Version classiqueVersion mobile
OpenEdition Books

The Mediterranean region under climate change

Jean-Paul Moatti
Stéphane Thiébault

Part 1. Mechanisms, observed trends, projections

Sub-chapter 1.4.4. Impact of atmospheric chemistry on the regional climate

Texte intégral

1In this section, we address the complex atmospheric-chemistry and climate interactions and feedback processes in the Mediterranean region.


2Projecting climate changes at regional scale remains a challenge. This is especially true in regions where aerosols play a significant role in the radiative balance with effects on the climate and water cycle (Nabat et al. 2014 and 2015) like in the Mediterranean region. Regional projections of this type are required for the design of adaptation and mitigation strategies and a deeper understanding of regional processes is thus needed.

3Atmospheric aerosol particles play a major role in the water cycle and also in the Earth’s energy balance through their radiative effects, including direct effects (by extinction of radiation and attenuation of surface illumination), semi-direct effects (due to the heating of turbid air layers by particles absorbing solar radiation), and indirect effects (through their influence on cloud properties). The uncertainty of these climate effects exceeds that of any other forcing because the physical, chemical and optical properties of aerosols are highly variable in space and time since their atmospheric lifetime is short and their emissions are very heterogeneous (Forster et al. 2007).

4The Mediterranean region is at the intersection of large scale circulation patterns with highly contrasted sources of particles. The region is subject to high particle loading as described in the sub-chapter ‘High concentrations of aerosols and pollutant and greenhouse gases’, high photochemical activity, with mixing of different origins: particles come from anthropogenic sources such as highly densified cities in Europe, Turkey, and the Middle East; desert dust from the Sahara and from the Middle East in the eastern part of the basin (Mamouri & Ansmann, 2015); maritime particles from the Mediterranean Sea; as well as organic particles from biomass burning (Lelieveld et al. 2002; Sciare et al. 2003, 2008). Expected impacts of climate change in the region include heat stress associated with poor air quality in the urban environment (Lelieveld et al. 2012, 2015). This is why the impact of particles in the Mediterranean region is high and is expected to change, and why each effect of different particles on the climate of the Mediterranean region, whether direct or indirect, needs to be understood and quantified.

Assessment of aerosol direct radiative forcing (DRF)

5Aerosols observed over the Mediterranean basin are known to be able to interact with both shortwave (SW) and longwave (LW) radiation (Nabat et al. 2012; Papadimas et al. 2012; Zanis et al. 2012; Sicard et al. 2014). In the SW spectral range, due to their variability in size and in chemical properties, they can scatter (sea salt, sulfates, nitrates, ammonium and secondary organic particles) or scatter plus absorb (black and brown carbon from combustion [smoke], certain types of mineral dust) solar radiation. Only mineral dust and sea salt can interact with LW radiation due to their large particle size. Consequently, Mediterranean aerosols significantly affect the radiative budget of the Mediterranean by (1) decreasing surface incoming shortwave radiation, (2) increasing/decreasing outgoing shortwave fluxes depending on the surface albedo, and (3) possibly by heating turbid atmospheric layers when the particles absorb solar light.

6Concerning pollution aerosols, SW direct shortwave radiative forcing (DRF) has been estimated at the local scale by many authors (Horvath et al. 2002; Markowicz et al. 2002; Meloni et al. 2003; Roger et al. 2006; Di Sarra et al. 2008; Di Biagio et al. 2009, 2010). These authors showed a significant decrease in surface solar fluxes of 20–30 W m-2 (daily mean) at different locations including Almeria (Spain), Finokalia (Greece), Lampedusa (Italy), Marseilles and Toulon (France). In parallel, the combination of surface and satellite remote-sensing observations performed at Lampedusa has been used to calculate the radiative effects, in both the shortwave (Di Biagio et al. 2010) and longwave (Di Sarra et al. 2011; Meloni et al. 2015) spectral ranges for different cases of Saharan dust intrusions. These studies emphasized that desert dust in the LW spectral range has a significant radiative effect, and offsets a large fraction of SW forcing (Di Sarra et al. 2011; Meloni et al. 2015). More recently, based on remote-sensing observations in Barcelona and 1-D radiative transfer calculations, Sicard et al. (2014) also estimated the LW radiative effect of dust. Only a few studies are available concerning the radiative impact over the Mediterranean region of intense biomass burning events. One estimate was proposed by Formenti et al. (2002) for an aged Canadian biomass-burning plume and revealed a significant SW surface dimming of 60 W m2. In addition, the radiative effect induced by smoke aerosols at Lampedusa between August 3 and 23, 2003, during an exceptionally hot and dry season when the Mediterranean atmosphere was affected by massive forest fires, was estimated by Pace et al. (2005) to be between + 22 and + 26 W m-2. For a complete review of local direct radiative forcing see Mallet et al. (2016).

