Version classiqueVersion mobile
OpenEdition Books

The Mediterranean region under climate change

Jean-Paul Moatti
Stéphane Thiébault

Part 1. Mechanisms, observed trends, projections

Sub-chapter 1.2.1. The water cycle in the Mediterranean

Véronique Ducrocq, Philippe Drobinski, Silvio Gualdi et Patrick Raimbault

Texte intégral


1The Mediterranean region has quite a unique character that results both from physiographic conditions and historical and societal developments. Because of the latitude range it covers, the Mediterranean region is a transition area under the influence of both the temperate mid-latitude climate and the hotter-drier North-African climate. In addition, the region features a nearly enclosed sea surrounded by highly urbanized littorals and mountains in which numerous rivers have their source. This results in interactions and feedback between ocean-atmosphere-land processes that play a prominent role in the climate and hydrological cycle, including in the high-impact weather events that frequently affect the region.

2The hydrological Mediterranean basin is also characterized by a strong coastal component. Only 21 catchments have an area of more than 10,000 km2, and they represent only 42% of the total Mediterranean basin. The remaining part of the hydrological Mediterranean basin is made up of many small to medium size watersheds. Most of these small rivers are intermittent as well as ephemeral but can become powerful torrential rivers during flash-flooding episodes in fall and winter. The water resource is a major concern for a large part of the Mediterranean basin. Freshwater is rare and unevenly distributed in time and space, and the current situation is worsening due to increasing water demands related to population growth and economic development as well as to climate change.

3The Mediterranean region in fact concentrates all the main natural risks linked with the water cycle, including heavy precipitation leading to flash floods, strong winds and associated large swells and storm surges, heatwaves and droughts accompanied by forest fires. Such natural hazards affect the populations living in the area. Accurate forecasting of such high-impact weather events is still challenging and there are large uncertainties in the prediction of their evolution under climate change, especially precipitation extremes.

4The hydrological cycle in the Mediterranean region is thus a key scientific, environmental and socio-economic issue in a large region that includes southern Europe, North Africa and the Middle East, which motivated the 2010 launch of the 10-year HyMeX experimental research program dedicated to the hydrological cycle in the Mediterranean region (see Box 1).

5The following sections of this sub-chapter describe the main characteristics of the Mediterranean climate, of the Mediterranean Sea, and of the continental hydrology, their variability and trends over recent decades. Sub-chapters 1.2.2 and 1.2.3 present the results of the future climate projections under emission scenarios for the Mediterranean climate and the Mediterranean Sea, respectively. The variability, trends and future changes in hydrometeorological extremes under climate change are discussed in Chapter 1.3.

Box 1
HyMeX – A research program on the water cycle in the Mediterranean
The overarching objectives of the Hydrological Cycle in Mediterranean Experiment (HyMeX) program are to improve our understanding of the water cycle, with emphasis on the predictability and frequency of high-impact weather events, and to evaluate the social and economic vulnerability and capacity of the Mediterranean territories and citizens to adapt to these extreme events (, Drobinski et al. 2014). HyMeX is an international and interdisciplinary program involving about 350 scientists from about 20 countries.
The 10-year HyMeX program consists in observing and modeling the atmosphere/sea/continental surface system from the event to the climate scales. The HyMeX observation component includes (i) heavily instrumented special observation periods of a few months to provide detailed and specific observations to analyze key processes and (ii) longer observation periods repetitively or routinely collecting observations to monitor long-term water cycle processes and rare events such as flash-floods over a few specific instrumented watersheds. Two special observation periods, involving research aircraft and airborne instruments, instrumented platforms at sea and numerous ground-based research instruments, took place in northwestern Mediterranean to document heavy precipitation and flash-floods in the fall, 2012 (Fig. 1, Ducrocq et al. 2014) and ocean response to strong regional winds in late winter, 2013. The exploitation of the rich database (more than 450 datasets) obtained thanks to successful field campaigns, and the modelling components, has already produced numerous results, with more than 300 peer-reviewed articles published in international scientific journals. The outcomes of the program include the improvement of the numerical weather forecasting systems, the forecasting of flash-floods and the production of MED-CORDEX regional climate projections for the Mediterranean area for impact studies (Ruti et al. 2016), among others.

Figure 1
Some examples of research instruments deployed in the fall, 2012 in France, Spain, and Italy. Top row: ATR42 (left) and Falcon20 (right) from the French operator SAFIRE of environmental research aircraft (CNRS/INSU, Météo-France, CNES). Middle row: X-band radar from ETHZ deployed in the Cévennes (left), instrumented Météo-France buoy in the northern Mediterranean (center); Radiosonde launches from the CNRM mobile station (right).
Bottom row: Non-contact discharge radar (left), CTD rosette deployment at sea (right).

