Introduction
The increasing frequency and severity of climate disasters globally necessitate enhanced efforts to strengthen the response and adaptation capacities of various entities, including national and local governments, industries, and public institutions. The effects and consequences of climate risks differ based on regional characteristics, highlighting the critical roles of central and local governments in light of the urgency and spatial significance of adaptation measures (Mitchell and Ibrahim, 2010). Local governments play a key role in adapting to the climate crisis (Koh, 2017; Lee, 2017; Park, 2023). In South Korea, local governments have implemented measures to address climate crises at the regional level, including the establishment of regional climate crisis adaptation measures aligned with the Framework Act on Carbon Neutrality and Green Growth for coping with the Climate Crisis. However, despite the institutional framework, many local governments still adopt adaptation measures in a formalistic manner, without sufficiently reflecting local climate risks or administrative capabilities. Effective implementation of these measures remains challenging in the absence of national support (Im et al., 2013). This indicates a structural gap between the policy necessity of localized adaptation and the actual effectiveness of its implementation. Furthermore, the limitations of existing evaluation systems aggravate this gap. Each year, progress toward established goals and the allocation of budgetary resources are reviewed by evaluating the implementation of adaptation measures by local governments; however, these reviews show minimal correlation with the direct effects of these measures and their findings are rarely applied in practice. Assessing whether actual climate risks are reduced or if appropriate responses are implemented through budget execution remains challenging (Goonesekera and Olazabl, 2022). Although national-level indicators and international indices such as ND-GAIN or FEMA’s NRI offer some reference points, they are not tailored to the Korean local context. Following more than a decade of national climate change adaptation efforts, a persistent need remains to enhance the capabilities and infrastructure of local governments. To address this, a practical evaluation system is essential for comprehensively assessing the adaptation levels and capacities of local governments, thereby overcoming the limitations of existing measures. Systematic methods are necessary to inform decision-making about reducing climate crisis risks and to determine whether additional measures can enhance adaptation levels, contingent upon the adaptation efforts of local governments.
In this study, we aimed to present an assessment system capable of comprehensively diagnosing potential regional climate risks and evaluating the extent of regional efforts to mitigate them. To this end, the concept of a "climate crisis adaptation index" was defined to comprehensively assess climate change impacts, damage, vulnerability, and response levels by examining relevant national and international cases. In addition, evaluation indicators and methodologies for calculating the index were set. Furthermore, a range of statistical and adaptation measure implementation inspection data were utilized for the pilot application of the adaptation index. The scope of the pilot application of the adaptation index encompassed 229 local governments in South Korea. The temporal scope was established for 2022, with heat waves selected as the target climate disasters. Heat waves were classified as natural disasters in 2018. Since then, measures to mitigate heat wave damage have been implemented in other regions. In fact, heat waves now account for approximately 48% of all adaptation measures implemented by local governments (Park, 2023), indicating a rising level of interest and urgency. Accordingly, a comprehensive assessment of heat wave risks and the implementation of mitigation measures are essential, guided by the novel adaptation index proposed in this study. The findings of this study are intended to support the evaluation of climate risk significance and vulnerability and to facilitate the development of effective, regionally tailored adaptation measures by local governments.
Research Methods
Fig. 1 illustrates the research method used in this study. First, the concept of the “climate crisis adaptation index” was established to comprehensively assess climate crisis impacts and damages, vulnerability levels, and the effectiveness of adaptation efforts. This index was formulated by analyzing diverse methodologies (drawing on both domestic and overseas case studies and previous research) to evaluate the adaptive capacities of national or local governments and outcomes of implemented measures. Second, in accordance with the climate crisis adaptation index concept defined in this study, evaluation indicators for the index were identified by compiling indicators used in related research, leading to the construction of a relevant database (DB). Third, the outliers were removed from the constructed DB. Scoring criteria were prepared based on the average (μ) and standard deviation (σ) of the evaluation indicators, excluding outliers, and the score was calculated accordingly. Cook’s distance, which is a regression diagnostic technique, was employed to remove outliers. When the Cook’s distance value reached or exceeded 4/n (n = 229), the case was classified as influential and subsequently removed (Lee and Noh, 2013). The scoring criteria were divided into ten categories using standard deviation intervals derived from the normal distribution curve and definitions of extreme climate, thereby ensuring coverage of more than 90% of the dataset (Table 1). Scores ranged from 0 to 1, with values closer to 1 indicating higher relevance and a median score anchored at 0.5.
