1. Introduction
Amid intensifying global climate change, climate risk and energy security have become critical factors shaping the stable functioning of human society and remain top priorities for governments around the world. Climate risks encompass the adverse consequences for ecological systems, socioeconomic activities, and financial stability that stem from climate change. These risks are characterized by the interplay of climate-related hazards (e.g., extreme weather events and gradual shifts in climate patterns) and the vulnerability and exposure of both natural and human systems. They can be classified into both physical risks and transition risks [
1]. Physical risks refer to the direct damage to the environment caused by natural disasters [
2], while transition risks arise from systemic changes implemented to mitigate climate change [
3], such as sunk costs associated with unsuccessful energy transitions.
Energy security refers to the continuous availability of energy at an affordable price. This involves mitigating risks of supply disruptions, drastic price volatility, or system failures that may arise from factors like resource scarcity, geopolitical conflicts, and environmental concerns [
4,
5]. Reasonable energy prices can help maintain price stability and improve people’s well-being. Stable energy supply can maintain the normal operation of production and consumption and promote sustainable economic growth. Especially in the current context of increasing geopolitical risks, increasing energy diversification and introducing alternative energy sources can effectively enhance the stability of energy supply. On the other hand, considering the negative impact of energy extraction on the environment and indirectly affecting the sustainability of energy acquisition, the sustainability of energy is important in ensuring the security of energy system.
In recent years, a series of extreme weather events disrupting energy supply have been reported globally. For instance, China’s Sichuan province was affected by a prolonged period of extreme high temperatures in 2022 with persistent power shortages [
6]. Similarly, the Texas state of the U.S. was affected by a winter storm in 2021 with widespread power paralysis and blackouts [
7]. These events not only highlight the threat of climate risk to energy security but also prompt their governments to re-examine their energy security strategies and explore the linkage between climate risk and energy security so that negative impacts of climate risk on energy security can be mitigated. China is highly susceptible to climate risks and faces considerable challenges in securing its energy supply while pursuing a low-carbon transition. The improvement of energy security can also help China achieve emission reduction in the energy industry to a certain extent, although China currently adopts various methods to reduce carbon emissions, such as industrial upgrading, promotion of CCUS (carbon capture, utilization, and storage) technology [
8], and use of alternative energy [
9]. But ultimately, without stable energy supply at reasonable prices, these measures to reduce carbon emissions cannot be implemented. Therefore, comprehensively understanding the impact of climate risk on energy security is of paramount importance.
The purpose of this study is to investigate the linkage between climate risk and energy security and figure out how climate risk affects energy security by using a panel dataset of 30 Chinese provinces covering a period from 2006 to 2022. We then use an instrumental variable generalized method of moments (IV-GMM) model to examine the marginal impact of climate risk on energy security. Panel quantile regression models and mediated effects models are also used to investigate the asymmetric linkages and internal influence mechanism.
There are several academic contributions in this study. First, although a number of scholars have studied energy security, few have investigated the relationship between energy security and climate risk. By revealing the linkage between climate risk and energy security, this study provides valuable insights to guide energy security-based investment and infrastructure construction. Second, this study demonstrates that climate risk can make a greater contribution to energy security, which helps policymakers identify regions with low levels of energy security. Third, we examine both direct and indirect impacts of climate risk on energy security. The direct impact is investigated from four dimensions of energy security (i.e., accessibility, affordability, sustainability, and technology and efficiency). The indirect impacts focus on how climate risk facilitates energy transition and improves energy efficiency. Unlike previous studies that utilized the International Energy Agency (IEA) framework to assess energy security, our research extends beyond current security considerations. We integrate potential upgrades and technological efficiency within the energy system, offering evaluations for the future development of regional energy systems. Finally, we propose several policy recommendations to reduce the impacts of climate risk on energy security. The whole paper is organized as follows: After this introduction section,
Section 2 presents a literature review related to climate risk and energy security.
Section 3 describes research methods and data.
Section 4 presents research results and discusses moderating and mediating effects in the climate risk–energy security nexus.
Section 5 proposes policy recommendations.
Section 6 draws research conclusions.
2. Literature Review
2.1. The Literature on Climate Risk
The escalating prevalence of extreme weather events coupled with the persistent challenges of climate mitigation and environmental degradation are key critical global climate risks [
10]. Consequently, the potential socio-humanitarian ramifications of climate risk have garnered more scholarly attentions.
Several studies have examined the nexus between climate risk and other relevant concepts, with a predominant focus on economic interrelationships. For example, Ren et al. [
10] analyzed the linkage between climate risk and corporate environmental performance, while Banerjee [
11] investigated the interconnections between climate, geopolitical risks, and global commodity markets. Wang et al. [
12] examined the impact of compound drought–heatwave events (CDHEs) on small nations, demonstrating the acute vulnerability of food security. Given China’s extensive landmass within the monsoon climate zone, the nation is particularly susceptible to substantial climate risks, prompting a surge in derivative research. In this regard, Rong et al. [
13] assessed the impact of climate risk on China’s renewable energy stock, and Ho et al. [
14] investigated the relationship between climate risk and interest rate spreads within China’s credit market. These scholars have combined climate risk with another physical concept to quantify the diverse impacts of climate risk, providing a solid foundation for a correct understanding of climate risk, especially its potential risks.
