Gamified Disaster Education. A Systematic Review of Serious Games on Dealing with Natural Hazards and of their Impacts and Outcomes for Users

Heidenreich, Anna; Kox, Thomas; Wankmüller, Franziska; Gerhold, Lars

doi:10.5194/egusphere-2026-2372

Preprints

https://doi.org/10.5194/egusphere-2026-2372

Preprints

20 May 2026

| 20 May 2026

Gamified Disaster Education. A Systematic Review of Serious Games on Dealing with Natural Hazards and of their Impacts and Outcomes for Users

Anna Heidenreich, Thomas Kox, Franziska Wankmüller, and Lars Gerhold

Abstract. Natural hazards pose a threat to human life and health. As a tool of risk communication and disaster education serious games are being used with increasing frequency. Even though an increase in research in the field is observable, little is known about serious games’ actual effectiveness. This systematic review gives an overview of existing publications on serious games dealing with natural hazards and evaluations of reported effects on users. We analyse 132 scientific articles and conference papers published between 2007 and 2025. Most publications describe digital serious games; analogue or hybrid formats were rarely explored. Flooding is the hazard addressed most frequently. 17 publications introduce and describe the serious games, while about half of the publications (n=73) report a usability evaluation of the games. This review focuses on the 42 publications that also include an evaluation of the serious games’ impact on the player. Most of these effectiveness evaluation studies do not apply any theoretical frameworks for the game development and/or evaluation processes. A noticeable concentration of findings concerns positive impacts on knowledge and self-efficacy. Behaviour, motivation, intention, further psychological impacts as well as specific competencies are explored in comparably few publications. Our findings show that theoretical frameworks play a subordinate role in the current research landscape examined. The quality of applied evaluation methods varied considerably. Most commonly reported impacts and outcomes for players were increased knowledge and self-efficacy.

Received: 24 Apr 2026 – Discussion started: 20 May 2026

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.

Download & links

Anna Heidenreich, Thomas Kox, Franziska Wankmüller, and Lars Gerhold

Status: final response (author comments only)

RC1:
'Comment on egusphere-2026-2372', Anonymous Referee #1, 10 Aug 2026

Please see comment attached

Citation: https://doi.org/10.5194/egusphere-2026-2372-RC1
- AC1:
  'Reply on RC1', Anna Heidenreich, 09 Oct 2026
  Dear Reviewer1,
  Thank you for your thorough and constructive feedback on our manuscript. Below we reply to every one of your points. For better readability our responses are given in italics.
  General assessment
  The topic addressed by this manuscript is both important and highly relevant to Geoscience Communication. Serious games are increasingly used in disaster-risk education and risk communication, yet the field still lacks a sufficiently clear and consolidated understanding of what these tools actually achieve, how their impacts are evaluated, and how strong the available evidence is. For this reason, I regard a systematic review of this literature as very much needed.
  Precisely because I think this review has the potential to become a useful reference for the field, some of the comments below are deliberately strict. They are intended to help the authors make the review methodologically stronger, more transparent, and more reliable as a basis for future research and practice.
  The manuscript addresses a relevant question: what evidence currently exists regarding the impacts of serious games used for natural-hazard education and risk communication? The distinction between studies that simply describe games, those evaluating usability, and those attempting to evaluate impacts on players is useful. The review also brings together a relatively large body of literature (132 publications) and highlights several patterns that are potentially valuable for researchers and practitioners, particularly the predominance of digital games, the strong representation of flood and earthquake applications, and the concentration of evaluation studies on knowledge and self-efficacy.
  I nevertheless have concerns about the methodological transparency of the review and, more importantly, about the strength of some conclusions drawn from the evidence collected. At present, the manuscript moves somewhat uneasily between a systematic mapping of the literature and an assessment of effectiveness. The former is reasonably convincing and potentially useful; the latter requires more caution. The 42 “effectiveness evaluation studies” are highly heterogeneous, most use relatively weak designs for causal inference, and the manuscript does not report a formal assessment of study quality or risk of bias. Yet several passages synthesise the evidence in terms that imply effectiveness, and the Discussion itself makes statements about the quality of the underlying studies without explaining how quality was systematically assessed.
  There are also important reproducibility issues in the search, screening and coding procedures. These are particularly relevant because the manuscript explicitly argues for more structured, standardised and transparent evaluation research. The review itself should meet a similarly high standard of methodological transparency. I consider the paper potentially suitable for Geoscience Communication, and the topic is clearly within the journal’s remit. However, I think major revision is required before the review can support the claims currently made.
  Major comments
  The literature search needs to be updated and described in substantially more detail
  
