Climate change is ambiguous, having significant effects on human and ecological systems throughout the globe [(Pandey et al., 2015); (Abid et al., 2023)]. The Intergovernmental Panel on Climate Change (I.P.C.C., 2014) has released its fifth assessment report, which highlights how climate change is affecting every part of the globe—from tropical to Arctic regions, islands, and continents, as well as affluent and poor nations. Along with its effects on local agro-climatic and ecological conditions, rainfall patterns, and other aspects of human well-being, climate change may also have significant consequences for livelihoods, health, and other facets of human welfare [(Connolly-Boutin & Smith, 2016); (Pagnani et al., 2021); (Sithole et al., 2023)]. Particularly vulnerable to climate change is the country Turkey through its geographic, climatic and socio-economic conditions, which make Turkey extremely susceptible to the effects of climate change and other environmental threats, making resilience and adaptation becoming key priorities [(Bayram & Ozturk, 2021); (Gumus et al., 2023)]. The degree to which a system is impacted by climate stressors connected to the environment, either negatively or positively, is known as its susceptibility. All aspects of climate change, such as mean climate characteristics, variability in the climate, and the frequency and intensity of extremes, are considered climate-related stressors. The term ‘adaptation’ describes changes made to ecological, social, or economic systems in reaction to real or predicted climate stressors and their consequences (Change, 2024). To mitigate potential harm or take advantage of opportunities brought about by climate change, it refers to modifications in procedures, customs, and organizational structures (Change, 2024). In addition to having a more fragile transportation system than other nations, Turkey is dealing with problems related to food security, rising water stress, and unprecedented natural disasters like the post-2020 forest fire outbreaks and the massive earthquake that struck the south-eastern region of Turkey in February 2023 [(Bank, 2022); (Ozkula et al., 2023)]. Climate conditions, population exposure (such as the percentage of the population affected by floods, forest fires, and destructive effects of the earthquake), and socioeconomic factors (such as the percentage of the economy that is based on agriculture, human and economic losses in the region that was affected by the earthquake) all contribute to this susceptibility (Bank, 2022). The unprecedented earthquakes that occurred in February 2023 caused damage to residential buildings, manufacturing facilities, bridges, transit systems, earth structures, harbors, and lifelines in densely populated areas. According to the (Bank, 2023), initial estimates of the direct infrastructure loss exceeded $34 billion and over 15 million of the population (about 17% of Turkey’s population) was severely affected by the earthquake. Planning adaptation measures and providing a justification for their adoption would be made possible by understanding the expected climate-related risks in specific regions and their interactions with current and future populations and varieties of assets [(Castells-Quintana et al., 2018); (Singh & Chudasama, 2021)]. Understanding the geography of climate change susceptibility has several benefits, including improved disaster risk reduction, a decreased exposure of ecological and human resources, and the identification of people who are more susceptible to climate change. Consequently, strategies for mitigating susceptibility and adapting to climate change have gained widespread attention from the public. However, Turkey’s changing environment is not well monitored, and it is still unclear exactly what the future repercussions will be and where they will occur. Furthermore, information about the degree and distribution of human susceptibility to climate change and its effects in the area is scarce.