In France, the program is mainly funded by the MISTRALS meta-program (CNRS, Météo-France, IRSTEA, INRA), the French National Research Agency (ANR), CNES, and the territorial collectivity of Corsica.

Mediterranean climate

Temperature and precipitation variability

6The climatology of the Mediterranean region is characterized by dry summers frequently associated with very long drought periods, followed by fall and winter rainfall events that are mostly very intense. It is not rare that total monthly precipitation at a specific location falls in only few hours (during thunderstorms). Mean summer temperatures show a gradient from north to south with mean temperatures exceeding 30 ° C in the southeast (Ulbrich et al. 2013). Summer in southern Mediterranean regions is characterized by high temperatures and lack of rain, leading to drought and marked arid conditions. Total precipitation values show high spatial and temporal variability. Mean annual precipitation amounts range from less than 200 mm/year (North Africa, Arabian Peninsula) up to 2,000 mm/year over some northern mountainous areas. Winter half-year precipitation accounts for between 30% (western and northern Mediterranean area) and 80% (eastern and southeastern parts) of the annual total precipitation (Xoplaki et al. 2004).

7The high spatial and temporal variability of the seasonal mean temperature and total precipitation is explained by several features. First, the Mediterranean region is located at the southeastern limit of the North Atlantic storm tracks, and is thus particularly sensitive to interannual displacement of the paths of mid-latitude cyclones that can affect precipitation over the region. The Mediterranean climate is also influenced by tropical and subtropical systems (tropical cyclones, Asian summer monsoon, etc.). All these influences result in marked variability. In addition, the complex morphology of the Mediterranean region, including high mountain ridges surrounding the coast and sharp orographic features, islands and peninsulas, leads to much sharper and smaller-scale climatic features than over other ocean basins. The Mediterranean Sea is also a source of moisture and heat for the mesoscale atmospheric circulation that can evolve into high-impact weather systems such as heavy precipitation thunderstorms, cyclogenesis and wind storms.

Observed trends

8Observed trends indicate a general tendency for annual mean conditions to become warmer and drier. Indeed, observations and the CMIP5 global climate simulations show progressive warming of the land surface air temperature since the 1960s. Using the CMIP5 multi-model and observational data, Mariotti et al. (2015) estimated the annual mean surface air temperature trend in the Mediterranean region over the 1960-2005 period to be 0.19-0.25 °C per decade. The trend for the summer months is higher, over 0.3 °C per decade. The annual mean precipitation trend has been estimated to be about-0.6 10-2 mm/day/decade for the last century over the entire Mediterranean region. These observed long term trends are combined with marked decadal and interannual variability.

Mediterranean Sea


9The Mediterranean Sea is a semi-enclosed basin connected with the Atlantic Ocean through the Gibraltar Strait and with the Black Sea through the Dardanelles Strait. The surrounding orography tends to produce cold dry northern regional winds (Mistral, Tramontane, Bora, etc.) over the Mediterranean Sea, leading to high losses of heat and of freshwater by evaporation and latent heat transfer. These specific features strongly influence the water budget and the thermohaline circulation of the Mediterranean Sea. Indeed, the Mediterranean Sea has a negative water budget over a multi-year period: the loss to the atmosphere by evaporation is larger than gains due to precipitation and runoff from the rivers. This freshwater deficit is offset by exchanges through the narrow and shallow Strait of Gibraltar, where the inflow is composed of relatively warm and low-salinity upper water, while the outflow to the Atlantic Ocean is relatively cooler and saltier. Light low salinity water from the Atlantic is transformed into denser water through interaction with the low-level atmosphere and deep ocean convection that renew Mediterranean waters at intermediate and deep levels, and generate the thermohaline circulation in the Mediterranean Sea. The formation of deep ocean convection takes place preferentially in the Gulf of Lion, the Adriatic, the south Aegean and the north-east Levantine, regions under the influence of regional winds like the Mistral in the Gulf of Lion. Ocean convection produces deep vertical mixing processes that provide oxygen to the deepest part of the water column and consequently have major impacts on marine ecosystems.

Observed trends

10Estimation of trends using observations of the ocean is much less certain than estimation of trends over land, as series of observations are fewer and not always sufficiently reliable due to sampling errors or temporal homogeneity issues. This justified the setting up of additional long-term ocean observatories such as the MOOSE research observatory (see Box 2). Still, analyses of the trends observed in the Mediterranean Sea in recent decades tend to show a marked increase in the temperature of the deep layer from the mid-1980s as well an increase in salinity. Observations of sea surface temperatures show an increase over the last half-century (Sevault et al. 2014), with rates of increase estimated from different observation datasets for the 1979–2006 period in the range of 0.2–0.3 °C per decade (Mariotti, 2010). Concerning surface salinity, which displays decadal variability, no significant change was observed in the eastern basin, with possibly a small increase in the western Mediterranean (Ulbrich et al., 2013).