7At the regional scale, Papadimas et al. (2012) proposed an estimation of the aerosol DRF using MODIS data from 2000 to 2007 for both all-sky and clear-sky only conditions. These authors derived a multi-year regional mean surface DRF of-19 W m-2, associated with a DRF of-4.5 W m-2 at the top of the atmosphere (TOA). It should be noted that such radiative forcing is regionally higher than that exerted by greenhouse gases. Regional modelling studies were also recently conducted by Nabat et al. (2012, 2015) using the coupled-chemistry RegCM and the CNRM-Regional Climate System Model (RCSM) models for multi-year simulations. These authors reported a mean SW regional surface (TOA) DRF of-13.6 W m-2 (-5.5 W m-2) and-20.9 W m-2 (-10.5 W m-2) for the RegCM and CNRM-RCSM (see Fig. 8) models, respectively. Zanis et al. (2012) also proposed a regional estimate of the DRF of anthropogenic particles for the 1996-2007 period using RegCM and showed a significant negative forcing of up to -23 W m−2 at TOA over Eastern Europe.

8In parallel with the radiative forcing exerted by aerosols, Richards et al. (2013) investigated the effect of reducing certain sources of emissions on the radiative effect of ozone over the Mediterranean (ozone shows a marked but localized summertime maximum, especially over the eastern basin). These authors reported a mean radiative effect at the top of the atmosphere of between -1 and -40 W m-2, depending on the types of emission tested. Hauglustaine & Brasseur (2001) reported mean radiative forcing associated with an increase in tropospheric ozone since the preindustrial era of around 0.70 W m-2. Such studies demonstrate that the radiative effect of ozone is significantly lower than that exerted by natural/anthropogenic aerosols over the Mediterranean region.

9The aerosol atmospheric load can also directly impact the production of secondary gaseous species (i.e. chemically produced from primary emitted species) by attenuating or scattering visible and UV radiation. Recently, Mailler et al. (2016) showed a net reduction in the photolysis rates of ozone and nitrogen oxides due to the absorbing effect of a mineral dust plume observed in Lampedusa. This effect led to a change of several ppbs in ozone surface concentrations. Over the Mediterranean Sea and continental Europe, close to the sources of NOx, the effect of dust leads to reduced ozone concentrations (but an increase occurs over remote areas such as the Sahara and the tropical Atlantic Ocean). By reducing photosynthetically active radiation (PAR), the aerosol may also reduce crop production, and hence biogenic emissions, leading to complex interactions in the formation of secondary organic aerosols that are not yet well understood.

Implication of aerosol direct radiative forcing (DRF) in the Mediterranean water cycle

10Concerning surface and TOA atmospheric forcings, Zanis et al. 2012; Spyrou et al. 2013; Nabat et al. 2014, 2015a, b recently investigated how changes in the radiative budget due to natural/anthropogenic aerosols influence the surface temperature (over both land and sea), relative humidity profiles, exchanges (latent heat fluxes) between ocean and atmosphere, cloud cover (semi-direct effect of absorbing particles), precipitation, and finally the whole Mediterranean hydrological cycle. Indeed, notable perturbations in sea surface-atmosphere fluxes are expected despite the relatively small size of the Mediterranean Sea, since the latter plays an important role at a much larger scale by providing moisture for precipitation to its surrounding land areas, which extend to northern Europe and northern Africa (Gimeno et al. 2010 and Schicker et al. 2010). In parallel, the absorbing particles over the Mediterranean Sea (Mallet et al. 2013) could have a semi-direct effect that could modify the vertical profiles of relative humidity and cloud cover.

