Observed Changes in Temperature Extremes over Greece: Warm and Cold Spells †
1
Hellenic National Meteorological Service, 167 77 Hellinikon, Greece
2
Laboratory of Atmospheric Physics, University of Patras, University Campus, 265 00 Patras, Greece
*
Author to whom correspondence should be addressed.
†
Presented at the 17th International Conference on Meteorology, Climatology, and Atmospheric Physics—COMECAP 2025, Nicosia, Cyprus, 29 September–1 October 2025.
Environ. Earth Sci. Proc. 2025, 35(1), 68; https://doi.org/10.3390/eesp2025035068
Published: 9 October 2025
Abstract
The daily maximum and minimum temperatures measured by the HNMS’s stations from 1960 to 2022, are used to compute percentile-based indices capturing the percentage of days below or above the 10th and 90th percentile, respectively (TN10p, TX10p, TN90p, TX90p), and event duration indicators (WSDI and CSDI). The climate extremes indices are evaluated assuming two different reference periods (1961–1990 and 1991–2020), and trend analysis is performed using the Mann–Kendall test. The results show a significant increase in the frequency of the warm days and nights. The magnitude and perceived timing of trends depend on the baseline chosen. Using the warmer 1991–2020 reference period dampens the upward trends in warm–extreme indices and amplifies the downward trends in cold extremes.
1. Introduction
Although the Mediterranean region is not warming faster than other regions, it is often referred to as a hotspot for climate change and biodiversity. The annual temperature in the Mediterranean has reached +1.5 °C compared to the pre-industrial period, and warming rates are highest in summer, particularly for maximum temperatures. This will increase the intensity, frequency, and duration of heatwaves.
Climate indices proposed by the Expert Team on Climate Change Detection and Indices (ETCCDI) monitor the occurrence of climate extremes of temperature (i.e., warm and cold spells) and can be grouped into three different categories: percentile-based, threshold-based, and absolute-value-based [1].
In the present study, homogenized daily maximum and minimum temperature data series covering the period from 1960 to 2022 were used. To identify and analyze temperature extremes in terms of warm and cold spells, percentile-based ETCCI indices were calculated, and trend analysis was performed.
2. Materials and Methods
2.1. Study Area, Data Source and Homogeneity
The study area is Greek region, which approximately extends from 34° N to 42° N and from 19° W to 29° E. The time series of daily maximum (TX) and minimum (TN) air temperatures collected over the period 1960–2022 by 57 weather synoptic stations belonging to the Hellenic National Meteorological Network (HNMS) were used. The selected stations are distributed all over Greece covering all peripheries. Table 1 includes the names and coordinates of the stations.
The daily data series of meteorological observations are of capital importance for the study of the variability of climate extremes. However, changes in the observation conditions or in the environment of the meteorological stations may introduce anomalous perturbations in the data series. Therefore, homogenization of data series is a fundamental step for climatological analyses [2]. There are several homogenization methods, most of which focus on monthly series. In this study, the homogenization method Climatol 4.1.0 [3] was applied using daily TX and TN series. The recommended procedure for homogenizing daily data is to perform homogenization at a monthly scale and to adjust the daily series with interpolated monthly corrections. The Climatol method does not need to interpolate monthly corrections since it just splits the daily series at the break points detected in monthly homogenization and then reconstructs all the series from their homogeneous sub-periods. More details on the homogenization procedure are described in Mamara et al. [4].
2.2. Extreme Temperature Indices and Trend Detection
In this study, six percentile-based threshold indices proposed by ETCCI, focusing on both cold and warm extremes of daily minimum and maximum temperature were calculated for two different base periods: 1961–1990 and 1991–2020 (Table 2) [5]. WMO recommends that the base 30-year period should be updated every decade to consider the no stationarity of climate change, which is 1991–2020 currently. However, the 1961–1990 period has been retained as a standard reference period for long-term climate change assessments [6].
Trend analysis is performed for the six ETCCI indices using the Mann–Kendall test and Sen’s slope estimator. The Mann–Kendall test is a non-parametric test, widely used in climate studies and is recommended by WMO because its power and significance are not affected by the actual distribution of the data. The Mann–Kendall test was used to assess whether climate indices are increasing or decreasing over time and the statistical significance of trend direction. The Sen’s slope estimator was used to compute the climate indices percentage of change over each year for the 63-year period (1960–2022).
