Physical Vulnerability of Buildings to Flooding in Lilongwe City, Malawi

Review
This study aims to investigate the physical vulnerability of buildings to flooding in the settlements of Malawi’s capital city, Lilongwe. It is found that exposure factors variably influenced the physical vulnerability of individual building types and that building typology and foundation height were also important factors. In particular, buildings constructed using fired bricks with cement mortar walls and cement floors had low vulnerability while buildings constructed using fired bricks in mud mortar walls and cement floors had high vulnerability.
The study collected first-hand data of 130 buildings and 52 interviews, which shows the potential to make original contribution to this topic. However, the whole manuscript just reports some simple statistic results of the data, and there is lack of in-depth analysis. My comments are as follows.
1. Abstract: Though the authors mentioned OLS regression and p values, there is no such information in the major contents.
2. Abstract: With regards to the sentence started with “Rather than..”, there are few other issues for the authors consideration before reaching such conclusion. First, stream reserves and better construction of buildings should refer to different scales of flood risk prevention measures. The former works for relatively larger scale, and the latter only works for building itself. Moreover, stream reserved may be more related to the reduction of inundation depth, which is a key parameter in the examination of physical vulnerability. However, the impact of inundation depth was not analysed in the study.
3. Data collection: Why chose these parameters, and what are the relationships between these parameters and the physical vulnerability? The authors should describe the rationality.
4. Data collection: It would be better if the authors can list a table to describe all the parameters and their units and data ranges etc.
5. Table 1&Table 2: It would be better if the authors could explain the link between these two tables. Are slight damage and moderate damage both derived from Half Collapse?
6. Table 2: How to categorise the degree of damages?
7. 3.3 Data Analysis: I cannot find the results related to this part.
8. 4.0 results: (1) Please correct the number of the Tables. (2) These are just simple statistical description of the data. It is hard to distinguish or isolate the key parameters in affecting the damage or physical vulnerability.

Note:
This review refers to round 1 of peer review.

Review
The paper addresses the critical issue of physical vulnerability to floods in informal settlements but lacks clarity and alignment in key areas. These include a need for a more detailed methodology section explaining research methods, a consistent approach to data analysis and interpretation. The abstract does not accurately reflect the study's scope and findings, and there is insufficient integration of existing literature on building classifications and vulnerability assessments. Parameters such as foundation depth and identified building types, as well as protective measures, are collected but their impact on vulnerability levels remains unclear. A more robust critical analysis of how vulnerabilities correlate with on-site observations is essential, along with a discussion on replicability. Consistency in formatting is necessary to enhance the paper's impact and potential for publication.

Detailed Review:
The author should ensure that the font is consistent throughout the text, that tables and sections are numbered correctly and sequentially, and that captions explain the content of the figures in detail rather than being generic
Abstract:
The abstract does not accurately reflect the content of the paper. It should mention that various studies on the physical vulnerability of informal settlements have been conducted for other types of hazards (e.g., earthquakes) but not for floods, therefore it is important to investigate this.
The use of HIS, OLS, and SPSS is mentioned in the abstract but is not clearly explained the methods and their application. More comments have been added for section 3.3.
It is mentioned that foundation type data and its depth are used to assess vulnerability, but there is no analysis in the paper considering foundation height and its impact on vulnerability.
The calculation and significance of the p-value are noted in the abstract but not discussed in the text. Specifically, p < 0.005 is not discussed in the paper; section 3.3 only refers to p < 0.001.
The conclusion lacks clarity regarding how the parameters collected on site impact vulnerability levels. It appears that vulnerability has been analyzed primarily based on damage levels, without clear evidence that different building typologies influence vulnerability. While this may not be the authors' intent, the data suggests otherwise.

