Climate Change Mitigation Behaviors in Tourists in Chinese Mountains

Reviewer #1

Comments

Revision

1. The results of the conducted research provide new insights, relevant to the context of climate change. This is an important and timely topic. It is worthwhile to deepen the research conducted, addressing the identified limitations of the conducted research (section 5.4 Limitations and recommendations for future research).

Thank you for the encouragement. Indeed, in tourist behavior research, climate change has provided a critical context to study how tourists perceive and respond to this globally pressing issue.

The authors appreciate the reviewer’s valuable suggestion regarding addressing the identified limitations of the conducted research. Beyond the previously identified limitations, the authors extend the discussion of limitations from the perspective of reliance on self-reports and measurement of constructs, and provide recommendations for future research. The supplementary content has been added to the “section 5.4 Limitations and recommendations for future research” section as follows:

“First, the reliance on self-reports raises potential endogeneity concerns. Although we have rigorously tested our hypothesized model, the risk of endogeneity cannot be fully ruled out due to omitted variable bias, reverse causality and measurement error. Future research should address these by incorporating longitudinal or experimental designs to establish temporal precedence and causality. Further, multi-group or interaction analyses should also be employed to compare groups based on prior visits, age or gender to test the model’s robustness.”

“In addition, future study could also distinguish between state and trait awe to provide greater measurement precision.”

“Future research should add objective climate knowledge construct in the model to test the different roles played by objectively assessed knowledge in this model.”

2. The manuscript outlines three main goals (lines 87-91). After reviewing the manuscript, it can be concluded that these goals were achieved.

Thank you for the encouragement. The three main goals raised in this study have been achieved and we hope to conduct further research to explore mountain tourist behaviors in the context of climate change.

3. The methodology used has been comprehensively described.

Thank you for the encouragement. This study employs Partial Least Squares Structural Equation Modeling (PLS-SEM) to test the proposed theoretical model. PLS-SEM proves particularly effective in identifying multiple key constructs and their influencing factors. It allows flexible specification of relationships between model structures and variables, while providing strong explanatory power for endogenous variables. These advantages have led to its widespread application in empirical research in recent years.

4. Figures and tables constitute a valuable part of the manuscript.

Thank you for the encouragement. To ensure the analysis results are clear to readers, the findings in this study have been visualized through a combination of tables and descriptive text.

5. The references are appropriate.

Thank you for the encouragement. We cite abundant and appropriate references to strengthen the foundation of our study.

Reviewer #2

Comments

Revision

1. This paper offers meaningful theoretical and applied contributions to sustainability and tourism psychology, particularly in understanding emotional drivers of pro-environmental behavior. The Introduction clearly identifies the research gap and presents the study’s aim in a logical and well-structured manner. The authors successfully explain why the topic is important and how their study contributes to filling the existing void in the literature. The data collection through 391 valid questionnaires meets the required sample size, providing credible statistical power. The manuscript is well structured, logically coherent, and thoughtfully designed. The authors present their arguments and empirical analysis in a systematic way, ensuring smooth transitions between sections. The methodology and analytical approach are appropriate and contribute to the overall rigor and clarity of the paper.

Thank you for the encouragement. Indeed, in tourist behavior research, climate change has provided a critical context to study how tourists perceive and respond to this globally pressing issue. This field has seen substantial investigation into each of these elements individually. However, it is observed that the role of emotional drivers such as “awe” have yet to be fully explored especially in the awe-eliciting contexts like highly mountainous areas. It is believed that this study will make a significant contribution to the literature on tourists’ climate mitigation behaviors by addressing this gap.

2. There is a minor typographical error in the first sentence of the Introduction section (“T During…”). The initial letter “T” should be removed.

The authors appreciate the reviewer’s valuable suggestion regarding typographical error in the first sentence of the Introduction section. We sincerely feel sorry for that typo issue. We have removed the initial letter “T” and the first sentenct of the Introduction section is revised as follows:

“During the period from June 23 to July 2 2025, soaring temperatures across Europe resulted in 2,305 deaths across 12 major cities, including Barcelona, Milan, and London. Among these, temperature changes attributable to climate change were responsible for 1,504 fatalities [1].”

3. Although the discussion and implications are well structured and comprehensive, the manuscript lacks a distinct Conclusion section. I recommend adding a concise concluding part that synthesizes the main findings, theoretical and practical contributions, and highlights the study’s overall significance. This would provide a clear sense of closure and align the paper with the journal’s structural standards.

·          

The authors appreciate the reviewer’s valuable suggestion regarding adding a concise concluding part. We sincerely agree with the reviewer that a conclusion part can provide a clear sense of closure and align the paper with the journal’s structural standards. The supplementary content has been added as section “6. Conclusion” as follows:

“6. Conclusion

This empirical investigation of tourists in China’s Hengduan Mountains provides robust evidence for a novel cognitive-emotional pathway to climate change mitigation and yields several distinctive theoretical and practical insights that advance the current understanding of pro-environmental behaviors in eco-sensitive tourism contexts. Our study makes a significant contribution by bridging the PMT and S-O-R frameworks, demonstrating that in awe-inspiring highly mountainous contexts, cognitive appraisals serve as distal catalysts, while awe acts as the proximal driver or behavioral intentions. This finding underscores the necessity of incorporating discrete emotions into cognitive models to fully understand pro-environmental action in spectacular natural settings. Another intriguing finding lies in the moderating role of subjective knowledge, which weakened the link between threat appraisal and awe, revealing a potentially counterproductive effect of overconfidence. These findings have significant implications for destination management practices, particularly regarding developing awe-elicited and fact-based intervention strategies. Destination management organizations should design strategies that collaboratively elicit awe through the grandeur of the landscape while simultaneously providing clear and feasible mitigation action. ”

Reviewer #3

Comments

Revision

1. Introduction

1.1 The problem framing is strong and the Hengduan context is relevant. Please sharpen the specific novelty claim: What does adding awe to a PMT account reveal that prior tourism/PMT papers did not (e.g., incremental variance explained, new boundary conditions in high-mountain settings)?

The authors appreciate the reviewer’s valuable suggestion regarding the need to further elaborate on the novelty claim in Introduction. This study chose Hengduan Mountain area as the research setting due to its vulnerability to climate change ang the vastness and grandeur in its nature from which awe is likely to be elicited. The authors sharpen the specific novelty claim of “What does adding awe to a PMT account reveal that prior tourism/PMT papers did not?” by extending the discussion from new boundary conditions. The supplementary content has been added to the “Introduction” section as follows:

“However, its application in high mountain contexts has remained in a cognitive-oriented framework. Prior studies have largely focused on how these cognitive appraisals directly led to behaviors or behavior intentions, overlooking the potential mediating role played by emotions that the appraisal processes might elicit. We posit that in high mountain settings that feature overwhelming landscape, threat and coping appraisal is not the sole driver of actions. Rather, it first elicits awe, a transcend emotion which is closely associated with stimuli that are both vast and fear-appealing, which in turn motivates climate change mitigation behavior intentions.”

