the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 07 Oct 2025
Mediterranean Sea Warming and Marine Heatwaves in 2024
Abstract. This study offers an assessment of Mediterranean Sea surface temperature (SST) trends (1982-2024) and marine heatwaves (MHWs) in 2024, using three MHW detection approaches: a fixed baseline (1982–2011), a 20-year moving baseline, and detrended SST data. Our trend analysis reveals a significant warming signal across the basin, with an average rate of 0.032 ± 0.001 °C/year and mean anomalies of 1.23 °C in 2024 relative to the 1982–2024 period. When evaluating MHW characteristics, 2024 stands out as the most extreme year of the past two decades in terms of total MHW days, regardless of the detection method. The Eastern Mediterranean (EMed) experienced record MHWs in 2024 across all three detection approaches. It recorded the highest number of MHW days and the longest mean event durations over the study period, and also showed the highest mean intensities under the fixed baseline. In contrast, in the Western Mediterranean (WMed), 2024 featured an exceptional number of MHW days; however, both the duration and intensity of events remained below those of 2003, 2022, and 2023, particularly when using the moving baseline and detrended methods, which reduce the influence of long-term warming. These results underscore the importance of MHWs in 2024 in the Mediterranean Sea, particularly in the EMed. In addition, this study highlights how methodological choices in MHW detection significantly shape the characterisation of extreme marine heat events in a warming climate.
- Preprint
(1408 KB) - Metadata XML
- BibTeX
- EndNote
Interactive discussion
Status: closed
-
RC1: 'Comment on sp-2025-13', Anonymous Referee #1, 08 Oct 2025
- AC1: 'Reply on RC1', Blanca Fernandez-Alvarez, 03 Feb 2026
-
RC2: 'Comment on sp-2025-13', Anonymous Referee #2, 05 Nov 2025
General Comments
The paper presents an assessment of SST and SST-based MHW trends (1982–2024) in the Mediterranean Sea, with particular focus on the significant 2024 MHWs. The authors apply three MHW detection approaches considering fixed, moving, and detrended baselines. The study examines how the three detection approaches affect MHW descriptions (in terms of total MHW days, duration and mean intensity of MHWs) and highlights the exceptional character of 2024 MHWs.
However, the manuscript would benefit from a deeper discussion of the physical meaning and interpretative implications of the three methods. In particular, I encourage the authors to expand the discussion beyond the comparison between fixed and alternative baselines and explicitly address the conceptual difference between the moving and detrended approaches. Although both reduce the imprint of long-term warming, they describe different physical frameworks. Moving baselines yield extremes relative to the contemporary climate state, while detrending isolates variability independently of the long-term trend. Clarifying this distinction is essential for discussing the implications and limitations of each method for interpreting MHW trends and associated impacts, and thus would considerably strengthen the paper.
Another concern is the lack of discussion of the 2024 MHWs which are meant to be an important part of the paper. Apart from comparisons with previous years in terms of the three basin-averaged metrics, the study lacks a comprehensive description of MHW conditions during this exceptional year (which areas were most affected, when did the most severe MHW conditions occur?) and a basic discussion of potential drivers. There is definitely enough space for such an addition, as the paper now includes figures (Fig. 3g-i) that are not even mentioned in the text. The inclusion of results should be balanced according to their relative importance.
The study should also be enriched by a broader integration of the existing literature in several parts of the paper. With an enriched discussion of methods interpretation and associated limitations, clearer rationale for certain choices, and some textual clarifications (all listed in specific comments below), the manuscript will provide a much improved contribution to the MHW literature.
I would like to kindly encourage the authors to submit a revised version of their work taking this review as constructive feedback aimed at improving their already interesting work.
Specific comments:
Abstract:
Line 1-2: Here you say that the study includes analysis of SST trends for 1982-2024 and MHWs in 2024, which suggests that only 2024 events are analysed; please, rephrase
Line 10: Why “In contrast”? No contrast reported for the WMed MHW days in this sentence.
