the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 09 Apr 2025
Increasing Daily Extreme and Declining Annual Precipitation in Southern Europe: A Modeling Study on the Effects of Mediterranean Warming
Abstract. Understanding the evolving patterns of intense rainfall in the Mediterranean under climate change is an urgent challenge. An analysis of the annual total and maximum one-day precipitation from 1955 to 2023 performed with the ERA5-Land dataset over the EURO-CORDEX domain reveals emerging patterns of contrasting trends along much of the northern Mediterranean coast, with heavy precipitation events increasing and total annual rainfall decreasing. An independent investigation on a ground-based dense monitoring network in southern Italy confirms the results. We focus on this representative sub-region of the study area to examine in detail the role of sea-atmosphere-orography interactions, particularly the impact of increasing sea surface temperature (SST), in enhancing heavy precipitation despite overall drying. Twenty consecutive precipitation events identified in a particularly intense rainy season (September–December 2019) are reproduced at a convection-permitting resolution (2 km) using the Weather Research and Forecasting (WRF) model with ERA5 reanalysis boundary conditions. Then, two scenarios are tested: one with past SST levels approximating 1980 and another with future SST increases in line with end-of-century Shared Socioeconomic Pathways (SSPs) projections. WRF simulations thoroughly describe cyclone tracks and precipitation patterns, showing that increased SST boosts the frequency of heavy rainfall events overland, though peak intensities remain mostly unchanged because the highest precipitations occur over the sea. The study demonstrates the unique capability of high-resolution, convection-permitting analyses to capture complex processes in orographically challenging regions and contributes to clarifying the seemingly contradictory trend of rising daily precipitation extremes despite falling annual precipitation totals in Southern Europe.
- Preprint
(21014 KB) - Metadata XML
-
Supplement
(11146 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-1567', Anonymous Referee #1, 20 May 2025
General Comments
This study investigates the evolution of heavy precipitation events in the Mediterranean basin under a changing climate, focusing on the role of sea-atmosphere-orography interactions. Using both observational data and numerical simulations, the authors assess the trends in annual and maximum precipitation, with an emphasis on the Calabrian peninsula. They also explore how sea surface temperature changes might impact precipitation patterns, particularly during intense rainy seasons. The study emphasizes the importance of high-resolution, convection-permitting models to capture key processes. While the topic is relevant and the approach is promising, several aspects of the manuscript require clarification and improvement to strengthen the overall impact and scientific contribution.
Introduction
This section is overall well-written and provides relevant background. However, it does not clearly address the specific research gaps this study aims to fill. The authors should explicitly state the study’s novelty (for example, the use of 20 real-case events at convection-permitting scale with calibrated SST perturbations) and how it builds on existing work.
Data and Methods
This section would benefit from greater clarity and conciseness. The SST perturbation approach is not clearly explained, details on the magnitude, spatial pattern, and implementation (e.g., uniform or spatially varying changes) are missing. The quadrant classification based on PRCPTOT and RX1day trends is also unclear, a clearer definition or simple diagram would help. Additionally, the event identification method (starting the event the day before precipitation begins) is unconventional, as most studies define event onset based on a precipitation threshold or objective criteria, further justification is needed.
Results and Discussion
Overall, the results and discussion sections offer valuable insights, but they lack clear physical explanations to support the findings.
In the Trend Analysis section, the authors present an analysis of observed rainfall trends in Calabria but fail to adequately link these results to the region's orographic features. While the orography is mentioned multiple times, there is no attempt to explain how it might be influencing the observed rainfall patterns, particularly in terms of total maximum rainfall. This lack of connection makes it harder to understand why different trends are observed.
In the Observed and Projected SST Warming section, the authors choose SSP changes of -1 and +3°C for their simulations, but the rationale behind these choices is unclear. Were these values based on region-specific data, or were they generalized from the broader Mediterranean basin? It would also be more logical for the authors to focus more on their study area (Domain 3) rather than a larger region.
For the WRF Simulation Evaluation, the authors chose to spatially interpolate the observations. However, it’s unclear whether this is the most appropriate method for comparison. Interpolating observations can introduce uncertainty and might not accurately reflect the spatial variability of the actual observations. It would be useful for the authors to justify why they chose this method.