Literature Review:
The author should mention previous studies on seismic risk and vulnerability assessment of informal settlements (Kloukinas, P., Novelli, V., Kafodya, I., Ngoma, I., Macdonald, J., & Goda, K. (2020). A building classification scheme of housing stock in Malawi for earthquake risk assessment. Journal of Housing and the Built Environment, 35, 507-537; Novelli, Viviana Iris, et al. "Fragility curves for non-engineered masonry buildings in developing countries derived from real data based on structural surveys and laboratory tests." Soft Computing 25 (2021): 6113-6138; Novelli, Viviana, et al. "Seismic mitigation framework for non-engineered masonry buildings in developing countries: application to Malawi in the East African rift." Resilient Structures and Infrastructure (2019): 195-223).
Why does the authors’ work not refer to existing building classifications? How does the authors' classification differ from those developed in previous research? The classification in this work is also based on wall type and floor type, similar to other building classifications. The need for a new building classification is implied in the paper, yet the reasons behind it are not highlighted.
On page 3, the authors should refer to the more recent census data rather than NSO.
The authors should clarify the terms "weak," "strong," and "very strong" and their relevance.
The authors should explain how building age data was used to assess vulnerability and the source of this information.
The authors should add a reference to the statement: “The foregoing studies suggest that different countries use different construction materials, making buildings’ vulnerability analysis difficult to compare.”

Section 3.0 Methodology:
The authors should clearly define the qualitative and quantitative aspects of the study.
The authors should ensure that it is clear where Figure 2 is sourced from and indicate that Figure 2 is close up from Figure 1.
The authors should differentiate between the building survey and household survey, Including forms or specific questionnaires used for the data collection if applicable.
The authors should justify the comparison of data from 2016 and 2017. It is not clear why this was needed and how the data gathered from this comparison was used for the vulnerability study.
The authors should clarify how the sample sizes of 80 buildings and 56 households have been calculated.
The authors should explain the proximity to the river and its significance and discuss how this has been used to assess vulnerability. Including this data as a parameter to assess vulnerability will be much more significant than basing the vulnerability assessment solely on the observed damage.
The authors should Discuss the importance of foundation height, building age, and construction details, and how these can impact vulnerability.
The authors should Specify the building protection measures. Define what you mean by structure type and provide pictures so that it is clear how these building have been classified and how these are compared with the typologies defined by previous studies such as Novelli et al. (2020) or those defined by the census (permanent, semi-permanent, traditional).
Use the term "roof" instead of "floor" since houses are single-story.
Provide clear definitions and ensure consistency between Table 1 and Table 2. Table 1 depicts three levels of damage, whereas Table 2 presents four levels of damage. It is unclear how the authors classified the damage observed on site using either three or four levels, indicating inconsistency between the two tables.
How can you be sure that the light damage is caused by floods and not pre-existing damage? Most of the time, the damage observed on these houses is not necessarily caused by floods, so how have different types of damage (flood-related or not) been distinguished? Or has this distinction been made?
Clarify the classification of damage and provide pictures taken on site to illustrate the levels of damage observed.
Provide criteria for determining vulnerability ratios and ensure consistency in terminology and explanations. Some time you use moderate and other you use medium vulnerability
Explain the methodology for classifying houses into types and provide visual examples (pictures taken on site) to help the reader understand the typologies.
Ensure that figures like Figure 3 and Table 7, which are not mentioned in the text, are clearly explained.
This distinction is crucial as it affects how buildings are classified. In the context of "cement and earth floor," it typically refers to concrete floors and earth-based materials for roofing, such as thatch. It's important to note that in Malawi, roofs are predominantly constructed from lightweight and flexible materials rather than concrete.

Sec 3.3 Data Analysis:
Clarify the purpose of population information and its correlation with vulnerability. Explain the calculation of p-values and the use of OLS and ArcGIS on 52 households. Justify the significance of these analyses in the context of the study.
It is not clear why you need information on the population and how this is correlated with vulnerability. How did you calculate the p-values, and what are you trying to demonstrate? How did you use OLS and ArcGIS on 52 households? This is not clear, and I do not understand why it is important in your study.