  The search was conducted on 16 January 2025, whereas the preprint entered discussion in May 2026. Given the sharp increase in publications documented by the authors themselves, this is a substantial gap for a review presented as an updated overview of the current research landscape. The manuscript should therefore be updated with a new search conducted as close as reasonably possible to submission/revision. This is not simply a matter of adding a few recent references. A search update may alter the frequencies reported throughout the Results, the composition of the 42 effectiveness studies, and potentially some of the conclusions.
  The search strategy also requires greater reproducibility. The manuscript states that Scopus, Web of Science and “EBSCOhost” were searched, but EBSCOhost is a platform hosting multiple databases. Please specify which EBSCO database(s) were used. Similarly, it is not clear which bibliographic fields were searched in each database (title, abstract, keywords, topic, etc.), or whether the displayed search string was used identically across all three platforms. Database-specific search syntax and field restrictions should be reported.
  The sensitivity of the search strategy also deserves consideration. The eligibility criteria include studies described using terms such as “disaster games” or simply “games”, yet the game-related part of the search is relatively restricted. The authors should explain how they assessed whether this strategy was sufficiently sensitive to capture studies not explicitly using “serious game”, “game-based learning”, “gamification”, or “educational game” terminology. I am not asking for an unnecessarily complex search strategy, but the manuscript needs enough information for another researcher to reproduce it and enough justification to demonstrate that the search is fit for the review’s stated scope.
  We appreciate the reviewer’s feedback regarding the time gap between our literature search and the current revision. While we understand the concern, we respectfully argue that a full re-run of the search is not essential for this review. Our analysis spans nearly two decades of research (2007–2025), and the addition of a single year’s publications is unlikely to substantially alter the overall trends or conclusions, which are based on long-term developments. Moreover, the most recent comparable reviews in this field date back to 2018 (Flood et al, Solinska-Nowak et al.), and our work already represents a significant advancement in scope and methodological rigor. Some lag between search and publication is unavoidable and does not undermine the validity of our findings. We will acknowledge the time gap as a limitation in the revised manuscript more clearly and clarify that our review provides a systematic baseline for the field up to early 2025.
  We will clarify the search process by specifying the EBSCO databases used and adding information about the search string.
  Given the high number of found records we argue that our search terms regarding serious games were not too narrow. Since the search was conducted in titles, abstracts, and keywords (see below), many papers were found that e.g., covered games and merely included the terms “serious” or “educational” in another context in their abstract.
  See also reply below: we clarified open questions in the description of the search strategy, following the reviewer’s suggestions.
  Screening, extraction and coding procedures are currently insufficiently described
  
  The manuscript states that the systematic review was conducted following the PRISMA 2020 statement. I suggest that the authors provide a completed PRISMA 2020 checklist, either as supplementary material or alongside the revision, indicating where each applicable item is addressed in the manuscript. At present, I could not identify complete reporting for several core items, including the information sources and database interfaces (Item 6), database-specific search strategies (Item 7), selection process (Item 8), data-collection process (Item 9), some elements of data-item definition and synthesis methodology (Items 10 and 13), registration/protocol information (Item 24), and availability of extracted data and review materials (Item 27). Not all PRISMA items will necessarily require an analysis-for example, meta-analysis-specific procedures are not relevant if no meta-analysis was undertaken-but the checklist would make it clear which items are applicable, which procedures were performed, and where they are reported. Where a procedure was not undertaken, the manuscript should state this rather than leaving the reader uncertain.
  Section 2.3 states that “at least one of the authors” read and documented each publication. This leaves several important aspects of the systematic review process unclear. I could not determine whether title/abstract screening and full-text screening were conducted independently by more than one reviewer; whether any subset of the records was double-screened; how disagreements about eligibility were resolved; whether data extraction or categorical coding was independently checked; whether a coding manual or operational definitions were established beforehand; or whether agreement between coders was assessed. This matters particularly because several central classifications require researcher judgement. Examples include deciding whether something constitutes a “complete” serious game rather than gamification, whether a study qualifies as a usability or effectiveness evaluation, assigning games to DRM phases, classifying learning objectives, identifying theoretical frameworks, and grouping measured outcomes into knowledge, competencies, motivation/intention/behaviour, and psychological effects. The manuscript does not necessarily need to have used two independent coders for every record, but it does need to explain the procedure actually followed and how coding reliability was protected. If no independent verification was performed, this should be transparently acknowledged as a limitation. I would also encourage the authors to provide the coding framework or extraction sheet as supplementary or repository material. This would substantially improve the reproducibility of the review.
  Thank you for pointing out the PRISMA checklist. We consulted the checklist again and worked through all points, especially those you mentioned. After implementing all the changed mentioned in both response letters we are positive that all points will be checked. As far as we have seen in other current systematic reviews, however, it is uncommon to publish the full checklist; therefore, we would not include it into the supplementary files.
  Thank you also for your comment on the review process; we will clarify this. We are planning to change the sentence to include that the first three publications were read by authors 1-3 and discussed afterwards. Disagreements between reviewers were discussed in the group in detail and the review strategy was improved in an iterative manner during the review process. Out of the total sample of publications, author 1 read half and authors 2 and 3 a quarter, each. In case one of the reviewing authors had open questions, these were discussed in the team and if necessary, another reviewer read the publication in question, too. Furthermore, we will add the following sentence: “The first author screened titles and abstracts and decided in favour of those texts that were on the brink of illegibility.”
  We will provide the list of questions used in our review protocol as a supplementary file and describe the key questions of the document in the text.
  The manuscript needs to distinguish more clearly between mapping reported outcomes and demonstrating effectiveness
  