3. Results
3.1. Behavior of Extreme Temperature Indices
The outcomes reveal that the investigated temperature indices are strongly affected by the base period. This is not unusual, especially for percentile-based temperature indices, since they are more sensitive to the changes in the base period than the precipitation indices, due to the prompt and intense ongoing global warming [7]. Figure 1 illustrates six percentile-based temperature indices (average of 57 stations) with respect to the two based periods distinctly. As expected, from various studies and the ongoing globally warming climate [8], the frequency of warm days (TX90p) has gradually increased from the end of the 1970s to the middle of the 1990s, and after that followed a steeper increase. The frequency of warm nights presents a similar pattern, albeit TN90p exceedances are more mitigated when they are evaluated against the most recent base period 1991–2020. On the contrary, the frequency of cold days (TX10p) presents the opposite pattern, having a dramatical decrease from early 1990s, which is amplified when using the 1991–2020 reference period.
WSDI and CSDI are more sensitive when using different reference periods. WSDI presents a steep rise after 1980, indicating an increase in the number of warm spells relative to 1961–1990 climatology. By using a later (warmer 1991–2020) reference period, a time shift of about 20 years is observed for the increase in the number of warm spells, and the rate of change is reduced compared to using the earlier (cooler, 1961–1990) reference period. Conversely, CSDI presents low variability over time when using the cooler reference period, while data are more dispersed and extreme values are more likely when using the later warmer reference period.
Figure 2 indicates the temporal evolution of warm spells for five geographical regions. Warm spells were classified into three categories according to their duration: short events (6–15 days), medium events (16–25 days), and long events (longer than 25 days). Short warm spells (blue circles) occur most often and regularly in all regions. A decline in short warm spells is observed over north Greece. An absence of occurrence of medium warm spells (yellow circles) is observed between 1970 and 1985 for all regions, but regularly almost every year since 2000. A general increase in the number of medium warm spells is obvious over south Aegean. Long warm spells (red circles) are rare before 2000 for all regions, occurring in only 3–6 events out of 40 years. On the contrary, they occur extremely often after 2010, mainly over North and West Greece.
3.2. Trend Analysis of Extreme Temperature Indices
A discrepancy in the magnitude of trends is observed by using two different based periods. The decreasing trends of 10th percentile indices (TN10p, TX10p) are amplified by 0.6 percent per decade, i.e., 2.2 days/decade, when using the reference period 1991–2020, and the opposite phenomenon is observed in the increasing trends of 90th percentile indices. Differences in trend magnitude are even louder for WSDI and CSDI. Overall trends for temperature indices are summarized in Table 3. The highest magnitude of change in increasing trend was recorded for WSDI (using the 1961–1990) and was noted for the region of North Greece (2.7 days/decade). The highest magnitude of change in decreasing trend was recorded for TN10p (using the 1991–2020) and was noted for the region of South Aegean (−1.9 percent/decade, i.e., 7 days/decade).
Assuming the 1961–1990 reference period, all stations showed an increasing annual trend in warm days and nights, with 89% and 98% (80 and 89% using 1991–2020) of the total stations showing a significant positive trend, respectively (Figure 3).
Conversely, the majority of stations showed a decreasing annual trend in cold days and nights, with 77% (90% using 1991–2020) of the total stations presenting statistically significant negative trends. In terms of seasonality, the increase in warm nights during the summer is undeniable; 95% of stations showed a statistically significant positive trend, regardless of the reference period. Also, during autumn, a large number of stations presented statistically significant positive trends on warm nights. The increase in warm days is pronounced during summer; more than 80% of stations presented a statistically significant positive trend regardless of the reference period, and spring follows with 70% (63% using 1991–2020) of stations.
4. Conclusions
Although temperature percentile indices are linked to the presence of extreme warm and cold events, the results need careful interpretation. The usage of two different base periods led to discrepant results, especially for the WSDI and CSDI indicators. Using the warmer 1991–2020 reference period dampens the upward trends in warm–extreme indices and amplifies the downward trends in cold extremes, effectively shifting the apparent onset of pronounced warming measured by WSDI, by about two decades. This underscores that WSDI, which helps to quantify the frequency and duration of heatwaves, is highly sensitive to the selected reference period. However, a strong increase in the frequency of warm nights, mainly during summer and autumn, and in the frequency of warm days, during summer and spring, is apparent regardless of the reference period.
Author Contributions
Conceptualization, A.M.; methodology, A.M.; software, A.M., A.A.A. and V.A.; data curation, A.M. and N.K.; writing—original draft preparation, A.M.; writing—review and editing, A.M., A.A.A., and N.K.; visualization, A.M.; supervision, A.M., A.A.A., and N.K. All authors have read and agreed to the published version of the manuscript.