11Sanchez-Gomez et al. (2011) compared the heat and water budgets of the Mediterranean Sea based on state-of-the-art observational datasets and revealed a wide range of uncertainty. However, the datasets showed an overall increase in evaporation from the sea during the last half-century, with a decrease up until the mid-1970s followed by an increase, thereafter associated with the increase in sea surface temperature (Mariotti, 2010). Observed precipitation estimates over the Mediterranean Sea are very uncertain, which makes it difficult to produce evidence for robust trends. Yet these estimates have shown no significant trend over the whole Mediterranean Sea in recent decades.

Box 2
MOOSE - An example of research observatories
Despite intensive research efforts in the Mediterranean Sea over more than a century, an integrated view of its evolution, in the framework of climate change and anthropogenic pressures is still lacking. In this context, a Mediterranean Ocean Observing System for the Environment (MOOSE) was set up as an interactive, distributed and integrated observatory of the north-western Mediterranean Sea to detect and identify long-term changes in the Mediterranean Sea and its ecosystems. MOOSE, built as a multi-scale observation network, is based on a multisite system of continental-shelf and deep-sea fixed stations as well as Lagrangian and mobile platforms to observe the spatio-temporal variability of interactions between the coastal-open ocean and the ocean-atmosphere components. The long term aims of the observatory (Fig. 2) are (i) to monitor the mesoscale circulation in the north-western Mediterranean Sea; (ii) to qualify and quantify river inputs; (iii) to observe and understand the nutrient and gas (O2 and CO2) dynamics in the NW Mediterranean Sea in relation with physical ocean processes and anthropogenic/natural inputs (atmospheric and continental);(iv) systematic end-to-end monitoring, including of prokaryotic communities, phytoplankton, micro-and meso-zooplankton, in relation with environmental parameters to provide an overview of the evolution of biodiversity.
Data obtained during the first phase of MOOSE (2011-2015) allowed us to obtain new insight into the interaction of dense shelf water cascading with open-sea convection, their connection with peculiar atmospheric circulation patterns, and their impact on the alteration of deep water characteristics at basin scale. In deep water, a significant increase was observed in the nitrogen: phosphorus ratio, together with a clear decrease in the concentration of oxygen (1.2 µ mol/dm3/yr) revealing some changes in biogeochemical components of the Mediterranean Sea.
The MOOSE network is a solid observation service for research into the environment, which is able to provide operational service for the timely, continuous and sustainable delivery of high quality environmental data and information products, related to the northwestern Mediterranean environment.
The MOOSE network is primary funded by the French National Research Center (CNRS-INSU) and the Alliance Allenvi.

Figure 2
MOOSE set-up.

Continental hydrological cycle


12Hydrological processes are highly variable in time and space, due to the high variability of the rainfall regime, the complexity of the topography, and the geological, soil and land use characteristics. Most Mediterranean rivers have maximum discharges between February and May and minimum discharges in summer, due to reduced precipitation, elevated temperatures and the associated evapotranspiration in the summer season. Evapotranspiration over land is a key process in the annual continental hydrological cycle, representing between 50% and 80% of the total annual rainfall. Ranked according to annual water discharge, the 10 largest rivers flowing into the Mediterranean Sea are the Rhône, Po, DrinBuna, Nile, Neretva, Ebro, Tiber, Adige Seyhan and Ceyhan rivers, which together account for about half the average freshwater input to the Mediterranean Sea by rivers (Ludwig et al. 2009). The differences between low and high water discharge can be extreme, in particular in the numerous small to medium watersheds most of whose water is collected during short-duration floods. In space, hydrological regimes depend on geographical and anthropogenic factors such as basin size, topographic position (mountainous versus plain), the hydrogeological and aquifer systems (e.g. karstic regions), urbanization, the role of dams and reservoirs (e.g. Asswan), lakes (e.g. the more than 1,000 artificial lakes in Tunisia), human activities (irrigation, farming or industrial activities, hydroelectricity), etc.

Observed trends

13Except for the two largest rivers in the north (the Rhône and the Po) for which no trends have been observed, Mediterranean river discharges are decreasing, with a reduction of at least 20% in freshwater inputs to the Mediterranean Sea between 1960 and 2000 (Ludwig et al. 2009). This decrease is mainly due to the reduction in annual precipitation associated with climate change, and the construction of dams, which may have further reduced discharge. The biggest decrease is indeed observed in rivers that have been affected by the construction of dams such as the River Ebro in Spain and the Moulouya River in Morocco. Lespisnas et al. (2010) studied changes in discharge of smaller coastal rivers and found that the mean annual water discharge in their entire study region (southern France) decreased by about 20% between 1965 and 2004. As they found no clear trend in annual precipitation over the region, the decrease in water discharge is likely to be the result of the rise in temperature causing the switch from snowfall to rainfall at high altitudes and from the drop in groundwater levels, which are also partly explained by the rise in temperature. In a recent study (Zampieri et al., 2015), a consistent earlier spring discharge peak of more than three weeks per century has been found for the Rhône and Po.