(Eq. 1)
Fourth, after calculating the scores for each evaluation indicator, the indices for the four areas (damage, hazard, vulnerability, and adaptation implementation performance) were determined using the arithmetic mean of each evaluation score. Finally, the climate crisis adaptation index was derived by integrating these four indices. Relevant previous studies were examined to determine the calculation formula for the adaptation index (Song and Lee, 2012; Lee et al., 2016; Cho and Choi, 2018). In addition, the Euclidean distance-based scoring formula, which facilitates the calculation of a comprehensive index by integrating multiple indicators, was referenced. This method quantifies the distance from 1 (the ideal state) for each indicator, facilitating a comprehensive assessment of the factors representing different attributes (Cho and Choi, 2018). The proposed climate crisis adaptation index ranges from 0 to 1, with values closer to 1 indicating higher relevance. The calculation formula employed a Euclidean distance-based scoring formula, as presented in Eq. 1.
Results and Discussion
The definition of the climate crisis adaptation index and the selection of evaluation indicators are presented in this section, rather than in the Research Methods, to emphasize that they are not simple applications of existing frameworks. Specifically, the index structure, calculation formula, and indicator composition were newly developed through the study's analytical process, based on comparative case analysis, and are therefore presented as key research outcomes.
Definition of the climate crisis adaptation index
Analysis of domestic and international cases and previous studies revealed that the regional safety index and regional safety level established by the Ministry of the Interior and Safety for evaluating the safety levels and capabilities of local governments in the field of disaster safety serve as representative indices in Korea. Another study evaluated the climate change risks and capabilities of industries and compared the results among countries (Myeong et al., 2012). At the local government level, the impacts of climate change and their responses have been evaluated using various methodologies, such as quantitative and narrative assessments, specifically for metropolitan governments (Lee and Choi, 2014; Kim et al., 2016). In Seoul and Gyeonggi-do, indices for assessing climate change causes and response efforts have been developed and applied (Ministry of Environment, 2014). However, few studies have evaluated the effects of adaptation measures in climate change-related areas in Korea. The scope of this research was limited to metropolitan areas and specific local governments. In international countries, the Notre Dame’s Global Adaptation Initiative (ND-GAIN), presented by the University of Notre Dame in the United States, serves as a representative index. It assesses the adaptation capacity of the country by comprehensively evaluating climate change vulnerability and preparedness. The Federal Emergency Management Agency (FEMA) of the United States has assessed a comprehensive risk index composed of natural disaster factors, disaster increase factors, and disaster reduction factor indicators using the National Risk Index (NRI). The Department for Environment, Food, and Rural Affairs (Defra) in the United Kingdom has qualitatively evaluated the existence of reports or documents associated with the implementation of adaptation measures using National Indicator 188 (N188). In international countries, integrated evaluation tools for comparisons among nations, rather than among local governments, are predominantly presented, while the scale and type of information provided differs from those presented in South Korea.
The above indicates the absence of criteria for comprehensively evaluating regional climate change risks and the effects of the establishment and implementation of adaptation measures applicable to the contexts of local governments in Korea. Despite different research purposes and fields, most studies have found that negative indicators linked to hazards or vulnerability serve to increase risks, whereas positive indicators related to efforts for adaptation or mitigation serve to reduce risks. Consequently, in this study, damage indicators represented the results of climate risks, hazard indicators denoted the disaster risks caused by climate change, while vulnerability indicators were sensitive to risks and could increase damage from a social perspective. In addition, adaptation implementation performance indicators representing the administrative efforts of local governments for climate crisis adaptation and performance were set as evaluation indicators for the climate crisis adaptation index. The concept of the climate crisis adaptation index was subsequently defined by differentiating between directions and characteristics, allowing damage, hazard, and vulnerability indicators to function in the negative direction while the adaptation implementation performance indicator operates in the positive direction (Eq. 2).