Although these studies are all conducted on the topic of climate risk, there are certain differences in their methodologies, especially in the definition of the concept of climate risk and data processing. The methodological approaches to quantifying climate risk exhibit considerable heterogeneity. Shen et al. [
15] utilized the EM-DAT database, which relies on hazard site reporting, to assess four climate phenomena: storms, floods, extreme temperatures, and wildfires. Lee et al. [
16] employed the Climate Risk Index (CRI) to assess regional climate risks, which covers floods, droughts, typhoons, freezes, and extreme temperatures. Moreover, He et al. [
17] focused exclusively on anomalous temperature scenarios. Based on the above summary, it can be observed that each scholar has a different approach to defining climate risk. This difference is specifically manifested in subjective data and variations in the scope of climate risk definition. The subjectivity issue is reflected in the fact that some scholars use datasets based on hazard site reporting, which are obtained through the cognition and understanding of disaster by site staff, thus carrying a certain degree of subjectivity. The difference in definition scope is reflected in the fact that some scholars focus on the temperature field, while others focus on climate phenomena such as storms. This has brought a certain degree of ambiguity to the advancement of research on climate risk-related topics.
2.2. The Literature on Energy Security
The conceptualization of energy security has been studied for a long time. Initially, this concept mainly centered on the secure provision of oil resources, as exemplified by Georgiou’s [
18] analysis of U.S. energy security within the context of the 1990s global oil supply system. However, the proliferation of novel energy technologies has broadened the scope of energy security, encompassing a diverse array of energy sources, including thermal, photovoltaic, wind, and hydroelectric power [
19]. The diversification of energy types has broadened the definition of energy security, shifting its focus beyond merely the reasonable price and stable supply of oil. Contemporary research on energy security primarily investigates price rationality and supply stability within the synergistic context of multiple regional energy systems. The International Energy Agency (IEA), for instance, articulates energy security within the European Union through three fundamental dimensions: (1) the physical availability and accessibility of supply sources [
4], (2) affordability, and (3) long-term environmental sustainability. In alignment with this framework, researchers increasingly employ the Energy Security Diversity Index to assess regional energy security [
20]. Globally, established models such as those developed by the Joint Study Group on Security of Energy Supply (JESS) in the UK, the Energy Research Center of the Netherlands (ECN), the IEA, and the Asia-Pacific Energy Research Center (APERC) are frequently utilized [
21]. Contemporary energy indicator systems often draw upon the IEA’s multidimensional perspective, extending the concepts of availability, affordability, and environmental sustainability. For instance, Zhang et al. [
22] developed an evaluation framework for energy security in Chinese provinces that covers five primary dimensions—energy supply, consumption efficiency, economic resilience, environmental impact, and policy support—and includes 20 measurable sub-indicators. Similarly, Iyke [
23] employed the 2021 edition of the International Energy Security Risk Index, developed by the U.S. Chamber of Commerce’s Global Energy Institute, which aggregates multiple risk dimensions—including fuel import exposure, energy prices, and geopolitical factors—to measure national energy security.
Based on the above summary, it can be found that most scholars’ research on energy security has focused on using entropy weight method to quantitatively analyze energy security. However, there are certain differences in the indicator construction and data range used in the entropy weight method. Specifically, some scholars have given higher weight to policy guidance in indicator construction, while others place greater emphasis on the promotion of renewable energy. Some scholars study the international or continental scope of data, while others specialize in the scope of cities and districts, which has significant differences in data selection and analysis. However, overall, most research on energy security has clear frameworks and mature analytical approaches.
2.3. Research Gaps
Based on our literature analysis on both climate risk and energy security, we identified several research gaps. Firstly, studies applying the concept of climate risk to the economic impacts of energy security is relatively limited. Secondly, within the domain of energy security, while indicator-based assessments have gained widespread acceptance, ambiguities persist in the selection and construction of specific indicators. Notably, a disproportionate emphasis is placed on the availability and affordability dimensions of energy security, as articulated by the International Energy Agency, while the environmental sustainability dimension receives less attention. Moreover, the integration of forward-looking technological and efficient indicators, crucial for long-term energy security, remains underdeveloped. Our research employs an indicator system that, while aligned with the traditional IEA framework, specifically aims to conduct an environmental sustainability analysis and a forward-looking technical efficiency analysis of energy security. This approach allows for us to offer regional development recommendations grounded in environmental sustainability and future foresight. Furthermore, while the application of indicator systems for energy security assessment is prevalent, the underlying drivers of regional disparities in energy security levels are often overlooked. In the Discussion Section, we include both direct and indirect impact analyses to partially delineate the causal link from climate risk to energy security, aiming for a more comprehensive understanding.
Consequently, this study aims to fill these research gaps and propose innovative solutions. Our research takes China mainland as the boundary to study the availability, affordability, sustainability, technology, and efficiency of energy in 30 provinces of China in order to determine the determining factors of energy security level and reduce the inherent uncertainty in energy security analysis. This will help provide evidence-based recommendations and theoretical insights.
3. Methods and Data
3.1. Methods
Energy security indicators are developed using the following estimation model, which takes energy security, climate risk, and control variables into consideration:
where
denotes the level of energy security, which is a dependent variable.
denotes climate risk, which is a key independent variable in the model. In addition, there are four control variables in the model, including
, which represent economic growth, industrial development, official financial support, and human capital, respectively. These control variables are closely related with energy security and help reduce the omitted variable bias. In addition,
i and
t represent a Chinese province
i (30 provinces in China) and year
t within the range 2006–2022.