Overall, please ensure the abstract reflects the revisions suggested in this review.
Sect. 1 - Introduction
Line 19: “sea surface” should be removed from the definition, MHWs are not necessarily surface events
Line 20: Marine ecosystem disruption is not only by altering species distributions and triggering mass mortality events. I suggest including additional examples.
Line 24: “on the event’s spatial and temporal scale and distribution” what do you mean by distribution here? Please, clarify.
Line 28: Consider rearranging and enriching (recent literature) the info now included in this paragraph, separating Mediterranean-specific information from general MHW background. Also, “complete review” on drivers?
Line 33: This paragraph should better be enriched and split into smaller ones. A short introduction on previously applied alternative detection methods should be included, ending with the gaps this study aims to fill.
Line 41: adapt to what? Explain better the purpose of employing alternatives, and then describe these alternatives. Also, refer to studies that apply such methods explaining the corresponding objectives.
Line 47: Why the 2 past decades? The rationale should be explained in Methods.
Sect. 2 – Dataset and Methodology
Line 53: Please specify the detection algorithm used (if not developed by the authors). Providing this reference allows you to omit some standard processing steps (e.g. threshold smoothing).
Line 55: I would suggest: “..using three different methods, as in...”
Line 62: Change paragraph
Line 64: What do you mean by “real” here? Maybe choose another wording
Line 67: The term cumulative intensity in MHW studies is usually considered to be the product of intensity and event duration. Mean cumulative intensity per year or the sum of the cumulative intensities over a year may then be examined, depending on the objective. Please, clarify what you compute here.
I do not see results for this metric in the study (same for maximum intensity). Is there a reason to introduce these metrics in Methods?
Sect. 3 – Results and Discussion
Line 74: Acronyms EMed/WMed should be used consistently; note that the EMed–WMed boundary on the map includes the Adriatic in the EMed, but here you separate it. Please ensure consistency.
Line 79: I suggest enriching this part with more recent studies based either on the same or other SST datasets reaching closer to present.
Line 85: The increase is gradual throughout the entire study period; please, rephrase.
Line 92: Have these studies considered exactly the same EMed/WMed separation? If not, rephrase accordingly.
Line 112: Why this reference period? So far, 1982-2011 is used for the climatology computation and the entire 1982-2024 period is used for computing 2024 anomalies. Why is this most recent period used here for averaging MHW days? This should be explained for clarity. If possible, keeping consistency within the study would be ideal to avoid confusions and allow for inter-comparing your different findings that are relevant to deviations from the same state.
Line 116: Maybe remove this, this sentence compares findings between the last decade and the rest of the dataset, inter-annual is not much accurate here
Line 117: MHWs
Line 117: I do not understand the selection of this specific year, it does not appear as a changing point.
Line 120: In the fixed baseline method we detect events relevant to a past, cooler period in order to account for the long-term warming. Therefore, the observed increase in MHW days is not an artifact, or unreal. It is associated with the physical meaning of choosing to detect events that account for the long-term warming.
Line 128: It would be interesting to discuss this differentiation of mean intensity. Do these results potentially suggest a weaker role of the long-term warming in shaping the magnitude of this metric? In this respect, check also Marin et al. 2021 (https://doi.org/10.1029/2020JC016708) and Denaxa et al. 2025 (https://doi.org/10.5194/sp-6-osr9-10-2025) which report a more complex behaviour (from different aspects) of the mean MHW intensity compared to other MHW metrics.
Line 134: Better keep MHW days consistently throughout the text
Line 137: See previous comment
Line 138-9: Here you compare results for the moving baseline against what? It is not clear, please rephrase
Line 138: Not sure what “still” means here
Line 140: This is also true for the WMed, so better rephrase to support your point in the rest of the paragraph.
Line 140: the EMed
Line 148: This sentence needs to be rearranged/split.