1. Introduction

1.2 Define early that “climate change knowledge” is subjective (self-assessed) and anticipate why it could dampen threat-appraisal effects (overconfidence), to reduce perceived contradiction later.

The authors appreciate the reviewer’s valuable suggestion regarding the need to define early that “climate change knowledge” is subjective and anticipate why it could dampen threat-appraisal effects to reduce perceived contradiction later in Introduction. The authors introduce the concept of “subjective knowledge about climate change” early in the introduction part by phrasing it explicitly as subjective knowledge to reduce perceived contradiction later. More detailed definition to subject knowledge is added in the literature review section. The supplementary content has been added to the “Introduction” section as follows:

“Although public perception and awareness of climate change in growing in China [10], it is worth discussing whether increased subjective knowledge in this field can affect people’s appraisal of threat and coping efficacy of climate change.”

The supplementary content has also been added to the “Literature review” section as follows:

“Nevertheless, the role of knowledge is complex, particularly when considering the distinction between subjective and objective knowledge. Subjective knowledge represents the degree to which an individual believes he or she understands a topic.”

1. Introduction

1.3 Tighten the motivation for Hengduan: beyond “under-studied,” specify ecological, cultural, or visitation features that make it theoretically diagnostic (not just practically important).

The authors appreciate the reviewer’s valuable suggestion regarding tightening the motivation for Hengduan. The authors extend the discussion in Introduction from ecological and cultural in addition to visitation features. The supplementary content has been added to the “Introduction” section as follows:

“With profound topographic variation and spectrum of climatic conditions, this area has been renowned for its rich biodiversity [7]. However, with the change in precipitation and temperature increase [8], the flora and fauna species are impacted severely. This area is also home to a variety of ethnic groups, primarily Tibetan people, which has contributed to its unique and rich cultural heritage, both tangible and intangible.”

2. Theoretical Framework and Hypotheses
2.1 Make the PMT→SOR bridge explicit. At present, “threat/coping appraisal as stimuli” is asserted; justify theoretically why appraisals (internal cognitions) can operate as S in SOR, and contrast with conventional S (external cues).

The authors appreciate the reviewer’s valuable suggestion regarding making the PMT→SOR bridge explicit. The authors extend the discussion in Literature Review to justify theoretically why appraisals can operate as S in SOR. Several prior studies are also added to provide stronger evidence. The supplementary content has been added to the “Literature Review” section as follows:

“In traditional S-O-R research, "stimuli" are regarded as external factors, such as the natural environment, retail ambiance, and quality of service. The S-O-R model suggests that particular environmental stimuli directly influence an individual's cognitive and emotional conditions, resulting in either approach or avoidance behaviors [47]. However, this research conceptualizes cognitive appraisals as internalized psychological stimuli. During the high mountain tourism experience, tourists receive lots of environmental clues and mentally process them. This internal assessment generates cognitive outcomes [48], i.e. threat appraisal and coping appraisal in our study. As direct reflection of the external reality of high mountain contexts, threat and coping appraisal function as psychological events that subsequently stimulate the emotional state of the “organism”.”

“In a study of online consuming behavior based on S-O-R model, it was found that potential customers appraised or evaluated online reviews, which then served as stimuli to foster different emotions, resulting into various empathy behaviors [52]. By extending the research model to the context of online travel agency apps such as Tripadvisor and Booking.com, the authors revealed the same stimulus-organism-response mechanism [53].”

2. Theoretical Framework and Hypotheses
2.2 Clarify construct hierarchy. Threat appraisal (severity, vulnerability) and coping appraisal (response-, self-efficacy) are treated as second-order factors, but the modeling approach (repeated indicators vs. two-stage vs. hybrid) is not specified. State the specification and justify it.

The authors appreciate the reviewer’s valuable suggestion regarding clarifying construct hierarchy. Treating threat appraisal and coping appraisal as second-order factors, the authors further state the specification and justify it in Methodology so as to ensure consistency of discussion. Reliability and validity of the second-order constructs are also assessed and presented in Table 1. The supplementary content has been added to the “Methodology” section as follows:

“According to the theoretical model, perceived vulnerability, perceived severity, self-efficacy, and response efficacy constitute the first-order latent variables. These first-order latent variables, in turn, formatively construct the two second-order latent variables: threat appraisal and coping appraisal. The measurement indicators, first-order latent variables, second-order latent variables, and the connections between them collectively form a reflective-formative second-order model. To operationalize these higher-order constructs, we employed the repeated indicators approach [74]. In this approach, all indicators of the lower-order components including all items measuring the first-order variables are repeatedly used as indicators for the higher-order components. This method is well-suited for our model as it allows for a direct estimation of the relationships among constructs in the model. As presented in Table 1, the assessment of threat appraisal revealed an AVE of 0.506, which demonstrated its validity. Although coping appraisal revealed an AVE of 0.458, slightly below the strict threshold of 0.5, its composite reliability was high, confirming the internal consistency of the measure.”

2. Theoretical Framework and Hypotheses
2.3 Position awe precisely: is it an appraisal-contingent affect or an environment-elicited emotion amplified by cognitive appraisals? This matters for interpreting mediation.

The authors appreciate the reviewer’s valuable suggestion regarding positioning awe precisely. This study measures both environment-elicited awe and awe amplified by cognitive appraisals since the high mountain context itself elicits the first type of awe and the cognitive appraisal process arouse the second type of awe. Therefore, the authors assessed two types of awe in one measure. The authors extend the discussion of two types of awe and justify the reason of assessing both of them in one measure in Literature Review. The supplementary content has been added to the “Literature Review” section as follows:

“As an emotion usually aroused by magnificent natural scenery, awe is increasingly seen as an important emotional experience in the tourism industry. This type of awe is usually associated with other positive emotions, such as gratitude and joy, and is classified as unthreatened awe [54]. When tourists visit high mountain areas, the vastness of nature is likely to elicit this type of awe. In the meantime, when tourists perceive and appraise the likelihood and severity of climate change on high mountains, threatened awe is likely to be aroused, which is characterized by increased feeling of fear [55]. Therefore, the current study theorizes and measures both threatened and unthreatened awe as one variable “awe”.”

2. Theoretical Framework and Hypotheses
2.4 For H4a–H4b, specify the form of moderation tested (interaction on awe only) and motivate tests of moderated mediation (conditional indirect effects via awe), which the current logic implies.