14This shift in the river discharge of these two largest rivers might, in turn, have implications for the hydrological cycle of the whole region, or at least a large part of it.


15Analyses of long-term trends in the Mediterranean region show that annual mean conditions tend to be warmer and drier, with an increase in evaporation and a decrease in runoff. However, there is high spatial and temporal variability of the atmospheric, ocean and hydrological conditions due to the climate, topographical and anthropogenic factors that are specific to the Mediterranean.



Drobinski P., Ducrocq V., Alpert P., et al., 2014
HyMeX: a 10-year multidisciplinary program on the Mediterranean water cycle. Bulletin of the American Meteorological Society, 95: 1063-1082.

Ducrocq V., Braud I., Davolio S., et al., 2014
HyMeX-SOP1, the field campaign dedicated to heavy precipitation and flash flooding in the northwestern Mediterranean. Bulletin of the American Meteorological Society, 95:1083-1100.

Lespinas F., Ludwig W., Heussner S., 2010
Impact of recent climate change on the hydrology of coastal Mediterranean rivers in Southern France. Climatic Change, 99:425-456.

Ludwig W., Dumont E., Meybeck M., Heussner S., 2009
River discharges of water and nutrients to the Mediterranean and Black Sea: Major drivers for ecosystem changes during past and future decades? Progress in oceanography. 80:199-217.

Mariotti A., 2010
Recent Changes in the Mediterranean Water Cycle: A Pathway toward Long-Term Regional Hydroclimatic Change? Journal of Climate. 23 (6): 1513-1525.

Mariotti A., Pan Y., Zeng N., Alessandri, A., 2015
Long-term climate change in the Mediterranean region in the midst of decadal variability. Climate Dynamics, 44: 1437-1456.

Ruti P. M., Somot S., Giorgi F., et al., 2016
MED-CORDEX initiative for Mediterranean Climate studies. Bulletin of the American Meteorological Society. 97: 1187-1208.

Sanchez-Gomez E., Somot S., Josey S., Dubois C., Elguindi N., Déqué M., 2011
Evaluation of Mediterranean Sea water and heat budgets simulated by an ensemble of high resolution regional climate models. Climate Dynamics, 37: 2067-2086.

Sevault F., Somot S., Alias A., Dubois C., Lebeaupin-brossier C., Nabat P., Adloff F., Déqué M., Decharme B., 2014
A fully coupled Mediterranean regional climate system model: design and evaluation of the ocean component for the 1980-2012 period. Tellus A, 66: 23967.

Ulbrich U., Xoplaki E., Dobricic S., et al., 2013
Past and Current Climate Changes in the Mediterranean Region. Regional Assessment of Climate Change in the Mediterranean: Volume 1: Air, Sea and Precipitation and Water. Ed. Navarra A. and Tubiana L.. Springer Netherlands. 9-51

Xoplaki E., Gonzalez-Rouco J. F., Luterbacher J., Wanner H., 2004
Wet season Mediterranean precipitation variability: influence of large-scale dynamics and trends. Climate Dynamics. 23(1):63-78.

Zampieri M., Scoccimarro E., Gualdi S., Navarra A., 2015
Observed shift towards earlier spring discharge in the main Alpine rivers, Science of The Total Environment, 503–504: 222-232

Table des illustrations

Légende Figure 1Some examples of research instruments deployed in the fall, 2012 in France, Spain, and Italy. Top row: ATR42 (left) and Falcon20 (right) from the French operator SAFIRE of environmental research aircraft (CNRS/INSU, Météo-France, CNES). Middle row: X-band radar from ETHZ deployed in the Cévennes (left), instrumented Météo-France buoy in the northern Mediterranean (center); Radiosonde launches from the CNRM mobile station (right).Bottom row: Non-contact discharge radar (left), CTD rosette deployment at sea (right).
Fichier image/jpeg, 334k
Légende Figure 2MOOSE set-up.
Fichier image/jpeg, 272k


Climatologist, Laboratoire de Météorologie Dynamique, CNRS & École Polytechnique, France

Climatologist, Centro Euro-Mediterraneo sui Cambiamenti Climatici & Istituto Nazionale di Geofisica e Vulcanologia, Italy

Oceanographer, Mediterranean Institute of Oceanography (MIO), Marseille Université, CNRS/INSU, Université de Toulon, IRD, France

© IRD Éditions, 2016

Conditions d’utilisation :