Version classiqueVersion mobile

The Mediterranean region under climate change

 | 
Jean-Paul Moatti
, 
Stéphane Thiébault

Part 1. Mechanisms, observed trends, projections

Sub-chapter 1.4.1. Sources of reactive species and source apportionment

Texte intégral

Figure 1
Variety of continental sources impacting the Mediterranean basin.

1Natural and anthropogenic emissions of gaseous and particulate pollutants are key factors in air quality degradation and climate change. As indicated by the Latin origin of the name Mediterranean (“the middle of the lands”), the Mediterranean basin is a receptor of anthropogenic emissions from eastern Europe and surrounding coastal urban areas combined with wind-driven dust from the Sahara and Arabian deserts, biogenic emissions from the surrounding vegetation, and sea salt [KANAKIDOU et al. 2011; Fig. 1].

Natural sources

2Natural sources play a key role in the exchange of compounds between the Earth’s surface, the oceans and the atmosphere. The species emitted are characterized by high diversity and potentially high chemical reactivity. Consequently they have direct and indirect impacts on climate and air quality.

Marine sources

3The sea surface produces airborne particles that contribute to the aerosol load, Earth’s albedo, climate, and air quality in marine environments. There are large gaps in our knowledge of marine emissions, which, in turn, are responsible for large uncertainty on our future climate (Carslaw et al. 2013). Marine primary aerosols are produced by bubble bursting processes, mostly under the influence of wind-driven wave breaking mechanisms. Primary marine organic particles consist of microbiological organisms (including viruses and bacteria), biological debris, exudates and by-products. Marine secondary aerosols are formed by condensation of gas-phase species emitted from the seawater. Secondary organic aerosol (SOA) particles are expected to result from the atmospheric oxidation of biologically driven emissions of volatile organic compounds (VOC), which have not yet been clearly identified. Among the formation pathways of secondary aerosol particles, nucleation is the process responsible for the formation of new nanoscale particles (as opposed to the process of condensation onto pre-existing particles). Triggered by photochemical processes, new particle formation (NPF) takes place as an “event” that lasts several hours, during which the concentration of the clusters of nanoparticles increases to high levels by nucleation. These clusters grow rapidly to a few nanometers in size when they can be detected. NPF is expected to generate a large number of aerosols, which, in turn, can affect climate by influencing cloud radiative processes (Spracklen et al. 2006). Questions are still open concerning the conditions that favor the occurrence of NPF and particularly the type and the origin of precursors. How are VOC emissions and nucleation events influenced by water composition (clean or polluted seawaters), biological composition and activities? In the framework of ChArMEx, marine aerosol emissions to the atmosphere have recently been studied in two research projects, MEDSEA (MEDiterranean Sea Acidification) and SAM (Sources of marine Aerosol in the Mediterranean atmosphere). Both used mesocosms, i.e. semi-opened chambers containing natural seawater and an atmospheric headspace. Results showed a clear correlation between the level of seawater chlorophyll a (Chl a), measured as the standard proxy for phytoplankton biomass, and the amount of organic compounds in airborne particles produced during the wave breaking process (Schwier et al. 2015). This result is in agreement with parametrizations obtained in the Atlantic Ocean combining Chl-a levels measured by satellite and the organic fraction of ambient marine aerosol measured at receptor sites (Rinaldi et al. 2013). Seawater biology also influences gas phase emissions, and the number of particles formed from some of these gas phase species by nucleation. SAM experiments showed that some VOC emissions increase during phytoplankton blooms. In parallel, several nucleation events were seen to be initiated from seawater emissions both in the mesocosms and in laboratory experiments. The experiments also identified iodine species that trigger the formation of new particle clusters after they reach a threshold concentration, and excluded the usually suspected precursor dimethyl sulfide (DMS) in the northwestern Mediterranean region investigated (Sellegri et al. 2016). Another result revealed that emitters of iodine species are not linked to Chl-a as expected, but to other biological tracers. Furthermore biologically driven emissions of amines appear to contribute strongly to early growth of the cluster in the 1-10 nm size range. These findings advance our understanding and our ability to model the complex climate feedback loop that involves temperature, biological populations in the seawater and marine aerosol emissions.