Based on Equation (1), we further adopt the logarithmic form of the above variables to analyze the nonlinear relationships between these variables and also to reduce the heteroskedasticity problem. This logarithmic form is shown in Equation (2).
where
β denote the marginal effects of each independent and control variable, with
as the intercept and
the error term. Specifically,
β1 reflects the marginal effect of
RISK on the dependent variable, the main focus of this study. As all variables are in natural logarithms, the coefficients
–
indicate elasticities, showing the percentage change in the dependent variable from a one-percent change in
RISK. The model includes province and year fixed effects (
), capturing unobserved heterogeneity. Traditional models (OLS, RE, and FE) were considered, but they inadequately address endogeneity from omitted variables and reverse causality, especially given the complexity of climate risk [
24]. Thus, we adopt the instrumental variable-generalized method of moments (IV-GMM), which merges IV (using instruments to resolve endogeneity) and GMM (applying orthogonality conditions for efficiency under heteroskedasticity and autocorrelation) [
25]. Lagged terms of
RISK serve as valid instruments to strengthen causal estimates.
This study investigates the internal mechanisms linking climate risk and energy security by assessing the technology and efficiency effects. Accordingly, renewable energy share and energy efficiency are selected as mediating variables. Renewable energy share serves as a key indicator of the energy technology transition process, while energy efficiency reflects the level of energy utilization. Based on these variables, we develop the following estimation framework; if the estimated coefficients for climate risk, renewable energy share, and energy efficiency are statistically significant, this indicates the existence of mediation effects via these channels, as detailed in Equations (3)–(6) [
26]:
where
lnET is the mediating effect of energy transition, which is measured by the ratio of renewable energy generation to the total electricity and thermal generation.
lnEE is the mediating effect of energy efficiency, which is the ratio of the total GDP to the total energy consumption. Equations (3) and (4) represent the channel of energy transition, while Equations (5) and (6) represent the channel of energy efficiency. It is worth noting that regarding certain variables, while other potentially relevant factors—such as energy prices or R&D investment—could influence energy security, they were excluded primarily due to data constraints and cross-provincial comparability. For example, energy prices in China are subject to central administrative control, resulting in limited variation across provinces and over time. Similarly, disaggregated and consistent data on provincial-level energy-related R&D investment are either unavailable or fragmented across sources, especially over long time horizons such as 2006–2022. Including such variables would compromise the temporal consistency and statistical robustness of the empirical model.
3.2. Data
3.2.1. Key Independent Variable
We use the climate risk dataset developed by Guo et al. [
27]. This dataset collects raw data from NOAA (National Oceanic and Atmospheric Administration of the U.S.) and is processed through missing value processing and secondary computation. This dataset defines the number of extreme low-temperature days (LTD), extreme high-temperature days (HTD), extreme rainfall days (ERD), drought days (EED), and takes a weighted approach to calculate physical climate risks. Compared to the subjective data based on disaster site reports used by many scholars, the dataset we use does not rely on the experience and judgment of disaster site reporters, but uses mathematical definitions to obtain climate risk data, which clearly avoids errors caused by subjective judgments. At the same time, meteorological data based on climate monitoring stations is higher-frequency and more accurate, which helps us improve the accuracy of our analysis. We selected the climate physical risk indices of 30 Chinese provinces with four types of extreme weather data from 2006 to 2022 and plotted the box plots and distribution maps of RISK, which is shown in
Figure 1.
3.2.2. Dependent Variables
The dependent variable
is a complex variable. Due to the large number of items that can affect energy security, it is necessary to fully consider all aspects of the energy system, including extraction, processing and conversion, capital investment, final consumption, and environmental impacts [
28]. Drawing upon the literature review, previous studies have predominantly employed the International Energy Agency’s (IEA) tripartite evaluation framework to quantify energy security, encompassing (1) the physical availability and accessibility of supply sources, (2) affordability, and (3) long-term environmental sustainability [
4]. Nevertheless, these antecedent studies have often placed a greater emphasis on accessibility and affordability, with comparatively less attention directed towards sustainability. This study endeavors to examine the influence of climate risks and the advancement of renewable energy technologies on energy security. Within this context, sustainability and the enhancements in technological efficiency engendered by the ongoing evolution of renewable energy assume paramount importance. To this end, our comprehensive indicator system draws inspiration from the IEA’s established evaluation dimensions, selecting four key sub-indicators—namely, availability, affordability, sustainability, and technology and efficiency—further delineated into twelve subordinate metrics, with the overarching aim of assessing China’s energy risk profile. The entropy weighting method was used to calculate the comprehensive score. The specific steps of the methodology are as follows:
Step 1: Normalize sub-indicator data (X:min–max scaling).
Step 2: Calculate proportion for each indicator across provinces.
Step 3: Compute entropy value and .
Step 4: Determine divergence and weight .
Step 5: Comprehensive energy security index calculation .
The first sub-indicator is accessibility. Energy accessibility encompasses security of supply, external dependence, trade dominance, and diversity. These indicators are directly related to a stable energy supply in the region. The second sub-indicator is affordability, where the price of energy and the affordability of the region’s energy needs are critical to regional energy system as a whole. The third sub-indicator is sustainability, which represents the environmental impact of energy consumption, which is recognized as a major source of various environmental pollutants [
29]. On the other hand, the development of new energy holds the potential to substantially bolster sustainability [
30]. The fourth sub-indicator is technology and efficiency, which is because technological development can represent, to a certain extent, the future trend of energy security since it contributes to the efficiency of energy production, conversion, transmission, and consumption [
29].
Table 1 lists the comprehensive indicator system, relevant definitions, and references for indicator construction. Furthermore, utilizing the entropy weighting method,
Figure 2 illustrates both spatial and temporal distribution of the energy security index and its constituent sub-indicators across 30 Chinese provinces. Specifically, panels (a), (b), (c), and (d) of
Figure 2 represent the data for years of 2006, 2012, 2017, and 2022, respectively, thereby elucidating the evolving trends of
RISK across different Chinese regions.