Funding
This research received no external funding.
Institutional Review Board Statement
Not applicable.
Informed Consent Statement
Not applicable.
Data Availability Statement
Climate indices and trends are available from the authors upon request. Daily temperature data are available from HNMS upon request.
Conflicts of Interest
The authors declare no conflicts of interest.
References
- Tank, A.M.G.K.; Zwiers, F.W.; Zhang, X. Guidelines on Analysis of Extremes in a Changing Climate in Support of Informed Decisions for Adaptation; Climate Data and Monitoring WCDMP-No. 72; World Meteorological Organization: Geneva, Switzerland, 2009. [Google Scholar]
- WMO. Guidelines on Homogenization; WMO-No 1245; WMO: Geneva, Switzerland, 2020; p. 54. [Google Scholar]
- Guijarro, J.A. User’s Guide of the Climatol R Package (Version 4.1.0). 2023. Available online: https://cran.r-project.org/web/packages/climatol (accessed on 20 January 2023).
- Mamara, A.; Argiriou, A.A.; Anadranistakis, M. Homogenization of mean monthly temperature time series of Greece. Ιnt. J. Climatol. 2013, 33, 2649–2666. [Google Scholar] [CrossRef]
- Zhang, X.; Alexander, L.; Hegerl, G.C.; Jones, P.; Tank, A.K.; Peterson, T.C.; Trewin, B.; Zwiers, F.W. Indices for monitoring changes in extremes based on daily temperature and precipitation data. Wiley Interdiscip. Rev. Clim. 2011, 2, 851–870. [Google Scholar] [CrossRef]
- WMO. Guidelines on the Calculation of Climate Normals; WMO-No. 1203; WMO: Geneva, Switzerland, 2017. [Google Scholar]
- Yosef, Y.; Aguilar, E.; Alpert, P. Changes in Extreme Temperature and Precipitation Indices: Using an Innovative Daily Homogenized Database in Israel. Int. J. Climatol. 2019, 39, 5022–5045. [Google Scholar] [CrossRef]
- Stocker, T.F.; Qin, D.; Plattner, G.K.; Tignor, M.; Allen, S.K.; Boschung, J.; Nauels, A.; Xia, Y.; Bex, V.; Midgley, P.M. 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: Cambridge, UK, 2014. [Google Scholar]
Figure 1.
Extreme temperature indices as average of 57 stations considering two base periods (red lines: 1961–1990 and blue lines: 1991–2020). Bold lines present Loess smoothing.
Figure 1.
Extreme temperature indices as average of 57 stations considering two base periods (red lines: 1961–1990 and blue lines: 1991–2020). Bold lines present Loess smoothing.
Figure 2.
Time series of warm spells (WSDI) using 1961–1990 climatology for five geographical regions of Greece. The three categories of WSDI events are color coded (blue circles indicate warm spells of 6 to 15 days, yellow indicates those of 16 to 25 days and red indicates those of more than 26 days). The radii of the circles are proportional to the percentage of stations found to have the corresponding warm events.
Figure 2.
Time series of warm spells (WSDI) using 1961–1990 climatology for five geographical regions of Greece. The three categories of WSDI events are color coded (blue circles indicate warm spells of 6 to 15 days, yellow indicates those of 16 to 25 days and red indicates those of more than 26 days). The radii of the circles are proportional to the percentage of stations found to have the corresponding warm events.
Figure 3.
Percentage of stations showing different annual and seasonal (winter-DJF, spring-MAM, summer-JJA, autumn-SON) trend characteristics for extreme temperature indices: (a) base period 1961–1990; (b) based period 1991–2020.
Figure 3.
Percentage of stations showing different annual and seasonal (winter-DJF, spring-MAM, summer-JJA, autumn-SON) trend characteristics for extreme temperature indices: (a) base period 1961–1990; (b) based period 1991–2020.
Table 1.
List of the stations adopted in this study. For each station, latitude and longitude (in decimal degrees) are reported.
Table 1.