The authors appreciate the reviewer’s valuable suggestion regarding specifying the form of moderation tested for H4a–H4b. The authors specify a first-stage moderated mediation model and add explanation for the model specification. The supplementary content has been added to the “Literature Review” section as follows:

“Thus, we propose a first-stage moderated mediation model. Specifically, we hypothesize that the relationship between cognitive appraisals, i.e. threat and coping, and awe is moderated by subjective knowledge about climate change. This theoretical positioning implies that the strength of the predictive power of threat appraisal and coping appraisal on behavior intentions through awe is conditional upon the level of subjective knowledge. ”

3. Data Analysis
3.1 Sampling: online recruitment viaWenjuanxing/XiaoHongShu/Weibo risks self-selection and urban/digital-literacy bias. Describe quotas/stratification, completion checks (beyond <10s/item), and any post-stratification weighting.

The authors appreciate the reviewer’s valuable suggestion regarding sampling. We sincerely agree that online recruitment risks self-selection and urban/digital-literacy bias and acknowledge this in limitation section. The authors extend discussion on quotas/stratification, completion checks (beyond <10s/item), and post-stratification weighting. The supplementary content has been added to the “3.2 data collection and sample descriptive analysis” section as follows:

“To enhance sample diversity, we employed soft quotas method by disseminating surveys across a diverse range of accounts and groups, such as regional community forums and interest groups pertaining to different age groups. We specifically monitored the responses in real-time against key demographic variables to avoid the predominance of one single demographic group.”

“In addition, an item for attention check was embedded in the questionnaire which required the respondents to do a simple calculation problem. Those with a wrong answer were also eliminated.”

3. Data Analysis
3.2 Justify PLS-SEM beyond “sample and distribution flexibility.” Link to prediction goals, higher-order constructs, and interaction modeling; cite current best-practice guidance.

The authors appreciate the reviewer’s valuable suggestion regarding justification of the adoption of PLS-SEM in the current study. The authors extend discussion by linking to prediction goals, higher-order constructs, and interaction modeling; current best-practice guidance is also cited. The supplementary content has been added to the “Methodology” section as follows:

“PLS-SEM is explicitly designed for predicting and explaining variance in the dependent variables, which is aligned with the purpose of the current study to predict tourists’ climate change mitigation behavior intentions and to explain the key antecedents of this outcome [72].”

“Thus, without imposing strict distributional assumption on the data, it provides a robust method for testing interaction effects. PLS-SEM can also estimate complex models with higher-order constructs due to its component-based approach, while avoiding issues such as model identification and estimation instability [73].”

3. Data Analysis

3.3 Report outer VIFs (indicator-level) in addition to inner/path VIFs; add reliability ρA; provide cross-loadings (or HTMT inference with CIs).
 

The authors appreciate the reviewers for their valuable suggestions. We have, in accordance with your opinions, added more data related to the models. We calculate and report the external VIF values, and all of them are well below 5. Statistics are provided in Table 4. We also integrate HTMT inference with CIs and other statistics into one table (Table 6) which demonstrate reliability, convergent validity and discriminant validity of the measurement model in a clearer and comprehensive way.

The supplementary content and table3 have been added to “4.1 common method bias” section, and Table 6 has been added to “4.3 measurement model” section.

3. Data Analysis

3.4 Common method bias: Harman’s test is insufficient. Add a marker-variable or full-collinearity VIF test; consider a latent method factor or temporal/psych separation acknowledgments.

The authors appreciate the reviewer’s valuable suggestions regarding common method bias test. We agree that the Harman one-factor test has limitations, and we sincerely appreciate your suggestion to use a more robust method to test for common method bias. Following your advice, we conducted the following supplementary analysis by conducting an indicator-level full collinearity VIF test. Table 4 presents the required data. We propose that common method bias is not a general concern in our model based on both Harman’s test and full-collinearity VIF test. The supplementary content and table3 have been added to the “4.1 common method bias” section as follows:

“Furthermore, Table 4 summarizes the item-level statistics, including collinearity assessed via the full VIF procedure. Cronbach’s α estimates of all the seven constructs were greater than the suggested cut-off values of 0.70, ranging from 0.757 to 0.913. A vast majority of items met Kock’s 3.3 rule [82], indicating the collinearity is not a general concern in the measurement model. Two items (AWE2 and AWE 3) exhibited VIF values above the threshold (4.667 and 3.962 respectively) but below 5.0. it is important to note that these items demonstrated high outer loadings and the corresponding construct demonstrated high composite reliability. Therefore, we retained these items for their theoretical relevance. Their inclusion did not pose a critical threat to the model’s interpretation.”

3. Data Analysis

3.5 Bootstrapping: 5,000 resamples are reported—also report CI type (percentile/bias-corrected), two-tailed significance, and effect sizes.

The authors appreciate the reviewer’s valuable suggestions for emphasizing these necessary reporting standards. In the current analysis, we adopted the bootstrap sampling method: 5000 replications were reported, the analysis was configured to generate bias-corrected and accelerated confidence intervals. Two-tailed tests were used for all significance tests, with a significance level of 0.05. The supplementary content and table3 have been added to the “4.3 measurement model” section as follows:

The analysis was configured to generate bias-corrected and accelerated confidence intervals, which provided a more robust estimate. Two-tailed tests were used for all significance tests, with a significance level of 0.05.

4. Results and Discussion

4.1 Interpret magnitudes, not only significance. Awe→CB (β=.588) dominates; discuss practical meaning (e.g., 1 SD change implications) and compare direct PMT paths (β=.173/.126).

The authors appreciate the reviewer’s valuable suggestions for interpreting magnitudes of different paths and discussing practical meaning. We sincerely believe your suggestion can enhance our results and discussion. The supplementary content has been added to the “4.4 hypothesis testing” section as follows:

“The strong positive relationship between awe and climate change mitigation behavior intention is not only the largest in the model but also dominates the direct effects. To appreciate the practical meaning, a one standard deviation increase in awe is associated with a 0.588 standard deviation increase in behavior intention. This suggests that interventions which successfully elicit the emotion of awe can be more effective to motivate climate change mitigation behavior intentions than those which solely focus on encouraging cognitive appraisal.”

4. Results and Discussion

4.2 Moderation: present interaction plots and simple-slope tests at ±1 SD knowledge; explore conditional indirect effects (index of moderated mediation) to test whether knowledge weakens TA→AWE→CB.

The authors appreciate the reviewer’s valuable suggestions for presenting interaction plots and exploring conditional indirect effects. We sincerely believe your suggestion can enhance our results and discussion. We have conducted simple-slope tests at ±1 SD knowledge and present a interaction plot. Figure 3 demonstrates that knowledge weakens TA→AWE→CB. The supplementary content and Figure 3 have been added to the “4.4 hypothesis testing” section as follows:

“After confirming the significant mediating effect of awe, we estimated the indirect effects at different levels of the moderator. The significant negative interaction between threat appraisal and subjective knowledge indicates that the positive relationship between threat appraisal and behavior intentions though awe is weakened as subjective knowledge increases. As illustrated in figure 3, this relationship was stronger when subjective knowledge was low (-1SD, simple slop=0.203) and the relationship became weaker when subjective knowledge was high (+1SD, simple slope=0.098).”

