the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 12 Feb 2024
The cascading effect of wildfires on flood risk: a study case in Ebro River basin Spain
Abstract. Climate change increases the risk of wildfires and floods in the Mediterranean region. Yet, wildfire hazards are often overlooked in flood risk assessments and treated in isolation even though they can amplify floods. Indeed, by altering the hydrological response of burnt areas, wildfires can lead to increased runoff and cascading impacts. This study aims to comprehensively assess flood risk using a multi-criteria GIS-based approach, considering both current conditions and future scenarios for the Ebro River basin in Spain in the year 2100. More specifically, this study investigates future flood risk under the Shared Socioeconomic Pathways (SPP) 1-2.6 and SSP5-8.5 scenarios, taking into account projected socio-economic conditions and the cascading impact of wildfires. An Analytical Hierarchy Process (AHP) approach is employed to assign weights to various indicators and components of flood risk, based on insights gathered from interviews with seven experts specializing in natural hazards. Results show that the influence of wildfires on baseline flood risk is not apparent. Under the SSP1-2.6 scenario, regions with high flood risk are expected to experience a slight risk reduction, regardless of the presence of wildfires, due to an expected substantial development in adaptive capacity. The highest flood risk, almost double compared to the baseline, is projected to occur in the SSP5-8.5 scenario, especially when considering the cascading impacts of wildfires. Therefore, this research highlights the importance of adopting a multi-hazard risk management approach, as reliance solely on single-risk analyses may lead to an underestimation of the compound and cascading impacts of multi-hazards.
- Preprint
(5338 KB) - Metadata XML
-
Supplement
(10158 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2024-153', Anonymous Referee #1, 27 Feb 2024
General Comments
This study presents an approach for considering related risks for wildfires and floods which is applied to produce baseline and future predictions for a single case study (Ebro river basin). This risk assessment method utilises the Fire Weather Index (FWI) and a number of indicators which were weighted according to expert feedback via an Analytical Hierarchy Process approach. The importance of considering the cascading, interlinked risks of fires and floods are clearly outlined and the chosen case study provides a useful demonstration and context.
However, further information about the expert panel decision-making (in addition to the information provided in the supplementary info) would be helpful given the important role it plays in the final risk assessment method and the chosen weightings. In particular, I would welcome more information regarding the variability in opinions offered by the experts (ideally quantified to show the variability around the final, chosen weightings at each stage). Additionally, it may be useful to provide further context in the results/discussion by explicitly comparing the expert panel opinions to the existing literature where possible.
Additionally, a large part of the fire activity is characterised by FWI predictions and it is important that the meaning of these predictions, the historical context of this index and it’s role as a fire danger prediction tool, along with limitations if looking to extrapolate expected fire occurrence from FWI are clearly outlined. This is addressed in some of the references cited in the manuscript as highlighted in the specific comments below.
Specific Comments
Line 21: Is there a more appropriate reference here than Wilby and Keenan which so far as I can tell does not address the link between fire and drought?
Lines 30-31: Double check this statistic in the provided reference. Is this based on the info given at the start of the introduction in this reference? If so this is actually only over the last 20 years which may be worth highlighting. Also the % of the population affected seems a bit higher for flooding with droughts having affected 25% of the population.
Lines 44-46: Perhaps rephrase this section since one of the conclusions of Versini et al 2013 study is that 'our assumptions can appear as a low hypothesis that should underestimate the impact of forest fire on the hydrological response'. This seems to contradict the claim here that this study in an exception in not underestimating the amplification effects. Perhaps this is more about being understudied or receiving little consideration in which case this could be clarified in-text.
Lines 73-74: Could clarify that Balasch et al. state that this was the mean figure for the period of 1920-2000.
Lines 82-83: Can you clarify/ rephrase for clarity here? I think from what's written in Terrado et al, that 38 people/km^2 is the average population density of the basin, rather than the average density for these 2 largest cities.
Lines 91-93: There may be specific motivations for dealing with wildfire management at various spatial scales. Is there other evidence that can also be provided here to support the statement that 'flood management appears to have a higher priority than fire management' e.g. a comparison of spending/funding?
Line 100: Are these previous literature reviews published and available to cite here?
Lines 104-105: Can you clarify that Fire Weather Index provides a prediction of fire danger? And perhaps accompanied by clarification of the distinction from 'probability' which will also be affected by other factors e.g. the limitations highlighted in Abatzoglou et al, 2019.
In this context, is 'probability of future fire events' a suitable term? As the probability (including likelihood of ignition) will vary with other factors e.g. location relative to population centers, public access, social dimensions which are not considered in a meteorological index which predicts the fire weather and danger were a fire to occur.
See also the discussion in Di Giuseppe et al 2018 which you have cited. e.g. pg 5360 'The FWI is already widely employed in fire management and control (Lee et al., 2002). However, it does not explicitly model fire evolution, but it is a measure of fire danger (Van Wagner, 1987). Even for extreme FWI values there is a need for a stochastic component, i.e. ignition, to start a fire. For this reason, situations in which FWI is high but no fire is recorded are not uncommon'
Line 116: As discussed later, were only highways considered? If so can you clarify this here when introducing the distance from roads parameter.