3.2.3. Control Variables
With respect to the control variables, we include four indicators: economic growth (GDP) [
36], value-added of the tertiary industry (TER) [
37], government fiscal expenditure (FE) [
25], and human capital (HC), represented by the number of college students per capita [
38]. These variables were selected based on their theoretical relevance to both climate risk and energy security, as well as the robustness and availability of data at the provincial level across time. GDP reflects the economic scale and energy demand intensity, influencing both the exposure and adaptive capacity to climate shocks. TER captures the structural composition of the economy, where a higher share of services often implies reduced energy intensity and different risk transmission mechanisms. FE serves as a proxy for government capacity to invest in energy infrastructure, disaster response, and climate adaptation. HC is included to reflect the region’s knowledge base and adaptive capability in the face of compound risks, which has been shown to influence both energy technology adoption and resilience. These variables have been widely adopted in the literature on energy–climate interactions and align with prior empirical models assessing climate risks, low-carbon energy transitions, and energy resilience [
39].
All the data were collected from China Statistical Yearbooks, China Provincial Statistical Yearbooks, and China Labor Statistical Yearbooks. Thirty Chinese provinces were investigated, excluding Tibet, Hong Kong, Macau, and Taiwan due to the lack of relevant data. The study period is from 2006 to 2022.
Table 2 lists the descriptions of these variables.
Table 2 presents descriptive statistics for the key variables. The logged energy security (ln ES) averages −0.932 (Std. = 0.211), reflecting moderate variability and a bounded original scale (0–1). Climate risk (ln RISK) shows significant heterogeneity across provinces (mean = 6.809; Std. = 1.66; range: 1.541–10.61), indicating uneven regional climate vulnerabilities. Economic development (ln GDP) also exhibits substantial disparities (mean = 9.237; Std. = 1.065; range: 5.953–11.59), consistent with known regional inequalities in China. The logged energy transition indicator (ln TER) varies widely (mean = −2.365; range: −4.937 to −0.525), capturing provincial differences in renewable energy penetration. Similarly, energy efficiency (ln FE: mean = 7.661; range: 4.658–9.758) and human capital (ln HC: mean = 4.022; range: 0.959–5.447) display pronounced regional variability.
4. Results
4.1. Baseline Regression Results
Table 3 presents baseline regression results from five econometric specifications: OLS, FE, RE, FGLS, and IV-GMM. While conventional panel data estimators face methodological limitations in addressing endogeneity, cross-sectional heterogeneity, and serial correlation [
30], the IV-GMM framework demonstrates superior econometric specification by systematically resolving these identification challenges through its generalized moment conditions. To rigorously validate the instrumental variable (IV) selection, we implement two diagnostic tests: the Kleibergen–Paap Lagrange Multiplier (KP-LM) test for under-identification and the Kleibergen–Paap Wald (KP-Wald) test for weak instrument assessment [
40]. The empirical diagnostics reveal a statistically significant KP-LM test statistic, with the KP-Wald statistic exceeding the Stock–Yogo critical thresholds. These results collectively validate the appropriateness of the IV-GMM methodology and confirm the robustness of our instrumental variable identification strategy in this empirical investigation.
The results in
Table 3 clearly show that the coefficient of
RISK is significant and negative in all models, indicating that there is a negative correlation between
RISK and
ES. The results in the last column show that 1% increase in
RISK can contribute to 0.0553% decrease in
ES, suggesting that climate risk amplification substantially undermines regional energy resilience. This relationship between
RISK and
ES can be explained by the following mechanisms. For instance, extreme droughts, a component of
, may reduce the availability of renewable energy sources like hydroelectricity, thereby reducing the diversity of regional energy supply [
41]. Additionally, extreme temperatures—both high and low—can exacerbate residential electricity demand, leading to surges in aggregate energy consumption, rising energy expenditures, and systemic degradation of energy affordability, all of which reduce energy affordability and thereby compromise
.
All control variables demonstrate statistically significant associations with ES. Specifically, economic development, tertiary industrial expansion, and governmental expenditure exhibit positive correlations with ES, whereas human capital shows a negative relationship. The three positively associated factors collectively enhance energy security through dual mechanisms of resource mobilization and systemic innovation. Economic growth generates fiscal capacity and technological spillovers that facilitate investments in renewable energy infrastructure, energy efficiency technologies, and low-carbon innovation ecosystems. Moreover, improvements in these areas foster social innovation, which in turn optimizes energy system efficiency, strengthens regional energy supply, and reduces external supply dependencies. The negative relationship between human capital and ES arises from consumption externalities. Increased population increase the total demand for energy, thus increasing the energy burden and reducing ES.
4.2. Robustness Tests: Replacing Control Variables
To address potential concerns regarding model specification sensitivity, we implemented a comprehensive robustness check by systematically substituting the original control variables. We replace the baseline controls—economic growth
, industrial development
, and official financial support
—with alternative indicators: economic growth per capita
PGDP, industrial structure index
IS, and foreign direct investment
FDI. Here, the industrial structure metric is calculated as the tertiary-to-secondary industry output ratio to measure sectoral composition shifts. These robustness tests are systematically presented in
Table 4. The
RISK coefficient maintains its statistically significant negative association with
ES, maintaining consistency with the baseline regression results. This demonstrates the resilience of our primary conclusion across alternative model configurations.