Box 1
ECCAD/ChArMEx regional Mediterranean emission database
The production of the ChArMEx state-of-the-art specific regional emission inventory for the preceding decade began in 2011 using the most recent literature, starting with yearly anthropogenic fossil and biofuel emissions in southern Europe from TNO (Kuenen et al. 2011; Liousse et al. 2014) completed by estimates for northern Africa (Assamoi & Liousse, 2010; Liousse et al. 2014). It was further completed with key emissions from agricultural and forest biomass burning (Turquety et al., 2014), soils for dust (Callot et al. 2000) and for NOx (Yienger & Levy 1995), aircraft (RIAHI et al. 2007), shipping, volcanoes (Andres & Kasgnoc, 1998), sea surface (Schwier et al. 2015), and vegetation (Guenther et al. 2006). The domain of interest extends from 10°N (tropical Africa) to 70°N (northern Europe) and from 20°W (Iceland) to 50° E (Caspian Sea) sometimes with a spatial resolution of 10 km. Fig. 2 summarizes the content of the database.

Figure 2
The different emission subsets and emission parameterizations available in the ECCAD/ChArMEx regional emission inventory (http://www.aeris-data.fr/​redirect/​eccad/​ChArMEx).

Emissions from vegetation

4Ecosystems are a notable source of a wide variety of reactive volatile organic compounds (VOCs), biogenic VOCs (BVOCs), such as isoprene, monoterpenes, methanol, and many others. BVOC emissions largely dominate anthropogenic VOC emissions at the global scale. BVOCs play a key role in atmospheric chemical processes particularly in the ozone cycle and in the formation of secondary organic aerosols (SOA). Due to high temperatures, high levels of solar radiation, and high biodiversity, to which BVOC emissions are very sensitive, the Mediterranean region is a major source of BVOCs (Owen et al. 2001), and has a significant impact on ozone and aerosol formation (Sartelet et al. 2012). Models of the correct meteorological and chemical conditions that are able to reproduce the diurnal variation in isoprene emissions (Guenther et al. 2012) are needed to understand the underlying atmospheric photochemistry.

5Large uncertainties persist in the composition, distribution and levels of BVOC emissions. Both experimental and modeling studies are thus crucial to improve our knowledge, especially in the context of climate change that is projected to severely impact the Mediterranean region. As part of the ChArMEx project, a thorough case study was performed at the OHP Oak Observatory (Haute Provence, France) in a coppice of downy oak, a widely represented tree species in the Mediterranean area. Measurements from the branch (Genard-Zielinski et al. 2015) to the canopy scale confirmed high isoprene emissions (up to almost 10 mg m-2 h-1), significant methanol emissions (up to 0.63 mg m-2 h-1), and negligible monoterpene emissions (Kalogridis et al. 2014). The isoprene degradation within the canopy was found to be very low (<3%) due to the low level of NOx and the low canopy height (Kalogridis et al. 2014). Measurements showed that isoprene emissions increased with an increase in radiation and air temperature, and latent heat flux was also shown to be a useful parameter to explain variations in isoprene emissions (Baghi, 2013). A study of the potential impacts of climate change on water resources (namely a 30% water deficit, as foreseen for the year 2100) suggests a significant increase in isoprene emissions in the future, irrespective of the warming scenario (Genard-Zielinski et al. submitted).

6Atmospheric chemistry is being modeled to simulate the fate of BVOC emissions on a typical hot sunny day (July 3) of the 2014 ChArMEx airborne experiment above the OHP. Preliminary results show that strong diurnal emissions of BVOC lead to a clear SOA formation event.