Lines 117-120: Is there any way to further assess the validity of this weighting approach? e.g. by further exploring the heterogeneity of economic activity in some of these regions e.g. by using population as a proxy for this or incorporating a distance from major town/city element?
Lines 122-123: Was there a particular reason for the choice of this cut-off length or is this choice arbitrary?
Lines 123-124: In relation to an earlier comment, were these the only roads considered? If so can this be clarified earlier when distance from roads is first mentioned (line 116). Perhaps this parameter could even just be labelled 'distance from highways' throughout or 'distance to major roads' as in Roy et al 2021.
Lines 141-143: Are the size/number of personnel at fire stations considered at all? Is number of fire stations a better indicator (and/or easier to analyse) than for example total spending on fire resources?
Indeed, McLennan and Birch outline some of the complexity involved in the prevention and management stages alone in their discussion of various factors including station staff size, average age of firefighting staff, degree of co-operation between staff, additional private firefighting resources.
Could you provide some further discussion of the suitability and limitations of number of fire stations as an indicator?
Lines 165-175: Can you clarify if all assumptions are listed here? If not could you provide a full list of unavailable exposure/vulnerability indicators and corresponding assumptions made e.g. in the Supplementary data.
Can you explain how these assumptions were chosen? Does this involve the previously mentioned expert interviews or is this an arbitrary choice?
Lines 191-192: Would danger be a better word here than effects? Or perhaps just refer only to fire weather index?
Line 205: Understand the potential need for anonymity around experts involved but is it possible to provide any further details about specific areas/extent of expertise?
Lines 227-231: How do these expert comments relate to any existing approaches in the literature?
Given these complex considerations, is the scenario well-defined enough for experts to pass judgment on relative importance? In the future, would more local scale analysis involving expert analysis be required? Or for example, could a further set of scenarios be designed for expert feedback in which additional factors could be incorporated e.g. different landscape types, distance from river. Understanding of course that these are considered in other parts of the model.
Line 232: For comparison, could you also show burnt area and runoff maps in this figure? So that the influence of the chosen weighting can be understood.
Lines 237-239: So is it an increase in burnt area as a result of increased FWI which results in the increased FHI?
Lines 242-244: Could more information about the variability in weighting assigned by the expert panel be provided?
Lines 266-267: Was this the area in which the greatest difference in expert opinion was observed?
Lines 272-273: How does this compare to findings in the existing literature as were outlined in the introduction to this study?
Lines 273-274: Would greater consensus have been reached by asking experts only to consider a smaller number of more relevant indicators?
Lines 283-286: Are there existing studies regarding the influence of hydraulic conductivity which can provide further context for the decisions of the expert panel?
Lines 296-297: Can you explain how decision fits within the context of understanding and mapping existing risk (prior to any other management interventions) and for predictions based upon assigned levels of societal intervention.
Lines 301-303: What was the break-down of expertise in the panel? How significant was this difference in perspective?
Lines 306-307: What was the break-down of expertise in the panel? How significant was this difference in perspective?
Lines 324-325: Can you clarify how notable this finding is? Or whether in fact this is just entirely due to the weightings assigned by the expert panel?
Lines 342-343: Given the role of FHI, how much is the effect of wildfires controlled by the choice of weighting for burnt area:runoff?
Lines 356-357: How much does it indicate this vs. indicating the perceived role cascading effect of wildfires given the role of the expert panel in determining the various weightings?
Lines 360-362: How possible would it be in future studies to further incorporate these previous findings to augment or replace the need for expert weightings?
Lines 364-366: As discussed in Bedia et al 2013, does fuel moisture (and/or linked meteorological conditions) also play a role in limiting these large fire events?
Lines 378-379: As per previous comments, can you explicitly address the limitations of using FWI as a proxy for fire probability?
Line 432: How did the annual timeframe affect the chosen FWI? Was this an average value for a whole year? Or a maximum value?
Section 4.4: Can you also discuss any limitations involved with the expert panel and the Analytical Hierarchy Process and associated data?
Lines 445-446: Can you also discuss any limitations involved with the expert panel and the Analytical Hierarchy Process and associated data?
Technical Corrections
Figure 2: Proofing comment - check figure quality/resolution as slightly blurred in places.
Line 149: Typesetting issue ‘1971 -2000’
Line 179: ‘USDA’ - Acronym needs to be defined on 1st use.
Line 312: Typo: ‘into intervals of 16,7% per class’
Line 365: Typo: ‘can be related to that the Mediterranean’
Citation: https://doi.org/10.5194/egusphere-2024-153-RC1 - AC1: 'Reply on RC1', Samuel Jonson Sutanto, 03 Apr 2024
-
RC2: 'Comment on egusphere-2024-153', Anonymous Referee #2, 13 Mar 2024
General comments:
This study shows an assessment of flood risk using a multi-criteria GIS-based approach incorporating wildfires and floods for the Ebro basin. The study tackles an important topic; it is a generally well written and well-argued paper, with a strong narrative and clear structure.