4.3. Asymmetric Analysis
Building upon the empirically validated negative association between
RISK and
ES, we further investigate the asymmetric nexus between the two variables. Given the substantial heterogeneity in
ES levels across Chinese provinces, we posit that RISK effects may exhibit nonlinear patterns across the
ES distribution. Essentially, there is a need to investigate how
RISK would have different impacts on different Chinese provinces with different quartile levels of
ES. To perform this asymmetric analysis, we employ the panel quantile regression model proposed by Koenker and Hallock [
42], examining the impact of
RISK on
ES at the 20th, 40th, 60th, and 80th percentiles. The results are presented in
Table 5 and
Figure 3.
According to the results listed in
Table 5, RISK is always negatively correlated with ES regardless of the quartile of ES. Moreover, as the
quartile increases, the coefficient of
becomes less negative, rising from –0.068 at the 20th percentile to –0.041 at the 80th percentile. This suggests that the marginal effect of
is more pronounced when
is at a relatively low level. This is due to the fact that regions with higher levels of energy security tend to have better energy infrastructure and are more resilient to potential climate risks. From a social development perspective, this also highlights that investments in energy systems to regions with lower levels of energy security are more efficient in coping with climate risks.
4.4. Direct Impact Analysis
The
ES indicator system comprises four dimensions, namely, accessibility, affordability, sustainability, and technology and efficiency. To dissect the direct associations between
RISK and
ES, we analyze its sub-dimensions through Equation (2), where each
ES component serves as an independent dependent variable. The results are shown in
Table 6, where the dependent variables in the four columns represent accessibility, affordability, sustainability, and technology and efficiency, respectively.
The results in
Table 6 reveal that a 1% increase in
RISK is associated with an approximate 0.19% reduction in energy accessibility, a 0.07% decrease in energy affordability, and a 0.23% decline in energy sustainability. Furthermore, such an increase in
RISK is also observed to diminish energy technology and efficiency by approximately 0.11%. These results indicate that climate risk is negatively correlated with all four sub-indicators of the energy security indicator system.
RISK can directly affect
ES by reducing the development of the four sub-indicators of
ES.
4.5. Indirect Impact Analysis
Exploring the influence mechanisms between climate risk and ES is critical to uncover their linkages. To this end, we use mediated effects to explore possible influence channels through which RISK contributes to ES. We use Equations (3)–(6) to measure this indirect impact. Equation (3) represents the impact of climate risk on energy transition, Equation (4) represents the impact of energy transition on energy security, Equation (5) represents the impact of climate risk on energy efficiency, and Equation (6) represents the impact of energy efficiency on energy security.
The findings presented in
Table 7 delineate two discrete yet statistically significant pathways through which
RISK exerts influence on energy security. Firstly, a 1% reduction in
RISK is correlated with a 0.27% increase in
ET, which subsequently results in a 0.10% enhancement in overall energy security. This suggests that a diminution of climate-related risks cultivates favorable economic and social circumstances, thereby facilitating a more effective and streamlined energy transition process. Secondly, a 1% decrease in
RISK corresponds to a 0.12% augmentation in
EE, which, in turn, contributes to a 0.05% improvement in energy security. This implies that a lower incidence of climate risk serves to mitigate energy losses by minimizing operational inefficiencies and transmission losses that are often exacerbated by extreme environmental conditions.
5. Conclusions
This study aims to uncover the linkage between RISK and ES based on a panel dataset of 30 Chinese provinces for a period of 2006–2022. We used the preferred IV-GMM model to investigate the impacts of RISK on ES. We also used a panel quantile regression model to investigate the nonlinear relationship between RISK and ES. In addition, a mediated effects model was used to identify the influence mechanism. The following findings were obtained:
The baseline regression results show a negative causal relationship between RISK and ES, implying that increased RISK reduces ES. Under a series of robustness tests, we confirm that RISK is negatively correlated with energy security ES.
Asymmetry analysis results present a nonlinear linkage between RISK and ES. Increased climate risk significantly leads to decreased ES for all quartiles. However, the effect of RISK on ES is more pronounced when ES is in the lower quartiles (20th and 40th).
Four sub-indicators—accessibility, affordability, sustainability, and technology efficiency—were employed to assess the impacts of climate risks on energy security. These indicators were selected to capture key dimensions of energy security that are directly influenced by climate variability, and they provide a comprehensive framework for understanding how climate risks affect the energy sector.
Increased climate risks reduce the capacity for energy transition and energy efficiency, thus indirectly affecting ES.
Our findings are situated within an expanding body of literature on climate–energy interactions. In domestic research, Lu et al. employed a mediation effect model to demonstrate that optimal redeployment strategies for power plants not only directly mitigate climatic impacts but also indirectly enhance system resilience by improving overall installed capacity efficiency, thereby providing empirical support for the mediation path analysis in this study [
43]. Lv et al. developed a machine learning model encompassing approximately 8000 hydropower, wind power, and photovoltaic plants in China over 2002–2017, and under SSP1-2.6 and SSP5-8.5, climate scenarios demonstrated that spatially optimized redeployment could reduce generation losses of existing plants by 6–8% between 2045 and 2060, while optimized layout of newly commissioned plants could increase national renewable generation by 24–28% and concurrently achieve 25–28% reductions in carbon emissions and 42–97% reductions in air pollutant emissions [
44]. The stronger adverse impact of climate risk in coastal provinces can be attributed to their greater exposure to typhoons and sea level rise, as highlighted by case studies of Guangdong and Fujian, where storm-induced grid failures are frequent [
44]. In contrast, inland provinces exhibit milder effects, likely due to fewer extreme storm events and diversified energy mixes dominated by coal and hydro. This regional heterogeneity aligns with vulnerability–resilience frameworks, which suggest that both hazard exposure and adaptive capacity jointly determine system sensitivity to climate risks [
6]. In international studies, Perera et al. conducted an empirical analysis of extreme-temperature events and grid resilience across multiple European countries, finding that each 1 °C increase corresponded to an average decline of 0.06 percentage points in grid resilience—comparable to the –0.067% observed in China’s coastal regions—thus confirming the cross-regional consistency of climate risk impacts on power infrastructure vulnerability [
45]. Moreover, Brás et al. applying a stochastic–robust optimization approach under low- and high-impact scenarios across 30 Swedish cities, showed that under extreme climate conditions, system integration could decline to as low as 16%, thereby underscoring the importance of incorporating uncertainty modelling [
46]. The IEA, in its World Energy Outlook, further emphasizes that the synergy between risk-sharing mechanisms and renewable-energy incentives within the EU’s Clean Energy Package provides valuable insights and policy models for enhancing global energy security and resilience [
47].