Line 154: MHW days, see previous comments
Line 156: As noted in previous comments, I suggest expanding this discussion and compare (further supporting) your findings with recent studies (several satellite- and model-based findings report stronger increases in MHW durations in the eastern basin).
Line 159: Figs 3g-i are not presented at all in the text, therefore they should either be removed or you should adjust the results section accordingly. Spatial differences for MHW days across the basin are now only discussed based on the EMed-WMed time-series. Also, if you decide to show spatial distributions of metrics, why limit this discussion to the MHW days? In any case, the selection of the figures to be shown and corresponding discussions should be based on the importance of your findings.
Sect. 4 Conclusions
Line 171: They do not account for the long-term warming, please correct and rephrase this conclusion
Line 176: The paper lacks discussion on the three methods. The last paragraph suggests different approaches as next steps (ensemble approach, absolute temperature thresholds) but the section skips discussing the methods being the focus of this study. I suggest discussing differences in the physical meaning especially between the moving vs detrending methods (this should also be mentioned in the introduction of these methods as noted before) and consequent suggestions for applying these methods.
An important advancement would also be including critical reflection on limitations associated with the physical interpretation of results from the different methods. For instance, whether, and to what extent, we can assume that detrending corresponds to removing the climate-change warming signal, and what assumptions are implied for the remaining variability such as what should be considered internal/physical or interannual variability. Moreover, the implications/limitations of applying linearly detrending should be discussed in the context of warming basins that also exhibit strong multi-decadal variability, such as the Med Sea.
Line 184: This paragraph is likely accidentally located here. It also largely overlaps with the beginning of this section and should be removed.
Citation: https://doi.org/10.5194/sp-2025-13-RC2 - AC2: 'Reply on RC2', Blanca Fernandez-Alvarez, 03 Feb 2026
Interactive discussion
Status: closed
-
RC1: 'Comment on sp-2025-13', Anonymous Referee #1, 08 Oct 2025
This short paper provides an assessment of SST evolution throughout the Mediterranean Sea in terms of 1982-2024 trends, and in terms of 2002-2024 Marine heatwave (MHW) statistics (duration, days per year, intensity) computed with 3 different methods. A special focus is put on year 2024, which is shown to have been exceptional since 2002 in terms of number of MHW days, as well as in terms of mean event durations and intensities in the Eastern basin. The other main conclusion concerns the sensitivity of the results to MHW detection methods.
The paper is well written and clearly reports about most diagnostics that were produced, although it provides barely any hypothesis (or discussion) about the Mediterranean dynamics or possible MHW drivers in 2024 (or before). I would thus suggest to discuss the potential reasons (atmospheric state, ocean preconditioning, etc) that may explain the 2024 MHW events (duration, intensity, location), on which the paper is focused.
Some readers might also regret that this study is not innovative in terms of methodology: the exact same 3 detection methods (fixed baseline, moving base line, detrending) were compared on a smaller region and on a slightly shorter dataset in the Ocean Science 2025 preprint by the same authors, and the present study basically provides the same methodological conclusion. The main added value of the present study over the latter is thus its extension to the whole Med Sea and the inclusion of year 2024.
Regarding methodology again, I am also a bit surprised that the authors do not prioritize their 3 detection methods, nor comment about their own purposes, their strengths and weaknesses. Instead, the expressions "the increase is real" (line 1120), "ensemble perspective" and "robust metrics" (line 177) suggest that the 3 methods are implicitly considered as equally valid or adequate for the study. However, they are quite different by construction in the way they deal with long-term SST evolution (in particular the fixed baseline approach compared with the other two). A good part of the differences between the results of these methods is expected a priori (based on the existing literature about them), but this is not commented. In order to improve the paper, better anchor it in the existing methodological MHW literature, and enrich it with respect to the Ocean Science preprint, I would suggest to document, explain and discuss the known/expected properties and strengths of each approach. It would be also be very valuable to provide a posteriori recommendation about the 3 methods for the present purpose, and discard the least (or designate the most) adequate, if any.
