Heat waves are among the most relevant extreme climatic events due to their effects on society, agriculture and environment. The aim of this work is to improve our understanding of heat waves over the Mediterranean basin during the 21st century from an ensemble of Regional Climate Models (RCMs). Focus has been placed on sensitivities to forcing global models, emissions scenarios and the RCM resolution, being the first work based on Euro-CORDEX simulations to fully analyze future heat waves in the Mediterranean. Heat wave features are studied with Warm Spell Duration Index (WSDI, duration) and Heat Wave Magnitude Index daily (HWMId, intensity). Results indicate a large increase by the end of the century in both intensity and length of heat waves from all emissions scenarios, global models, and regional models at any resolution. Exceptional heat waves observed early on the century could then become normal by the end of this period. Forcing global models and emissions scenarios play a major role. Clear added value on spatial distribution and heat wave indices are obtained from global to regional models dynamical downscaling, related to the important coastal or orographic aspects widely present over the Mediterranean.
By the end of current 21st century, the last IPCC report1 indicates that the mean global temperature is projected to increase more than 2 °C for the more extreme greenhouse gases emissions scenario hypotheses (the so-called Representative Concentration Pathways, RCPs). Focusing on extreme climatic features, present and recent past conditions studies2 have already measured important temperature changes. Larger increases than the ones obtained for the mean temperatures are expected due to global warming for future conditions3,4,5. Extreme events and, in particular, heat waves such as the recent ones (2003, 2010, 2015 or 2018) over Europe6,7 are likely to be attributed to this warming conditions8. Heat waves, due to the societal vulnerability and their effects over human health9,10,11,12,13,14, infrastructures15,16, agriculture6 or natural ecosystems1,17, are considered among the most relevant extreme climatic events. For climate change projections these phenomena can become even more relevant, pointing to a clear increase in their frequency and duration throughout the 21st century1,18,19,20.
Heat wave is defined, according to the World Meteorological Organization21, as an extreme weather event with marked warming of the air over a large area that usually lasts from few days to few weeks. It has to be clearly above the usual values and thus, high percentiles are needed to characterize such events22. In23, a heat wave is defined as a period where at least 6 consecutive days exceeds their respective calendar-day 90th percentile of daily maximum temperature. In24, it is considered 3 consecutive days over the 98th percentile of maximum temperature. There is plenty of specific definitions with slight differences for thresholds or number of consecutive days. In22,25,26, it is concluded that there is no perfect or complete method to fully characterize heat waves. Each definition allow us to partially describe heat waves through different magnitudes: the number of days (episode length), the number of events (to distinguish between several short and very long ones that cover similar total days), the magnitude (accumulated temperature exceedance) or the hottest day on the event (peak temperature). Any accurate heat wave definition should include both intensity and duration aspects. Some of the most commonly used indices are HWMId (Heat Wave Magnitude Index daily27) for intensity and WSDI (Warm Spell Duration Index25) for duration.
From a meteorological perspective, a heat wave event is produced when a high pressure system remains in the same place for a prolonged period of time, that make a heat wave event to last by advecting warm dry air to the affected region22,28,29. This situation is enhanced over dry soils or low humidity regions due to the extreme temperatures probability amplification30,31,32,33,34,35,36,37. High pressure systems are also related to clear-sky conditions and above-average radiative heating. These typical blocking conditions during summer season can be caused by specific conditions38 in the previous spring.
The Mediterranean basin is a region where hot conditions and particularly heat waves are common and relevant extreme climatic features during summer39,40,41. Several studies have already analyzed this region for present climate conditions and for future climate projections6,23,27,32,33,42,43,44,45,46,47. Using an ensemble of up-to-date Global Climate Models (GCMs), an increase in heat wave days, number of events and peak intensities for the Mediterranean basin have been projected for future conditions48. This region is among the ones with largest projected changes. Over the Mediterranean region, several projects have been focused on the regional climate analysis by downscaling procedures in the last two decades, starting from the pioneering projects PRUDENCE49,50 and ENSEMBLES51 to TiPES52. For a region such as the Mediterranean, and for a complex procedure such as heat waves, Regional Climate Models (RCMs) seems to be an interesting tool to complement the studies performed with GCMs. In recent years, CORDEX initiative53 and, specifically, the Euro-CORDEX54,55 and MedCORDEX56 projects offer a new set of higher resolution simulations that allow us to improve our understanding on how extreme events can be described for present and future conditions. The analysis of43 and57 (specifically for the center of Europe) have shown the ability of the Euro-CORDEX simulations to represent heat waves for present conditions, as forced by the ERA-Interim reanalysis58. These works indicate also a large spread among RCMs. It is important to notice that modelled extreme temperatures can suffer from temperature biases59, and even more in the case of the Mediterranean basin60. Bias-correction methods have been proposed to help on this issue61,62. Nevertheless, future climate projections focused on heat waves description over certain parts of Europe have been analyzed: over France26, Portugal63, Central Europe64 and eastern Mediterranean65.