4. Results and Discussion

4.3 Reconcile the negative CK×TA effect with theory using the subjective-knowledge lens (overconfidence/Dunning–Kruger) and propose tests with objective knowledge.

The authors appreciate the reviewer’s valuable suggestions for reconciling the negative CK×TA effect with theory using the subjective-knowledge lens. We caution against overconfidence and the supplementary content has been added to the “2.4. Subjective knowledge about climate change as a moderator” section as follows:

“Some studies have found that elevated subjective knowledge can paradoxically weaken risk perception [65, 66], often attributed to the development of overconfidence [67], which leads individuals to underestimate threats.”

We also extend discussion on the possible reasons for the negative interaction effect from the lens of

information avoidance and risk normalization in addition to attribution to Dunning–Kruger effect. The supplementary content has been added to the “5.1 general discussion” section as follows:

“Furthermore, alternative theoretical perspectives can also elucidate this finding. Information avoidance [94] might stop individuals with high subjective knowledge from deeply processing the threatening information. This defensive avoidance could be a coping strategy to manage “eco-anxiety” [95] associated with fully confronting the crisis. In addition, the phenomenon of risk normalization [96] could provide another explanation. Repeated exposure to climate-related information though both media and personal travel experience may reduce its novelty or salience. This normalization process would dampen the emotional impact necessary for awe to emerge, as the stimulus is gradually appraised as a familiar and normalized risk.”

4. Results and Discussion

4.4 Ensure label consistency (CK vs. “Previous Knowledge/PK”; “CB8” in Table 6) and clean minor formatting/typos that currently reduce credibility.

The authors appreciate the reviewer’s valuable suggestions for label consistency and cleaning minor formatting/typos. We are sincerely sorry for those careless typos and inconsistency in construct labeling. We truly believe credibility can be largely increased by making those corrections. We label the construct “subjective knowledge about climate change” as “SK” and climate change behavior intention as “CB” consistently. We correct and double check formatting issues and typos. All those corrections have been marked in the read color for your kindly reference.

4. Results and Discussion

4.5 Strengthen the theoretical contribution: precisely state how PMT+SOR with awe advances pro-environmental behavior theory in tourism (e.g., emotions as proximal drivers; appraisals as distal catalysts), and specify boundary conditions (high-altitude, spectacle-rich contexts).

The authors appreciate the reviewer’s valuable suggestions for strengthening the theoretical contribution. We extend discussion on how PMT+SOR with awe advances pro-environmental behavior theory in tourism and specify boundary conditions. The supplementary content has been added to the “5.2 theoretical implication” section as follows:

“We posit that threat and coping appraisal function as distal catalysts, not immediate drivers of behavior intentions. Instead, mountain tourists generated awe experience from cognitive processes responding to risks and challenges brought by climate change, which acts as the proximal driver to directly motivate behavior intention. Furthermore, this study specifies specific boundary conditions for this theoretical model. The emotional pathway is likely to be particular to awe-eliciting settings, especially high-altitude and spectacle-rich contexts like Hengduan Mountains. In such settings, the vastness and grandeur of nature can overwhelm cognitive processing and make emotion a more direct motivator to behavior intention.”

5. Practical Implications
5.1 Condense into actionable strategies: (1) design awe-evoking interpretation (scale cues, perspective framing, soundscapes) paired with clear severity/likelihood information; (2) convert “awareness” to “actionable competence” (context-specific behaviors, implementation intentions, defaults) to bolster coping appraisal; (3) beware overconfidence from subjective knowledge—use quizzes/feedback to calibrate; (4) reduce friction for low-carbon choices on-site (transport, waste, energy).

The authors appreciate the reviewer’s valuable suggestions for condensing into actionable strategies. Your kindly suggestions can largely enhance our practical implication. The authors extend discussion from aspects kindly suggested by the reviewer. The supplementary content has been added to the “5.3 practical implication” section as follows:

“The findings indicate that placing greater emphasis on the threats posed by climate change to mountain tourism destinations can enhance the positive effect of awe in promoting behavioral change.”

“In addition, audio and visual cues facilitated by AR or VR technologies could translate the severity of climate change impact into tangible tourist experience. For example, an AR application can enable tourists to view the contemporary landscape with an overlay that shows how it looked like decades ago to demonstrate the process of glacier retreat. The application of soundscapes can also create an immersive and emotionally resonate context for severity information. Soundscape stations can be built at pivotal locations, playing natural sounds that are diminishing with climate change.”

“Our findings caution against simply increasing tourists’ subjective knowledge, which may foster overconfidence. To address this, educational interventions should aim to calibrate the gaps between tourist’ subjective knowledge and actual or objective knowledge. One effective way of doing this could be the use of interactive quizzes with immediate feedback on DMO websites or in visitor centers. Quizzes should cover effective and actionable climate change behaviors rather than general climate science. Completion of the quizzes with a satisfactory score could earn “green points”, which can be converted into a small monetary donation to the climate change cause or redeemed as low-carbon experiential rewards.”

5. Practical Implications
5.2 Discuss fear-appeal ethics and efficacy: emphasize efficacy and agency to avoid reactance.

The authors appreciate the reviewer’s valuable suggestions for discussing fear-appeal ethics and efficacy and emphasizing efficacy and agency to avoid reactance. We sincerely agree with the reviewer that overreliance on fear-appeal communication might raise ethical issues. Thus, the authors propose distinct behavioral domains to tourists, which emphasizes efficacy and agency to avoid reactance. The supplementary content has been added to the “5.3 practical implication” section as follows:

Next, highlighting accessible pathways to implement these pro-environmental actions is also meaningful. When tourists begin a hike, an in-app nudge could push a notification which clearly communicates the sensitivity of the ecosystem and the effectiveness of personal emission-reduction behaviors like carpooling or choosing public transportation as coping responses. During trip planning process, official websites or OTAs could provide low-carbon itinerary options rather than conventional routes. The prompt on the website could suggest that theses itineraries prioritize electric shuttle transport, local-food based meal options, and certified ecolodges.

6. Limitations and Future Research
6.1 Acknowledge self-report intentions (not behaviors), single-country, online convenience sampling, and potential unmeasured confounds.