List of the stations adopted in this study. For each station, latitude and longitude (in decimal degrees) are reported.
| Name (Lat Lon) | Name (Lat Lon) | Name (Lat Lon) |
|---|---|---|
| Aghialos (22.793 39.224) | Kalamata (22.023 37.069) | N.Filadelfeia (23.746 38.045) |
| Aktio (20.769 38.922) | Karpathos (27.147 35.428) | Naxos (25.373 37.101) |
| Alexandroupoli (25.947 40.857) | Karystos (24.391 38.001) | Paros (25.115 37.022) |
| Andravida (21.287 37.923) | Kasteli (25.329 35.189) | Rhodes (28.089 36.402) |
| Araxos (21.422 38.149) | Kastoria (21.274 40.448) | Samos (26.916 37.691) |
| Argos (22.760 37.633) | Kerkyra (19.914 39.608) | Santorini (25.474 36.402) |
| Argostoli (20.505 38.120) | Kithira (22.989 36.149) | Serres (23.529 41.077) |
| Astros (22.719 37.406) | Konitsa (20.738 40.046) | Sitia (26.103 35.216) |
| Chios (26.142 38.345) | Kos (27.091 36.801) | Skiathos (23.502 39.175) |
| Chrysoupoli (24.620 40.920) | Kozani (21.842 40.289) | Skyros (24.491 38.963) |
| Desfina (22.5298 38.4208) | Lamia (22.436 38.877) | Souda (24.145 35.529) |
| Doxato (24.252 41.066) | Larisa (22.460 39.646) | Soufli (26.291 41.194) |
| Edessa (22.041 40.809) | Leonidio (22.892 37.151) | Spata (23.931 37.921) |
| Elefsis (23.552 38.068) | Limnos (25.233 39.922) | Syros (24.949 37.427) |
| Florina (21.428 40.805) | Macedonia (22.970 40.529) | Tanagra (23.563 38.335) |
| Helliniko (23.742 37.890) | Methoni (21.705 36.825) | Tatoi (23.780 38.107) |
| Heraklio (25.182 35.335) | Milos (24.429 36.739) | Tripoli (22.397 37.525) |
| Ikaria (26.345 37.683) | Mykonos (25.346 37.436) | Xanthi (24.903 41.137) |
| Ioannina (20.819 39.695) | Mytilini (26.604 39.054) | Zakynthos (20.888 37.751) |
Table 2.
List of the climate indices investigated in this study.
| Name | Units | Definition |
|---|---|---|
| Warm Spell (WSDI) | day | Annual count of days with at least 6 consecutive days when daily max temperature > 90th percentile centered on a 5-day window |
| Cold Spell (CSDI) | day | Annual count of days with at least 6 consecutive days when daily min temperature < 10th percentile centered on a 5-day window |
| Warm Days (TX90p) | % | Percentage of days with max temperature above the corresponding calendar day 90th percentile for a 5-day moving window |
| Warm Nights (TN90p) | % | Percentage of days with min temperature above the corresponding calendar day 90th percentile for a 5-day moving window |
| Cold Days (TX10p) | % | Percentage of days with max temperature below the corresponding calendar day 10th percentile for a 5-day moving window |
| Cold Nights (TN10p) | % | Percentage of days with min temperature below the corresponding calendar day 10th percentile for a 5-day moving window |
Table 3.
Overall trend for two based periods expressed in units per decade. Bold indicates significant trend at the 95% confidence level.
Table 3.
Overall trend for two based periods expressed in units per decade. Bold indicates significant trend at the 95% confidence level.
| Index | 1961–1990 | 1991–2020 |
|---|---|---|
| WSDI | 2.08 | 0.24 |
| CSDI | −0.01 | −1.24 |
| TN10p | −0.78 | −1.41 |
| TX10p | −0.63 | −1.18 |
| TN90p | 1.84 | 1.07 |
| TX90p | 1.43 | 0.94 |
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2025 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
MDPI and ACS Style
Mamara, A.; Argiriou, A.A.; Karatarakis, N.; Armaos, V. Observed Changes in Temperature Extremes over Greece: Warm and Cold Spells. Environ. Earth Sci. Proc. 2025, 35, 68. https://doi.org/10.3390/eesp2025035068
AMA Style
Mamara A, Argiriou AA, Karatarakis N, Armaos V. Observed Changes in Temperature Extremes over Greece: Warm and Cold Spells. Environmental and Earth Sciences Proceedings. 2025; 35(1):68. https://doi.org/10.3390/eesp2025035068
Chicago/Turabian StyleMamara, Anna, Athanasios A. Argiriou, Nikolaos Karatarakis, and Vasileios Armaos. 2025. "Observed Changes in Temperature Extremes over Greece: Warm and Cold Spells" Environmental and Earth Sciences Proceedings 35, no. 1: 68. https://doi.org/10.3390/eesp2025035068
APA StyleMamara, A., Argiriou, A. A., Karatarakis, N., & Armaos, V. (2025). Observed Changes in Temperature Extremes over Greece: Warm and Cold Spells. Environmental and Earth Sciences Proceedings, 35(1), 68. https://doi.org/10.3390/eesp2025035068