Here are a few other remarks :
- L 62: I would suggest to write ".. remove the mean seasonal cycle" and to explain how this is done.
- Fig 1: is the 365-day moving average (L 107) applied on Fig1d as well? Please clarify. Would it be possible to see the original (unfiltered) SST timeseries as well in the background of panels c and d?
- L 149: I suggest to split this long sentence after "(respectively)".
- L 150: It indeed seems to me that "... the moving baseline its total was surpassed by 2022", but not by 2023.
- L 161: Are panels g,h,i shown for 2024? Please clarify.
- L 184-187. Is this paragraph the end of the conclusion? It repeats statements that are given before...
- Please consider providing the link toward the actual dataset that is used, and to the tools that were used for this study.
Citation: https://doi.org/10.5194/sp-2025-13-RC1 - AC1: 'Reply on RC1', Blanca Fernandez-Alvarez, 03 Feb 2026
-
RC2: 'Comment on sp-2025-13', Anonymous Referee #2, 05 Nov 2025
General Comments
The paper presents an assessment of SST and SST-based MHW trends (1982–2024) in the Mediterranean Sea, with particular focus on the significant 2024 MHWs. The authors apply three MHW detection approaches considering fixed, moving, and detrended baselines. The study examines how the three detection approaches affect MHW descriptions (in terms of total MHW days, duration and mean intensity of MHWs) and highlights the exceptional character of 2024 MHWs.
However, the manuscript would benefit from a deeper discussion of the physical meaning and interpretative implications of the three methods. In particular, I encourage the authors to expand the discussion beyond the comparison between fixed and alternative baselines and explicitly address the conceptual difference between the moving and detrended approaches. Although both reduce the imprint of long-term warming, they describe different physical frameworks. Moving baselines yield extremes relative to the contemporary climate state, while detrending isolates variability independently of the long-term trend. Clarifying this distinction is essential for discussing the implications and limitations of each method for interpreting MHW trends and associated impacts, and thus would considerably strengthen the paper.
Another concern is the lack of discussion of the 2024 MHWs which are meant to be an important part of the paper. Apart from comparisons with previous years in terms of the three basin-averaged metrics, the study lacks a comprehensive description of MHW conditions during this exceptional year (which areas were most affected, when did the most severe MHW conditions occur?) and a basic discussion of potential drivers. There is definitely enough space for such an addition, as the paper now includes figures (Fig. 3g-i) that are not even mentioned in the text. The inclusion of results should be balanced according to their relative importance.
The study should also be enriched by a broader integration of the existing literature in several parts of the paper. With an enriched discussion of methods interpretation and associated limitations, clearer rationale for certain choices, and some textual clarifications (all listed in specific comments below), the manuscript will provide a much improved contribution to the MHW literature.
I would like to kindly encourage the authors to submit a revised version of their work taking this review as constructive feedback aimed at improving their already interesting work.
Specific comments:
Abstract:
Line 1-2: Here you say that the study includes analysis of SST trends for 1982-2024 and MHWs in 2024, which suggests that only 2024 events are analysed; please, rephrase
Line 10: Why “In contrast”? No contrast reported for the WMed MHW days in this sentence.
Overall, please ensure the abstract reflects the revisions suggested in this review.
Sect. 1 - Introduction
Line 19: “sea surface” should be removed from the definition, MHWs are not necessarily surface events
Line 20: Marine ecosystem disruption is not only by altering species distributions and triggering mass mortality events. I suggest including additional examples.
Line 24: “on the event’s spatial and temporal scale and distribution” what do you mean by distribution here? Please, clarify.
Line 28: Consider rearranging and enriching (recent literature) the info now included in this paragraph, separating Mediterranean-specific information from general MHW background. Also, “complete review” on drivers?
Line 33: This paragraph should better be enriched and split into smaller ones. A short introduction on previously applied alternative detection methods should be included, ending with the gaps this study aims to fill.
