| 16 Oct 2025
Flood vulnerability models for damage assessment of artworks and cultural heritage buildings
Abstract. The assessment of flood-related losses to cultural heritage (CH) remains one of the most underexplored areas in flood risk management, largely due to the complexity of CH assets and their multiple, often intangible, values. In this study, extensive field data collection and archival research on artwork restoration costs were undertaken to support a synthetic approach for developing vulnerability models for both the building envelope and internal artworks across three primary CH asset types: places of worship, museums, and libraries. The methodology was applied to the historic city of Florence (Italy), enabling the derivation of mean and percentile vulnerability curves from a sample of 48 inspected CH buildings. For a 500-year flood scenario, estimated average losses amount to approximately €2.5 million for the building envelope and €3 million for artworks per asset, with total damages to CH in the city reaching approximately €550 million. While direct monetary loss estimates are subject to considerable uncertainty, the model results align well with available ex-post data, particularly for places of worship. These findings demonstrate that flood-related monetary losses to CH assets are far from negligible when compared to other damage categories, such as residential buildings, and therefore warrant increasing attention from the scientific and policy-making communities.
The study provides an interesting study that addresses the quantification of flood-related damage to cultural heritage assets in Florence. In this approach, the authors develop vulnerability functions for cultural heritage buildings and artworks, combining empirical data with synthetic modelling. While the topic is relevant and timely to current debates on heritage resilience and disaster risk reduction, the following comments highlight a few points that should be addressed in a revised version of the manuscript to increase methodological clarity and robustness:
- In Section 2.1, the authors refer that “the "content features" section of the survey focused primarily on the most common artwork types: panel or canvas paintings, frescoes, stone or wooden sculptures, and paper-based items”, and that everything else is assigned to other "Other artworks". Assuming the authors are proposing a general methodology that could be applied in other scenarios outside the one they are using to illustrate their proposal, this seems to be an arbitrary decision that should be clarified or better discussed, at least. If the authors consider that in their case study, damage to metal- or textile-based artworks are negligible, this should be said, but it should also be highlighted that the methodology can be applied to any type of material as long as data to characterize the cost-related factors are available. Addressing this issue needs to also consider what is written in Section 2.4 on this matter. Moreover, if the authors decided not to include these other artworks because there are no data available to quantify these cost-related factors, this must be made clear, and the topic must be revisited in the discussion section to highlight the limitations of the presented results.
- In Section 2.2, the authors use the term “building's internal envelope”, which is a rather unusual terminology. Do they mean “interior finishes”? This unusual terminology should be replaced across the manuscript.
- In Section 2.2, when introducing Eq. (1), it should be referred that it is for each artwork k or the kth artwork, for clarity (otherwise k remains undefined).
- In Section 2.2, the authors should clarify the practical use implications of their restoration cost assessment framework. From my understanding of the proposed framework, it is not supposed to be used based on the availability of an estimate of the physical damage to a given artwork. This issue comes from the fact that 1) it is difficult to predict the actual level of physical damage that will occur if an asset is exposed to a certain water level (namely due to some of the uncertain factors listed by the authors such as floodwater contamination, the amount of water the artwork was submerged in, the duration, etc.) and 2) it is also difficult to estimate the level of physical damage of the damaged assets for which the authors have the repair cost values. In light of these issues, the practical application of the proposed framework seems to establish the repair cost of a given asset by assessing the expected water level that can be reached in a flood at the location of the asset, and it then assigns minimum and maximum values of the repair cost (which accounts for the referred uncertain factors) to the asset in case the water level reaches it. If this interpretation is correct, it should be clearly described in the manuscript. If not, the authors should clearly describe how the repair cost is established based on physical damage.
- In Section 2.3, line 175, the parameter for opening heights (ho,i) has an index i that is not defined. Note that, later, the letter i is used for the ith cultural heritage asset.
- In the assumptions of Section 2.3, the authors should include that material uniformity is assumed along the height of a given floodable floor.
- In Section 2.4, more details should be provided (to ensure replicability of the procedure that was used) about the statistical analysis that was carried out to determine the average surface area potentially exposed to floodwater. What is a statistically significant sample in this context? What were the artwork typologies that were selected? Paintings on canvas and painting on panels, or others? Can the data that was used actually be accessed thought the link that was provided?
- In Section 2.4, explain the reasoning behind the seven levels of exposure heights that were considered for each asset. Why 7?
- In Section 2.4, additional clarification is necessary regarding Eq. (7). 𝑅𝐶k,j is said to be the maximum unit restoration cost for artworks and, for the case of large of large artworks, it is also said the cost depends on the water level w. This means that, in Eq. (7), 𝑅𝐶k,j should be a function of w. As it is, Eq. (7) only works if 𝑅𝐶k,j a restoration cost per meter or if w is a water level normalized by a maximum water level value, but this is not said. Similarly, the component associated with small-sized artworks does not seem to depend on the water level. As it is, the expressions just adds a constant value, which does not seem right. In addition, parameter P is also not clear. This parameter is said to be the perimeter of the cultural heritage asset in Eq. (2) but its use in Eq. (7) implies that all large artworks cover the entire perimeter of a given building, but what is said at the beginning of this section is that “the affected area (or linear meters of damaged material) is quantified based on the building’s perimeter”, which is not the same. Detailed clarifications of these issues need to be included in the revised version of the manuscript to help the reader understand the proposed approach.
- In Section 2.4, the role of the CRI (Eq. 9) is not clear. How is it expected to be used with the proposed methodology for developing vulnerability functions? It seems that its role is that of an uncertainty factor that provides extra information about the reliability of the predicted damage or vulnerability, but this is not clearly explained when the CRI is introduced. Moreover, the qualitative scale shown in Fig. 7 d) is not explained and justified.
- In Section 3.2, clarify what criteria were used to select the 48 sites (among the 175) to which the proposed methodology was applied to.
- Regarding the costs reported in Section 4.1, it is assumed that they are corrected for inflation, but this should be said, nonetheless.
- Section 4.4 should discuss/refer a few aspects that underline the proposed application: the applicability of the proposed methodology relies on the availability of enough detailed post-disaster data to establish the repair cost components that are considered in the method. While this is something that should be clearly emphasized to highlight the importance of collecting and sharing such data. In cases where such data are not available, the authors should discuss how to proceed. What proxies or what approach could be used to determine these cost components? This is particularly important in order to clarify the scalability of the proposed methodology.
- In the Conclusions the authors refer that one of their contributions is the “monetary assessment of direct total flood-related losses to CH”. This is a bold statement that implies a level of comprehensive loss/damage analysis that is not achieved. Even if the “other artworks” are indeed negligible, what the authors achieve is an estimate of the quantifiable direct losses represented by the repair costs. These are not the total direct losses. For example, losses due to destroyed assets (i.e. irreparable) are not accounted for, as well as the loss in market value of assets that may have a market value. In light of this, this final statement should be rewritten accordingly.