The authors appreciate the reviewer’s valuable suggestions for acknowledging self-report intentions (not behaviors), single-country, online convenience sampling, and potential unmeasured confounds. The supplementary content has been added to the “5.4 limitations and recommendations for future research” section as follows:

“First, the reliance on self-reports raises potential endogeneity concerns. Although we have rigorously tested our hypothesized model, the risk of endogeneity cannot be fully ruled out due to omitted variable bias, reverse causality and measurement error. Future research should address these by incorporating longitudinal or experimental designs to establish temporal precedence and causality. Further, multi-group or interaction analyses should also be employed to compare groups based on prior visits, age or gender to test the model’s robustness. Second, the adoption of online data collection method in a single country can be subject to limitations of sample representativeness.”

6. Limitations and Future Research
6.2 Add multi-group or interaction tests (e.g., prior visits, age, gender, altitude sickness risk) and longitudinal/experimental designs to establish temporal precedence and causality.

The authors appreciate the reviewer’s valuable suggestions for adding multi-group or interaction tests and longitudinal/experimental designs to establish temporal precedence and causality. The supplementary content has been added to the “5.4 limitations and recommendations for future research” section as follows:

Future research should address these by incorporating longitudinal or experimental designs to establish temporal precedence and causality. Further, multi-group or interaction analyses should also be employed to compare groups based on prior visits, age or gender to test the model’s robustness.

6. Limitations and Future Research
6.3 Include objective climate-knowledge scales and state-vs-trait awe; test dual-process models (cognitive vs. affective routes) and compare high- vs. low-spectacle settings.

The authors appreciate the reviewer’s valuable suggestions for including objective climate-knowledge scales and state-vs-trait awe. The supplementary content has been added to the “5.4 limitations and recommendations for future research” section as follows:

“In addition, future study could also distinguish between state and trait awe to provide greater measurement precision.”

“Future research should add objective climate knowledge construct in the model to test the different roles played by objectively assessed knowledge in this model.”

Reviewer #4

Comments

Revision

1. Theory framing and hypothesis logic

1.1 The rationale for treating threat appraisal and coping appraisal as “stimuli” in SOR is interesting; please anchor this move more firmly with sources that explicitly conceptualize subjective appraisals as stimuli rather than outcomes. This will help justify H3a–H3c conceptually, beyond analogies to virtual tourism or destination credibility. (Section 2.3.)

The authors appreciate the reviewer’s valuable suggestion regarding making the PMT→SOR bridge explicit. The authors extend the discussion in Literature Review to justify theoretically why appraisals can operate as S in SOR. Several prior studies which explicitly conceptualize subjective appraisals as stimuli are also added to provide stronger evidence. The supplementary content has been added to the “Literature Review” section as follows:

“In traditional S-O-R research, "stimuli" are regarded as external factors, such as the natural environment, retail ambiance, and quality of service. The S-O-R model suggests that particular environmental stimuli directly influence an individual's cognitive and emotional conditions, resulting in either approach or avoidance behaviors [47]. However, this research conceptualizes cognitive appraisals as internalized psychological stimuli. During the high mountain tourism experience, tourists receive lots of environmental clues and mentally process them. This internal assessment generates cognitive outcomes [48], i.e. threat appraisal and coping appraisal in our study. As direct reflection of the external reality of high mountain contexts, threat and coping appraisal function as psychological events that subsequently stimulate the emotional state of the “organism”.”

“In a study of online consuming behavior based on S-O-R model, it was found that potential customers appraised or evaluated online reviews, which then served as stimuli to foster different emotions, resulting into various empathy behaviors [52]. By extending the research model to the context of online travel agency apps such as Tripadvisor and Booking.com, the authors revealed the same stimulus-organism-response mechanism [53].”

1. Theory framing and hypothesis logic

·         1.2 Distinguish threatened awe vs unthreatened awe; your discussion at times implies both. Because PMT cues can evoke fear alongside awe, clarifying which awe variant you theorize (and measure) would sharpen predictions. (Section 2.3.)

The authors appreciate the reviewer’s valuable suggestion regarding positioning awe precisely. This study measures both threatened awe and unthreatened awe since the high mountain context itself elicits the first type of awe and the cognitive appraisal process arouse the second type of awe. Therefore, the authors assessed two types of awe in one measure. The authors extend the discussion of two types of awe and justify the reason of assessing both of them in one measure in Literature Review. The supplementary content has been added to the “Literature Review” section as follows:

“As an emotion usually aroused by magnificent natural scenery, awe is increasingly seen as an important emotional experience in the tourism industry. This type of awe is usually associated with other positive emotions, such as gratitude and joy, and is classified as unthreatened awe [54]. When tourists visit high mountain areas, the vastness of nature is likely to elicit this type of awe. In the meantime, when tourists perceive and appraise the likelihood and severity of climate change on high mountains, threatened awe is likely to be aroused, which is characterized by increased feeling of fear [55]. Therefore, the current study theorizes and measures both threatened and unthreatened awe as one variable “awe”.”

1. Theory framing and hypothesis logic

·         1.3 Consider whether moral obligation (a common extension of PMT) belongs in your framework, or at least acknowledge it as a theoretically adjacent driver (you cite Chen, 2020 in refs but it is not positioned in the main text).

The authors appreciate the reviewer’s valuable suggestion regarding considering whether moral obligation (a common extension of PMT) belongs in your framework. The authors acknowledge it as a theoretically adjacent driver as in Chen (2020). The supplementary content has been added to the “Literature Review” section as follows and the citing of Chen (2020) is properly positioned in the main text as citation [33]:

Following the core components of PMT, recently moral obligation, the feeling of personal responsibility to act, has been frequently integrated as an extension of PMT [33]. While our study focuses on testing the fundamental structure of PMT, it is well recognized that moral obligation represents a potent parallel driver, which suggests that individuals participate in climate change mitigation not only because they feel that they are in danger or believe that their actions are effective, but also because they believe they have a moral responsibility to protect the environment and future generations [33].

2. Hypotheses numbering and consistency

·         In Section 4.4 you state “coping appraisal … had a significant positive impact on climate change mitigation behavior intention. H1 is thus supported,” but H1 (Section 2.2) is threat appraisal → behavior. Swap these labels or correct the text to avoid confusion. (Pages 5 and 13.)

The authors appreciate the reviewer’s valuable suggestion regarding the labels of coping appraisal and threat appraisal. We are sincerely sorry for this careless mistake. We have corrected the text to avoid confusion. The corrected content in page 13 has been marked in red.

3. Conceptual model presentation

Figure 1 (p.7) clearly depicts the higher-order structure (PV, PS → Threat appraisal; RE, SE → Coping appraisal) and the moderated path to awe. Please specify in Methods how you implemented the higher-order constructs (repeated indicators, two-stage, hybrid?). Also report reliability/validity for the higher-order constructs, not only their first-order facets. (Figure 1, Section 3.)