11In that context, using simulation ensembles of the direct radiative effect of aerosols (Fig. 8a) carried out with the regional climate system model CNRM-RCSM, Nabat et al. (2015) showed annual average cooling of the Mediterranean sea surface caused by aerosols of-0.5 °C, up to-1 °C locally, and higher in spring and summer when aerosol loads (mainly sulfate and dust particles) are maximal (see summer average in Fig. 8b). This cooling also affects the surrounding land, not only due to negative aerosol DRF over land, but also because of the advection of maritime cooler air over land regions. For example, in summer, northern winds over the eastern basin favor additional cooling over northern Libya and Egypt.

Figure 8
Average (JJA 2003-2009) surface aerosol direct radiative forcing (DRF, W m-2) (a) and the resulting impact of the aerosol on surface air temperature (°C) (b), latent heat loss (W m-2) (c) and precipitation (mm day-1) (d), simulated by the CNRM-RCSM regional model. Hatched areas are not significant at the 0.05 probability level.

12In addition, aerosols are responsible for a decrease in the latent heat loss by evaporation over the Mediterranean Sea (Fig. 8c) due to these lower sea surface temperatures (SST). Consequently a reduction in the whole hydrological cycle has been attributed to aerosols, resulting in an average 10% decrease in specific humidity in the lower troposphere, in cloud cover and in precipitation (Fig. 8d). In addition, dust aerosols also reduce atmospheric convection in the region by warming the lower troposphere by absorbing solar radiation, and hence stabilizing the atmosphere. The comparison of these results with the model response in atmosphere-only simulations shows that feedback is reduced if SST cannot be modified by aerosols, highlighting the essential role of the Mediterranean SST and the need to use atmosphere-ocean coupled regional models in regional aerosol-climate studies. In addition, the decrease in SST causes an increase in the density of surface Mediterranean waters that favors ocean convection in sub-basins where deep water masses are formed (Gulf of Lion, Adriatic Sea and Aegean Sea), and reinforces the Mediterranean thermohaline circulation.

13It is also worth mentioning that since the 1980s, aerosol loads have been dramatically reduced over the Mediterranean basin (Nabat et al., 2013), because of the reduction in anthropogenic emissions of sulfate precursors in Europe (improved air quality standards plus economic crises). As a consequence, aerosol DRF decreased over the region between 1980 and 2012, leading to an increase in surface solar radiation (brightening period) and additional warming of land and ocean surface temperature: aerosol changes explain 23% of the warming in the region over this period (Nabat et al., 2014).

Aerosol-cloud interactions and impact on precipitation

14The Mediterranean region is particularly vulnerable to climate change due to its location at the interface of the hot dry North African climate and the cooler wetter European climate. The water cycle in this region is very sensitive to the position of the descending branch of the Hadley cell, which is expected to move poleward following global warming, leading to increasingly scarce fresh water (Giorgi & Lionello, 2008). Understanding the formation of clouds and precipitation and its link with atmospheric composition is particularly important in this region. Clouds and precipitation strongly influence tropospheric composition and radiative properties, water transport and energy redistribution. At the cloud scale, clouds and the precipitation life cycle are controlled by the prevailing meteorology and aerosol particles, particularly the presence of cloud condensation and ice nuclei (CCN/IN) (Flossmann Wobrock, 2010). Indeed, hydrometeors cannot form spontaneously in the thermodynamic conditions encountered in the atmosphere. They need a substrate (a nucleus) during the early stage of formation. An aerosol particle that serves as a nucleus for a cloud droplet is called a cloud condensation nucleus (CCN), and for an ice crystal, an ice nucleus (IN).