The Figure 12 showing the eastward shift of extreme rainfall events is visually appealing, but the authors do not explain the underlying mechanisms driving this shift. While they focus on a single event, there is no physical explanation for the observed trend, which limits the depth of the physical understanding gained from the analysis.
In the Comparison with Previous Studies section, the authors compare their findings to studies that account for global warming, rather than focusing only on SST. This comparison could lead to misleading conclusions because the underlying drivers of rainfall changes could differ between global warming and SST warming alone. More clarity on the boundaries of their analysis and comparison with relevant studies focusing on SST would enhance the discussion.
Lastly, Figures 7 and 8 present an overwhelming amount of information, making them difficult to interpret. The authors might consider finding an alternative way to present these results, perhaps by simplifying the figures or breaking them down into more digestible parts.
Conclusions
The conclusions section lacks impact and doesn’t clearly tie the study’s findings together. It doesn't explain how the research advances our current understanding. While the authors summarize their results, they could better highlight the practical implications of their work. For example, they could link their findings to how the study might help predict or mitigate future storms. It would also be useful to mention the key implications of these trends for climate adaptation or urban planning, especially in terms of how extreme precipitation affects flood risks. Lastly, the statement that "only high-resolution, convection-permitting analyses can accurately capture key processes" is too strong and could benefit from further context.
Citation: https://doi.org/10.5194/egusphere-2025-1567-RC1 - AC1: 'Reply on RC1', Alfonso Senatore, 09 Jul 2025
-
RC2: 'Comment on egusphere-2025-1567', Anonymous Referee #2, 26 May 2025
- AC3: 'Reply on RC2', Alfonso Senatore, 09 Jul 2025
-
RC3: 'Comment on egusphere-2025-1567', Anonymous Referee #3, 27 May 2025
This is an interesting study on understanding better the precipitation paradox in the Mediterranean (decreased mean precipitation/increased extremes). It falls within the scope of HESS, and I believe it could attract the interest of the scientific community. While there are different approaches and sound datasets used, there are no straightforward linkages between the various parts of the analysis.
The overall presentation is well structured and clear. One exception is the inclusion of discussions in the results. I would include any discussion in a separate section or present it with the conclusions. High-quality visualizations are used for the presentation of results. Some more effort could be put into making the language style more fluent.
In the title, please mention that you refer to the “Mediterranean Sea Warming”.
Overall, the introduction section is informative, however, the list of references is not exhaustive. Topics such as future extreme precipitation trends in the Mediterranean or why the region is characterised as a climate change hotspot could be better covered. Any research gaps and main objectives of the present analysis could be more emphasized.
EURO-CORDEX is mentioned several times in the text, however, the only linkage with this regional initiative is the selection of the domain. Please mention this only once in the definition of your domain of analysis.
Since the findings presented in Figures 7 and 8 are mostly used for estimating a reasonable warmer-SST scenario, I strongly recommend moving these two figures to the Appendix. The number of visualizations is already large.
In the methods section, it is not clear how future SSTs were taken into account in the WRF simulations. Some information is presented in the results (L262-269), however, this approach should be demonstrated in more detail.
Some additional explanation of the methods used to derive Figure 12 should also be included in the methods.
For increased confidence, I strongly recommend repeating the analysis of Figure 13 for an additional event. For example, for event 15, which is characterised by extreme rainfall, underestimated by the SST0 simulation.
Minor comments are provided in the attached PDF document.
- AC4: 'Reply on RC3', Alfonso Senatore, 09 Jul 2025
-
RC4: 'Comment on egusphere-2025-1567', Anonymous Referee #4, 27 May 2025
In their study "Increasing Daily Extreme and Declining Annual Precipitation in Southern Europe: A Modeling Study on the Effects of Mediterranean Warming" the authors investigate how projected changes in SST would unfold with respect to precipitation extremes of the Mediterranean north shore and in particular for the region of Calabria. In addition they elaborate on the hypothesis that there is a general trend of decreasing total annual precipitation along with increasing daily maximums, based on an analysis of reanalysis data and local observations. The paper addresses an important topic and the rationale is reasonable. The analysis is sound and well structured. Some arguments and conclusions may require a deeper analysis and discussion than currently presented, also given the limited significance of the trend analysis.