Notwithstanding this, I have some questions about selected aspects of the paper and the approach taken. The following general comments may be helpful in strengthening some aspects of the paper.
At points in the paper, some terms are used somewhat interchangeably and not always defined. For example, in the title the authors use “effects”, then later “impacts”, then a mix. Similarly, what is meant by a cascade in this paper? Is it a cascade in terms of a trigger or something that increases risk (i.e. a mechanism or process that links these two hazards even if temporally), or a cascade in terms of impacts, or both? I feel as those terms are being used somewhat interchangeably in this paper. For example, in L12 in the abstract, the authors say “…especially when considering the cascading impacts of wildfires”. I would question here what the cascade is? Or indeed, whether this is an impact? To me, this perhaps is more of knock on effect or something that increased the risk of something else through changing vulnerabilities (as the authors note later on) – a wildfire affects the flood risk through burnt area and so forth, which in turn may cause impacts for example – but is this process an impact? I would suggest that clearly defining these terms and then staying with them throughout would benefit the understanding for the reader.
The abstract mentions indicator only once, but a large part of the study is actually focused towards the integration of socio-economic indicators and land-use change information with ‘conventional’ hydrological properties and the cascading effects of wildfire to assess flood risk. This is a complicated endeavour within a multi-hazard/multi-risk approach, which is good to see, but I think the fact that this is approach should be made much clearer in the abstract and title so it is clear that the story is not solely about the cascading effects of wildfires and flood, it is more about better flood risk assessment as a whole, incorporating wildfires. Later, in section 4.3 in the discussion, the authors state “Therefore, this research provides an example of how to integrate multiple hazards into risk evaluation by conducting comprehensive assessments that consider numerous drivers and indicators that will contribute to increased flood risk in the future.” – this is a, I believe, a better (more correct?) framing for the study. I would suggest multi-hazard/risk be included in some way in the title, perhaps even removing the word cascade, and a reduction on the focus of the wildfire and flood cascades and more towards flood risk using multiple inputs from the start.
The methods are comprehensive but section 2.2 is framed around flood risk indicators, however here this is where wildfire risk and indicators (FWI for example) is employed. Related to my point above, this section to me should be given a clear multi-hazard focus towards flood risk to make it clear that fires are part of this. Some subtle reframing and – importantly – including wildfire in the subtitle may be beneficial to guide the reader.
The use and placement of the equations based on expert judgement FR, FEI etc is very confusing. FR is at the end of the methods section but if not defined until section 3.4. Then, additional equations, such as FVI, appear later on. Lines 213-4 says “This process allows us to calculate the Flood Hazard Index (FHI), the Flood Exposure Index (FEI), and the Flood Vulnerability Index (FVI), as denoted by Equation 1”, however equation 1 shows the equation for FR. Equations 2 and 3 are not referenced from the text. Some terms, such as FS, are really hard to find the definitions of (one has to go looking in the text), and no units are provided. The use and presentation of these needs a rethink. I would suggest that the equations are all placed in the methods and defined there, leaving the results to focus on the weighting by expert judgment and, therefore, the outcomes of the study. Indeed, many of the sub-sections 3.3 and 3.4 for example, stray into methods rather than results. Some careful reordering would really help the readability and accessibility.
Related to my above point, there is a large emphasis on expert judgement of the indicator weightings. It is not clear though quite how much emphasis they have on the results. Is seven people enough? Are they all from the Erbo region? Does this matter? The earlier phases of the study are quite analytical, but then the focus moves to a judgement based approach. Additional details on this process, perhaps in section 2.5, would in turn help understand and interpret the later results section.
The discussion is good and very readable. It provides some excellent additional information. I do think though that perhaps a bit more work may be needed to separate the location-specific findings based on local expert judgement and the wider findings that can be employed elsewhere. The authors don’t really attempt to do this; instead the assumption is that the results shown here for Erbo would hold elsewhere. This localised expert judgement is not mentioned in the limitations in 4.4. The title at the start calls this paper a “study case” but really Erbo is the study here primarily. Making it clear that the findings for the Erbo maybe separate from broader interpretations would be beneficial in the discussion, including some detail on how this can be done (and the limitations of doing so), would really elevate this to be a usable example more broadly.
The conclusions in section 5 state “The research underscores the need for interdisciplinary collaboration…” in relation to the experts. While I agree with this statement, this is not a focus of the study and isn’t presented up to this point. If the authors which to explore this narrative, this would be better placed in the discussion above.
Specific/minor comments:
Line 191-2 Please clarify or reword what is meant by future wildfire effects (FWI) and the burnt area? FWI is already defined, but what is meant by effects?
Citation: https://doi.org/10.5194/egusphere-2024-153-RC2 - AC2: 'Reply on RC2', Samuel Jonson Sutanto, 03 Apr 2024
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 252 | 86 | 14 | 352 | 35 | 6 | 9 |
- HTML: 252
- PDF: 86
- XML: 14
- Total: 352
- Supplement: 35
- BibTeX: 6
- EndNote: 9
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