The authors appreciate the reviewer’s valuable suggestion regarding conceptual model presentation. Treating threat appraisal and coping appraisal as second-order factors, the authors further state the specification and justify it in Methodology so as to ensure consistency of discussion. Reliability and validity of the second-order constructs are also assessed and presented in Table 1. The supplementary content has been added to the “Methodology” section as follows:

“According to the theoretical model, perceived vulnerability, perceived severity, self-efficacy, and response efficacy constitute the first-order latent variables. These first-order latent variables, in turn, formatively construct the two second-order latent variables: threat appraisal and coping appraisal. The measurement indicators, first-order latent variables, second-order latent variables, and the connections between them collectively form a reflective-formative second-order model. To operationalize these higher-order constructs, we employed the repeated indicators approach [74]. In this approach, all indicators of the lower-order components including all items measuring the first-order variables are repeatedly used as indicators for the higher-order components. This method is well-suited for our model as it allows for a direct estimation of the relationships among constructs in the model. As presented in Table 1, the assessment of threat appraisal revealed an AVE of 0.506, which demonstrated its validity. Although coping appraisal revealed an AVE of 0.458, slightly below the strict threshold of 0.5, its composite reliability was high, confirming the internal consistency of the measure.”

4.Measures and content validity

4.1 Climate-change knowledge (CK) uses subjective knowledge items only (“I have knowledge…”). Given your moderation findings hinge on CK, please (a) explicitly label CK as “subjective knowledge,” (b) add/acknowledge an objective knowledge scale in future work, and (c) discuss how subjective overconfidence might explain the negative moderation (you begin to do this in 5.1–5.2; make it more explicit up front).

The authors appreciate the reviewer’s valuable suggestion regarding the construct “climate change knowledge”. To increase clarity and keep consistent with later discussion, the authors explicitly phrase this construct as “subjective knowledge about climate change” and label it as “SK” and introduce this construct early in the introduction part to reduce perceived contradiction later. More detailed definition to subject knowledge and discussion on how subjective overconfidence might explain the negative moderation is added in the literature review section. The supplementary content has been added to the “Introduction” section as follows:

“Although public perception and awareness of climate change in growing in China [10], it is worth discussing whether increased subjective knowledge in this field can affect people’s appraisal of threat and coping efficacy of climate change.”

The supplementary content has also been added to the “Literature review” section as follows:

“Nevertheless, the role of knowledge is complex, particularly when considering the distinction between subjective and objective knowledge. Subjective knowledge represents the degree to which an individual believes he or she understands a topic [63]. Although both types of knowledge play important roles in people’s cognitive process, subjective knowledge has been proved to strongly associated with consumer behaviors and is easier to be measured than objective knowledge [64]. Thus, the present study measures subjective knowledge only. Some studies have found that elevated subjective knowledge can paradoxically weaken risk perception [65, 66], often attributed to the development of overconfidence [67], which leads individuals to underestimate threats. In contrast, other research reported a positive association, aligning with the conventional view that knowing more makes one feel more vulnerable [68]. Yet another body of work has found non-significant relationships, suggesting the influence of contingent factors [28, 64]. This inconsistency underscores a critical gap of the specific mechanism through which subjective knowledge influences the cognitive appraisals within protection motivation frameworks and moderates the formation of awe. ”

4.Measures and content validity

4.2 The awe scale includes semantically overlapping items (“evoked my awe,” “made me feel awe”); consider reporting an item-level EFA/CFA or dropping redundancy for parsimony.

The authors appreciate the reviewer for the valuable comments on semantic overlap. The scale for “awe” was adapted from Tsaur et. al (2024), which was tested to be a strong measurement for this construct in prior study. The two semantically overlapping items, to our understanding, can be used to double check respondents’ consistent perception of awe. Nonetheless, we still conducted exploratory factor analysis to distinguish the construct validity. The results of the EFA clearly identified 7 different factors, with the four items loaded on the same factor, which were in complete agreement with our expectations. The results of the EFA are presented here for your kindly review. Thus, we carefully considered your suggestions and ultimately did not delete the duplicate items.

5.. Sampling, inclusion criteria, and generalizability

·         5.1 Data collection was via Wenjuanxing and social media (July 11–Aug 9, 2025). Clarify the recency of the mountain visit (you screen for “visited high mountain areas in the Hengduan,” but not when). Temporal distance can affect awe recall (you acknowledge this as a limitation; consider adding it to Methods as a measured covariate). (Section 3.2; Section 5.4.)

The authors appreciate the reviewer’s valuable suggestion regarding temporal distance. We sincerely agree that this is a very important factor that can affect awe recall. However, during questionnaire development stage, we did not include it as a measured item. Due to the time limit, we are not able to add it as a measured covariate for the current study. As we acknowledged in the “limitations and recommendation for future research” section, we will add this temporal variable to our study as a moderation to test whether the level of awe will be influenced by different levels of temporal distance based on construal level theory.

5. Sampling, inclusion criteria, and generalizability

·         5.2 Specify how you verified visits to one of the “eight most renowned 5A-level scenic spots” (self-report only? list provided?). Also discuss coverage bias from recruiting on Xiaohongshu/Weibo (skews to younger, digitally active travelers). (Section 3.2.)

The authors appreciate the reviewer’s valuable suggestion regarding verification of visits to one of the “eight most renowned 5A-level scenic spots” in Hengduan Mountains. We are sincerely sorry for failing to make it clear and explicit. We did verify visits to one of the “eight most renowned 5A-level scenic spots” by providing a list of the 8 scenic spots which were carefully selected according to certain criteria. The supplementary content has been added to the “3.2 data collection and sample descriptive analysis” section as follows:

“These sites were selected based on a comprehensive set of criteria, including their officially certified 5A level, annual visitor numbers, and their prominence in regional tourism marketing. The finalized list comprised the following eight destinations: Daocheng Yading, Yulong Snow Mountain, Siguniang Mountain, Zheduo Mountain, Genie Sacred Mountain, Gongga Snow Mountain, Haizi Mountain, and Shiziwang Peak. To verify visitation, the first question of the survey explicitly presented this list to respondents and asked them to confirm whether they had ever visited at least one of these specific scenic areas.”

As kindly suggested by the reviewer, we extend discussion on possible coverage bias from recruiting on social media platforms. The supplementary content has been added to the “3.2 data collection and sample descriptive analysis” section as follows:

“The respondents were relatively evenly distributed across the three age groups of under 25, 25-40, and 40-65 years old, which was consistent with the physical profile of typical adventure tourists to high-altitude destinations, for whom such travel was often more accessible.” “Concurrently, it was also possible that this distribution mirrors the user base of the social media platforms, such as Xiaohongshu and Weibo, employed for recruitment, which are predominantly utilized by younger, urban demographics in China.”