The analysis of the ERA5-Land data for the EURO-CORDEX area does show negative trends but they are only significant for the Iberian peninsula and some regions of northern Africa. The findings for RX1day are even weaker. This lack of robustness should be addressed in more detail in the discussion. From ERA5-Land it looks like for the Calabrian region, most pixels are seen in zone I of Fig. 3 and only a few in zone II which would contradict your statement about the match of ERA5-Land and the local observations. If you skipped the non significant values in Fig. 5 only a few would remain. What's the reason for the non-significance? Are there recurrent outliers in the observations?
For the simulations with the regional atmospheric model it is assumed that just the SSTs are changing according to certain SSP scenarios. The atmospheric properties remain unchanged which creates an inconsistency for the described future conditions. In your WRF configuration, the GHG settings seem to be constant across your simulations. Later WRF versions, e.g., with the CAM radiation scheme allow for an adjustment of GHG concentrations and also support different SSPs. With these updated atmospheric settings, you should obtain a more realistic interplay between sea surface and the atmospheric boundary layer, mostly due to radiation effects.
Another concern is the influence of large scale dynamics and patterns. From the text it is not clear whether you applied the updated SSTs to both domains. I assume you did so also for the outer domain. Adding this much energy to the full extent of domain one could considerably change the larger scale dynamics, patterns and feedback. Therefore, to corroborate your findings of translocated precipitation events, I recommend to create another set of simulations to apply some spectral nudging to the outer domain, at least for the geopotential, to ensure consistency of the large scale structures also with respect to ERA5.
Moreover, the effect of sea surface salinity should also be considered in your discussion or limitations section. Would the increase in SST and a decrease in total precipitation lead to increased salinity levels and how could that potentially diminish evaporation and consequently reduce severe precipitation events?
How relevant are gradients between the SSTs of the Tyrrenian and Ionian Sea for the emergence of extreme precipitation events in the region?
How are the main horizontal wind fields for the precipitation events (Fig. 12) organized? Is there any obvious clustering for the big events, e.g. all originate from the south? Is it possible to annotate the dots of the events with an arrow that shows the direction of the storm path? Are the obtained spatial shifts of the event centers consistent for small perturbations (e.g. another PBL scheme or varied initial conditions)? Was there any spin-up performed to exclude impacts of imbalanced soil moisture? How do these extreme events look like in the outer domain: Are the centers of precipitation mass identically located to what was found for the inner domain?
Minor:
The title is probably not so ideal since your main focus is on SST sensitivity. Maybe better: "… on the Effects of Mediterranean Sea Warming"?
L33: Give the actual years instead of "In the last two years"
L324: From Fig. 13 it seems that the storm system of event 12) travels along the coast in northward direction and that the high precipitation over the sea occurs after it traveled over the east part of the Calabrian coast rather than "exploding before".
L391: I think the investigation should not be called "comprehensive" since many real-world aspects had been left out in this PGW experiment.
Figure 2: add the term ERA5-Land to the caption.
Figure 3: add the term ERA5-Land to the caption. Increase image resolution.
Figure 4: "Mann-Kendall and Sen's slope test for observations of a) ..."
Figure 5: Increase image resolution. It would be good to scale the pointsize by the trend values. It's also hard to distinguish the significant values. A different color might be better show them. Add "observations" to figure caption.
Figure 7 & 8: annotate the periods in the figure and add "GCM ID" or similar to the y-axis.
Figure 10: Add "precipitation" somewhere in the figure caption
Figure 13: Add red cross section line also to c) and e) and add the center of mass points for event 12; what is the unit of Omega?Citation: https://doi.org/10.5194/egusphere-2025-1567-RC4 - AC1: 'Reply on RC1', Alfonso Senatore, 09 Jul 2025
- AC2: 'Reply on RC4', Alfonso Senatore, 09 Jul 2025
Data sets
Daily precipitation (mm) simulated with WRF on the D02 domain by varying the SST scenarios Alfonso Senatore et al. https://zenodo.org/records/14848874
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 545 | 129 | 23 | 697 | 31 | 17 | 39 |
- HTML: 545
- PDF: 129
- XML: 23
- Total: 697
- Supplement: 31
- BibTeX: 17
- EndNote: 39
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1