as the elderly, children, illiterate, number of divorces, low-income population, aboriginal, and physical and mental disabilities are regarded as the indicators of population vulnerability based on the indicators published in the ‘Taiwan Demographic Bulletin’. Population resilience indicators are expressed by young and middle-aged population and high-income population. Specific indicators are shown in Table 1 (the calculation method of the weight value is introduced in the entropy method below).

Based on past experience, setting the threshold to 0.7 can better reflect the spatial matching of flood risk and population vulnerability (Liu et al. 2022). When 0 < τ ≤ 0.7, it means that the flood risk is less than the population vulnerability, which is a relative reflection of environmental justice. When 0.7 < τ ≤ 1.4, it means that the flood risk and population vulnerability appear in space at the same time, which is relative to the environmental injustice. When τ > 1.4, it means that the population vulnerability is less than the flood risk, and the vulnerable groups have not migrated to the floodplains relatively, which also reflects environmental justice relatively.

Based on the coupling between population vulnerability and flood risk, this research explored the reasons for such coupling from the perspective of urbanization. Urbanization has changed the original living environment, causing many people to actively or passively adjust their living environment to adapt to the impact of urbanization. Chen et al. (2018) and Xia et al. (2020) show that urbanization is affected by four aspects: economic development, population growth, built environment, and life improvement. Economic development indicators are represented by per capita income, agricultural output value, industrial output value, and commercial output value (Gu et al. 2020; Patri et al. 2022), reflecting the economic development of each region. Previous studies on population growth indicators used population density to reflect the population congestion in the region and natural population growth rate to highlight the population structure of the region (Clar et al. 2023). The built environment indicator is the physical space that reflects the level of urbanization. Previous studies used per capita public land, housing quality, per capita road area, and spatial development intensity (Chen et al. 2018; Xia et al. 2020; Bernardini et al. 2024) to highlight the security level of the physical space. The life improvement indicator mainly focuses on the impact of social conditions on people's lives. Xia et al. (2020) believe that government capacity, housing price, and medical institutions are the three major indicators that affect the quality of life. In summary, the specific impact indicators of the coupling in this study are summarized in Table 2.

It can be seen from the second column of Table 3 and Figure 3(b) that flood disaster risk is mainly distributed in coastal areas and areas on both sides of the river. The Qingshui district has the highest flood risk. This district is located in the lower reaches of the river and in the coastal area, and thus it is easily affected by flooding and seawater intrusion, but the population vulnerability of this district is at an average level. Other districts with high risk value include the Wuri, Wufeng, and Dajia districts, all of which have a risk value greater than 0.6. These districts are also mainly located in coastal areas and areas on both sides of rivers. It can be seen that the development of coastal areas and areas on both sides of rivers is vulnerable to flooding. The districts with low risk are mainly located in the central areas of the city, including Central, West, East, North, and South districts, all of which have a risk value of less than 0.1. This shows that the main land development in Taichung City avoids areas with high flood risk, which is conducive to the development of the city and the protection of personnel. In general, the overall flood risk value of Taichung City is 0.368, which is at a relatively low level.

From the perspective of population vulnerability (the third column of Table 3), the average population vulnerability of Taichung is 0.366, which is at a relatively low level. However, the Taiping, Beitun, and Dali districts have relatively high population vulnerability, indicating that compared with other administrative districts, these districts have more vulnerable groups. Similarly, the administrative districts located in the central area of the city do not have high population vulnerability, and the values are all less than 0.5.

From the perspective of the CCD (the fifth column in Table 3), the overall CCD of Taichung City is 0.391, indicating that the flood risk and population vulnerability are not highly coupled in space, and the coupling is at a low level. However, there are still several districts where the CCD exceeds 0.5, including the Beitun, Xitun, Taiping, and Qingshui districts, demonstrating that these districts are in a relatively coordinated state. From the perspective of the relative climate justice, the districts where flood risk is as high as population vulnerability are the Shengang, Daya, Nantun, and Shalu districts. Therefore, these districts were regarded as the climate injustice districts because they need to face flood risk and population vulnerability simultaneously. Other districts only need to deal with one of the two problems. For example, the flood risk value of the Fengyuan district is only 0.149, while the population vulnerability reaches 0.576. Although the population vulnerability is high, the flood risk value is small; therefore, this district can effectively avoid disaster losses. Another example is the Wufeng district which has a very high flood risk (0.730), but its population vulnerability is only 0.246; thus, when a disaster occurs, the personnel and property losses can be effectively avoided.