15An additional consequence of the enlargement of the Hadley cell (ICCP, 2014b) is the increasing frequency of dust particles transported over the eastern Mediterranean. Rosenfeld et al. (2001) reported that an excess concentration of atmospheric dust can prevent cloud precipitation. This argument is often referred to as the second indirect effect in the Twomey classification (1980): an excess of CCN causes an increase in the number of droplets and consequently the cloud droplet size distribution shifts to smaller sizes, thereby narrowing the droplet size distribution. This hampers precipitation. Indeed, precipitation is triggered by a mechanism that forms hydrometeors of millimetric sizes in a matter of minutes. Only the presence of a few large hydrometeors in the population that will fall and collect the population of small hydrometeors can lead to precipitation. The narrowing of the size distribution eliminates the largest hydrometeors and reduces precipitation ability. This will inherently amplify dryness due to extension of the Hadley cell, and is referred as the “desertification positive feedback loop”.

16Our understanding of the processes that take place inside clouds has considerably increased in recent decades (Flossmann and Wobrock, 2010, Lohmann et al. 2010, Levin and Cotton, 2008). Microphysical features of clouds in the Mediterranean region have already been characterized together with their link with the loading of aerosol particles (Levin et al. 1996, 2005; Rosenfeld et al. 2001; Teller et al. 2012). The impact of aerosols is uncertain but potentially maximal when aerosol particles can cause the formation of ice crystals in clouds. This is referred to as the ice indirect effect by Lohmann (2002). Indeed, global precipitation is predominantly produced by clouds containing ice crystals (DEMOTT et al. 2011). Aerosols that can act as IN are mostly insoluble particles that often mimic ice lattice structure. Only a few types of aerosols have this property, including mineral dust, volcanic ash, and bioaerosols such as bacteria, fungal spores, and pollen (Vali, 1985; Hoose & Möhler, 2012). Their IN ability, their likelihood of precipitating the ice phase, and to what extent anthropogenic aerosols can play a role in cloud ice formation are still not fully understood although the number of laboratory studies on the IN properties of different kinds of aerosol particles has been continuously increasing (e. g., Möhler et al. 2007, for bacteria; Connolly et al. 2009, and Ardon-Dryer & Levin, 2014, for dust).

17Even though air quality and water resources in the Mediterranean region have been the center of scientific interest, and have been studied intensively, many questions remain unanswered, in particular in the eastern part i.e., how and to what extent natural and anthropogenic aerosols can increase or prevent rainfall (e.g. Levin et al. 1996, Teller & Levin, 2006).

Ongoing research and recommendations

18Recent studies clearly demonstrated the added value of using regional climate models including ocean-atmosphere coupling to study the impact of the aerosol direct radiative effect on the Mediterranean climate. One priority should now be to conduct an intercomparison of such models to check the robustness of individual models, as scheduled in the next phase of the MedCORDEX program (​). In addition, the impact of the anthropogenic and natural aerosol DRF on future climate needs to be investigated using different RCP scenarios, with the 8.5 scenario as a priority. The second important point concerns the role of aerosols in cloud microphysical and macro-physical properties at climatic scale. To this end, regional climate simulations should be built including the first (aerosol-cloud albedo) and second (aerosol-cloud precipitation) indirect effects of anthropogenic and natural aerosols on warm clouds. A better representation of dust-IN interactions is also required over this region especially for the eastern Mediterranean. Specific experimental campaigns focused on this aspect associated with improvements in its representation in climate models should be encouraged.

Table des illustrations

Légende Figure 8Average (JJA 2003-2009) surface aerosol direct radiative forcing (DRF, W m-2) (a) and the resulting impact of the aerosol on surface air temperature (°C) (b), latent heat loss (W m-2) (c) and precipitation (mm day-1) (d), simulated by the CNRM-RCSM regional model. Hatched areas are not significant at the 0.05 probability level.
Fichier image/jpeg, 346k

© IRD Éditions, 2016

Conditions d’utilisation :