Line 41: adapt to what? Explain better the purpose of employing alternatives, and then describe these alternatives. Also, refer to studies that apply such methods explaining the corresponding objectives.
Line 47: Why the 2 past decades? The rationale should be explained in Methods.
Sect. 2 – Dataset and Methodology
Line 53: Please specify the detection algorithm used (if not developed by the authors). Providing this reference allows you to omit some standard processing steps (e.g. threshold smoothing).
Line 55: I would suggest: “..using three different methods, as in...”
Line 62: Change paragraph
Line 64: What do you mean by “real” here? Maybe choose another wording
Line 67: The term cumulative intensity in MHW studies is usually considered to be the product of intensity and event duration. Mean cumulative intensity per year or the sum of the cumulative intensities over a year may then be examined, depending on the objective. Please, clarify what you compute here.
I do not see results for this metric in the study (same for maximum intensity). Is there a reason to introduce these metrics in Methods?
Sect. 3 – Results and Discussion
Line 74: Acronyms EMed/WMed should be used consistently; note that the EMed–WMed boundary on the map includes the Adriatic in the EMed, but here you separate it. Please ensure consistency.
Line 79: I suggest enriching this part with more recent studies based either on the same or other SST datasets reaching closer to present.
Line 85: The increase is gradual throughout the entire study period; please, rephrase.
Line 92: Have these studies considered exactly the same EMed/WMed separation? If not, rephrase accordingly.
Line 112: Why this reference period? So far, 1982-2011 is used for the climatology computation and the entire 1982-2024 period is used for computing 2024 anomalies. Why is this most recent period used here for averaging MHW days? This should be explained for clarity. If possible, keeping consistency within the study would be ideal to avoid confusions and allow for inter-comparing your different findings that are relevant to deviations from the same state.
Line 116: Maybe remove this, this sentence compares findings between the last decade and the rest of the dataset, inter-annual is not much accurate here
Line 117: MHWs
Line 117: I do not understand the selection of this specific year, it does not appear as a changing point.
Line 120: In the fixed baseline method we detect events relevant to a past, cooler period in order to account for the long-term warming. Therefore, the observed increase in MHW days is not an artifact, or unreal. It is associated with the physical meaning of choosing to detect events that account for the long-term warming.
Line 128: It would be interesting to discuss this differentiation of mean intensity. Do these results potentially suggest a weaker role of the long-term warming in shaping the magnitude of this metric? In this respect, check also Marin et al. 2021 (https://doi.org/10.1029/2020JC016708) and Denaxa et al. 2025 (https://doi.org/10.5194/sp-6-osr9-10-2025) which report a more complex behaviour (from different aspects) of the mean MHW intensity compared to other MHW metrics.
Line 134: Better keep MHW days consistently throughout the text
Line 137: See previous comment
Line 138-9: Here you compare results for the moving baseline against what? It is not clear, please rephrase
Line 138: Not sure what “still” means here
Line 140: This is also true for the WMed, so better rephrase to support your point in the rest of the paragraph.
Line 140: the EMed
Line 148: This sentence needs to be rearranged/split.
Line 154: MHW days, see previous comments
Line 156: As noted in previous comments, I suggest expanding this discussion and compare (further supporting) your findings with recent studies (several satellite- and model-based findings report stronger increases in MHW durations in the eastern basin).
Line 159: Figs 3g-i are not presented at all in the text, therefore they should either be removed or you should adjust the results section accordingly. Spatial differences for MHW days across the basin are now only discussed based on the EMed-WMed time-series. Also, if you decide to show spatial distributions of metrics, why limit this discussion to the MHW days? In any case, the selection of the figures to be shown and corresponding discussions should be based on the importance of your findings.
Sect. 4 Conclusions
Line 171: They do not account for the long-term warming, please correct and rephrase this conclusion
Line 176: The paper lacks discussion on the three methods. The last paragraph suggests different approaches as next steps (ensemble approach, absolute temperature thresholds) but the section skips discussing the methods being the focus of this study. I suggest discussing differences in the physical meaning especially between the moving vs detrending methods (this should also be mentioned in the introduction of these methods as noted before) and consequent suggestions for applying these methods.