Vulnerability indicators offer a potentially valuable way to track vulnerability over time and space, pinpoint the processes that lead to vulnerability, rank the strategies that reduce vulnerability, and assess how well these strategies work in various ecological and social contexts (Adger et al., 2009). They make it possible to compare the systems under evaluation and make information about what has to be changed in an easy-to-understand manner more visually appealing. Numerous authors have used this technique in different ways, either locally or nationally. (Birkmann et al., 2022), (Edmonds et al., 2020), (Formetta & Feyen, 2019), and (Abid et al., 2023) developed a set of critical indicators using a quantitative method to evaluate the vulnerabilities and adaptive capabilities of nations worldwide. On a regional level, (Busby et al., 2014) showed how climate change indicators can be used to identify communities that are most vulnerable to climate impacts in Africa; (Fawcett et al., 2017) indicated how two longitudinal methodologies—cohort and trend studies—are employed in the evaluation of climate change susceptibility through an analysis of case studies from the Arctic region; (Malik et al., 2012) assessed Pakistan’s climate change vulnerability by combining evaluations of climatic risk, environmental and population sensitivity, and adaptive capacity; and (Gupta et al., 2020) developed a spatial social vulnerability index for evaluating the socio-environmental vulnerability in four altitude zones in the Indian Himalayas. (Buchir & Detzel, 2023) considered the governance factor besides exposure, sensitivity, and adaptive capacity which constitute the vulnerability index to assess the climate vulnerability in Mozambique. At the local level, (Notenbaert et al., 2013) examined how Mozambican agro-pastoralists manage their sensitivity to climatic pressures and their coping mechanisms, as well as the reliability of several widely used indicators of vulnerability. (Mainali & Pricope, 2017) created maps of Nepal’s climate change vulnerability by combining evaluations of climatic risk, and geo-physical and socio-economic factors sensitivity, and analysed how these different variables interact to shape climate vulnerability in Nepal. (Kumar et al., 2016) is another local study that applied a method for assessing the susceptibility of cities to climate change by evaluating different indicators, demonstrating the vulnerability assessment approach, and presenting a spatial assessment of climate change sensitivity patterns. (Omerkhil et al., 2020) conducted questionnaires to assess vulnerability and adaptation strategies for smallholder farmers in the hills of Yangi Qala District in Takhar province, Afghanistan. A final example is (Phuong et al., 2023) which applied a livelihood vulnerability index to examine the vulnerability of minority communities in upland regions of Thua Thien Hue province, central Vietnam. The literature on climate change in Turkey is multi- variate including climate change policy applied in Turkey between 1992-2015 (Turhan et al., 2016); renewable energy used for climate change mitigation (Keles & Bilgen, 2012); impacts of climate change on precipitation climatology and variability (Turkes et al., 2020), evaluation of climate change simulations over climate zones of Turkey (Onal & Unal, 2014), climate change and its consequences in Turkey (Bayram & Ozturk, 2021), and impacts of climate change on agricultural yield (Chandio et al., 2020). Despite there is increasing literature on climate change, particularly in recent periods [(Keles & Bilgen, 2012); (Onal & Unal, 2014); (Turhan et al., 2016); (Turkes et al., 2020); (Chandio et al., 2020); (Bayram & Ozturk, 2021)], to the knowledge of the authors, no study in Turkey focuses on developing composite climate change vulnerability indices representing the climate vulnerability of the cities in Turkey.

This article represents the first attempt to rate Turkey's cities based on how vulnerable they are to climate change. This is accomplished by creating an index of climate change sensitivity that accounts for the natural exposure of these cities to climate change, in addition to their socioeconomic fragility and occupants' ability to adapt. Sectors generally identified with socio-economic and environmental sustainability i.e. social, public utility and transport, economics, land cover, meteorology, and atmospheric conditions, and natural disasters experience differences in vulnerability. Data from the official sources obtained for Turkey contain vulnerability indices across these six sectors over 81 regions (cities) were used to construct a composite climate change vulnerability index (CCVI) for each region. The CCVI was constructed through the application of weights that were developed from expert-based systems including the BWM (Gomez-Limon et al., 2020) and the SWARA method (Hashemkhani Zolfani et al., 2018). The former is a multi-criteria decision-making process in which pairwise comparisons are conducted when the best and the worst options are selected from a list of options. The latter method, on the other hand, was introduced in 2010 as a different paradigm in the multiple attributes decision-making field (Hashemkhani Zolfani et al., 2018). It was designed to be used in decision-making procedures where policy-making is given greater weight than in traditional decision-making processes (Haghnazar Kouchaksaraei et al., 2015). The SWARA is a newly proposed method; however, it has been used for solving many problems such as selection of green suppliers (Yazdani et al., 2016), selection of hotel projects based on environmental sustainability (Zolfani et al., 2018), evaluation of sustainable design of the household furnishing materials (Zolfani & Chatterjee, 2019), evaluation of investment alternatives of microgeneration energy technologies (Dinçer et al., 2022), and assessment of medical waste treatment techniques (Patel et al., 2023)). SWARA's advantage is that it evaluates the accuracy of experts and weighs each criterion by taking into account the opinions of several experts. The most significant rank criterion is the first one, and the least significant rank criterion is the last one. The final rankings are decided by a panel of experts using the mean of their ratings. Other weight assessment methods need extensive computations with low accuracy rates; in contrast, SWARA is intuitive and time-efficient. The multiple weighs of pairwise comparisons resulting from a high number of criteria or alternatives would make addressing Multi-Criteria Decision Making (MCDM) problems more complex and would also cause pairwise comparisons to become less consistent. (Rezaei, 2015) who introduced the BWM noted that the unstructured method of performing pairwise comparisons is the cause of the constraint, and that the many workloads and expert complexity are not required. Rezaei came up with a novel method-the BWM-to close this gap by conducting pairwise comparisons in an organized manner. Since the BWM's appearance, it has drawn the interest of numerous academics, and numerous studies on the BWM have been published. Some examples include: [(Ahmadi et al., 2017); (Salimi & Rezaei, 2018); (Kheybari et al., 2019); (Wankhede & Vinodh, 2021); (Uyan & Ertunç, 2023); (Amjad et al., 2024)].