The main objective of this study is to analyze the capability of the largest available ensemble of RCMs to reproduce future heat wave events over the whole Mediterranean basin. Work novelty is related to the focus over the whole Mediterranean basin, with all the available Euro-CORDEX simulations, with both RCP4.5 and RCP8.5 climate change greenhouse gases emissions scenarios and with both 0.11° and 0.44° spatial resolutions (around 50 and 12 km size, respectively). Uncertainties, limitations and robust features are analyzed by comparing these different forcing global models, horizontal resolutions, emissions scenarios and the ensemble of regional climate models, and the computation of two indices to describe heat waves. Specific focus is set to the comparison between GCMs and RCMs with the aim to quantify the added value of regional climate dynamical downscaling methods related to heat wave events. The interest is focused on the climatological description of heat waves, as a needed first step for a later study of the mechanisms that could be responsible of the obtained changes.
Mean heat wave indices for the whole Mediterranean basin: 2071–2100
Mean annual values of the relative increments of both indices (HWMId, a-b panels, and WSDI, c-d panels) averaged over the whole Mediterranean basin for the 2071–2100 future period, together with the annual standard deviation is shown for each model for both RCPs and horizontal resolutions in Fig. 1. Much larger HWMId and WSDI increments for the RCP8.5 scenario (b-d panels) are obtained when compared with RCP4.5 (a-c panels), which is an expected result64 due to its higher radiative forcing. Heat waves are more intense and last longer under the RCP8.5 than in the RCP4.5 scenario, related to the larger increase of mean temperatures. Smaller differences between resolutions (0.44° gray bars and 0.11° green/pink bars) can be observed, which can be mainly related to the averaging result over such a large region (the whole Mediterranean). The global forcing model seems to be of importance. Thus, IPSL and HadGEM exhibit higher heat wave days and magnitude, for both emissions scenarios, as it can be clearly seen when looking at RCA4 regional model, which is forced by all the GCMs. The variability among RCMs show a big spread for both emissions scenarios, in agreement with the results already shown for present climate conditions43. Even though, an overall concordance among all RCMs is obtained for the more severe scenario (second column), with more intense and longer heat waves (as in26 only focused over France). Both resolutions, 0.11° and 0.44°, show very similar results for both indices, HWMId and WSDI, a result also obtained in57. A Mann-Whitney-U test for the averaged values for the whole domain revealed no statistical significance in the difference between median values when compared both resolution results, meanwhile it was significant when compared the two RCPs. This was the result for each model and both indices. With the aim to focus on the main relevant features of heat waves, in the following only higher resolution (0.11°) and emissions scenario (RCP8.5) simulations will be analyzed.
Mean climatic spatial patterns distribution
The spatial structure of both indices for RCMs at 0.11° resolution and RCP8.5 scenario (averaged for the last 30 years of the 21st century) is then presented on Fig. 2 (for HWMId) and Fig. 3 (for WSDI), together with the forcing GCMs. Figures scale has been truncated to 125 for HWMId, as exceptional heat waves66, and to 360 for WSDI. The overall patterns of each RCM seem to be first mainly related to their forcing GCM, as column panels (RCMs forced by same GCM) are closer to each other when compared with figures in rows (same RCM forced by different GCMs) on both indices. This is in agreement with the mean values for the whole basin described on previous section with Fig. 1. Among GCMs, CNRM-CM5 is the driving model that shows a less increase for both indices and HadGEM2-ES is the global model with future highest values in the heat wave indices. At the same time, a clear added value from the RCMs when compared with their forcing GCMs can be seen. Many regional features, such as orographic effects over mountainous regions and coastal effects over the Mediterranean basin are obtained4. In mountainous regions, heat wave indices exhibit larger increases, meanwhile coastal areas show less heat waves duration and magnitude at the end of century. These results might be due to a better representation of local sea-breeze circulation, and so the transport of cooler and moisture air inland4. This potential added value of regional models as a downscaling procedure to improve the representation of extreme events, such as heat waves, have already been pointed by43,55. Some studies even indicate that RCMs can help to reduce biases coming from the GCMs including heat waves67. When looking at the spatial patterns obtained by the RCMs, related to the low resolution forcing GCMs (first row of Figs. 2 and 3), more detailed patterns and features can be seen, for instance: coastal or orographic effects, which could be of interest for a more precise description of these type of extreme events on regional scales. This has been already pointed by several authors when analyzing heat waves as seen from regional climate model outputs26,27,30,64, being these references only focused on RCMs. Furthermore, in68 over Africa, it is shown that RCMs were not always able to improve the results of the driving GCMs. Therefore, this work can be seen in these sense as a step forward on RCM/GCM compared analysis, and so, on the studies of potential added value of RCMs. Thus, several regions such as the Iberian Peninsula, exhibit an important north/south and east/west gradients, as the result of the combined influence of the Atlantic ocean and the Mediterranean sea. A relative maximum of the indices is observed in Alps region compared with their surrounding areas69 both towards the Italian peninsula and Central Europe. The African coast also exhibits a clear difference between the eastern, central (with a relative minimum over the central African Mediterranean coasts of Libya and Tunisia) and western parts, as already seen in70. Sahara region, with the largest values, and mountainous areas over Europe or Turkey with relative maximum values can also be distinguished. All these patterns can be seen as a common signature of a warmer Mediterranean. HWMId and WSDI values would indicate, following previous studies using these indices, that the strongest present climate conditions heat waves could become almost the average values for the end of the century. The 2003 summer heat wave was the second strongest one in Europe since 19506 and reached a peak of HWMId equal to 44.727. Most models agree that this value would be exceeded in west Africa, Turkey and some parts of the Iberian Peninsula and Greece (>100) by the end of the 21st century under the RCP8.5 scenario. In70, mean values of