the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Comment on “Classification and mapping of European fuels using a hierarchical, multipurpose fuel classification system” by Aragoneses et al. (2023)
Abstract. Classifying and mapping vegetation as fuel is essential for various fire research and management applications. Aragoneses et al. (2023) proposed a hierarchical fuel classification system for Europe and allocated standard fuel models to the resulting fuel types, producing a continental fuel map. We examine the methods involved and their outcomes. The reasoning behind their system is misguided, as the proposed set of fuel types does not reflect fuel-complex characteristics and the inherent fire behaviour. In their categorization of shrublands and grasslands, fuel depth is a key variable; however, the bioclimatic modelling approach used to map it is unreliable, as it is based on local empirical relationships. The adopted 1-km mapping resolution is one to two orders of magnitude lower than the needs of spatially-explicit fire behaviour simulation, and implied up-scaling procedures adding uncertainty due to loss of thematic detail. Finally, a simplistic aridity-based rule was applied to assign fuel models to fuel types, limiting the options available. This, in combination with fuel-depth overestimation and untenable fuel model choices, contributed to a substantial fire-hazard overestimation across the large portion of Europe occupied by low-flammability cover types.
- Preprint
(5602 KB) - Metadata XML
- Corresponding article
- BibTeX
- EndNote
Status: final response (author comments only)
-
CC1: 'Comment on essd-2024-587', Elena Aragoneses, 28 Mar 2025
Dear Editor and authors,
We (Aragoneses, E; García, M; Salis, M; Ribeiro, LM and Chuvieco, E) appreciate the opportunity to publicly discuss the comments provided by Fernandes et al. regarding our paper, "Classification and mapping of European fuels using a hierarchical, multipurpose fuel classification system" (Aragoneses et al., 2023). While we welcome scientific discussions, we find it necessary to clarify several points and address inaccuracies presented in the critique, as we believe that it merely points out what they consider as flaws our work without any constructive criticism and/or scientifically sound alternative to the methods proposed. See the attached pdf.
-
RC1: 'Referee comment', Anonymous Referee #1, 23 May 2025
General Comments
This manuscript provides an invaluable deep assessment of the fuel classification system and subsequent fuel maps proposed at European level by Aragoneses et al. (2023), pointing at the potential limitations that may hinder their operational use as a base for future fire research and fire management.
The high-quality presentation, concisely written and well structured, is coupled with science-based content, finely argued and supported by empirical data as well as by an exhaustive list of references.
Specific comments
First, despite the large-scale mapping (continental), the objection to the low spatial resolution (1 km pixel) is pertinent, especially considering the original aim by Aragoneses et al. (2023) for providing and accurate information on fuel characteristics at European level to be used for operational estimations of fire danger, propagation and emissions. In addition to the general requirements to provide reliable fire behaviour predictions with commonly used wildfire simulations tools (that are one or two orders of magnitude lower, as argued by the authors), the highly populated and heterogeneous landscapes in Europe clearly add to the need for a higher spatial resolution in fuel mapping.
Second, the shortcomings mentioned regarding the proposed European Fuel Classification System (EFCS) and the estimation of fuel characteristics are sound. The authors propose alternative methods that could have been used to the current approach by Aragoneses et al. (2023), demonstrating the inaccuracy of the shrub fuel depth estimation (a key variable in the EFCS) with field data from different locations (n=164). This is not surprising, but yet relevant, given the fact that locally calibrated empirical models previously available were used by Aragoneses et al. (2023) to infer input variables for fuel characterisation in all Europe. This is even more critical when these previous studies were based on limited vegetation types (mainly Mediterranean shrubland, which are only dominant in certain parts of South Europe), and with limited model robustness and reliability when extrapolating to a continental scale from very regional results (especially considering that these studies already had originally low model accuracy).
Finally, the issues reported by the authors in the cross-walk proposed by Aragoneses et al. (2023) to convert the fuel types, previously mapped with the new EFCS approach, into fuel models according to Scott & Burgan (2005) classification systems is warranted. They highlight specific inconsistencies in several fuel model assignments in different fuel types associated to dominant vegetation in Europe, providing quantifications of potential fire behaviour derived from simulation examples that demonstrate significant overestimations.
Technical comments
The main cited paper, originating the authors comments (Aragoneses et al. 2023), is missing in the reference list.