Furthermore, the results of coupling between flood risk and population resilience were analyzed and compared with the coupling relationship between flood risk and population vulnerability. It can be seen from Table 4 that the population resilience of Taichung City is not high, at only 0.323. Only Nantun, Beitun, Xitun, and Dali districts have a population resilience higher than 0.6. The overall CCD is 0.366, indicating that the flood risk and population resilience of each district in Taichung City are not concentrated at the same spatial location at the same time. In other words, a relatively resilient population effectively avoids flood risk areas. However, Beitun, Xitun, and Taiping districts have the CCD greater than 0.5, indicating that the resilient population in these districts pays more attention to economic development than other districts and relatively ignores the problem of flood risk. From the perspective of the relative climate justice relationship, wealthy people will relatively avoid areas with flood risk, and the population resilience group can better reflect climate justice. However, climate injustice can still be observed in the Daya and Taiping districts. The relatively resilient population in these districts prefers flood risk areas to find opportunities for economic development.

In general, the CCD between flood risk and population vulnerability is 6.5% higher than that between flood risk and population resilience, indicating that vulnerable groups are more likely to live in areas with high flood risk than those who are rich. This makes the area face the dual problems of flood risk and population vulnerability at the same time, which further raises man-made risks and indirectly reflects climate injustice. In Taichung City, there are four districts with injustice in both flood risk and population vulnerability, two more districts than the districts with population resilience injustice. This also indirectly shows that the resilient population tends to choose areas with low flood risk, and vulnerable groups are forced to transfer to areas with high flood risk.

In this study, the value of CCD was used as the dependent variable, and the indicators of urbanization were used as the independent variables. The analysis results of the Tobit model are shown in Table 5. The standard deviations of the model are all less than 0.4, and the R² values are 0.78 and 0.82, respectively, indicating that the Tobit model has a better fitting effect. It can be found from the results that the coupling between flood risk and population vulnerability is quite different from the coupling between flood risk and population resilience. The main driving factors for the coupling between flood risk and population vulnerability are government abilities and population growth. In addition, the coefficients of population density, housing quality, and spatial development intensity of the city are negative, suggesting that in areas with great population density, housing quality, and spatial development intensity, the vulnerable groups are likely to leave. In other words, vulnerable groups tend to migrate to rural areas with low population density or areas with weak development intensity, and the more average the housing quality in an area, the easier it is for the vulnerable groups to gather. The coupling between flood risk and population resilience is more susceptible to economic development and the built environment. The economic growth and the improvement of the built environment can attract wealthier people. However, the results showed that the coefficient of per capita road area and housing prices is negative, indicating that wealthy people are also susceptible to the impact of housing prices and road noise. The analysis of driving factors in this research is helpful for handling the problems that the disadvantaged or wealthy groups cannot discover and solve.

In the face of the increasingly uncertain climatic conditions, how people pay more attention to climate justice in the urban environment has attracted increasing attention. Especially, people are paying increasing attention to the unequal distribution of flood risk (Steele et al. 2015). Our research results have identified the areas where flood risk and population vulnerability are highly spatially coupled, which can provide a realistic basis for the analysis of climate injustice.

The combined effects of urbanization and climate change have exerted a disproportionate negative impact on those vulnerable groups in the city. Our research results showed that the CCD between flood risk and population vulnerability is 6.5% higher than the CCD between flood risk and population resilience, suggesting that vulnerable groups are more susceptible to flood risk than wealthy groups. This makes some areas face the dual risks of flood disaster and population vulnerability simultaneously. Such double-vulnerability greatly increases the risk of water-borne diseases of vulnerable groups and threatens their property and even their lives.

The coupling results showed that urbanization has made historically marginalized populations more vulnerable and less secure, so that these populations are more likely to move to areas with high flood risk, and at the same time, more privileged residents will benefit from the built environment in the city. It can be seen from the analysis of the driving factors of coupling that vulnerable groups prefer economically favorable areas where the local government has strong abilities to take care of the vulnerable groups. The high prices of land and the high intensity of spatial development will further intensify this coupling. The consequence is that vulnerable groups will stay away from the areas with better urban built environment since many do not have access to reliable and affordable basic services. The rich are more affected by economic development and the built environment of the city, and they prefer to live in areas with low flood risk. The results provide a new understanding of urban climate justice and the spatial distribution of injustice.

