Community-Level, Participatory Co-Design for Landslide Warning with Implications for Climate Services
Round 1
Reviewer 1 Report
The paper presents the development of a landslide warning system for a relatively small community in Southern Alaska. The system's development was triggered by a critical landslide episode in 2015, which caused three fatalities.
The key feature of the study was its co-design with the local stakeholders and the attempt to involve the local community as producers and users of the final product to support individual relocation decisions in case of potential problems.
The manuscript is rather long and wordy but lacks precise details on the development of the system that can be useful for the readers to understand the transferability of the experience. For instance, the total cost of the study is never mentioned not it is the human resources actually used. Furthermore, the number of community members involved is not precisely detailed, and the exact schedule of the meetings is only reported in the appendix.
Also, the risk determination process is not reported in the paper because it will be published elsewhere. Finally, the authors state that a quantitative evaluation of the system's performance was impossible because no severe episode happened after its deployment. Moreover, it may never be possible because other similar initiatives took place at the same time and in the same community as the one at hand.
From a technical viewpoint, the main drawback of the system, according to the authors themselves, is the lack of spatial resolution of the proposed information system. This appears not coherent with the current scientific knowledge since the DEMs are globally available with high resolution, and geological and land use maps are usually precise enough to allow a spatially detailed risk analysis. If this was not the case for some reason, it should be explained in the text. In quite the same way, it would be helpful to know why deploying the sensors' network, particularly rain gauges, was so limited. The cost of perfectly working web-connected meteo stations is well below thousand dollars. Why do not equip more sites with rainfall measurements to better determine the precipitation distribution and directly involve the residents? The sentence in line 365 "For instance, data from the tipping bucket rain gauges emphasize the spatial variability of rainfall intensity across Sitka's relatively small spatial area, thus supporting the decision to issue landslide risk warnings for the entire area and not specific hillsides" seems contradictory. The spatial variability of rainfall should suggest specific warnings and not generic ones. Indeed, line 666 notes that "For community members, however, successful prediction required much more spatially specificity, differentiating landslides that endangered people and property from those that did not".
The paper dedicates many pages to the description of the approach but much less to the result. For example, is the system providing the information in fig. 4 only under request? Are the maps also readable on mobile phones? Is the information on the current rainfall (the only actual data, not filtered by the system) in comparison with past values available to the users? How many times was the system used, at least in the last months? How many unique users? Again, the paper details how the workshop participants were guided to evaluate the risk of missed alarms compared to false alarms but does not provide any evaluation of the system's performance in past cases. Consequently, it does not attempt to evaluate and communicate the uncertainty of the forecast.
My understanding is that a critical element that made the study possible and successful is the relative homogeneity of the community that allowed the determination and responsibilization of representatives recognized and accepted for their authority. Transferring the experience to other situations where such homogeneity is missing (e.g., to larger towns) would be prohibitive because, as noted by the authors, individual relocating decisions may lead to chaos.
Overall, I think the paper must be revised by cutting down the theoretical part on co-design and risk while concentrating on the experience with more detailed information. Probably, the part on land insurance can also be skipped.
I also suggest the following.
- Change the title because "climate services" seems too broad. The experience is only related to a landslide warning system.
- Delete the sentence in line 40 "Addressing climate change will often require significant changes in lifestyles and daily routines informed and enabled by new and evolving science and technology." While this is true, there is nothing about climate change or lifestyles in the paper.
- Define SSSC where it first appears (line 313)
- Provide a URL that the readers can access. Apparently, https://sitkalandslide.org/ "is configured to block access from your country".
- Delete or clarify how "a family whose home is [in] an area exposed to landslides might decide to sleep that night at a friend's house which has less exposure" given the explained lack of geographical resolution.
- Explain the meaning of "The dashboard suggests that even medium and high-risk days are not likely to occur more than several times a year." (line 375)
- Delete the question marks in table 1. Either it happens or not.
- Explain the meaning of "atmospheric rivers" (line 575).
- Correct "Sita" into "Sitka" in line 707.
- Change LWS or define it at lines 997 and 1005
- Delete or modify the sentence "The geoscientists would regard a warning to be accurate if a landslide occurred within many miles of Sitka." (line 1012) I do not think this represents the current level of geoscience and geoengineering.