Aeolian erosion/soil dust emissions

7Dust emission results from the erosion of soil by wind (Bagnold, 1941; Gillette, 1981), which mainly occurs in the arid and semi-arid regions of the Earth, the Sahara desert and its fringes being considered as the main source region in the world. The resulting aerosols both scatter and absorb solar and Earth radiations, which affect the Earth’s radiative budget. When deposited on the Earth’s surface, mineral dust contributes to the input of growth-limiting macro and micro nutrients to oceanic surface waters (see the dedicated sub-chapter hereafter) and terrestrial ecosystems.

8In North African countries, the rapid increase in population has led to a growing demand for agricultural products. As a result, the pressure on natural resources is increasing steadily with the expansion of cultivated areas stimulated by the introduction of modern plowing techniques. Beginning in the 1960s, the disc plow pulled by powerful tractors has progressively replaced the traditional mold board plow pulled by draft animals, which affects dust emissions.

9As can be seen in Fig. 3, wind erosion is more than one order of magnitude greater on land tilled with a disc plow than in fields tilled with a mold board plow, land prepared with a tiller being between the two. These results strongly suggest that new tillage techniques such as using a disc plow drastically increase soil erosion by wind in agricultural fields with loose soils. They also confirm that traditional tillage tools like the mold board plows are the most suitable tillage tools to preserve soils in semi-arid agricultural regions. Finally, these results suggest that dust emissions from North African countries are increasing because of changes in land use management rather than because of the extension of the cultivated area. For example, Yoshioka et al. (2005) suggest that this type of land preparation could now be responsible for up to 25% of North-African dust emissions.

10A monitoring station was recently set up close to Medenine (southern Tunisia; http://193.95.22.108/​) for the long term monitoring of the surface atmospheric concentration and atmospheric column load of soil dust, and of the dust deposition flux.

Figure 3
Normalized horizontal flux (G) of sediments eroded by wind in fields prepared by three different types of plow used in Tunisia (adapted from Labiadh et al. 2013).

Source apportionment

Urban areas

11The coastal areas of the Mediterranean include megacities like Cairo (12 million inhabitants), Istanbul (12 M), Athens Great Area (5 M), and Barcelona (5 M). All these cities are subject to heavy gaseous and particulate pollution. The population of these regions will continue to increase, especially in the eastern part of the basin, leading to a higher anthropogenic pressure in a context of climate change.

12Satellite images of nitrogen dioxide columns from SCIAMACHY identified coastal urban areas in the Middle East as hot-spots of pollution in the region (Lelieveld et al. 2009). Global emission inventories all agree on a marked increase in anthropogenic emissions of major pollutants (NOx, VOC and PM2.5) in the Middle East Area (MEA), in contrast to what is observed in post-industrialized regions like Europe and the USA. In the coming decade, anthropogenic emissions from MEA are projected to be even larger than those from Europe and the USA (Salameh et al. submitted).

13Anthropogenic emission inventories provide key input data for the atmospheric models used for the prediction of air quality and for the study of the most efficient regulations to reduce air pollution. The quality of emission inventories from rapidly growing megacities in the southern and eastern Mediterranean is of concern because local emission data are sparse. A new road traffic emission inventory was built for Algiers, where road traffic is a major source of atmospheric pollution. It was validated by comparing high resolution (4 km) simulations with the regional atmospheric chemistry and transport model CHIMERE and observed air quality measurements of NOx and CO (Rahal et al. 2014). Some highly resolved inventories have also recently been developed at the regional scale for Beirut (Waked et al. 2012) and Istanbul (Markakis et al. 2012), but their uncertainties are unknown. As part of the Transemed initiative supported by the MISTRALS/ENVIMED program, a source-receptor methodology was developed for emission inventory evaluation. The approach consists in combining existing and newly collected observations and complementary source-receptor approaches (i.e., urban enhancement emission ratios, multivariate models like positive matrix factorization, PMF) in large urban areas like Beirut (Lebanon), Istanbul (Turkey), and more recently Athens (Greece) and in the near future, Cairo (Egypt). Very detailed databases of ambient and near-source observations are being built with a focus on the composition of gaseous organic carbon. The results recently obtained for Beirut (SALAMEH et al. 2014, 2015, 2016) showed (i) the extremely high levels of pollution for organics, (ii) the dominant effect of traffic emissions on concentrations of VOC, (iii) the poor spatial variability of speciated NMHC traffic emissions regardless of the region, and (iv) the high uncertainty and discrepancies between large scale emission inventories compared to observational constraints and local scale inventories (see also Abdallah et al. 2016).