Selection of evaluation indicators for the climate crisis adaptation index
This study focused on heat waves to calculate the climate crisis adaptation index, and accordingly, heat wave-related literature was reviewed. Final evaluation indicators were selected from those actively used in previous studies, prioritizing indicators with quantitative data avail ability and clear, reliable sources. Based on the literature review, the direct human casualty status, including heat-related illnesses (e.g., heat stroke and sunstroke caused by high temperatures), was primarily used as a damage indicator (Kim et al., 2016). For hazard indicators, the average value or change rate of the daily maximum temperature, those of the daily minimum temperature, and weather statistics (e.g., heat wave days, tropical nights, relative humidity, heat index, heat wave duration index, and discomfort index) were utilized (Myeong et al., 2012; Lee and Choi, 2014; Kim et al., 2016; Kim et al., 2022). The vulnerability indicators were demographic or socioeconomic factors that exacerbated damage, while unemployed individuals, elderly individuals living alone, outdoor workers, public aid recipients, population density, population in hazardous areas, people with disabilities, people with low education levels, the elderly, infants, and individuals with cardiovascular diseases were utilized (Myeong et al., 2012; Kim et al., 2022). After reviewing for overlap and making necessary adjustments, eight indicators—comprising one damage indicator, two hazard indicators, and five vulnerability indicators—were finalized based on their data availability for at least three years and their applicability across 229 local governments within the study scope (Table 2). The adaptation implementation performance indicator employed the implementation achievement rate (%), reflecting the degree of achievement compared to the achievement plan. This was achieved by referring to the items related to "local climate risk adaptation measure implementation inspections" to identify those applicable across all local governments. This approach acknowledged the variability in the nature, budget, and performance target types or units, which differed based on the sizes and circumstances of local governments (Park, 2023). Based on this, a "heat wave"-related keyword search was used to classify the adaptation measures established by local governments into eight categories: "heat wave shelter measures," "heat wave reduction facility measures,” "green space and park measures," "vulnerable population care and support measures," "energy (air conditioning) welfare and facility (e.g., daycare centers and senior centers) support measures," "chronic diseases (e.g., heat-related and cardiovascular diseases) and medical service measures," "climate adaptation education/promotion measures," and "other measures." When the implementation achievement rate exceeded 100% (overachievement), a cap of 100% was applied, and the comprehensive average of the eight measures was utilized as the final adaptation implementation performance indicator.
Application of the climate crisis adaptation index
Index calculation by area
Before the climate crisis adaptation index was calculated, 91 outliers were excluded from the analysis. The main status of each evaluation indicator and the results of calculating the index for each sector based on the arithmetic mean of the evaluation scores were as follows. First, the heat-related illness occurrence rate, representing a damage indicator, averaged 2.53 persons/100,000 people with a standard deviation of 1.87 persons/100,000 people. It was high in Gurye-gun in Jeonnam (36.50 persons/100,000 people), Cheolwon-gun in Gangwon (28.40 persons/100,000 people), and Jinan-gun in Jeonbuk (24.44 persons/100,000 people) and lowest (0.00 person/100,000 people) in Gangdong-gu in Seoul, Jung-gu in Busan, Jung-gu in Daegu, Dong-gu/Seo-gu in Gwangju, Gwacheon-si in Gyeonggi, Goseong-gun in Gangwon, Jeungpyeong-gun in Chungbuk, Wando-gun in Jeonnam, and Yeongcheon-si/Yeongyanggun/Goryeong-gun in Gyeongbuk (Fig. 2). The damage index scored 1.00 in 26.6% of the area and 0.00 in 5.2% of the area (Fig. 3).
Second, the number of heat wave warnings, a hazard indicator, exhibited an average of 17.67 and a standard deviation of 6.96. It was highest in Seo-gu in Daegu (37.70), Jun-gu in Daegu (37.45), and Seo-gu in Gwangju (33.95), and lowest in Ulleung-gun in Gyeongbuk (0.00), Pyeongchang-gun in Gangwon (0.65), and Taebaek-si in Gangwon (0.85). This indicated that it was typically low in Gangwon (Fig. 4). The average number of tropical nights averaged 24.90 and a standard deviation of 8.88. Although it was high in Seoguipo-si in Jeju (43.46), Jeju-si in Jeju (42.38), and Yeongdeungpo-gu in Seoul (39.11), it was low in Taebaek-si in Gangwon (0.05), Jeongseon-gun in Gangwon (0.10), and Pyeongchang-gun in Gangwon (0.16) (Fig. 5). The hazard index, calculated as the arithmetic average of the evaluation scores of the two hazard indicators, was highest in Nam-gu/Seo-gu in Gwangju and Seo-gu/Jung-gu in Daegu (0.90) and lowest in Yanggugun/Jeongseon-gun/Hwacheon-gun/Inje-gun/Taebaek-si/Pyeongchang-gun in Gangwon (0.10) (Fig. 6).