Citation: https://doi.org/10.5194/essd-2024-587-RC1 -
CC2: 'Reply on RC1', Emilio Chuvieco, 30 Jun 2025
It seems that the comments of RC1 were posted before reading the Aragoneses et al.'s reply to Fernandes et al., critical paper. Most of the concerns of this reviewer were already answer by the authors, particularly what regards the spatial resolution of our fuel databaset, which can be soundly used for fire danger and risk assessment, as we have demonstrated in the FirEUrisk project, in close collaboration with endusers. Standard spatial resolutions used for fire propagation potential and fire emissions within DGVM are much coarser (0.1 to 0.25 d cells) than what was proposed in Aragoneses et al.'s (2023) paper.
Citation: https://doi.org/10.5194/essd-2024-587-CC2 -
AC1: 'Reply on RC1', Paulo Fernandes, 16 Jul 2025
We thank the reviewer for the positive comments and are glad that he/she agree with the shortcomings that we have detected.
The reference to Aragoneses et al. in our text was accidentally deleted, which will be rectified in the revised manuscript.Citation: https://doi.org/10.5194/essd-2024-587-AC1
-
CC2: 'Reply on RC1', Emilio Chuvieco, 30 Jun 2025
-
RC1: 'Referee comment', Anonymous Referee #1, 23 May 2025
-
RC2: 'Comment on essd-2024-587', Anonymous Referee #2, 14 Jul 2025
The manuscript by Fernandez et al (2025) is written from an author team clearly with expert knowledge on fire behavior, specifically in Mediterranean landscapes (more specifically for Portugal), given the publication and project list of the involved author team.
However, re-reading Fernandez et al. (2025), it still remains unclear why the authors are questioning the validity and performance of the FirEUrisk_Europe_fuel_map, European fuel map at 1 km resolution by Aragoneses et al. (2022, 2023). It is clear that the author team of Fernandez et al. (2025) has high expertise specifically for fire fuels in Portugal. However, I do not agree with the authors and Reviewer 1 that the shortcomings described in Fernandez et al. (2025) are sound, as all limitations that may occur related to a European-wide approach have been transparently discussed in depth by Aragoneses et al. 2023 in their publication. In contrast, Fernandez et al. (2025) use highly locally-regionally restricted viewpoints only, argumentation and ideas on methods that would not be realistic and operational for the generation of an European-scale classification system and product.
Given the European spatial scale of the FirEUrisk product, also the performance assessment by Fernandez et al (2025) solely focusing on Portugal seems inadequate. As Aragnoeses et al are stating in their reply to Fernandez et al (2025) in this ESSD discussion, it would only be adequate for an ESSD manuscript publication in case a full alternate assessment on European scale would have been performed. Similar to Aragnoeses et al. (2023) approach who applied the LUCAS Copernicus data set (Andrimont et al., 2024) for their assessment.
Therefore, I cannot support the publication of this comment paper.
Detailed comments:
Comments by Fernandez et al (2025) to fuel depth estimation in Aragoneses et al. (2023) using the Saglam et al. (2008) empirical relationship established in shrubland in a mediterranean region in Turkey. Deviations can be expected for regional assessments of the vegetation height derivation from biomass using the Saglam et al. (2008) empirical relationship that is the base for the fuel depth estimation in Aragoneses et al. 2023. However, interestingly, the description in Fernandez et al (2025) still shows a good, robust performance with only some scatter and deviation from the 1:1 relationship from the Saglam et al. (2008) predicted shrub height versus 128 observed shrub heights in Central Portugal.
Also related to the suggestion of Fernandez et al (2025) to Aragnoeses et al. (2023) to derive the vegetation height for different fuel depths from spaceborne LiDAR remote sensing, I do not agree. For example, Fernandez et al (2025) suggest to use spaceborne LiDAR for the derivation of vegetation height for the different land cover-related fuel classes. However, spaceborne LiDAR-derived vegetation height is only reliable for medium to closed-canopy forest height layers, not for grassland, shrubland, agricultural land and open woodland.
Citation: https://doi.org/10.5194/essd-2024-587-RC2
Viewed
HTML | XML | Total | BibTeX | EndNote | |
---|---|---|---|---|---|
576 | 63 | 17 | 656 | 15 | 26 |
- HTML: 576
- PDF: 63
- XML: 17
- Total: 656
- BibTeX: 15
- EndNote: 26
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1