Source apportionment at the regional scale

14Sources impacting the Mediterranean basin can be apportioned at regional scale using models that exhaustively account for all the processes (emission intensity, chemistry, air mass transport) that occur between the sources and the receptor zones. This approach has been applied for the summer (JJA) 2012 period at a resolution of 50 km using the chemistry-transport model CHIMERE dynamically forced by the model WRF (Rea et al. 2015). The contributions from different sources of particles and gaseous precursors to both the surface particulate air quality (PM2.5 and PM10) and the aerosol optical depth (a proxy of the aerosol concentration in the vertical column) were determined by eliminating particle sources one by one (anthropogenic, fire, soil dust, vegetation, sea), and comparing the results with the reference simulation with all the sources activated. Results showed that desert dust had the most influence on surface PM concentrations in the Mediterranean basin (up to 86% of PM10) followed by anthropogenic aerosols (up to 75% of PM2.5) in Western Europe. Sea salts also had a significant influence (up to 29% of PM10) in Atlantic and Mediterranean coastal regions.

15Another approach to apportioning sources at regional scale uses receptor oriented methods focusing on the chemical composition of pollutants measured at a representative receptor site. Two intensive observation campaigns were conducted as part of the ChArMEx research program in Corsica (2013) and in Cyprus (2015). Statistical analysis of the chemical composition of both gaseous and particulate pollutants measured at Cap Corsica combined with the residence time analysis of air mass trajectories pointed to the contribution of anthropogenic sources located in regions characterized by intense anthropogenic activities (e.g. the Po valley and south-eastern France, both located in the north-western Mediterranean basin). In addition to primary (i.e. directly emitted) volatile organic compounds (VOCs) emitted by biogenic sources (BVOCs), a group of secondary pollutants composed of first-generation oxidation products of BVOCs was also identified, while another group was characterized by more oxidized VOCs (OVOCs) of both biogenic and anthropogenic origin. The combined analysis of VOC and aerosol compositions in PM1 showed that during periods under a dominant biogenic influence, aerosol composition was dominated by the secondary organic fraction, whereas during periods of long range transport of anthropogenic emissions, the relative contributions of inorganic and organic fractions were the same. These results underline the importance of considering the roles of both anthropogenic and natural emissions in particulate pollution. The same approach is being applied using the observations acquired in Cyprus as representative of the eastern part of the Mediterranean basin.

Table des illustrations

Légende Figure 1Variety of continental sources impacting the Mediterranean basin.
URL http://books.openedition.org/irdeditions/docannexe/image/23370/img-1.jpg
Fichier image/jpeg, 341k
Légende Figure 2The different emission subsets and emission parameterizations available in the ECCAD/ChArMEx regional emission inventory (http://www.aeris-data.fr/​redirect/​eccad/​ChArMEx).
URL http://books.openedition.org/irdeditions/docannexe/image/23370/img-2.jpg
Fichier image/jpeg, 333k
Légende Figure 3Normalized horizontal flux (G) of sediments eroded by wind in fields prepared by three different types of plow used in Tunisia (adapted from Labiadh et al. 2013).
URL http://books.openedition.org/irdeditions/docannexe/image/23370/img-3.jpg
Fichier image/jpeg, 94k

Le texte et les autres éléments (illustrations, fichiers annexes importés) sont sous Licence OpenEdition Books, sauf mention contraire.

Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search