An important advancement would also be including critical reflection on limitations associated with the physical interpretation of results from the different methods. For instance, whether, and to what extent, we can assume that detrending corresponds to removing the climate-change warming signal, and what assumptions are implied for the remaining variability such as what should be considered internal/physical or interannual variability. Moreover, the implications/limitations of applying linearly detrending should be discussed in the context of warming basins that also exhibit strong multi-decadal variability, such as the Med Sea.
Line 184: This paragraph is likely accidentally located here. It also largely overlaps with the beginning of this section and should be removed.
Citation: https://doi.org/10.5194/sp-2025-13-RC2 - AC2: 'Reply on RC2', Blanca Fernandez-Alvarez, 03 Feb 2026
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 354 | 0 | 0 | 354 | 0 | 0 |
- HTML: 354
- PDF: 0
- XML: 0
- Total: 354
- BibTeX: 0
- EndNote: 0
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This short paper provides an assessment of SST evolution throughout the Mediterranean Sea in terms of 1982-2024 trends, and in terms of 2002-2024 Marine heatwave (MHW) statistics (duration, days per year, intensity) computed with 3 different methods. A special focus is put on year 2024, which is shown to have been exceptional since 2002 in terms of number of MHW days, as well as in terms of mean event durations and intensities in the Eastern basin. The other main conclusion concerns the sensitivity of the results to MHW detection methods.
The paper is well written and clearly reports about most diagnostics that were produced, although it provides barely any hypothesis (or discussion) about the Mediterranean dynamics or possible MHW drivers in 2024 (or before). I would thus suggest to discuss the potential reasons (atmospheric state, ocean preconditioning, etc) that may explain the 2024 MHW events (duration, intensity, location), on which the paper is focused.
Some readers might also regret that this study is not innovative in terms of methodology: the exact same 3 detection methods (fixed baseline, moving base line, detrending) were compared on a smaller region and on a slightly shorter dataset in the Ocean Science 2025 preprint by the same authors, and the present study basically provides the same methodological conclusion. The main added value of the present study over the latter is thus its extension to the whole Med Sea and the inclusion of year 2024.
Regarding methodology again, I am also a bit surprised that the authors do not prioritize their 3 detection methods, nor comment about their own purposes, their strengths and weaknesses. Instead, the expressions "the increase is real" (line 1120), "ensemble perspective" and "robust metrics" (line 177) suggest that the 3 methods are implicitly considered as equally valid or adequate for the study. However, they are quite different by construction in the way they deal with long-term SST evolution (in particular the fixed baseline approach compared with the other two). A good part of the differences between the results of these methods is expected a priori (based on the existing literature about them), but this is not commented. In order to improve the paper, better anchor it in the existing methodological MHW literature, and enrich it with respect to the Ocean Science preprint, I would suggest to document, explain and discuss the known/expected properties and strengths of each approach. It would be also be very valuable to provide a posteriori recommendation about the 3 methods for the present purpose, and discard the least (or designate the most) adequate, if any.
Here are a few other remarks :
- L 62: I would suggest to write ".. remove the mean seasonal cycle" and to explain how this is done.
- Fig 1: is the 365-day moving average (L 107) applied on Fig1d as well? Please clarify. Would it be possible to see the original (unfiltered) SST timeseries as well in the background of panels c and d?
- L 149: I suggest to split this long sentence after "(respectively)".
- L 150: It indeed seems to me that "... the moving baseline its total was surpassed by 2022", but not by 2023.
- L 161: Are panels g,h,i shown for 2024? Please clarify.
- L 184-187. Is this paragraph the end of the conclusion? It repeats statements that are given before...
- Please consider providing the link toward the actual dataset that is used, and to the tools that were used for this study.