WSDI between 15–25 days to year in present climate (1981–2010) were obtained for the Mediterranean coastal countries. Our results show mean RCM values between 60–90 heat wave days per year for the northern half of Europe and above 90 days per year for this future period are commonly obtained for most of the coastal areas, no matter which regional model and forcing GCM is used to compute WSDI. This result is in concordance with54, where Euro-CORDEX ensemble overview results were presented for the whole continent, including a brief description of heat waves. This agreement can be seen also over subregions such as Central Europe64, France26 or Africa70, despite the fact that a strict comparison is somewhat difficult, as each study uses slightly different indices to describe heat waves. In64, it is interesting to notice that in their analysis of heat waves over Central Europe pointed that future heat waves could be correlated with the distribution of maximum temperatures. Here it is shown that both HWMId and WSDI indices reproduce similar features and spatial patterns. Previously, ENSEMBLES regional models simulations, using the former SRES greenhouse gases emissions scenarios, already indicated a similar pattern of large increases23.
Despite the described aspects of RCM projections about heat waves increases, some spread among RCMs (when forced with the same GCM) are also obtained, which is consistent with43 for present conditions. Some examples of this behavior can be seen in CCLM RCM, that shows less warming in the eastern part of Africa, RACMO22E and WRF331F, that show less warming in France than in the rest of Europe, and ALADIN53, that presents higher warming than the other RCMs driven by the CNRM-CM5 model. Or in the comparison among all the RCA4 RCM combinations, as it is forced by all the GCMs.
Regional annual timeseries
To inspect the time evolution of heat waves throughout the whole 130 period (1971–2100) available from Euro-CORDEX RCM simulations, and with the aim to also study in more detail the subregional features seen on previous section, nine subregions have been defined covering the entire basin, separating the northern, half Mediterranean and African areas, and also the eastern, central and western sides of the basin (Supplementary Fig. S1). These subregions follow roughly what was proposed in the pioneering work about the Mediterranean climate of42. In54, just the northern side of the Mediterranean basin was considered for a similar analysis. Figure 4 for HWMId and Fig. 5 for WSDI compute the annual time series, in a 5-years running window, for each of the nine regions, for each of the 0.11º-RCP8.5 simulations, together with the ensemble mean of all the RCMs.
HWMId and WSDI values increase during the whole 21st century in all regions for all simulations, showing their maximum values by the end of the period. Statistical Mann-Kendall trend analysis with a positive and statistically significant trend (p-value < 0.05, see Supplementary Tables S1 and S2) is obtained for all the regions and simulations. One major result is that the temporal evolution of the indices is different among regions. Each region experience a different evolution both in terms of the magnitude of the change, as well as at the rhythm in which that increase occurs. The increasing trend is complex and non linear with time for each subdomain. Therefore, spatial and temporal heat wave features seem to be quite inhomogeneous. In that sense, the usage of RCMs instead of GCMs could then help on the analysis of the spatial and temporal features more in detail. Regions with a more pronounced increase of both indices are those situated at the east and south of the Mediterranean sea: EAST, WAFR, CAFR and EAFR. All regions exhibit larger heat waves duration increase, from values around 10 days/year during the first part of the period, to values larger than 50. In particular, Western African coast (WAFR) is the region with largest increases (up to 200 days on region average of WSDI, and values higher than 100 of HWMId, in agreement with66 findings, where it reaches values of 400 in HagGEM-RCA). This is probably due to the fact that the Saharan desert is part of this area. France area (FR) is the one with smallest increases (about 100 days in WSDI and less than 50 in HWMId), probably related to the fact that includes regions that are far away from the Mediterranean area. These results are coherent with what was seen in Fig. 2 for the 30 years (2071–2100) average. Changes in HWMId values also show a quick increase, similar to what has just been described for WSDI results. It is important to notice also that year-to-year variability is large, being specially clear for the end of the 21st century. This result indicates a large interannual variability of heat waves in future climate, as it is the case for current climate. It is remarkable to point that there are also differences among simulations, around what is shown with the curve of the mean value increases, which is logically consistent with the results spread already mentioned before, here in terms of the annual evolution over subregions. Thus, HadGEM-RCA4 and IPSL-WRF tend to show the highest HWMId increases for most of the regions. This is, however, not the case for all the regions. For example, IPSL-WRF clearly presents the highest magnitude values in EAFR and NEAST regions (almost 200 and 80, respectively, at the end of the century), while in WAFR or CAFR just reach medium values (about 60) comparing with the other RCMs. Also, RCM models forced by HadGEM GCM show an increment higher than the others in the three african regions. For WSDI, the same behavior is observed. As an example, HadGEM-CCLM shows the highest trend in annual heat wave days in FR region (with an increment of more than 150 days of heat wave per year to the end of the century), but it shows a smaller trend in the EAST one in comparison with the other RCMs. Regions with lower spread among RCMs are FR and ALPS. On the other side, EAST and WAFR are among the regions with larger dispersion. The comparison between both indices indicate very similar trends with increases in the main characteristics of heat waves, although some small differences have also been detected. WSDI shows a more smoothed trend than HWMId, that begins to increase later and more sharply. Heat waves magnitude (HWMId) begins to increase with higher rates around the year 2071, meanwhile annual heat wave days seem to have already started such increases with respect to the reference period (1971–2000), and so they can be evolving in a slightly different way and rhythm. This is a reasonable result, as they both do not describe exactly the same aspects of heat waves. These results then point to the interest of using more than one index to analyze heat waves. Heat waves are a complex phenomenon, so their complete characterization should be inspected based on the different aspects that define and complement the description of this extreme climatic process.