  This is my main conceptual concern. The review labels 42 papers as “effectiveness evaluation studies” and RQ3 asks which impacts and outcomes serious games “have” for players. However, the methodological summary shows that only eight studies included a controlgroup comparison, whereas most relied on pre-post designs. Follow-up assessment was rare. Some outcomes were based on self-reported knowledge or behavioural intentions rather than demonstrated knowledge or actual behaviour. These studies can certainly provide useful evidence, but they do not all provide the same level of evidence for effectiveness. The manuscript does not report a systematic risk-of-bias or methodological-quality assessment. Nevertheless, the Discussion states that evaluation quality varied considerably and sometimes describes approaches as close to “pseudoevaluations”. Without predefined appraisal criteria, it is difficult for the reader to understand how these judgements were reached.
  I see two reasonable routes here. One option would be to introduce an appropriate and transparent methodologicalquality appraisal of the effectiveness studies. Given the heterogeneity of designs, this does not need to become a metaanalysis, nor do I think a meta-analysis is necessarily appropriate. Alternatively, the authors could more explicitly frame the work as a systematic evidence map/descriptive synthesis of reported evaluation outcomes, and revise the language throughout accordingly. In that case, statements such as “playing a serious game can indeed improve the players’ knowledge” should be replaced by formulations that reflect what the reviewed studies actually demonstrate, for example that multiple studies reported pre-post increases in knowledge, while the strength of causal evidence varies substantially across designs. This distinction is especially important for the conclusion that serious games have the potential to increase knowledge and self-efficacy. I think this conclusion is plausible from the reviewed literature, but the confidence that can be placed in it should be explicitly related to the quality and design of the underlying studies.
  Thank you for these points. We agree that in some parts the language of the manuscript was imprecise and we plan changes here for clarification. We indeed understand our work as a synthesis of reported evaluation outcomes, and following your suggestion we will frame this explicitly as such. Therefore, we will rewrite statements on the reported effectiveness to be clear. We will not use the term “pseudo-evaluations”, but explain our concerns about the heterogeneity among studies in more appropriate words. By revising the language in these relevant parts, we hope to clarify these questions. As mentioned above, we will provide the questions we used in our review protocol and add a more detailed results table as a supplementary file. We will add an outlook on further methodological questions in the field which points out the need for a review of studies on effectiveness.
  Operational definitions and outcome classifications need to be made more reproducible
  
  The manuscript rightly points out that terminology is inconsistent across the field, but some of this ambiguity also enters the review’s own analytical framework. For example, the authors state in Section 3.2.3 that some studies treat risk awareness, perception and understanding similarly to knowledge. This is an important observation, but it raises the question of how the reviewers themselves classified such outcomes. The manuscript should explain the decision rules used to distinguish factual versus self-perceived knowledge; knowledge versus risk awareness; risk awareness versus risk perception; competencies/skills; motivation; behavioural intention; actual behaviour; and psychological outcomes. This is particularly important because the frequency with which each outcome category appears is one of the main findings of the review.
  A similar issue applies to the central distinction among no evaluation, usability evaluation, and effectiveness evaluation. The conceptual explanation given earlier in the manuscript is helpful, but the coding criteria used in practice should be stated more explicitly. What minimum evidence was necessary for a publication to be classified as an effectiveness evaluation? How were studies containing both usability and impact outcomes classified? Were categories mutually exclusive by design? The manuscript would become considerably stronger if these classifications could be reproduced by another research group using the same definitions.
  We appreciate these very relevant points and will make a number of changes in order to foster the study’s reproducibility. We will add more precise definitions on the terms we applied into the methods section. More specifically, we will explain the differentiation of usability evaluations and effectiveness evaluations in section 2.2 to clarify these questions. We will explain that we categorised publications as evaluation studies if they reported carrying out any kind of tests, interventions, experiments or sessions (regardless the terminology) of games with people (regardless the sample size). Those publications reporting that they explored the games’ usability or functionalities were categorised as usability evaluation studies while those reporting research questions and methods exploring impacts and/or outcomes for users as effectiveness evaluation studies. As mentioned in 3.1, many effectiveness evaluation studies also included a usability evaluation; we will move this explanation to 2.2, as well. Furthermore, we will move the section explaining the adapted Connolly framework from the results section (3.2.3) to 2.2. We will clarify that we report the terms given in the pertinent publications (e.g., knowledge, risk perception) and discuss in section 4 possible implications that may arise from different use of terms. Thus, our “decision rule”, so to say was if a term was used in the publication, we categorized it as mentioned by the publications’ authors.
  There are numerical inconsistencies in the core review results that need to be resolved
  