People know little about the position of vulnerable groups in climate justice and the potential relationship between flood risk and population vulnerability. It seems that there is no acceptable empirical method to test the relationship between vulnerable groups and climate justice. Previous studies mainly focus on how flood risk causes great social vulnerability (Yang et al. 2021), but the relationship between flood risk and population vulnerability has rarely been discussed from the perspective of spatial coupling. The difference between this research and related studies is that it constructed a coupling and coordination relationship between flood risk and population vulnerability from the perspective of climate justice. In this framework, the indicators of population vulnerability and population resilience were established, and the entropy method was used to objectively evaluate the weight of each indicator. In addition, the CCD model was used to compare the relationship between flood risk and population vulnerability with the relationship between flood risk and population resilience, thereby revealing their spatial differences. Finally, the driving factors of coupling were used to reveal how the vulnerable groups and the rich strengthen climate injustice in the process of urbanization. This framework helps better understand the spatial inequality in flood risk and population vulnerability and the phenomenon of climate injustice, and can also help explore the factors of urbanization leading to unbalanced and unfair results. The results establish the spatial and statistical link between flood risk and population vulnerability, which can provide a theoretical basis for future urban planning, spatial design, government policy implementation, and governance.

There are still some limitations in this research. (i) It explored the relationship between flood risk and population vulnerability from a regional perspective, but the scale of urban design was not included. Subsequent research can combine urban land development and functional zoning to make the phenomenon of climate injustice concrete in space. (ii) The entropy method in this study does not consider the correlation between criteria. Since it only assigns weights based on entropy (variability), it cannot resolve redundant information. Subsequent research can further refine the weights of indicators in combination with the Criteria Importance Through Intercriteria Correlation (CRITIC) method (Manikanta et al. 2023).

Climate change has intensified the vulnerability of cities and may have a great impact on vulnerable groups. Our research results have verified this hypothesis. The results make clear the geographic and statistical relationships between flood risk and population vulnerability and illustrate how these relationships have shaped the unbalanced and unfair urban development. The specific conclusions can be summarized as follows:

• (I)

  This research constructed a framework to explore the coupling relationship between flood risk and population vulnerability from the perspective of climate justice. The framework studied which are the real disadvantaged and wealthy groups and applied the information entropy method to objectively assign the weight of the indicators. The flood risk was identified using the relationship of inundation potential. Furthermore, the CCD model was adopted to calculate the coupling between population vulnerability and flood risk, and the driving factors of coupling were analyzed using the Tobit model.

• (II)

  The research results showed that vulnerable groups in Taichung City are more likely to live in areas with high flood risk, 6.5% higher than wealthy groups. Therefore, the areas with high risk often face the double crisis of flood risk and population vulnerability. In Taichung City, there are four administrative districts with injustice in both flood risk and population vulnerability, which double the result of population resilience. This also indirectly reflects the potential connection between social vulnerability and environmental injustice.

• (III)

  Judging from the analysis results of the Tobit model, vulnerable groups tend to move to rural areas with low population density or areas with low development intensity, and the ability of the government is a major consideration for the settlement of vulnerable groups. Relatively wealthy groups are easily affected by the level of economic development and the built environment of the city and tend to settle in areas with low flood risk. This result can help solve the problems that the disadvantaged or wealthy groups cannot discover and solve, which is vital to the future pursuit of climate justice.

After the Paris Climate Agreement was established, increasing attention has been paid to fair and just process, which requires the identification of the risk relationship between the characteristics of vulnerable population and climate change (Hughes & Hoffmann 2020). The advancement of urbanization and a highly adaptable system are not necessarily fair, nor are they beneficial to vulnerable groups. Although our research results revealed the phenomenon of climate injustice to a certain extent, a broader knowledge system involving social vulnerability, poverty, political economy, and different social powers (Friend & Moench 2013) needs to be established to promote the true understanding and resolution of climate justice.

ZJ contributed to the conceptualization, methodology, reviewing the writing – reviewing, writing the original draft, visualization, and data curation. QS contributed to the conceptualization, methodology, software, data curation, writing the original draft, visualization, and reviewing and editing the writing. YC contributed to reviewing and editing the writing, software, and data curation.

This study was supported by Projects of Fujian Social Science Foundation (No. FJ2022C097), Projects of Fuzhou Philosophy and Social Science Foundation (No. 2023FZC32), Projects of Fujian Natural Resources Science and Technology Innovation (No. KY-020000-04-2022-035), Projects of National Natural Science Foundation of China (No. 52008110), and Projects of Fujian Natural Science Foundation (No. 2020J05195).

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.