Third, the proportion of individuals aged 65 years and older, representing a vulnerability indicator, demonstrated an average of 21.18% with a standard deviation of 7.78%. Although it was high in Euiseong-gun in Gyeongbuk (44.26%), Goheung-gun in Jeonnam (43.18%), and Gunwigun in Daegu (43.12%), it was low in Hwaseong-si in Gyeonggi (9.81%), Buk-gu in Ulsan (9.83%), and Gwangsangu in Gwangju and Sejong-si (10.48%) (Fig. 7). The proportion of individuals aged 14 years and younger averaged 10.49% with a standard deviation of 2.46%. It was high in Gangseo-gu in Busan (19.16%), Sejong-si (19.16%), and Hwaseong-si in Gyeonggi (17.07%) and low in Gunwi-gun in Daegu (4.44%), Jung-gu in Busan (4.56%), and Euiseong-gun in Gyeongbuk (5.46%) (Fig. 8). The proportion of people with disabilities averaged 5.91% with a standard deviation of 2.17%. It was high in Euiseong-gun in Gyeongbuk (12.26%), Yeongyang-gun in Gyeongbuk (12.10%), and Hampyeong-gun in Jeonnam (11.85%) and low in Seocho-gu in Seoul (2.51%), Gangnam-gu in Seoul (2.86%), and Gwacheon-si in Gyeonggi (2.87%) (Fig. 9). The proportion of workers in agriculture, forestry, and fisheries averaged 0.15% with a standard deviation of 0.23%. It was high in Jung-gu in Busan (2.85%), Cheongdo-gun in Gyeongbuk (1.59%), and Naju-si in Jeonnam (1.18%) and lowest in nine districts in Seoul, Michuhol-gu/Yeongsugu in Incheon, and Seo-gu/Suseong-gu in Daegu (0.00%) (Fig. 10). Finally, the proportion of public aid recipients averaged 4.97% with a standard deviation of 1.67%. Although it was high in Dong-gu (11.98%), Yeongdo-gu (11.87%), and Jung-gu in Busan (11.02%), it was low in Gyeonggi (1.32%), Yongin-si in Gyeonggi (1.70%), and Hwaseong-si in Gyeonggi (1.73%) (Fig. 11). The vulnerability index, calculated as the arithmetic mean of the evaluation scores of the five vulnerability indicators, was highest in Gimje-si in Jeonbuk (0.86) and Imsil-gun in Jeonbuk (0.84) and lowest in Mapo-gu in Seoul (0.30) and Gwangjingu in Seoul (0.32) (Fig. 12).
Fourth, the comprehensive average implementation achievement rate, representing the adaptation implementation performance indicator, averaged 51.26% with a standard deviation of 16.46%. It was the highest in Gwanak-gu in Seoul (100.00%), with an implementation achievement rate of 100.00% for all eight items (Fig. 13). It was also high in Gimcheon-si in Gyeongbuk (98.25%), with implementation achievement rates of 86.00% for one item (energy welfare and facility support) and 100.00% for the other seven items. However, it was lowest in Danyang-gun in Chungbuk (0.00%). Among the eight items, only one (vulnerable population care support) had an achievement plan, with an achievement rate of 0.00%. Seocheon-gun in Chungnam (6.25%) also had achievement plans for only three items (heat wave reduction facilities, vulnerable population care support, and climate adaptation education/promotion), while the achievement rate was 50.00% for one item (vulnerable population care) and 0.00% for the rest. The adaptation implementation performance index that applied the scoring criteria was highest in Gwanak-gu/Songpa-gu in Seoul, Suyeong-gu in Busan, Dong-gu in Ulsan, Suseong-gu in Daegu, Asan-si in Chungnam, Gimcheon-si in Gyeongbuk, and Yeonggwang-gun in Jeonnam (1.00) and lowest in Danyang-gun in Chungbuk (0.00), Ongjin-gun/Ganghwa-gun in Incheon, Seosan-si/Seocheon-gun in Chungnam, Gyeongju-si in Gyeongbuk, Gapyeong-gun in Gyeonggi, Iksan-si in Jeonbuk, Hamyang-gun in Gyeongnam, and Wando-gun/Goheung-gun/Jangheung-gun in Jeonnam (0.10) (Fig. 14).