Using this framework, we assess the findings from the composite indicators estimated in the study by contrasting two different weighting techniques that allow us to estimate weights external to the data. Our composite indicator of CCVI is different from other studies as we considered these two novel methods of multi-criteria decision analysis in the construction of the composite indicators. Also it represents the local conditions of the regions in Turkey which makes it different from the other indices that included different sub-criteria in the construction of the climate vulnerability index. Specifically, we want to examine the composite indicator values that come from different weighting strategies that are calculated at the regional level for Turkey. We selected a wide variety of indicators that are relevant to describe relative socio-economic and environmental vulnerability to construct a CCVI informing vulnerability to climate change. However, different input construction methods exist to generate composite climate change vulnerability indices, resulting in different outcomes. Literature providing guidance on the assessment of the methods used and composite indicator outcomes are still scarce. Therefore, our study aims to quantify the differences between climate change vulnerability outputs from the two different approaches (BWM and SWARA methods). Through constructing CCVI from these approaches, the developed composite indicators could assist in climate practitioners in evaluating and policy making regarding the climate change vulnerability in Turkey.

The assessment of vulnerability to climate change includes utilizing many indicators and criteria based on exposure, sensitivity, and adaptive capacity components to estimate the potential impacts and consequences of climate change (Žurovec et al., 2017). To assess the climate change vulnerability of the cities in Turkey, a comprehensive dataset consisting of 55 indicators across 6 themes such as economy, land cover change, meteorological conditions, natural disasters, public utility, and transport and social characteristics were utilized, as outlined in Table 1. The selection of indicators is based on a literature review [(Pandey et al., 2015); (El-Zein & Tonmoy, 2015); (Zanetti et al., 2016); (Balaganesh et al., 2020); (Edmonds et al., 2020); (Birkmann et al., 2022)], and the availability of data in the Turkish case was also influential.

The large and narrow range in the variance between the variables in multivariate statistical analysis necessitates normalizing the variables. Thus, before they are aggregated, indicators must be converted into dimensionless numbers. The variables listed in Table 1 were normalized using the min-max approach, which rescales the data linearly by placing a value of 0 at the worst and a value of 1 at the best. Eq. (1) was utilized to represent a positive indicator, and Eq. (2) was utilized to represent a negative indicator.

Assessing climate change vulnerability in cities is a complex and crucial task that necessitates the application of comprehensive and robust approaches. Multi-criteria decision Analysis (MCDA) approaches play a key role in this assessment since they allow for the creation of composite indicators and index computations [(El-Zein & Tonmoy, 2015); (Žurovec et al., 2017); (Truong et al., 2023)]. The BWM and SWARA are two powerful MCDA methods that offer a structured and quantitative way to calculate the weights of different indicators and combine information to create composite indicators and indexes.

(Rezaei, 2015) introduced the BWM, which offers a methodical way to assess the significance of different criteria by comparing them pairwise and pinpointing the most and the least crucial factors to establish reliable weightings. The method is praised for its simplicity, efficiency, and capacity to generate dependable weights with little cognitive load on respondents (Bovkır et al., 2023). SWARA complements BWM by providing a detailed method for adjusting initial weights based on expert judgment. (Keršuliene et al., 2010) introduced the SWARA method, which enables experts to iteratively adjust weights by hierarchically evaluating criteria and considering differences in importance among successive indicators. This method is an MCDA method and does not involve complex and time-consuming calculation processes.

The BWM methodology involves multiple steps (Rezaei, 2015):

The outstanding characteristic of the SWARA method is that it can estimate the opinions of experts or interest groups regarding the importance of variables in the weight-determination process (Mardani et al., 2017). In our study, six professionals in the fields of public policy, environmental science, geomatics engineering, and urban development and planning made up the panel of experts. They were asked to rank the importance of particular criteria based on the weights given in Table 2. Because each pair of criteria was represented as a single number in the pairwise comparison matrix, they assigned relative weights to the subject criteria based on their level of expertise. The resulting outcomes were then averaged to obtain a single value for each pair of criteria. The application methodology of the SWARA method is generally described in six steps [(Alrasheedi et al., 2023); (Jalilian et al., 2022)].