Summary and Conclusions
An analysis of heat waves for the Mediterranean basin as seen from Euro-CORDEX ensemble of Regional Climate Model database for future climate conditions in the late 21st century has been presented. As described in the introduction, current heat wave events for present climate are becoming more and more important, frequent and relevant. Results shown here are relative increments to the baseline period, so absolute current heat waves are not shown, but only increments related to those events. Therefore further increases obtained from our analysis point to the importance and relevance of these future climate extremes. The results point towards a clear increase by the end of the century, in both intensity and length, of heat waves over the Mediterranean basin in the RCP4.5 and RCP8.5 from any global model and regional model used (including both 0.11º and 0.44º resolutions). In terms of relative importance, greenhouse gases concentration emissions play a major role. The more severe the emissions scenario analyzed, the more intense and lasting heat waves projections will be. As shown by our results, in agreement with previous studies27,54, the selection of the GCM that forces the regional models is also relevant. Despite the spread among them, each of the RCMs have a role to describe heat waves, but is less important than the forcing GCM. This spread among RCMs was already indicated by a former study with ERA-Interim reanalysis forcing boundary conditions simulations43. Regional climate model resolution (0.11 vs 0.44 degrees size) seems to play a minor role, at least when looking at the overall picture of the whole Mediterranean basin. The analysis presented here is a step forward in a deeper, more complete and detailed analysis compared with the former results of only one index (number of heat waves)54, which was made for a sub-sample of the Euro-CORDEX ensemble and analyzed a sub-region of the whole Mediterranean basin. Another relevant feature of the work presented is the comparison between the forcing global models and the regional models, with the aim to look at the potential added value of such dynamical downscaling methodology in terms of both intensity and duration of heat waves. Overall values for large subregions are relatively similar, as it can be expected, but interesting regional features are obtained when going to smaller scales that the RCM are capable to simulate. The Mediterranean basin presents complex orographic characteristics, large land-surface heterogeneities and key and vast coastal processes over the whole basin that are likely to be quite relevant. The obtained results point clearly on that direction, and so to the added-value of using regional climate models related to a more precise description of heat waves over the Mediterranean region.
It has been proved that both WSDI and HWMId heat wave indices are relevant in the analysis and characterization of extreme heat events. The two indexes exhibit consistent results, but also some differences both on the spatial structures and in the time evolution throughout the whole period (1971–2100). For example, annual duration of heat waves (WSDI) seems to start rising earlier than their maximum intensity (HWMId), which would start later, in the middle of the century, and more abruptly. These aspects have not been described before and should be further analyzed in future works. This result point to the interest of using more than one index to better describe heat waves in a more precise way, as duration, magnitude and intensity are all relevant aspects. The subregional analysis indicate that the maximum heat wave increases are obtained in the western African coast (WAFR), probably because part of the Sahara desert is included in the studied domain. France region is the one with less projected heat waves increase at the end of 21st century. In summary, the results shown here indicate that by the end of the century the exceptional heat waves measured during the first years of this period in previous works27,43,71, could become almost the average values for the end of the century. A better understanding of heat waves over the Mediterranean basin regional environment will be a key aspect on the adaptation strategies needed under the current anthropogenic climate change conditions. Also, the uncertainties and limitations of the current regional climate models on their description are a crucial point to take into account. The next step of the work proposed here would be a detailed analysis of the climatic and atmospheric mechanisms and processes that could be responsible of the obtained changes.