  There is a clear inconsistency between Figure 1 and the Abstract/Results regarding the evaluation categories. Figure 1 reports: No evaluation: n = 23; Usability evaluation: n = 67; Effectiveness evaluation: n = 42. The Abstract and Section 3.1 report: No evaluation: n = 17; Usability evaluation: n = 73; Effectiveness evaluation: n = 42. Both sets sum to 132, so this appears to be a classification/version inconsistency rather than a simple arithmetic error. Because this categorisation determines which publications enter the detailed effectiveness analysis, it should be corrected throughout and the underlying dataset checked.
  There is a second issue in the description of study countries. For example, the manuscript reports USA n = 20 as 14.4% and Italy n = 12 as 8.6%. Those percentages are not percentages of the 132 publications. They appear instead to use a denominator of approximately 139 country occurrences, presumably because multi-country studies were counted more than once. If so, this is perfectly legitimate, but the denominator needs to be explicitly stated. At present, the text refers to percentages of “publications”, which makes the calculation difficult to interpret. Please conduct a consistency check of all frequencies, percentages, figures and the Appendix after the database is finalised.
  We will carry out a check for consistency once all revisions are done. The different numbers of the subsamples were indeed an error that we will correct. Regarding the countries of study: Table A1. shows that some studies were carried out in multiple countries while other studies did not report where the study was carried out. We will add a sentence explaining this in the results section, right where you pointed it out, in order to avoid further confusion.
  Several statements in the Discussion go beyond what can be established from the review
  
  There are a number of places where interpretation should be separated more clearly from evidence. Most importantly, the statement that “publication bias can be assumed” because negative and null findings were rarely reported is too strong. The observed pattern is certainly compatible with publication or selective-reporting bias, but this review has not formally assessed publication bias. I suggest stating that the scarcity of null findings raises the possibility of publication/selective-reporting bias. Similarly, the manuscript states that none of the 132 studies “had been pre-registered”. If this means that none of the publications reported preregistration, please say exactly that. Absence of a preregistration statement in the article does not necessarily demonstrate that a study was not preregistered elsewhere. The ethics discussion has the same problem. The manuscript correctly reports whether an ethics approval or ethics statement was mentioned, but subsequently states that the majority of EE studies “do not have ethics approvals”. From the information analysed, the defensible conclusion is that they did not report an ethics approval, unless the authors independently verified this elsewhere.
  Finally, some explanations for temporal publication patterns are highly speculative. The proposed connections between the 2016 increase and the Sendai Framework, and between the post-2020 pattern and COVID-19/distanced learning, are possible interpretations but are not tested in this review. They should be explicitly presented as hypotheses rather than explanations of the observed pattern. These changes would not weaken the paper; they would make the distinction between review evidence and reviewer interpretation much clearer.
  Following your suggestions we will change the phrases regarding ethics appraisal and pre-registration accordingly to be more precise. We will also change the wording in this discussion point to make clear that our assumptions are merely hypotheses.
  The Data availability statement is not consistent with the study that is reported
  
  The manuscript states: “No data sets were used in this article.” However, Section 2.3 explains that information from the publications was systematically collected in SoSciSurvey and analysed in SPSS and Microsoft Excel. That extracted and coded information is itself a research dataset underpinning the numerical results of the systematic review. I therefore do not think the current Data availability statement is appropriate. At minimum, the authors should provide the extracted review dataset underlying the figures, frequencies and classifications, together with enough documentation to interpret its variables. Ideally, the coding framework and, where possible, the search/export information should also be archived. This is particularly relevant because reproducibility and transparency are central to the manuscript’s own recommendations for the field.
  Part of the data set was provided in the appendix, Table A1. We will provide a more detailed list of the results of our reviewing process, alongside with the review protocol. We will adapt the data availability statement accordingly.
  Minor comments
  Title: “Gamified Disaster Education” is broader than the actual scope and may also create conceptual confusion because the manuscript explicitly excludes interventions consisting only of gamification elements. A title referring directly to serious games for natural-hazard education/risk communication would more precisely reflect the review.
  
  Following your and reviewer 2’s feedback we will adapt the title to “Serious Games about Natural Hazards and their Impacts and Outcomes for Users: A Systematic Review”.
  Abstract: Please correct the evaluation-category numbers once the discrepancy with Figure 1 has been resolved.
  
  As stated above, we will correct this error.
  Introduction, Section 1.1: “Recent educational discourses, reinforced by the COVID-19 pandemic underscores…” should be grammatically corrected (“underscore”).
  
