The determinants affecting the intention of urban residents to prevent flooding in China

General remarks

This is a very interesting article, related to the topics of interest of the Journal. Overall, literature review and methodology are thorough, and there are only few issues regarding presentation, comprehension, structure, and use of language. I think it's worth accepting after addressing these issues.

Significant remarks

First of all, the title can be improved as it is not accurate. In my opinion, the residents are not expected to prevent flooding, but to prepare against flood risk, to be able to cope with flooding…etc. I would suggest to chose another expression.

The major issue concerns model hypotheses: In the illustrated model it appears that attitude is considered a mediator between trust and intention. Also, that PC and PL mediate the effects of trust on worry and intention. These are not reported as hypotheses, which makes reading of results confusing. In general, I would suggest to remove the words direct/indirect in hypotheses, and to refer to effects and specify the mediations. The subject of the paper is what affects preparedness intention, and which are the underlying mechanisms (here worry, risk perception and attitude). This need to be better expressed. Direct/indirect could be directly used in Results.

Results: A lot of information provided here belong to Methods. I suggest the methodological approach to be analysed in the previous section, while ‘Results’ to present only the SEM results and reliability/validity/significance results, not the theoretical background. This makes it difficult to follow results and understand the effects.

L289: I thought PMT was the basis for the proposed model. Why are the authors referring to a PMT model, which has not been previously illustrated as a stand alone model analysed within this paper? This is confusing, as the reader does not recognize which part of the Fig1 model is the PMT. In Methods, there was not any reference to analysis and comparison of 2 models. Please be clear on these issues.

Specific remarks

References are ok, except Holub et al., 2012 (L32) that does not seem relevant.

L50: Please consider being more specific: ‘communication’ of what, and/or with who?

L54: Section XX?

L112: Please correct ‘have’ to ‘has’

L134: Please consider using a singular form: a direct effect.

Table 1: a) in Oper.def. of Self-efficacy: please consider changing the wording ‘Personal…your ability’ to e.g. ‘perceived personal ability…’ to be consistent with wording used in the rest of the definitions. B) in definition of Trust: Please correct as: trust in the government’s…

L202: Please merge the two sentences.

L207: Please consider rephrasing, as ‘hypothesis test the survey data’ does not sound correct.

L208-211: Please consider rephrasing. It is difficult to read these sentences. Also, too many ‘and’, even in the beginning of a sentence. SEM could better introduced.

L224: This explication about women needs clarification. What was the title and introduction of the questionnaire, and how they affected the attraction of female respondents. Otherwise, this difference is commonly statistically expected (based on population statistics).

L225: Please don’t start a sentence with ‘And’. This is the case also in other parts of the paper.

L230: I would suggest to add in parenthesis income in USA dollars or Euros.

L241: Replace comma with fullstop.

Table S2: Please consider adding notes for what CR and AVE signify.

L247: Please use small letter for p (value)

L275. Please rephrase to produce complete sentences. VIF was lower than 1.7. The critical value is 3.3 (not less, the critical value is one, and the acceptable ones lie below this).

L284. Please don’t use >< within the flow of the written speech. These are usually used in parentheses, eg for p-value. Replace with lower/less/more/higher than.

L287: it is probably better to use the word threshold than categories and rephrase. Actually, as I understand there are four levels defined by three thresholds. I assume that below 0.1 it is unacceptable.

Table 4: Add note to explain the initials. Why do you report R2 for each construct, and not the overall R2 of the model? The same question applies for the preceding text.

L321-323. These 2 sentences, should be somehow merged. ‘Because’ is a conjunction.

L331-L334: Very confusing. Attitude here is considered a mediator?? This was not reported in the model introduction or among hypotheses. This is an issue that needs to be addressed properly.

L334-335: what do you mean? Which effects were insignificant? The point is that worry and attitude mediated trust.

L338: As I suggested previously, in my opinion wording here is inaccurate. Residents are not expected to prevent flooding but prepare against flood risk.

L349: Please correct: consequence ‘from, due to….’, not ‘to’.

L355: Please merge the sentences, don’t cut them in the middle with fullstop (! this is a recurrent issue throughout the article).

L356: Please replace the word ‘poor’ with a more suitable one, eg negative.

L368-369: Again, merge the sentences.

L388: Another social context factor to consider is the protection of public flood protection measures: Please rephrase, it is not understandable.

L405-407: This needs to be rephrased. Especially ‘how to do’ and ‘extreme floods exceeding standards’.

L407-408: Please rephrase and merge (again the same issue with fullstop in the middle of nowhere.)

L409: Please rephrase (what will be delayed??)

L416-417: Please rephrase. The sentences sound awkward, they need correction.