The analysis uses the ensemble of RCM simulations available from Euro-CORDEX initiative54,55. The Euro-CORDEX simulations consists on multiple dynamical downscaling regional climate models forced by multiple GCMs from the Coupled Model Intercomparison Project Phase 5 (CMIP5)72, which is the base of the main climate change projections of the fifth IPCC assessment report1. GCMs have a horizontal resolution of about 2 degrees in longitude and latitude for the whole globe. Whilst RCMs from Euro-CORDEX describe the regional climate of Europe at a resolution from 0.44 to 0.11 degrees. Two RCP scenarios are used in this work. They correspond to the stabilization of the radiative forcing by the end of the 21st century at certain watts per square meter73. The emissions scenarios used here are 4.5 and 8.5 W/m2, being the first related to a moderate emissions scenario and the last one related to the highest forcing scenario. The large amount of Euro-CORDEX available RCM simulations matrix allows to compare different sensitivities and uncertainties due to all the climate modelling chain parts. The simulations cover the 1971–2100 time period, with the European domain fixed by the CORDEX protocol procedures (https://www.cordex.org/domains/).
The combinations of seven RCMs and six GCMs simulations is described in Table 1, where regional models, resolution, emissions scenarios and forcing global models are indicated. A brief description of global climate models can be seen at54,72. Daily maximum temperature of the simulations is the used magnitude, as the appropriate variable to compute indices related to temperature extremes22. Specific periods are: from 1971 to 2005, defined as historical runs, whereas from 2006 to 2100 are projections forced by the RCPs scenarios. The selected area for this study is the whole Mediterranean basin as shown in Supplementary Material Fig. S1. Models capability to reproduce Euro-CORDEX observed heat waves have already been described in several papers before. In55 the main climatic fields are analyzed when forced with ERA-Interim reanalysis and, specifically, for temperature extremes and heat waves on43 and57. An overview of the main fields for climate change scenarios using the RCPs and the GCM/RCM matrix described above can be seen in54. The study is focused directly on climate change results signal, with indices that are related to baseline periods corresponding to present climate conditions.
Heat wave indices
Two indices, Warm Spell Duration Index (WSDI)25 for duration and Heat Wave Magnitude Index Daily (HWMId)27 for magnitude have been used. WSDI was proposed by an international committee from the World Meteorological Organization, named ETCCDI (Expert Team on Climate Change Detection and Indices74), that proposed this magnitude to quantify in days the extension of warm spells in a general sense. It is defined as the number of days per year with at least 6 consecutive days in which the maximum daily temperature is higher than the 90th percentile of the maximum daily temperature in a 5 days moving window during the reference period. This period is here defined as 1971–2000. This index considers only the duration of heat waves, therefore, two heat waves with the same duration are considered equally severe, despite having different temperature exceedance from the reference period70.
Meanwhile, the Heat Wave Magnitude Index Daily (HWMId)27,66,71, is a dimensionless magnitude that was designed to consider both heat wave duration and intensity. It is described as the maximum magnitude of the heat waves in a year. A heat wave is defined from the occurrence during at least 3 consecutive days with daily maximum temperature above the calendar 90th percentile centered on a 31 day window related to the reference period (1971–2000), and the maximum magnitude is the sum of the daily magnitudes of each day that compose a heat wave. Both indices show the relative increment for future climate with respect the baseline period of each model employed. More details of the computational procedure of both indices can be found on the references, and on the manual of the free software R packages75 used: extRemes76 for HWMId and climdex.pcic for WSDI77.
Literature is plenty of proposals to characterize heat waves, all of them with advantages and shortcomings. The inspection of how consistent the description from two of the more commonly used indices is also a relevant objective of the work. Multi-annual mean of both indices have been calculated for the whole Mediterranean basin in the far future 30 year period (2071–2100). This is made for each model at both resolutions and both climate change scenarios. Statistical evaluation of the data was carried out with a Mann-Whitney-U test (p < 0.05) in R version 3.6.1 (2019-07-05)75 to test the differences in resolution in both scenarios (RCP4.5 and RCP8.5) and the differences between scenarios for both resolutions (0.11° and 0.44°). Annual temporal series are presented for nine subregions, to analyze temporal evolution of the heat waves properties with this spatial detail. The non-parametric test for monotonic trend detection, known as the Mann-Kendall test (from the R package Kendall78), have been computed to analyze the statistically significance of the trend of those time series, with a 95% confidence level.
Stocker, T. Climate change 2013: the physical science basis: Working Group I contribution to the Fifth assessment report of the Intergovernmental Panel on Climate Change (Cambridge University Press, 2014).
Donat, M. et al. Updated analyses of temperature and precipitation extreme indices since the beginning of the twentieth century: The HadEX2 dataset. J. Geophys. Res. Atmospheres 118, 2098–2118 (2013).
Seneviratne, S. I. et al. Changes in Climate Extremes and their Impacts on the Natural Physical Environment, 109–230 (Cambridge University Press, 2012).