Calculation of the final climate crisis adaptation index
The final climate crisis adaptation index was derived by combining the indices by area (damage, hazard, vulnerability, and adaptation implementation performance). Grades were assigned for evaluation by local government as follows: A (15%), B (20%), C (30%), D (20%), and E (15%). This was based on the ratio of the calculated value. In addition, the results of reflecting versus not reflecting the adaptation implementation performance indicator were compared to analyze variations in evaluation results based on adaptation implementation performance.
First, the climate crisis adaptation index calculation results, which reflected the adaptation implementation performance, ranged from 0.2385 to 0.7850 (Fig. 15). Table 3 presents the top 20 and bottom 20 regions according to the index. The areas exhibiting high adaptation index values were Goseong-gun in Gangwon (0.7850), Yeongyang-gun in Gyeongbuk (0.7315), and Jung-gu in Daegu (0.6967). They primarily exhibited very low heat-related illness occurrence rates, which corresponded to the damage indicators. In particular, the adaptation index was evaluated as high in the Gangwon region because of the limited occurrence of heat wave warnings and tropical nights based on its geographical conditions. Meanwhile, Gwacheon-si in Gyeonggi (0.6564) and Jung-gu in Busan (0.6149) were classified as grade A owing to their low damage, hazard, and vulnerability indices despite a low adaptation implementation performance index of 0.40. Suseong-gu in Daegu (0.6748) and Jung-gu in Daegu (0.6967) were assigned grade A due to their very high adaptation implementation performance indices despite high hazard index values of 0.80 and 0.90, respectively. In addition, the evaluation of some areas was positively affected by various adaptation measure items (six or more) and a high implementation achievement rate (90% or higher) despite the high hazard index values. However, the adaptation index was the lowest in Changnyeong-gun, Gyeongnam (0.2385), followed by Iksan-si in Jeonbuk (0.2388), Nonsan-si in Chungnam (0.2450), and Hampyeong-gun in Jeonnam (0.273). The adaptation index values were predominantly low in the Jeonnam region. The proportion of vulnerable populations (e.g., people aged 65 and over; workers in agriculture, forestry, and fisheries; and public aid recipients) was high, coinciding with a large number of heat wave warnings. In particular, Yesan-gun in Chungnam (0.4301) and Hapcheon-gun in Gyeongnam (0.4243) were classified as grade D due to the high damage, hazard, and vulnerability indices despite high their adaptation implementation performance index values of 0.70 and 0.60, respectively. Most low-ranked areas contained merely two to three adaptation measure items. For areas classified as grade E, the comprehensive average implementation achievement rates did not reach 30%.
The calculation results, which employed the damage, hazard, and vulnerability indicators solely without accounting for adaptation implementation performance, ranged from 0.1409 to 0.7799 (Fig. 16). Goseong-gun in Gangwon exhibited the highest value, whereas Gimje-si in Jeonbuk recorded the lowest value. And, for specific comparisons, local governments were classified into groups based on grade changes (those with an increase, a decrease, and no change) considering the adaptation implementation performance indicator. The group exhibiting an increase in grade comprised 60 areas (26.2%). Two areas improved by three grades (E→B), ten areas improved by two grades (C, D, E→A, B, C), and 48 areas improved by one grade (B, C, D, E→A, B, C, D) (Fig. 17). For areas with an increase of three grades, the adaptation implementation performance index was recorded at 1.00, despite a high damage index and moderate hazard and vulnerability indices. For areas with an increase of two grades, both the hazard and average adaptation implementation performance indices were high (0.85). For areas with an increase of one grade, both the average adaptation implementation performance and damage indices were high (0.74). In contrast, the group exhibiting a decrease in grade comprised 71 areas (31.0%). Three areas decreased by two grades (A, B, C → C, D, E) and 68 areas decreased by one grade (A, B, C, D → B, C, D, E) (Fig. 18). The areas with a decrease of two grades were identified as the lowest levels, as evidenced by an average adaptation implementation performance index of 0.16, despite very high damage and vulnerability indices. For the areas with a decrease of one grade, the decrease in grade was linked to a very low average adaptation implementation performance index (0.30) despite high damage, hazard, and vulnerability indices. The group exhibiting no change in the evaluation results comprised 98 areas (43.0%). The analysis of variance revealed no significant differences in the damage index (p = .094) or vulnerability index (p = .209) based on the presence/absence of grade changes; however, significant differences were observed in the hazard index (p = 0.000) and adaptation implementation performance index (p = 0.001).