Determination of the decision criteria (j) on the relevant problem.

Decision-makers rank the criteria from the best to the worst based on their own knowledge and experience.

Considering the created ranking, the criteria are scored comparatively, starting from the second-ranked criterion, between 0-1, consistent with multiples of 5%. In this way, sj values are calculated.

The kj coefficient expressed in Eq. ( 3) is calculated for each criterion.

1 \end{cases} \end{document} ]]> (3)

The weight coefficient qj expressed in Eq. ( 4) is calculated for each criterion.

1 \end{cases} \end{document} ]]> (4)

The relative weights (wj) of the criteria are obtained using Eq. 5.

(5)

Because the weights we obtained from BWM and SWARA were close to each other (Table 3), the calculated composite vulnerability indicators were also similar. Therefore, we jointly presented the results as given in Figure 4. The dark colors in the Figure present the cities that are the most vulnerable to climate change in terms of exposure, susceptibility, and adaptive capacity. These cities are mainly located in the northern, western, and southern coasts of Turkey as well as the cities in the southeast region. This suggests two important observations that need to be highlighted: First, coastal areas and climate risk are interlinked, and the closest cities to the coast are also the most vulnerable to climate change. Second, the highest temperature extremes are related to impacts of the climate change. The highest seasonal temperatures particularly in summer are observed in south-eastern cities which were found to be most vulnerable to climate change. Forecasts of poor rainfall and extreme temperatures in the southeastern region suggest that a high percentage of population and livelihoods could be at risk. Based on the continuation of this trend, food security could be affected due to a lower amount of arable land, draught, and low yields of agricultural production. These regions are also prone to wildfires, among which Şanlıurfa and Mardin are the most vulnerable cities with annual share of burned land averaging 6.6 and 5.4 percent, respectively between 2011 and 2021 (Bank, 2022). These cities are among those with large shares of poor population in Turkey implying that the poor population had to cope with the devastating impacts of wildfire in the past 10 years.

The coastal cities are vulnerable to climate change due to heavy rainfalls resulting in rising sea levels, floods, and windstorms. The estimations by (Kerblat et al., 2021) indicated that a 200-year flood along the Mediterranean coast could push 100,000 people into poverty. There were 935 extreme events in 2019 with heavy rain/floods (36%), windstorms (27%), and hail (18%) that pose risks to people's lives and infrastructure (Bank, 2022). It was estimated by the (Bank, 2023) that a 100-year flood would affect more than 3 percent of GDP (or $20 billion) and 3 million people. And it was shown that both rich and poor people are affected by floods, particularly in the coastal areas.

The vulnerability index can be utilized to evaluate the possible effects of a program, initiative, or policy by changing the value that is anticipated to change and adjusting the index through recalculation (Nguyen et al., 2016). The new predicted proportion might be entered into the formula to determine the new vulnerability score, for instance, if the goal of the policy is to increase access to social security benefits. Any policy intervention's true direct and indirect consequences, however, are hard to quantify and require an additional assessment procedure (Pandey et al., 2017).

As an alternative, the index can be applied to determine which components are most important in causing vulnerability when guiding policy. Our results from weight calculation analysis indicated that the most significant factors contributing to climate change vulnerability are meteorological conditions and land cover change. Therefore, policy applications should address these issues to increase the resilience and adaptability of societies and reduce their vulnerability. Meteorological factors and land cover were also highlighted in the research by (Yoo et al., 2011) which focused on the development of climate vulnerability indices in the selected coastal city in Korea; (Pandey & Jha, 2012) constructed a climate vulnerability index by considering multiple criteria to assess the vulnerability of communities in rural Himalaya, India; (Krishnamurthy et al., 2014) which provided a methodological framework to evaluate the impacts of climate risk on food security at the national level in the globe; and (Balaganesh et al., 2020) which developed composite climate index to assess climate change-induced drought in India. Other factors were also mentioned in the literature such as ecosystem services, food, human habitat, health, governance, electricity and waste, infrastructure, and water [(Edmonds et al., 2020); (Kling et al., 2021); (Al Mamun et al., 2023); (Buchir & Detzel, 2023); (Phuong et al., 2023)].