Cardoso, R. M., Soares, P. M. M., Lima, D. C. A. & Miranda, P. M. A. Mean and extreme temperatures in a warming climate: Euro cordex and wrf regional climate high-resolution projections for Portugal. Clim. Dyn. 52, 129–157, https://doi.org/10.1007/s00382-018-4124-4 (2019).
Fischer, E. M., Rajczak, J. & Schär, C. Changes in European summer temperature variability revisited. Geophys. Res. Lett. 39 (2012).
Barriopedro, D., Fischer, E. M., Luterbacher, J., Trigo, R. M. & García-Herrera, R. The hot summer of 2010: redrawing the temperature record map of Europe. Science 332, 220–224 (2011).
Hoy, A., Hänsel, S. & Maugeri, M. An endless summer: 2018 heat episodes in Europe in the context of secular temperature variability and change. Int. J. Climatol. (2018).
Beniston, M., Stoffel, M. & Guillet, S. Comparing observed and hypothetical climates as a means of communicating to the public and policymakers: The case of European heatwaves. Environ. Sci. & Policy 67, 27–34 (2017).
Dunne, J. P., Stouffer, R. J. & John, J. G. Reductions in labour capacity from heat stress under climate warming. Nat. Clim. Chang. 3, 563 (2013).
Amengual, A. et al. Projections of heat waves with high impact on human health in Europe. Glob. Planet. Chang. 119, 71–84 (2014).
Rohat, G. et al. Influence of changes in socioeconomic and climatic conditions on future heat-related health challenges in Europe. Glob. planetary change (2018).
Díaz, J. et al. Mortality attributable to high temperatures over the 2021–2050 and 2051–2100 time horizons in Spain: Adaptation and economic estimate. Environ. Res. 172, 475–485, https://doi.org/10.1016/j.envres.2019.02.041 (2019).
Matzarakis, A., Laschewski, G. & Muthers, S. The heat health warning system in Germany—Application and warnings for 2005 to 2019. Atmosphere 11, 170 (2020).
Tomczyk, A. M., Bednorz, E. & Matzarakis, A. Human-biometeorological conditions during heat waves in Poland. Int. J. Climatol. (2020).
Colombo, A. F., Etkin, D. & Karney, B. W. Climate variability and the frequency of extreme temperature events for nine sites across Canada: implications for power usage. J. Clim. 12, 2490–2502 (1999).
McEvoy, D., Ahmed, I. & Mullett, J. The impact of the 2009 heat wave on Melbourne’s critical infrastructure. Local Environ. 17, 783–796 (2012).
Hobday, A. J. et al. A hierarchical approach to defining marine heatwaves. Prog. Oceanogr. 141, 227–238 (2016).
Frich, P. et al. Observed coherent changes in climatic extremes during the second half of the twentieth century. Clim. Res. 19, 193–212, https://doi.org/10.3354/cr019193 (2002).
Ballester, J., Rodó, X. & Giorgi, F. Future changes in Central Europe heat waves expected to mostly follow summer mean warming. Clim. Dyn. 35, 1191–1205 (2010).
Horton, R. M., Mankin, J. S., Lesk, C., Coffel, E. & Raymond, C. A review of recent advances in research on extreme heat events. Curr. Clim. Chang. Reports 2, 242–259 (2016).
Data, C. Guidelines on the defintion and monitoring of extreme weather and climate events, tt-dewce. Tech. Rep., World Meteorological Organization (2015).
Perkins, S. E. A review on the scientific understanding of heatwaves:their measurement, driving mechanisms, and changes at the global scale. Atmospheric Res. 164, 242–267 (2015).
Fischer, E. M. & Schär, C. Consistent geographical patterns of changes in high-impact European heatwaves. Nat. Geosci. 3, 398–403 (2010).
Schoetter, R., Cattiaux, J. & Douville, H. Changes of western European heat wave characteristics projected by the CMIP5 ensemble. Clim. dynamics 45, 1601–1616 (2015).
Zhang, X. et al. Indices for monitoring changes in extremes based on daily temperature and precipitation data. Wiley Interdiscip. Rev. Clim. Chang. 2, 851–870 (2011).
Ouzeau, G., Soubeyroux, J.-M., Schneider, M., Vautard, R. & Planton, S. Heat waves analysis over France in present and future climate: Application of a new method on the EURO-CORDEX ensemble. Clim. Serv. 4, 1–12 (2016).
Russo, S., Sillmann, J. & Fischer, E. M. Top ten European heatwaves since 1950 and their occurrence in the coming decades. Environ. Res. Lett. 10, 124003 (2015).
Tomczyk, A. M. & Bednorz, E. Heat waves in Central Europe and tropospheric anomalies of temperature and geopotential heights. Int. J. Climatol. 39, 4189–4205 (2019).
Suarez-Gutierrez, L., Müller, W. A., Li, C. & Marotzke, J. Dynamical and thermodynamical drivers of variability in European summer heat extremes. Clim. Dyn. (2020).
Sillmann, J. et al. Understanding, modeling and predicting weather and climate extremes: Challenges and opportunities. Weather. climate extremes 18, 65–74 (2017).