Empirical studies and authoritative reports, both domestic and international, consistently demonstrate that spatially optimized deployment, enhancements in energy efficiency and technology, and the coordinated integration of renewable energy technologies with storage significantly enhance energy system resilience, thereby mitigating the adverse effects of climate risks on energy security.
Although this study makes a substantive contribution, it nevertheless exhibits several limitations. First, the use of provincial-level data constrains our ability to capture finer-scale spatial heterogeneity, particularly regarding the resilience of energy infrastructure at the city or county level. Second, due to data availability constraints, variables such as energy prices and R&D investments were excluded from the model, which may have led to the omission of certain influence pathways. Third, because the findings are grounded in China’s specific energy system and institutional context, their direct applicability to other countries may be limited. Future research could extend this framework to cross-country comparisons, incorporate high-frequency climate data, and leverage micro-level firm or plant data to explore nonlinear threshold effects and adaptive dynamics.
6. Policy Implications
The findings of this study indicate that climate risk exerts a significantly adverse effect on energy security and amplifies this effect through energy transition and energy efficiency. Consequently, policymakers are urged to implement comprehensive strategies to mitigate the impacts of climate risk and bolster the resilience of energy systems. Considering the specific context of China, the following policy recommendations are proposed:
First, strengthening climate risk management is essential to improve the resilience of energy systems. Since climate risks can undermine energy security, governments at all levels should prioritize climate risk management to enhance the resilience of energy infrastructure. Specific measures include improving climate risk prediction and early warning systems, enhancing the capacity to respond to extreme weather events, and providing stricter safety regulations for critical energy infrastructure. In addition, investment in energy infrastructure should be increased to improve its stability and resilience under extreme weather conditions, thereby ensuring an uninterrupted energy supply.
Second, accelerating energy transition and decarbonization is vital. Our results show that energy transition plays a key mediating role between climate risk and energy security. Therefore, policymakers should further promote the development of clean energy and expedite the optimization of the energy structure. Governments can encourage firms and households to consume renewable energy through fiscal incentives (e.g., tax reductions, subsidies, and low-interest loans). Additionally, increased investments in grid modernization, including enhanced grid connection capacity for renewables and improved grid flexibility and dispatchability, are needed to mitigate climate-induced energy supply disruptions.
Third, improving energy efficiency and facilitating structural transformation are crucial. The Chinese government should release stricter energy efficiency standards and promote the application of energy-saving technologies in the industrial, building, and transportation sectors. For example, it should encourage enterprises to adopt energy-efficient equipment, promote smart grid technologies, and enhance end-use energy efficiency. Furthermore, rigorous regulation on energy-intensive industries and the promotion of green technological transformation are vital for reducing energy consumption and improving the efficiency of energy utilization.
Fourth, implementing region-specific energy security strategies is essential. Acknowledging the asymmetric impact of climate risk on energy security, particularly in regions with lower energy security levels, policymakers should adopt tailored policy interventions. This includes prioritizing smart grid development, optimizing energy supply chains, and providing targeted financial assistance in vulnerable regions. Additionally, fostering interregional energy cooperation through cross-regional energy dispatch can enhance energy supply stability.
Finally, the development of an integrated energy security guarantee system is essential. A comprehensive policy framework, encompassing carbon markets, environmental tax policies, and green financial instruments, is crucial for ensuring long-term energy security. This includes refining carbon trading markets to incentivize emissions reductions and improvements in energy efficiency, as well as promoting green bonds and sustainable investment funds to facilitate low-carbon energy projects.
In summary, reducing the negative impacts of climate risk on energy security requires an integrated approach that accounts for climate risk management, energy transformation, energy efficiency enhancement, and region-specific policies. Only through coordinated and synergistic policy arrangements can energy security be effectively enhanced and sustainable energy development goals be realized.
Author Contributions
Z.Z.: Investigation, Supervision, Writing—original draft; X.L. (Xiaokai Liu): Data curation, Methodology; R.S.: Formal analysis, Visualization; M.W.: Validation, Writing—original draft; X.L. (Xianli Liu): Investigation, Resources; P.H.: Validation, Writing—review and editing; Z.G.: Methodology, Writing—review and editing; P.X.: Visualization, Writing—original draft; Y.Z.: Writing—original draft; Y.G.: Funding acquisition, Project administration, Supervision, Writing—review and editing. All authors have read and agreed to the published version of the manuscript.
Funding
This study was supported by the technology project (NeiDianKeChuang [2024] Number 5) from Inner Mongolia Power (Group) Co., Ltd. and National Natural Science Foundation of China (72088101).
Data Availability Statement
The raw data supporting the conclusions of this article will be made available by the authors on request.