Meteorological conditions and land cover change make the Country highly vulnerable to the impacts of climate change and other environmental hazards and bring adaptation strategies and resilience to the forefront. Compared to similar nations, Turkey's transportation infrastructure is more susceptible, and the Country is dealing with problems related to food security, rising water stress, and unprecedented natural disasters like the 2021 forest fire season and earthquakes in February 2023 in the southeastern region. Climate conditions, population exposure (such as the percentage of the population exposed to floods and forest fires), and socioeconomic factors (such as the percentage of agriculture in the economy) all contribute to this susceptibility. Another important factor contributing to climate susceptibility is the energy sector-which includes power and different land use sectors (transport, industry, residential uses, etc.)-is the country's largest contributor to greenhouse gas emissions. Therefore, land use changes observed in the Country such as conversion of agricultural land to urban uses or conversion of forests to agricultural uses all contribute to increasing energy use resulting in higher greenhouse gas emissions. The air pollutants such as methane and carbon contribute to climate change and are associated with serious health problems. For the decarbonization of the power sector, some strategies can be introduced such as promoting energy efficiency and electrification for zero carbon transport and buildings, industrial energy efficiency and technological innovation to mitigate emissions, emission reduction in waste and agricultural sectors, and optimizing the carbon sequestration potential of forests and agricultural landscapes. Additionally, there are chances for more focused adaptation measures to control long-term climate change and natural risks. These steps can lessen the damage caused by natural disasters, but they can also delay the effects of climate change on agriculture, water availability, and human habitation.

The index developed in this study can be adopted to different contexts and spatial scales by following the same methodology that is introduced in the paper. At the more detailed spatial levels, there can be different indicators representing more spatial details that can be included in the analysis to compute the CCVI. If the study is repeated in a different national context, indicators that are relevant for a specific nation may provide vulnerability indices that are different from the ones shown here. Weights that are assigned to indicators and sub-indicators can be different at the national context as these will be specified by experts to highlight the importance of country specific factors. The index can be calculated at different time periods based on the availability of the indicator data to monitor the impact of government policies that aim to alleviate the impacts of climate change. This can be useful to compare vulnerability of cities at the benchmark year with the year when the policy is applied. The change in climate according to different climate scenarios have an influence on the CCVI in that the value of the index can be modified and the new vulnerability score can be re-computed. Within this framework, climate models can serve as a valuable input for the construction of the vulnerability index. Through the prediction of expected climate-related occurrences, together with the corresponding financial consequences and mortality under various scenarios, an understanding of how climate affects vulnerability can be gained. Another important step in the development of a composite indicator is the sensitivity analysis. The main goal with sensitivity analysis is to measure the impact of input elements on the model result. Sensitivity analysis methods allow the whole range of values for each input variable to be explored without assuming anything about the structure of the model (e.g., additivity or linearity). Sensitivity analysis can be specified to be performed as a future research focus.

The findings offered here provide a novel assessment of the Turkish population's susceptibility to the physical, social, and environmental effects of climate change. The creation of a set of indicators specific to the Turkish context and the mapping of the state of Turkish cities' vulnerabilities from a climatic standpoint are the study's two main accomplishments. The findings indicate that since socio-economic status, infrastructure transportation, and meteorological conditions were among the variables influencing the communities' existing susceptibility, particularly in the coastal cities and the cities in the south-eastern region of Turkey, improving people's quality of life, health, and infrastructure and alleviating the impacts of meteorological conditions would be a key policy toward preparing the Turkish population to deal with the effects of climate change. Investing in urban infrastructure, such as health, physical infrastructure, communication networks, and disaster preparedness, is another crucial step towards enhancing adaptive capacity, since drought, forest fires, and flood events may become more regular in the regions that are associated with high vulnerability. An increase in the severity of climate events suggests a threat to the food security of the area, while there is also an increased risk of death for the local population which may be influenced by severe floods, windstorms, and forest fires. The employed methodology was designed to be straightforward to understand and apply for policy making, allowing decision-makers at the local and provincial levels to adopt it. The development of CCVI at the province level made it possible to target resources and policies based on geographical differences and assess vulnerability comprehensively. Based on the availability of data, this approach can be used in the future in more local areas with the use of local indicators that show different socio-spatial features.

The abstract of this paper was presented at the Geographic Perspectives on Climate Change Mitigation in Urban and Rural Environments (GCUE) Conference -1st Edition which was held on the 25^th -27^th of April 2024.