Conclusion
In response to the increasing number of climate disasters globally, this study was conducted to evaluate and strengthen the adaptation capacity of local governments to establish adaptation measures. The main results and implications of this study are outlined below.
First, this study introduces an adaptation index concept to diagnose climate risks and adaptation levels in a relatively simple and comprehensive manner at the regional level by examining both domestic and international cases. Based on this, an evaluation system was developed to comprehensively reflect the directions and characteristics of each evaluation indicator associated with climate risks. Most importantly, the results of annual inspections of local climate crisis adaptation measures were directly utilized to inform the qualitative assessment of regional adaptation efforts in conjunction with a statistics-based quantitative evaluation. This approach differs from that used in previous studies. Second, this study intuitively presented areas with strong heat wave risk adaptation capabilities and those lacking such capabilities. In addition, we analyzed the evaluation results through the pilot application of the adaptation index to local governments in Korea. Areas with a high adaptation index demonstrated considerable adaptation efforts in response to climate risks. They exhibited low damage and vulnerability levels and high adaptation measure implementation achievement rates. However, areas with a low adaptation index exhibited numerous vulnerable populations and low adaptation measure implementation achievement rates. In particular, some areas received high evaluation scores owing to high adaptation measure implementation achievement rates despite facing challenging climate factors. Conversely, other areas received relatively low evaluation scores owing to low adaptation measure implementation achievement rates despite minimal vulnerable populations. Moreover, variations in grades due to local government efforts depended on the hazard and adaptation implementation performance indices. This suggests that climatic factors alone are insufficient for risk assessment, underscoring the need for both long-and short-term monitoring to assess the effects of measures implemented through local administrative efforts. Third, regarding the adaptation implementation performance of local governments, the implementation achievement rate of vulnerable population-related measures was high in areas with a high proportion of such populations; however, specific measures (e.g., heat wave shelters and energy welfare) were inadequate. This indicates that key measures are important in establishing adaptation measures; nevertheless, it is imperative to implement more practical measures tailored to regional characteristics by incorporating necessary enhancements and relationships.
While this study provides a meaningful attempt to construct a Climate Crisis Adaptation Index reflecting implementation performance, several limitations remain. First, although the selected indicators are grounded in national policy goals and previous literature, the focus on data availability at the local level led to the exclusion of key ecological variables like land surface temperature and vegetation cover. To enhance the robustness of the evaluation framework, engage local governments and experts to refine indicators and incorporate landscape and urban environmental factors such as green space and land use. These spatial-environmental factors are increasingly recognized as vital for understanding vulnerability and resilience to climate extremes, including heat waves(Meerow and Newell, 2017; Eo et al., 2021), and their inclusion could strengthen the index’s explanatory power from a human–environment systems perspective. Second, this study applied equal weighting to all indicators to ensure transparency and methodological simplicity. Future research could explore differentiated weighting schemes based on expert judgment or analytic approaches such as the Analytic Hierarchy Process (AHP). Overall, a forward-looking discussion on indicator enhancement and weighting strategies will be essential to improve the robustness, applicability, and long-term relevance of the Climate Crisis Adaptation Index. Future research on continuous improvement, particularly in relation to adaptation measure implementation inspections and expanded application to additional climate risks (e.g., floods and droughts), could enhance the utility of the proposed adaptation index concept. Such advancements are expected to foster greater engagement in climate adaptation among local governments and support evidence-based decision-making.
PDF Links
PubReader
ePub Link
Full text via DOI
Download Citation