Vautard, R. et al. Summertime European heat and drought waves induced by wintertime mediterranean rainfall deficit. Geophys. Res. Lett. 34, https://doi.org/10.1029/2006GL028001 (2007).
Della-Marta, P. M. et al. Summer heat waves over western Europe 1880–2003, their relationship to large-scale forcings and predictability. Clim. Dyn. 29, 251–275 (2007).
Fischer, E. M., Seneviratne, S. I., Lüthi, D. & Schär, C. Contribution of land-atmosphere coupling to European summer heat waves. Geophys. Res. Lett. 34, https://doi.org/10.1029/2006GL029068 (2007).
Coumou, D. & Rahmstorf, S. A decade of weather extremes. Nat. climate change 2, 491 (2012).
Lau, N.-C. & Nath, M. J. Model simulation and projection of european heat waves in present-day and future climates. J. Clim. 27, 3713–3730, https://doi.org/10.1175/JCLI-D-13-00284.1 (2014).
Brunner, L., Schaller, N., Anstey, J., Sillmann, J. & Steiner, A. K. Dependence of present and future European temperature extremes on the location of atmospheric blocking. Geophys. Res. Lett. (2018).
Schaller, N. et al. Influence of blocking on Northern European and Western Russian heatwaves in large climate model ensembles. Environ. Res. Lett. 13, 054015 (2018).
Brunner, L., Hegerl, G. C. & Steiner, A. K. Connecting atmospheric blocking to European temperature extremes in spring. J. Clim. 30, 585–594 (2017).
Lionello, P. et al. The Mediterranean climate: an overview of the main characteristics and issues (2006).
Schär, C. et al. The role of increasing temperature variability in European summer heatwaves. Nature 427, 332–336 (2004).
Stefanon, M., D’Andrea, F. & Drobinski, P. Heatwave classification over Europe and the Mediterranean region. Environ. Res. Lett. 7, 014023 (2012).
Giorgi, F. & Lionello, P. Climate change projections for the Mediterranean region. Glob. planetary change 63, 90–104 (2008).
Vautard, R. et al. The simulation of European heat waves from an ensemble of regional climate models within the EURO-CORDEX project. Clim. dynamics 41, 2555–2575 (2013).
Christidis, N., Jones, G. S. & Stott, P. A. Dramatically increasing chance of extremely hot summers since the 2003 European heatwave. Nat. Clim. Chang. 5, 46 (2015).
Dong, B., Sutton, R., Shaffrey, L. & Wilcox, L. The 2015 European heat wave. Bull. Am. Meteorol. Soc. 97, S57–S62 (2016).
Dong, B., Sutton, R. T. & Shaffrey, L. Understanding the rapid summer warming and changes in temperature extremes since the mid-1990s over Western Europe. Clim. Dyn. 48, 1537–1554 (2017).
Sánchez-Benítez, A., García-Herrera, R., Barriopedro, D., Sousa, P. & Trigo, R. June 2017: The Earliest European Summer Mega-heatwave of Reanalysis Period. Geophys. Res. Lett. 45, 1955–1962 (2018).
Perkins-Kirkpatrick, S. & Gibson, P. Changes in regional heatwave characteristics as a function of increasing global temperature. Sci. Reports 7, 12256 (2017).
Sánchez, E., Gallardo, C., Gaertner, M., Arribas, A. & Castro, M. Future climate extreme events in the Mediterranean simulated by a regional climate model: a first approach. Glob. Planet. Chang. 44, 163–180 (2004).
Jacob, D. et al. An inter-comparison of regional climate models for Europe: model performance in present-day climate. Clim. Chang. 81(S1), 31–52 (2007).
der Linden, P. V. & Mitchell, J. F. B. ENSEMBLES: Climate Change and its impacts: Summary of research and results from the ENSEMBLES project. Tech. Rep., Met Office Hadley Centre, FitzRoy Road, Exeter EX1 3PB, UK (2009).
Goswami, B. et al. Tipping points in the earth system: An introduction to the tipes project. Geophys. Res. Abstr 21, 1 (2019).
Giorgi, F. & Gutowski, W. J. Jr. Regional dynamical downscaling and the CORDEX initiative. Annu. Rev. Environ. Resour. 40, 467–490 (2015).
Jacob, D. et al. EURO-CORDEX: new high-resolution climate change projections for European impact research. Reg. Environ. Chang. 14, 563–578 (2014).
Kotlarski, S. et al. Regional climate modeling on European scales: a joint standard evaluation of the EURO-CORDEX RCM ensemble. Geosci. Model. Dev. 7, 1297–1333 (2014).
Ruti, P. M. et al. Med-cordex initiative for mediterranean climate studies. Bull. Am. Meteorol. Soc. 97, 1187–1208 (2016).
Plavcová, E. & Kyselý, J. Temporal characteristics of heat waves and cold spells and their links to atmospheric circulation in euro-cordex rcms. Adv. Meteorol. 2019, 1–13, https://doi.org/10.1155/2019/2178321 (2019).