  Thank you for pointing this out. We will correct this accordingly.
  The phrase “a long examination of the topic in a motivating way” in Section 1.1 is difficult to understand and should be rephrased.
  
  We will rephrase this part of the sentence.
  Section 1.3: “survey are commonly used” should be corrected.
  
  We will correct this (“surveys”).
  The authors use both “impact evaluation” and “effectiveness evaluation” for the same subset of studies. I suggest selecting one term and using it consistently, particularly because “effectiveness” carries a relatively strong causal implication.
  
  We will correct this and consistently use the term “effectiveness evaluation studies (ee studies)”.
  Figure 1: “Recorts” appears several times and should be corrected to “Records”.
  
  We will correct this to “Records”. Furthermore, we had a linguistic mix-up of records and reports, which we will fix following the PRISMA flow chart example.
  Figure 2: because the literature search took place on 16 January 2025, the 2025 bar represents only a very small fraction of the year. This should be stated directly in the figure or caption; otherwise the apparent fall in 2025 can easily be misread as a publication trend.
  
  We will correct the misleading figure title to make sure that the date of literature search is visible. We will add an asterisk in the figure above the 2025 bar and add an explanation in a note below the figure.
  Figure 3: the meaning of the colour intensity is not explained by a legend, and several country labels/numbers are difficult to read at manuscript scale. Please clarify the visual encoding and improve readability. The counting rule for multi-country studies should also be included in the caption.
  
  As stated above, we will add an explanation of the county of study counting. We will adapt the caption to clarify these points.
  Figure 4: “adressed” should be corrected to “addressed”. Since individual publications may address multiple hazards, the caption should explicitly state that categories are non-exclusive.
  
  We will implement this suggestion and correct the axis label.
  Table 1: theories, theoretical models, instructional-design frameworks and research approaches are combined in one list. Given that RQ1 specifically asks about behavioural theories or models, it may help the reader if these different types of frameworks are distinguished rather than presented as equivalent.
  
  Thank you for pointing this out. We will correct the wording of our RQ1 to “theoretical frameworks”, to be consistent with our results section. A more detailed discussion on different theoretical frameworks and their applications has to be carried out in future studies. We do not treat the reported ones as equivalent; we rather present them without any kind of rating in quality. We point out the disciplinary backgrounds of the theoretical frameworks that were applied.
  Section 3.2.2: I suggest revising “the majority of EE studies … do not have ethics approvals” to “do not report ethics approval”, for the reason discussed above.
  
  We will change this accordingly.
  Section 3.2.3: “participants perceptions” requires a possessive apostrophe.
  
  Thank you for pointing this out. We will correct this accordingly.
  Discussion: “In a several theoretical approaches” should read “In several theoretical approaches”; “meta-analysis show” should be “meta-analyses show” or “a meta-analysis shows”.
  
  Thank you for pointing this out. We will correct this accordingly.
  Section 4.1: “Although ethics approvals required…” is missing “are”.
  
  We will add the missing word.
  “metanalytical” should be corrected to “meta-analytical”.
  
  We will change this as suggested.
  Section 4.2: “games themself” should be “games themselves”.
  
  We will change this as suggested.
  Appendix Table A1: “Conclusive list” is unusual here; “Complete list” or “List of all publications included…” would be clearer.
  
  We will correct this accordingly.
  A final language edit would be worthwhile. The English is generally understandable and the argument can be followed, but there are enough grammatical and idiomatic issues to distract occasionally from the scientific content.
  
  The final manuscript will be subject to a language check.
  The manuscript cites at line 365 “Gargiulo et al. (2024)” in the discussion of risk perception, although two Gargiulo et al. publications from 2024 are present in the reference list and elsewhere distinguished as 2024a and 2024b. The relevant citation should be disambiguated consistently.
  
  We will correct this accordingly to make transparent which paper is addressed.
  “Feng et al. (2020, 2021)” appears (line 342) in the discussion of preparedness training although two Feng et al. publications from 2020 are listed as 2020a and 2020b. The relevant 2020 publication should be specified.
  
  Thank you for pointing this out. As said above, we will correct this accordingly.
  The Bareford et al. (2021) DOI is written as “https://doi.org/https://doi.org/...”, resulting in a duplicated DOI prefix.
  
  We will correct this as suggested.
  The Witte (1992) reference contains the same duplicated “https://doi.org/https://doi.org/...” construction.
  
  We will correct this as suggested.
  Bandura (1977a), “Self-efficacy: Toward a Unifying Theory of Behavioral Change”: the DOI currently given in the manuscript does not correspond to the Psychological Review citation listed. The DOI links to Bandura,A. .Self-efficacy: Toward a unifying theory of behavioral change, Advances in Behaviour Research and Therapy,Volume 1, Issue 4, Pages 139-161,ISSN 0146-6402, 1978 and not Bandura, A.: Self-efficacy: Toward a Unifying Theory of Behavioral Change, Psychological Review, 84, 191–215, 1977.The authors should verify which version they intend to cite and make the journal, year, pagination and DOI internally consistent.
  