- Change (or delete) the table titles in appendix A and improve their resolution.
- Change line 1060 "FAA/NWS Airport Station – we aren't responsible for this" since "we" is unclear.
- Rephrase line 1063, "Organize upgrades to a data feeds and algorithms". It is unclear and may represent an intolerable amount of work since studies on landslides will probably continue for decades worldwide.
- Rephrase line 1075, "Planning Commission – planted a seed here, applying landslide hazard to city planning, zoning etc." since this is a to-do list for the future.
Specify the meaning of "indirect costs" at 1104.
Author Response
Please see the attachment
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Dear co-author,
I welcome this manuscript. There are two sides of the lesson that can be deduced from your ideas. First is the novelty of the article. On a positive note, methods involving stakeholders to evaluate and formulate policies are seen as a competent package. Many studies, especially those that apply empirical models, do not necessarily invite stakeholders to participate in decision-making. Uniquely, here the authors include a practical lens. Second, in my opinion, the observation of this study is relative to the community to manage services in reducing climate, but there is no final instrument that focuses on the object of research. Is this implemented successfully or vice versa?. Furthermore, there are many writing systematics that are contrary to the rules and instructions in MDPI. I think that this paper tends to be thick on working papers from a practitioner's perspective. Ideally, in a scientific article, also understand the prevailing academic context. For both reasons, I recommend that there be a complex fix referring to "review status". It is worth encouraging the authors to continue working on the paper. Despite all the remarks, this is a valuable research initiative.
Kindly regards,
Anonymous reviewer
Comments for author File: Comments.pdf
Author Response
Please see attachment
Author Response File: Author Response.pdf
Reviewer 3 Report
Thank you for an interesting description of a case study of the development of a climate service. However, the article is not set up as a scientific article, and this makes it unclear what the method and findings really are. Is this research, or is it just a description of a climate service development?
Please check that the abstract has information about the background, research question/aim, methods, findings and conclusion.
The introduction and framework section is good, make sure that you use the theories from the framework section when you discuss the results.
Please state the research question and its limitations and framing. Please make clearer sections on the methods and findings.
The paper lacks a clear method-section. When you evaluated this climate service, how did you do it? Interviews, group interviews, surveys, document analysis? How did you perform these methods? How did you analyze it?
The paper lacks a clear results-section. What were the findings from the evaluation? Summarize the results, and present them topically - do you have examples or quotes to document the findings?
Discuss the findings in relation to the theoretical framework you have provided. Repeat the research question and answer it.
Author Response
Please see attachment
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
I disagree with the authors regarding various aspects, particularly the final result they propose. The dashboard developed by the project is very simplistic. It does not consider the spatial distribution of the rainfall (which the authors measured in another paper) or the user's specific location, which has been extensively studied (see the TerrainWorks landslide map and report). So, it appears that the (relevant and expensive) co-design effort was far more important in improving community awareness and preparedness than in developing the final product. I suggest the authors add this aspect to the conclusions.
Author Response
We thank the reviewer for their comments. As noted by the reviewer, we may disagree on some questions related to the availability and role of warning at high spatial resolution. We have, however, added some additional text in support of our views. First, in response to the reviewer’ question about the availability of spatial information on the dashboard, we have added text to Section IV to make clear that the TerrainWorks modeling results are available on the dashboard in one of the pull-down menus. Second, in response to the reviewer’s question about the importance of the co-design effort on warning system (rather than just on community awareness), we have added a more detailed explanation in Section VI on how the statistical landslide prediction algorithms and the thresholds chosen were influenced by the co-design process. We'd like to emphasize that while we may continue to disagree, we are very grateful to the reviewer whose comments have significantly improved our manuscript.
Reviewer 2 Report
Dear authors,
You and your colleagues have worked optimally to improve this manuscript. All corrections commented on by reviewers have been revised. I welcome your enthusiasm. Besides that, I also received the results of this repair well.
Best wishes,
Anonymous reviewer
Author Response
We thank the reviewer for their kind words.
Reviewer 3 Report
This has been approved. Thank you.
Author Response
Thank you!