Dee, D. P. et al. The ERA-Interim reanalysis: configuration and performance of the data assimilation system. Q. J. R. Meteorol. Soc. 137, 553–597, https://doi.org/10.1002/qj.828 (2011).
Christensen, J. H. & Boberg, F. Temperature dependent climate projection deficiencies in CMIP5 models. Geophys. Res. Lett. 39 (2012).
Boberg, F. & Christensen, J. H. Overestimation of Mediterranean summer temperature projections due to model deficiencies. Nat. Clim. Chang. 2, 433 (2012).
Dosio, A. Projections of climate change indices of temperature and precipitation from an ensemble of bias-adjusted high-resolution EURO-CORDEX regional climate models. J. Geophys. Res. Atmospheres 121, 5488–5511 (2016).
Dosio, A. & Fischer, E. M. Will half a degree make a difference? Robust projections of indices of mean and extreme climate in Europe under 1.5 C, 2 C, and 3 C global warming. Geophys. Res. Lett. 45, 935–944 (2018).
Parente, J., Pereira, M., Amraoui, M. & Fischer, E. Heat waves in Portugal: Current regime, changes in future climate and impacts on extreme wildfires. Sci. The Total. Environ. 631–632, 534–549, https://doi.org/10.1016/j.scitotenv.2018.03.044 (2018).
Lhotka, O., Kyselý, J. & Farda, A. Climate change scenarios of heat waves in Central Europe and their uncertainties. Theor. Appl. Climatol. 131, 1043–1054 (2017).
Zittis, G., Hadjinicolaou, P., Fnais, M. & Lelieveld, J. Projected changes in heat wave characteristics in the eastern Mediterranean and the Middle East. Reg. environmental change 16, 1863–1876 (2016).
Dosio, A., Mentaschi, L., Fischer, E. M. & Wyser, K. Extreme heat waves under 1.5c and 2c global warming. Environ. Res. Lett. 13, 054006 (2018).
Sørland, S. L., Schär, C., Lüthi, D. & Kjellström, E. Bias patterns and climate change signals in GCM-RCM model chains. Environ. Res. Lett. 13, 074017 (2018).
Dosio, A., Panitz, H.-J., Schubert-Frisius, M. & Lüthi, D. Dynamical downscaling of CMIP5 global circulation models over CORDEX-Africa with COSMO-CLM: evaluation over the present climate and analysis of the added value. Clim. Dyn. 44, 2637–2661 (2015).
Gobiet, A. et al. 21st century climate change in the European Alps - a review. Sci. Total. Environ. 493, 1138–1151 (2014).
Dosio, A. Projection of temperature and heat waves for Africa with an ensemble of CORDEX Regional Climate Models. Clim. Dyn. 49, 493–519 (2017).
Russo, S. et al. Magnitude of extreme heat waves in present climate and their projection in a warming world. J. Geophys. Res. Atmospheres 119 (2014).
Taylor, K. E., Stouffer, R. J. & Meehl, G. A. An overview of CMIP5 and the experiment design. Bull. Am. Meteorol. Soc. 93, 485–498 (2012).
Moss, R. H. et al. The next generation of scenarios for climate change research and assessment. Nature 463, 747 (2010).
Zhang, X. et al. Indices for monitoring changes in extremes based on daily temperature and precipitation data. Clim. Chang. 2, 851–870, https://doi.org/10.1002/wcc.147 (2011).
R Core Team. R: A Language and Environment for Statistical Computing. (R Foundation for Statistical Computing, Vienna, Austria, 2019).
Gilleland, E. & Katz, R. W. extRemes 2.0: An extreme value analysis package in R. J. Stat. Softw. 72, 1–39, https://doi.org/10.18637/jss.v072.i08 (2016).
Bronaugh, D. climdex.pcic: PCIC Implementation of Climdex Routines, R package version 1.1-9 (2018).
McLeod, A. Ian. Kendall rank correlation and Mann-Kendall trend test. R Package Kendall (Western Univ. 2005).
This work has been supported by a pre-doctoral fellowship by the University of Castilla-La Mancha (UCLM), financed by Fondo Social Europeo (ESF).
The authors declare no competing interests.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
Molina, M.O., Sánchez, E. & Gutiérrez, C. Future heat waves over the Mediterranean from an Euro-CORDEX regional climate model ensemble. Sci Rep 10, 8801 (2020). https://doi.org/10.1038/s41598-020-65663-0
This article is cited by
Time series clustering using trend, seasonal and autoregressive components to identify maximum temperature patterns in the Iberian Peninsula
Environmental and Ecological Statistics (2023)
Investigating the representation of heatwaves from an ensemble of km-scale regional climate simulations within CORDEX-FPS convection
Climate Dynamics (2023)
Scientific Reports (2022)
Climate change implications for olive flowering in Crete, Greece: projections based on historical data
Climatic Change (2022)
Scientific Reports (2021)