  We will verify the correct DOI for these sources.
  The DOI formatting for Witte (1992) should be corrected by removing the duplicated DOI prefix.
  
  See 23.
  Because the search predates the manuscript by more than one year, the reference set should be re-generated after an updated database search rather than supplemented manually. I did not identify evidence that would justify questioning the existence of the broader body of references. The main bibliographic concerns are consistency, DOI accuracy and the need to update the review period.
  Recommendation
  Minor revisions
  The manuscript addresses a worthwhile question and contains a potentially valuable synthesis, but a revision is needed before the conclusions about serious-game effectiveness are sufficiently supported. The essential points are to update and fully document the literature search, clarify and strengthen the transparency of screening and coding procedures, resolve the numerical inconsistencies, distinguish descriptive evidence on reported outcomes from stronger claims of effectiveness, and either formally appraise methodological quality or substantially temper conclusions according to the strength of the underlying study designs. The data generated through the systematic review should also be made available, or the authors should provide a clear justification if this is not possible. Once these issues are addressed, I think the manuscript could make a useful contribution to Geoscience Communication.
  Thank you again for taking the time and reviewing our manuscript.
  
  Citation: https://doi.org/10.5194/egusphere-2026-2372-AC1
- AC3: 'Reply on RC1', Anna Heidenreich, 09 Oct 2026
  
  Thank you very much for your review. Please find our response in the attached document.
  
  Citation: https://doi.org/10.5194/egusphere-2026-2372-AC3
RC2:
'Comment on egusphere-2026-2372', Anonymous Referee #2, 01 Sep 2026

Review of manuscript egusphere-2026-2372
Thanks for the opportunity to review this interesting manuscript.
General comments

The manuscript provides a review of serious games and game-based methodologies in the field of natural hazards, with particular focus on evaluation of potential impacts of participating in game-based sessions. This is a highly needed and welcome overview of the use of serious games in natural hazards research, and could fit well in GC.
The manuscript is overall well written and does a good job pointing out current challenges around using gamified approaches in DRR research and education. While the conclusions may not be novel, the manuscript adds valuable systematic insights to the field, especially given the temporal scope covering almost two decades, and will help future research make sensible methodological choices. Doing so, it positions itself well among some other recent review papers with a related, but somewhat different, scope (Flood et al., 2018; Nowak et al., 2018; Forrest et al., 2022). Below are a few major, and some small comments that the authors may consider when submitting a revised manuscript.
While the manuscript provides an overview of relevant approaches and definitions related to serious games, a key general comment that I have pertains to the limited extent to which authors expand on their definitions of ‘education’ and ‘risk communication’ and how they guide the analysis. Section 1.3 partly addresses this by noting a range of indicators for evaluation (e.g. based on the Competence Model for Environmental Education), but they pertain first and foremost to ‘learning’ outcomes, and the reader still has to make its own understanding of how they relate to education, risk communication or both. Would it be more straightforward to frame the review around learning and knowledge outcomes within DRR instead?
Specific comments

The title reads somewhat constructed (“Review of Serious games on dealing with natural hazards…”) - could this be rephrased more concisely?
Line 139-144: The description of inclusion/exlusion criteria are sufficient. However, instead of leaving implicit, would it be worth describing briefly (if relevant) how authors dealt with compound hazards that combine the natural hazard scope of the manuscript, with one or more hazards (e.g. health, chemical) that would fall outside of the scope?
Line 145-151: there are several papers out there that discuss different understandings of (serious) games, which is a quite relevant debate, the existence of which in itself shows that there may be a need for different methodological and theoretical perspectives in the field. Perhaps add one or two, to strengthen this paragraph?
Data analysis Section 2.3: There is potential to expand this section, and describe more thoroughly how authors interacted with the data throughout the analysis. Who did what, how did you deal with diverging insights. Overall, please provide more details on the analysis protocol.
Line 208-209: “Regarding the concept of serious games, many authors refer to the definitions by Abt (1970), or Michael and Chen (2011); others propose their own definitions without referring to specific authors.” Can this finding of varied theoretical embeddedness be described more thoroughly? I think it would add value as to position how the DRR field has been building on existing strands of theoretical definitions.
Line 438: “Based on our findings, we conclude that there is a clear need for a more structured approach to serious game research in the future.” This concluding sentence could be framed more specifically. For example, do you mean research on the use of serious gaming, research applying serious gaming as a theoretical and/or methodological approach, research specifically aimed at evaluation impact/outcomes of using serious gaming approaches?
Line 463: “To achieve better outcomes for education and communication an interdisciplinary approach on improving serious games is needed.” This sentence seems to function as a conclusion of the argument made in the paragraph. However, what about transdisciplinary approaches? Furthermore, the discussion could be strengthened by giving some examples of good practices (not necessarily within SG contexts, but perhaps from co-production approaches in DRR that successfully strengthen engagement of and with societal stakeholders)?
Minor comments