NHESS review

This is an interesting paper asking the question of how we explain people’s intentions to limit the impacts of flooding. Overall, I think the paper took several sensible choices. One top of the previous reviewers comments (many of which I also agree with), I would add the following (in no order of importance):

• What is the research gap that this paper is trying to fill? I ask because there is a large amount of research on behaviour intentions, and it is unclear in what direction the paper seeks to advance this literature. Is it for example is it, the application of this model to China[1, 2], or the application of a structural model of PMT to urban residents[3]? I raise this last paper as a comparison, as I think you follow a similar approach to those authors in that you start with PMT as a baseline model but greatly extend it into one that can start to be seen as moving towards the PADM[4], or other studies extending PMT[5].
  1. Similarly, as the model contains more variables than the commonly used version of PMT used, is it that surprising that it has a better predictive capability? It would be strengthened if you could validate how much of a better predicative capability your proposed model has, and if the increased survey complexity is worth this trade-off?
• The statement made on page 3, lines 81-82. This is quite a strong statement to make, especially as you state that you can model the heterogeneity as differences in risk coping attitude, as you define from data collected in your survey. I would think this raises the question of how this can deal with the heterogeneity in a more complete way than previous studies, as are not the different responses to the questions also looking at heterogeneity in the respondents?
• The argumentation on page 4, I think this section needs more nuance introduced into it. This is because the relationship can also be explained by the temporal dynamics of adaptive actions being mismatched with the cross-sectional nature of the data commonly connected[6-13], which when accounted for can explain the relationships seen to be missing.- This could also be seen as explaining the positive relationships with coping approval as people can increase their perceived coping appraisal after successfully implementing measures as well.
• The findings in relation to trust and social norms can also have additional nuance. This is because trust/social normal can impact multiple avenues of PMT both positively and negatively[14-17]. This has two further comments:
  1. Be more careful in using the word “trust” as you are explicitly talking about trust in governmental investment in protective measures
  2. Given that there is a literature explaining potential interconnections between “trust” and other social factors to coping appraisals, why is it not connected in your model as well?
• Could you place the 9.50 RMB incentive for completing the survey in the social context – is this a large incentive, was it reasonable to assume it would help people answer the survey.
• Following the information presented on the living conditions of the respondents described on page 10, I think you should mention the type of measures directly studied on the survey into the main paper itself. This is because if all the measures being asked for are not reasonable for employment by the households surveyed, I would question the fundamental validity of their stated intentions to employ the measures.
• I might be missing but for table 4, I only see that one of the variables has been defined elsewhere in the text.

• Noll, B., T. Filatova, and A. Need, One and done? Exploring linkages between households' intended adaptations to climate-induced floods. Risk Analysis, 2022. n/a(n/a).
• Huang, J., et al., Affect Path to Flood Protective Coping Behaviors Using SEM Based on a Survey in Shenzhen, China. International Journal of Environmental Research and Public Health, 2020. 17(3): p. 940.
• Dillenardt, L., P. Hudson, and A.H. Thieken, Urban pluvial flood adaptation: Results of a household survey across four German municipalities. Journal of Flood Risk Management. n/a(n/a).
• Lindell, M.K. and R.W. Perry, The Protective Action Decision Model: Theoretical Modifications and Additional Evidence. Risk Analysis, 2012. 32(4): p. 616-632.
• Botzen, W.J.W., et al., Adoption of Individual Flood Damage Mitigation Measures in New York City: An Extension of Protection Motivation Theory. Risk Analysis, 2019. 39(10): p. 2143-2159.
• Hudson, A.H. Thieken, and P. Bubeck, The challenges of longitudinal surveys in the flood risk domain. Journal of Risk Research, 2020. 23(5): p. 642-663.
• Bubeck, P., et al., Using Panel Data to Understand the Dynamics of Human Behavior in Response to Flooding. Risk Analysis, 2020. n/a(n/a).
• Bubeck, P., W.J.W. Botzen, and J.C.J.H. Aerts, A review of risk perceptions and other factors that influence flood mitigation behavior. Risk Analysis, 2012. 32(9): p. 1481-1495.
• Osberghaus, D., The effect of flood experiences on household mitigation – Evidence form longitudinal and insurance data. Global Environmental Change, 2017. 43: p. 126-136.
• Mondino, E., et al., Exploring changes in hydrogeological risk awareness and preparedness over time: a case study in northeastern Italy. Hydrological Sciences Journal, 2020. 65(7): p. 1049-1059.
• Seebauer, S. and P. Babcicky, (Almost) all Quiet Over One and a Half Years: A Longitudinal Study on Causality Between Key Determinants of Private Flood Mitigation. Risk Analysis, 2020. n/a(n/a).
• Siegrist, M., Longitudinal Studies on Risk Research. Risk Analysis, 2014. 34(8): p. 1376.
• Siegrist, M., The necessity for longitudinal studies in risk perception research. Risk Analysis, 2013. 33(1): p. 50-51.
• Hudson, P., L. Hagedoorn, and P. Bubeck, Potential Linkages Between Social Capital, Flood Risk Perceptions, and Self-Efficacy. International Journal of Disaster Risk Science, 2020. 11: p. 251-262.
• Babcicky, P. and S. Seebauer, The two faces of social captial in private flood mitigation: opposing effects on risk perception, self-efficacy, and coping capacity. Journal of Risk Research, 2017. 20(8): p. 1017-1037.
• Lo, A.Y., The role of social norms in climate adaptation: Mediating risk perception and flood insurance purchase. Global Environmental Change, 2013. 23(5): p. 1249-1257.
• Lo, A.Y., et al., Social capital and community preparation for urban flooding in China. Applied Geography, 2015. 64: p. 1-11.