Conflicts of Interest
Authors Zhiyong Zhang, Xiaokai Liu, Rula Sa, Meng Wang, Xianli Liu and Peiji Hu were employed by the company Inner Mongolia Power Research Institute Branch, Inner Mongolia Power (Group) Co., Ltd. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The authors declare that this study received funding from Inner Mongolia Power (Group) Co., Ltd. The funder was not involved in the study design, collection, analysis, interpretation of data, the writing of this article or the decision to submit it for publication.
Abbreviations
The following abbreviations are used in this manuscript:
IV-GMM | Instrumental variable generalized method of moments |
OLS | Ordinary least squares |
RE | Random effect |
FE | Fixed effect |
References
- Pinkse, J.; Gasbarro, F. Managing physical impacts of climate change: An attentional perspective on corporate adaptation. Bus. Soc. 2019, 58, 333–368. [Google Scholar] [CrossRef]
- Caldecott, B.; Dericks, G.; Pfeiffer, A.; Astudillo, P. Stranded Assets: The Transition to a Low Carbon Economy; Emerging Risk Report 2016; Lloyd’s: London, UK, 2016. [Google Scholar]
- Curtin, J.; McInerney, C.; Gallachóir, B.Ó.; Hickey, C.; Deane, P.; Deeney, P. Quantifying stranding risk for fossil fuel assets and implications for renewable energy investment: A review of the literature. Renew. Sustain. Energy Rev. 2019, 116, 109402. [Google Scholar] [CrossRef]
- IEA. Energy Supply Security 2014: Emergency Response of IEA Countries; International Energy Agency: Paris, France, 2014. [Google Scholar]
- Ang, B.W.; Choong, W.L.; Ng, T.S. Energy security: Definitions, dimensions and indexes. Renew. Sustain. Energy Rev. 2015, 42, 1077–1093. [Google Scholar] [CrossRef]
- Gao, H.J.; Guo, M.H.; Liu, J.Y. Challenges and prospects of new power system to maintain supply and respond to the power limitation event from Sichuan high temperature and drought. Chin. J. Electr. Eng. 2023, 43, 4517–4538. [Google Scholar]
- Busby, J.W.; Baker, K.; Bazilian, M.D.; Gilbert, A.Q.; Grubert, E.; Rai, V.; Rhodes, J.D.; Shidore, S.; Smith, C.A.; Webber, M.E. Cascading risks: Understanding the 2021 winter blackout in Texas. Energy Res. Soc. Sci. 2021, 77, 102106. [Google Scholar] [CrossRef]
- Fan, J.-L.; Li, Z.; Huang, X.; Li, K.; Zhang, X.; Lu, X.; Wu, J.; Hubacek, K.; Shen, B. A net-zero emissions strategy for China’s power sector using carbon-capture utilization and storage. Nat. Commun. 2023, 14, 5972. [Google Scholar] [CrossRef]
- Yang, T.; Li, F.; Du, M.; Huang, M.; Li, Y. Impacts of alternative energy production innovation on reducing CO2 emissions: Evidence from China. Energy 2023, 268, 126684. [Google Scholar] [CrossRef]
- Ren, X.; Li, Y.; Shahbaz, M.; Dong, K.; Lu, Z. Climate risk and corporate environmental performance: Empirical evidence from China. Sustain. Prod. Consum. 2022, 30, 467–477. [Google Scholar] [CrossRef]
- Banerjee, A.K. Second-order moment risk connectedness across climate and geopolitical risk and global commodity markets. Econ. Lett. 2024, 235, 111551. [Google Scholar] [CrossRef]
- Wang, A.; Tao, H.; Ding, G.; Zhang, B.; Huang, J.; Wu, Q. Global cropland exposure to extreme compound drought heatwave events under future climate change. Weather Clim. Extrem. 2023, 40, 100559. [Google Scholar] [CrossRef]
- Rong, X.; Chen, H.; Liu, S. Nonlinear impact of climate risks on renewable energy stocks in China: A moderating effects study. Int. Rev. Financ. Anal. 2024, 96, 103613. [Google Scholar] [CrossRef]
- Ho, K.-C.; Yan, C.; Kong, X. The impact of climate change on credit cycles: Evidence from China’s bond market. Technol. Forecast. Soc. Change 2024, 206, 123490. [Google Scholar] [CrossRef]
- Shen, Y.; Sun, X.; Ji, Q.; Zhang, D. Climate events matter in the global natural gas market. Energy Econ. 2023, 125, 106887. [Google Scholar] [CrossRef]
- Lee, C.-C.; Song, H.; An, J. The impact of green finance on energy transition: Does climate risk matter? Energy Econ. 2024, 129, 107258. [Google Scholar] [CrossRef]
- He, F.; Hao, J.; Lucey, B. Effects of climate risk on corporate green innovation cycles. Technol. Forecast. Soc. Change 2024, 205, 123447. [Google Scholar] [CrossRef]
- Georgiou, G.C. US energy security and policy options for the 1990s. Energy Policy 1993, 21, 831–839. [Google Scholar] [CrossRef]
- Sun, H.; Nie, F.F.; Hu, X.Y. Evaluation of China’s regional energy security and difference analysis based on entropy weight TOPSIS method. Resour. Sci. 2018, 40, 477–485. [Google Scholar]
- Zhou, X.J. Research on energy security issues: A literature review. Contemp. Econ. Manag. 2017, 39, 1–5. [Google Scholar]
- Liu, L.T.; Shen, R.; Gao, T.M. Characteristics of energy security assessment and spatial and temporal evolution in China. J. Geogr. 2012, 67, 1634–1644. [Google Scholar]