6. Analysis transparency and correctness

6.1Please report means, SDs, and inter-construct correlations for all latent variables (beyond Fornell–Larcker/HTMT), plus item descriptive stats and normality, to help readers evaluate the measures. (Section 4.3.)

The authors appreciate the reviewer’s valuable suggestion regarding statistics report. We sincerely believe adding these statistics can enhance the transparency and assessability of our data. We have now reported means, standard deviations, and inter-construct correlations for all latent variables (beyond Fornell–Larcker/HTMT), along with item descriptive statistics and normality.

We add Table 4 in “4.1 common method bias” section which presents item descriptive statistics and full collinearity VIF. The supplementary content has been added to the “4.1 common method bias” section as follows:

Furthermore, Table 4 summarizes the item-level statistics, including collinearity assessed via the full VIF procedure. Cronbach’s α estimates of all the seven constructs were greater than the suggested cut-off values of 0.70, ranging from 0.757 to 0.913. A vast majority of items met Kock’s 3.3 rule [82], indicating the collinearity is not a general concern in the measurement model. Two items (AWE2 and AWE 3) exhibited VIF values above the threshold (4.667 and 3.962 respectively) but below 5.0. it is important to note that these items demonstrated high outer loadings and the corresponding construct demonstrated high composite reliability. Therefore, we retained these items for their theoretical relevance. Their inclusion did not pose a critical threat to the model’s interpretation.

We add Table 6 in the “4.3 measurement model” section which presents means, SDs, and inter-construct correlations for all latent variables together with statistics describing model reliability and validity. In Table 6, the diagonal shows the square root of the AVE, the values below the diagonal are the correlation coefficients, and the values above the diagonal are the HTMT values (in italics).

6. Analysis transparency and correctness

·         6.2 Measurement model. Table 6 appears to have formatting/labeling inconsistencies (e.g., “CB8,” “PK” vs “CK”). Ensure construct acronyms are consistent with earlier sections. (Table 6.)

The authors appreciate the reviewer’s valuable suggestions for label consistency and cleaning minor formatting/typos. We are sincerely sorry for those careless typos and inconsistency in construct labeling. We truly believe credibility can be largely increased by making those corrections. We label the construct “subjective knowledge about climate change” as “SK” and climate change behavior intention as “CB” consistently. We correct and double check formatting issues and typos. All those corrections have been marked in the red color for your kindly reference.

6. Analysis transparency and correctness

6.3 Structural model. In Table 6, several 95% CIs are impossible (e.g., CA→CB β=0.126 but CI reported as [0.316, 0.506]; H3b β=0.213 with lower CI 0.224, which exceeds the point estimate). Re-run/bootstrap and correct the CI outputs; if using bias-corrected intervals in SmartPLS, state that explicitly. (Table 6.)

The author thanks the reviewer for their careful review and pointing out this crucial issue. We sincerely apologize for the mistake shown in Table 6. As you have pointed out, the data in the report is indeed incorrect and does not conform to the regulations. This is due to our carelessness in the data export processing. We have now recalculated and corrected the error, and have replaced it with the correct data. CA→CB β=0.126 and CI reported as [0.073,0.181]; H3b β=0.213 and CI reported as [0.127,0.299].

6. Analysis transparency and correctness

6.4 Research Methodology

Common method variance. Harman’s single-factor test is insufficient on its own. Please add a marker-variable approach or full collinearity VIFs (Kock’s 3.3 rule) and/or a latent method factor check. (Section 4.1.)

The authors appreciate the reviewer’s valuable suggestions regarding common method bias test. We agree that the Harman one-factor test has limitations, and we sincerely appreciate your suggestion to use a more robust method to test for common method bias. Following your advice, we conducted the following supplementary analysis by conducting an indicator-level full collinearity VIF test. Table 4 presents the required data. We propose that common method bias is not a general concern in our model based on both Harman’s test and full-collinearity VIF test. The supplementary content and table3 have been added to the “4.1 common method bias” section as follows:

“Furthermore, Table 4 summarizes the item-level statistics, including collinearity assessed via the full VIF procedure. Cronbach’s α estimates of all the seven constructs were greater than the suggested cut-off values of 0.70, ranging from 0.757 to 0.913. A vast majority of items met Kock’s 3.3 rule [82], indicating the collinearity is not a general concern in the measurement model. Two items (AWE2 and AWE 3) exhibited VIF values above the threshold (4.667 and 3.962 respectively) but below 5.0. it is important to note that these items demonstrated high outer loadings and the corresponding construct demonstrated high composite reliability. Therefore, we retained these items for their theoretical relevance. Their inclusion did not pose a critical threat to the model’s interpretation.”

6. Analysis transparency and correctness

6.5 Given cross-sectional self-reports, consider a Gaussian copula or 2SLS/latent instrumental approach for key paths (especially AWE→BI), or at minimum discuss endogeneity risks.

The authors appreciate the reviewer’s valuable suggestions regarding endogeneity risks due to cross-sectional self-reports. We agree with the reviewer on the possible endogeneity risks which is a common concern with more cross-sectional self-reports. The authors extend discussion on endogeneity risks and

the supplementary content and table3 have been added to the “5.4 limitations and recommendation for future research” section as follows:

“First, the reliance on self-reports raises potential endogeneity concerns. Although we have rigorously tested our hypothesized model, the risk of endogeneity cannot be fully ruled out due to omitted variable bias, reverse causality and measurement error. Future research should address these by incorporating longitudinal or experimental designs to establish temporal precedence and causality. Further, multi-group or interaction analyses should also be employed to compare groups based on prior visits, age or gender to test the model’s robustness.”

6. Analysis transparency and correctness

6.6 Moderation reporting. For CK×TA→AWE, provide simple-slope plots, report whether predictors were mean-centered, and consider Johnson–Neyman regions to interpret the negative moderation. (Table 6, Figure 2.)

The authors appreciate the reviewer’s valuable suggestions for presenting interaction plots and exploring conditional indirect effects. We sincerely believe your suggestion can enhance our results and discussion. We have conducted simple-slope tests at ±1 SD knowledge and present a interaction plot. Figure 3 demonstrates that knowledge weakens TA→AWE→CB. The supplementary content and Figure 3 have been added to the “4.4 hypothesis testing” section as follows:

“After confirming the significant mediating effect of awe, we estimated the indirect effects at different levels of the moderator. The significant negative interaction between threat appraisal and subjective knowledge indicates that the positive relationship between threat appraisal and behavior intentions though awe is weakened as subjective knowledge increases. As illustrated in figure 3, this relationship was stronger when subjective knowledge was low (-1SD, simple slop=0.203) and the relationship became weaker when subjective knowledge was high (+1SD, simple slope=0.098).”