4.1
It is stated that buildings are built according to recommendations for low-income earners; however, the houses in Figure 3, particularly the third and fourth pictures, do not appear to be built based on recommendations. These buildings have poor materials, lack connections, and have many deficiencies.
When the authors state that houses are designed based on how informal houses should be built, are they implying that there are existing guidelines and recommendations for flood resilience? To my knowledge, there are no standards available to create flood-resistant informal houses; there is likely only limited information in general safety guidelines.
On page 6, the authors mention having only one type of roof (metallic sheet), but later discuss roofs made of cement and earth. This inconsistency needs revision for clarity. Additionally, as said earlier it's unclear what is meant by floors made of earth and cement. Visual examples of these roof types would help define the different typologies observed on site.
Substitute the word “building” with “structure.”
Referring to my earlier comment, where I questioned the assertion that buildings are constructed according to guidelines for low-income housing, the building types described in the paper do not conform to these standards. Houses with sun-dried bricks and cement roofs (Type 4) do not typically align with expectations for informal settlements, as Type 4 structures are traditionally built with inappropriate roofing materials.

Section 4.2 and 4.3
Table 6: How was it determined that the houses had been flooded twice, and what significance does this hold for the study? What is the author attempting to demonstrate regarding vulnerability through this information?
Why are you considering whether the house is within 30 meters from the river? This suggests that some regulation sets this distance. Please specify why these 30 meters is important.
In Table 6, the numbers per row are incorrect: 94% + 7% = 101% (impossible, this should be 100); 5% + 91% + 5% = 101% (impossible, this should be 100). Each row should total 100%.
Table 6 is meaningless if the exposure variables are not combined with the classification of building type and vulnerability level associated to the buildings.
Then it is stated that 54% of the buildings were affected by floodwater with a height of 60 cm. According to the paper, this information is attributed to Table 6, but I am unable to locate it within the table. How was this specific data obtained?
How do the authors define a low foundation?
Figure 4a: Delete the "a".
Table 7 is not clear and not explained in the text.
How are the vulnerability levels and their ranges defined? Clarifying these definitions with visual examples would enhance understanding of how vulnerability is assessed. The current information and definitions in the paper make it challenging to apply the proposed method effectively.
The vulnerability level is sometimes called "medium" and other times "moderate." Please be consistent.
Type 4 is considered equivalent to Type 1 in terms of vulnerability. However, Types 2 and 3 show vulnerability levels that are not significantly different from Types 1 and 4. The author needs to clarify how different building types influence vulnerability. As defined in the paper, building typologies do not appear to strongly affect vulnerability, suggesting a gap in understanding. The data does not convincingly demonstrate that better-built houses respond better than weaker ones.

Section 4.4
What would authors recommend based on your data and results? What’s the protective measure that authors would suggest
The paper discusses structural and non-structural protection measures. However, it remains unclear how these measures affect vulnerability. Despite their presence, the measures do not appear to significantly reduce observed damage. This raises the critical question: what factors truly influence the vulnerability of these houses?
Table 9: The numbers per row do not make sense; they should total 100%.
How do these protective measures relate to building types? Is there a correlation suggesting that buildings with such measures are less vulnerable? The current data does not demonstrate this, raising concerns about the completeness of the correlation between building types and vulnerability.
Schools cannot be included in this study, as they are generally built better than houses.

Conclusion
The conclusion fails to acknowledge that Type 4 seems to perform similarly to Type 1, which is perplexing.
Furthermore, it overlooks the consideration of foundation depths and the potential impact of preventive measures on vulnerability.
There is ambiguity regarding whether certain building typologies outperform others, particularly as the first four types appear remarkably similar, which seems implausible.
Additionally, the proper definition of vulnerability levels remains uncertain. Lastly, suggesting a foundation raised 1 meter above ground as a viable preventive measure to reduce vulnerability is impractical.

Note:
This review refers to round 1 of peer review.