- Zhang, L.; Yu, J.; Sovacool, B.K.; Ren, J. Measuring energy security performance within China: Toward an inter-provincial prospective. Energy 2017, 125, 825–836. [Google Scholar] [CrossRef]
- Iyke, B.N. Climate change, energy security risk, and clean energy investment. Energy Econ. 2024, 129, 107225. [Google Scholar] [CrossRef]
- Taghizadeh-Hesary, F.; Dong, K.; Zhao, C.; Phoumin, H. Can financial and economic means accelerate renewable energy growth in the climate change era? The case of China. Econ. Anal. Policy 2023, 78, 730–743. [Google Scholar] [CrossRef]
- Zhao, C.; Dong, K.; Wang, K.; Dong, X. Can low-carbon energy technology lead to energy resource carrying capacity improvement? The case of China. Energy Econ. 2023, 127, 107092. [Google Scholar] [CrossRef]
- Kim, J.; Jaumotte, F.; Panton, A.J.; Schwerhoff, G. Energy security and the green transition. Energy Policy 2025, 198, 114409. [Google Scholar] [CrossRef]
- Guo, K.; Ji, Q.; Zhang, D. A dataset to measure global climate physical risk. Data Brief 2024, 54, 110502. [Google Scholar] [CrossRef]
- Jiang, W.; Zhang, L. Analysis of energy consumption change process and its spatial and temporal effects in China. J. Liaoning Univ. Eng. Technol. 2005, 24, 612–615. [Google Scholar]
- Wang, Q.; Zhou, K. A framework for evaluating global national energy security. Appl. Energy 2017, 188, 19–31. [Google Scholar] [CrossRef]
- Zhao, C.; Wang, J.; Dong, K.; Wang, K. How does renewable energy encourage carbon unlocking? A global case for decarbonization. Resour. Policy 2023, 83, 103622. [Google Scholar] [CrossRef]
- Aslam, N.; Yang, W.; Saeed, R.; Ullah, F. Energy transition as a solution for energy security risk: Empirical evidence from BRI countries. Energy 2024, 290, 130090. [Google Scholar] [CrossRef]
- Yang, B.; Ding, L.; Zhan, X.; Tao, X.; Peng, F. Evaluation and analysis of energy security in China based on the DPSIR model. Energy Rep. 2022, 8, 607–615. [Google Scholar] [CrossRef]
- Zhang, J.; Wang, J.Q.; Zhang, L.L.; Zhao, L. Impact of industrialization on China’s regional energy security in the New Era. Environ. Dev. Sustain. 2022, 24, 8418–8440. [Google Scholar] [CrossRef]
- Chuang, M.C.; Ma, H.W. An assessment of Taiwan’s energy policy using multi-dimensional energy security indicators. Renew. Sustain. Energy Rev. 2013, 17, 301–311. [Google Scholar] [CrossRef]
- Sovacool, B.K. The methodological challenges of creating a comprehensive energy security index. Energy Policy 2012, 48, 835–840. [Google Scholar] [CrossRef]
- Sadorsky, P. Trade and energy consumption in the Middle East. Energy Econ. 2011, 33, 739–749. [Google Scholar] [CrossRef]
- Dong, F.; Li, J.; Wang, Y.; Zhang, X.; Zhang, S.; Zhang, S. Drivers of the decoupling indicator between the economic growth and energy-related CO2 in China: A revisit from the perspectives of decomposition and spatiotemporal heterogeneity. Sci. Total Environ. 2019, 685, 631–658. [Google Scholar] [CrossRef]
- Wang, S.S.; Zhou, D.Q.; Zhou, P.; Wang, Q.W. CO2 emissions, energy consumption and economic growth in China: A panel data analysis. Energy Policy 2011, 39, 4870–4875. [Google Scholar] [CrossRef]
- Lin, B.; Li, X. The effect of carbon tax on per capita CO2 emissions. Energy Policy 2011, 39, 5137–5146. [Google Scholar] [CrossRef]
- Acheampong, A.O.; Dzator, J.; Shahbaz, M. Empowering the powerless: Does access to energy improve income inequality? Energy Econ. 2021, 99, 105288. [Google Scholar] [CrossRef]
- Sharifi, A. Resilience of urban social-ecological-technological systems (SETS): A review. Sustain. Cities Soc. 2023, 99, 104910. [Google Scholar] [CrossRef]
- Koenker, R.; Hallock, K.F. Quantile regression. J. Econ. Perspect. 2001, 15, 143–156. [Google Scholar] [CrossRef]
- Lu, R.; Zhao, X.; Liu, G. Quantifying global energy resilience and its carbon reduction effect: The role of technological innovation. Energy Strategy Rev. 2025, 60, 101783. [Google Scholar] [CrossRef]
- Lv, L.; Liu, M.; Shan, Y.; Xu, J.; Yang, J.; Fang, W.; Ma, Z.; Bi, J. Mitigating climate risk to China’s renewable energy generation by power plants redeployment using a machine-learning model. Earths Future 2025, 13, e2024EF005641. [Google Scholar] [CrossRef]
- Perera, A.; Nik, V.; Chen, D.; Scartezzini, J.; Hong, T. Quantifying the impacts of climate change and extreme climate events on energy systems. Nat. Energy 2020, 5, 150–159. [Google Scholar] [CrossRef]
- Brás, T.A.; Simoes, S.; Amorim, F.; Fortes, P. How much extreme weather events have affected European power generation in the past three decades? Renew. Sustain. Energy Rev. 2023, 183, 113494. [Google Scholar] [CrossRef]
- International Energy Agency (IEA). World Energy Outlook 2023; IEA: Paris, France, 2023; Available online: https://www.iea.org/reports/world-energy-outlook-2023 (accessed on 20 July 2025).
| Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2025 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).