7. Interpretation and claims

·         7.1 Please temper causal language (e.g., “triggering,” “drives”) to reflect correlational design. Emphasize intentions rather than “behaviors,” except where you truly measured behavior. (Abstract; Sections 1, 5.1.)

The authors appreciate the reviewer’s valuable suggestions for appropriate language and claims. We change causal language such as “trigger” to “foster” or “drivers” to “what can motivate” to reflect correlational design. The revised words and phrase are marked in red and two examples of these revisions in line 17 and 69 are as follows:

“The results indicate that threat appraisal and coping appraisal are significantly associated with stronger tourist intentions for climate change mitigation, by fostering positive emotional responses.”

“It is crucial to understand what can motivate tourists to behave in a way that fosters pro-environmental action.”

Intentions rather than “behaviors” are now emphasized where we just measure intentions rather than behaviors. The revised words and phrase are marked in red and three examples of these revisions in line 148 and 207 and 209 are as follows:

“2.2. PMT and climate change mitigation behavior intentions”

“H1. Threat appraisal is positively related to climate change mitigation behavior intentions.

H2. Coping appraisal is positively related to climate change mitigation behavior intentions.

”

7. Interpretation and claims

·         7.2 The negative CK moderation is intriguing. To avoid over-attributing to Dunning–Kruger, discuss alternative explanations (e.g., information avoidance, risk normalization among better-informed travelers).

The authors appreciate the reviewer’s valuable suggestions for avoid over-attributing to Dunning–Kruger effect. We extend discussion on the possible reasons for the negative interaction effect from the lens of information avoidance and risk normalization in addition to attribution to Dunning–Kruger effect. The supplementary content has been added to the “5.1 general discussion” section as follows:

“Furthermore, alternative theoretical perspectives can also elucidate this finding. Information avoidance [94] might stop individuals with high subjective knowledge from deeply processing the threatening information. This defensive avoidance could be a coping strategy to manage “eco-anxiety” [95] associated with fully confronting the crisis. In addition, the phenomenon of risk normalization [96] could provide another explanation. Repeated exposure to climate-related information though both media and personal travel experience may reduce its novelty or salience. This normalization process would dampen the emotional impact necessary for awe to emerge, as the stimulus is gradually appraised as a familiar and normalized risk.”

8. Practical implications

·         Useful suggestions for communication design (emphasize severity/likelihood, show feasible coping paths). Consider adding message examples (e.g., signage, in-app nudges at trailheads, low-carbon itinerary prompts) and how DMOs could experimentally test awe-eliciting media (A/B tests) at specific Hengduan 5A sites. (Section 5.3.)

The authors appreciate the reviewer’s valuable suggestions for useful suggestions for communication design. On the one hand, we emphasize severity/likelihood to arouse awe from tourists. The supplementary content has been added to the “5.2 practical implication” section as follows:

“Specifically, instead of a sign that simply reads “Protect the Glaciers”, on-site interpretive signage can play a more effective educational role, by containing information that focuses on clarifying the severity of climate change threats to mountain ecosystems such as glacial retreat or biodiversity loss and their anthropogenic impacts. In addition, audio and visual cues facilitated by AR or VR technologies could translate the severity of climate change impact into tangible tourist experience. For example, an AR application can enable tourists to view the contemporary landscape with an overlay that shows how it looked like decades ago to demonstrate the process of glacier retreat. The application of soundscapes can also create an immersive and emotionally resonate context for severity information. Soundscape stations can be built at pivotal locations, playing natural sounds that are diminishing with climate change.”

On the other hand, we propose more specific examples to show more feasible coping paths. The supplementary content has been added to the “5.2 practical implication” section as follows:

“Next, highlighting accessible pathways to implement these pro-environmental actions is also meaningful. When tourists begin a hike, an in-app nudge could push a notification which clearly communicates the sensitivity of the ecosystem and the effectiveness of personal emission-reduction behaviors like carpooling or choosing public transportation as coping responses. During trip planning process, official websites or OTAs could provide low-carbon itinerary options rather than conventional routes. The prompt on the website could suggest that theses itineraries prioritize electric shuttle transport, local-food based meal options, and certified ecolodges.”

9. Ethics and compliance statements

·         You indicate Institutional Review Board Statement: Not applicable and Informed Consent Statement: Not applicable despite collecting identifiable human survey data (demographics, attitudes). Please clarify local requirements and confirm that informed consent was obtained; MDPI typically expects an ethics statement even for minimal-risk surveys.

The authors appreciate the reviewer’s valuable suggestions for ethics and compliance statements. Regarding the ethical review of this study, the MDPI editorial office also recognized its importance when we submitted the manuscript. Therefore, the editorial office requested that we provide documentation of ethical approval. On October 24, we submitted to the editorial office the official statement from Sichuan International Studies University granting an exemption from ethical review, along with a sample of the consent form used for participation in this research. In fact, the ethical review application for this study was submitted to Sichuan International Studies University prior to data collection and was formally approved for exemption on July 10, 2025. Data collection commenced only after receiving this approval. Accordingly, the ethical review process of this study fully complies with academic standards.

10. Presentation and formatting

·         10.1 Replace relative time phrases (“during … of this year”) with calendar dates. (Introduction.)

The authors appreciate the reviewer’s valuable suggestions for replacing relative time phrases with calendar dates. We make a revision to that and the revised content is marked in red in the Introduction section as follows:

“During the period from June 23 to July 2 2025, soaring temperatures across Europe resulted in 2,305 deaths across 12 major cities,…”

10. Presentation and formatting

·         10.2 Standardize acronyms (CK vs PK), fix table typos, ensure figure/table captions follow journal style, and verify all references are complete (e.g., Kruger & Dunning missing venue/year; some entries lack full bibliographic details). (References.)

The authors appreciate the reviewer’s valuable suggestions for presentation and formatting. We standardize acronym for subjective knowledge on climate change as SK since we have made it explicit as subjective knowledge.

We fix table typos and mark them in red color.

We revise and ensure figure/table captions follow journal style

We verify references are complete. For example, Kruger & Dunning (1999) has been completed in the Reference section.

“Kruger, J.; Dunning, D. Unskilled and Unaware of It: How Difficulties in Recognizing One’s Own Incompetence Lead to Inflated Self-Assessments. J. Personal. Soc. Psychol. Attitudes Soc. Cogn. 1999, 77 (6), 1121–1134. https://doi.org/10.1037/0022-3514.77.6.1121.”

11. Comments on the Quality of English Language

A careful language edit by a fluent academic editor will improve readability and professionalism.

The authors appreciate the reviewer’s valuable suggestions for language edit. We sincerely agree that a careful language edit by a fluent academic editor will improve readability and professionalism. We submit our manuscript for English language editing by MDPI. The text has been checked for correct use of grammar and common technical terms, and edited to a level suitable for reporting research in a scholarly journal.