Figure 2 (line 181): were there any conference papers that subsequently were published as full papers?
Line 310-312: As the framework of Connolly is (partly) used to structure a major part of the research findings, this sentence could fit well in the Methodology section.
Line 426 - what is meant with ‘borderline pseudo-evaluations?
General: the manuscript needs a thorough proofread to check minor spelling/formatting issues across several tables as well as reference formats.

Citation: https://doi.org/10.5194/egusphere-2026-2372-RC2
- AC2: 'Reply on RC2', Anna Heidenreich, 09 Oct 2026
  
  Thanks for the opportunity to review this interesting manuscript.
  Dear Reviewer,
  Thank you for your thorough and constructive feedback on our manuscript. Below we reply to every one of your points. For better readability our responses are given in italics.
  General comments
  
  The manuscript provides a review of serious games and game-based methodologies in the field of natural hazards, with particular focus on evaluation of potential impacts of participating in game-based sessions. This is a highly needed and welcome overview of the use of serious games in natural hazards research, and could fit well in GC.
  The manuscript is overall well written and does a good job pointing out current challenges around using gamified approaches in DRR research and education. While the conclusions may not be novel, the manuscript adds valuable systematic insights to the field, especially given the temporal scope covering almost two decades, and will help future research make sensible methodological choices. Doing so, it positions itself well among some other recent review papers with a related, but somewhat different, scope (Flood et al., 2018; Nowak et al., 2018; Forrest et al., 2022). Below are a few major, and some small comments that the authors may consider when submitting a revised manuscript.
  While the manuscript provides an overview of relevant approaches and definitions related to serious games, a key general comment that I have pertains to the limited extent to which authors expand on their definitions of ‘education’ and ‘risk communication’ and how they guide the analysis. Section 1.3 partly addresses this by noting a range of indicators for evaluation (e.g. based on the Competence Model for Environmental Education), but they pertain first and foremost to ‘learning’ outcomes, and the reader still has to make its own understanding of how they relate to education, risk communication or both. Would it be more straightforward to frame the review around learning and knowledge outcomes within DRR instead?
  Thank you for pointing this out. We will add a definition on the relevant terms into chapter 1 to clarify this open question. We aimed at integrating all mentioned aspects but sometimes faced the difficulty, that the study’s authors used different wording in their papers. In our understanding based on the cited literature, (disaster) education focuses on recognizing risks, understand them and empowering to react appropriately. Disaster education aims at competency-oriented long-term goals. In comparison, risk communication is more situation specific and focuses on information transfer between stakeholders, e.g., institutions, governments, and the public. Some interventions seemed to focus on information transfer, and were thus closer to the concept of risk communication. Others named specific learning goals and framed their interventions as educational approaches. Again, other papers did not reference any of these concepts. However, we expect the audience of publications on natural hazard serious games may have an interest in learning more about the given serious games in order to draw their own conclusions for own planned risk communication or disaster education interventions.
  Specific comments
  
  The title reads somewhat constructed (“Review of Serious games on dealing with natural hazards…”) - could this be rephrased more concisely?
  Following both reviewers’ feedback we will adapt the title to “Serious Games about Natural Hazards and their Impacts and Outcomes for Users: A Systematic Review”.
  Line 139-144: The description of inclusion/exlusion criteria are sufficient. However, instead of leaving implicit, would it be worth describing briefly (if relevant) how authors dealt with compound hazards that combine the natural hazard scope of the manuscript, with one or more hazards (e.g. health, chemical) that would fall outside of the scope?
  As stated in the inclusion criteria (line 136), one or multiple hazards were a prerequisite. In lines 141 -143 we write “Studies which focused only on general health risks/public health (e.g., pandemics, viruses, COVID-19), general climate change adaptation, police work including cybercrime, or industrial and other man-made disasters (nuclear safety, fire unrelated to natural hazards) did not fulfil the inclusion criteria.“ In our view this makes explicit that a study with focuses on a serious game on one specific natural hazard and also another health risk would be eligible.
  Line 145-151: there are several papers out there that discuss different understandings of (serious) games, which is a quite relevant debate, the existence of which in itself shows that there may be a need for different methodological and theoretical perspectives in the field. Perhaps add one or two, to strengthen this paragraph?
  We find this point very relevant, too. In Chapter 1.1 (Definitions and Categorizations of Serious Games and Related Concepts), lines 38-60 we point out the variety of understandings of serious games in general. Within the eligibility criteria we find a deeper discussion of this relevant debate a bit misplaced. In lines 201ff we report on the usage of terms. We plan to include a stronger discussion of this finding in the results section.
  Data analysis Section 2.3: There is potential to expand this section, and describe more thoroughly how authors interacted with the data throughout the analysis. Who did what, how did you deal with diverging insights. Overall, please provide more details on the analysis protocol.
  Following your suggestions we will provide the analysis protocol that we used. Furthermore, we will elaborate on the review and analysis process, explaining the different reviewers’ activities in the process following both reviewers’ feedback.
  Line 208-209: “Regarding the concept of serious games, many authors refer to the definitions by Abt (1970), or Michael and Chen (2011); others propose their own definitions without referring to specific authors.” Can this finding of varied theoretical embeddedness be described more thoroughly? I think it would add value as to position how the DRR field has been building on existing strands of theoretical definitions.
  See also our response to your question regarding Serious Game definitions above. We plan to take up your point and discuss this further in the discussion section.
  Line 438: “Based on our findings, we conclude that there is a clear need for a more structured approach to serious game research in the future.” This concluding sentence could be framed more specifically. For example, do you mean research on the use of serious gaming, research applying serious gaming as a theoretical and/or methodological approach, research specifically aimed at evaluation impact/outcomes of using serious gaming approaches?
  Thank you for pointing this out. We will rephrase this sentence for clarification and suggest this wording: “Based on our findings, we conclude that there is a clear need for a more structured approach to serious game research in the future, particularly through the standardisation of evaluation protocols to facilitate comparisons across studies.”
  Line 463: “To achieve better outcomes for education and communication an interdisciplinary approach on improving serious games is needed.” This sentence seems to function as a conclusion of the argument made in the paragraph. However, what about transdisciplinary approaches? Furthermore, the discussion could be strengthened by giving some examples of good practices (not necessarily within SG contexts, but perhaps from co-production approaches in DRR that successfully strengthen engagement of and with societal stakeholders)?
  Thank you for pointing this out. We agree that transdisciplinary approaches would be beneficial. We will add this point here regarding SG research.
  Minor comments
  
  Figure 2 (line 181): were there any conference papers that subsequently were published as full papers?
  It is possible that the same research project produced a conference paper first and a full article later. However, we did not observe fully redundant analyses, therefore we saw no need in excluding any papers based on this.
  Line 310-312: As the framework of Connolly is (partly) used to structure a major part of the research findings, this sentence could fit well in the Methodology section.
  Thank you for pointing this out. We will move the sentence to chapter 2.2.
  Line 426 - what is meant with ‘borderline pseudo-evaluations?
  Following both reviewers’ feedback we will change the wording and avoid this term. We will include a more detailed description of the review process in order to enhance reproducibility.
  General: the manuscript needs a thorough proofread to check minor spelling/formatting issues across several tables as well as reference formats.
  The final manuscript will be subject to a language check. We will also double-check the formatting and numbers.
  Thank you again for reviewing our manuscript!
  
  Citation: https://doi.org/10.5194/egusphere-2026-2372-AC2

Anna Heidenreich, Thomas Kox, Franziska Wankmüller, and Lars Gerhold

Viewed

Total article views: 718 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
405	277	36	718	31	30

HTML: 405
PDF: 277
XML: 36
Total: 718
BibTeX: 31
EndNote: 30

Views and downloads (calculated since 20 May 2026)

Month	HTML	PDF	XML	Total
May 2026	127	45	10	182
Jun 2026	40	22	1	63
Jul 2026	53	52	2	107
Aug 2026	39	34	3	76
Sep 2026	99	99	14	212
Oct 2026	47	25	6	78

Cumulative views and downloads (calculated since 20 May 2026)

Month	HTML	PDF	XML	Total
May 2026	127	45	10	182
Jun 2026	40	22	1	63
Jul 2026	53	52	2	107
Aug 2026	39	34	3	76
Sep 2026	99	99	14	212
Oct 2026	47	25	6	78

Viewed (geographical distribution)

Total article views: 657 (including HTML, PDF, and XML) Thereof 657 with geography defined and 0 with unknown origin.

Country	#	Views	%

Cited

Latest update: 10 Oct 2026

Short summary

Serious games gained popularity in communicating about natural hazards and educating people on relevant behaviour. This work gives an overview on 132 publications on mostly digital serious games which often focussed on flooding next to other natural hazards. Some evaluation studies show positive effects on users' knowledge gain and other impacts. While promising, more rigorous research is needed to ensure these games' effectiveness and potentially build resilient communities.


Total:	0
HTML:	0
PDF:	0
XML:	0

Gamified Disaster Education. A Systematic Review of Serious Games on Dealing with Natural Hazards and of their Impacts and Outcomes for Users

Viewed

Viewed (geographical distribution)

Cited

1 citations as recorded by crossref.