The Impact of Desert Dust Across the Mediterranean

Saharan dust generation, transport and deposition over the Mediterranean depends on the prevailing atmospheric circulations and climatic conditions over the region. Different sets of models can be used to simulate these circulations and their possible response to changes in climatic forcings (e.g. greenhouse gas concentration): General Circulation Models (GCMs) and Regional Climate Models (RegCMs). GCMs provide numerical representations of the global atmosphere and are the primary tools today available for climate simulation. However, for an area such as the Mediterranean Basin, characterized by complex topographical and coastline features, their resolution (typically a few hundred km) is too coarse to capture the details of regional circulations. A finer model resolution (a few tens of km) can be obtained using RegCMs driven by initial and lateral boundary conditions either from analyses of observations or from GCM output. Alternatively, climate simulation on finer scales can be achieved through the use of variable resolution global models. This paper presents a brief review of the performance of GCMs over the Mediterranean Basin and of the application of RegCMs and variable resolution models to study climatic conditions over the region. The implications concerning the use of GCMs and RegCMs to study the effects of African dust across the Mediterranean are discussed.

The atmospheric dust life causes numerous impacts to the Mediterranean environment and climatology. Significant amount of dust from the Sahara is transporting toward remote regions causing deposition of the mineral material into the Mediterranean Sea. There are different factors that determine the nature of atmospheric dust processes: state and type of soil, surface atmospheric conditions, mountain effects, turbulent and large-scale atmospheric motions. The complexity and mutual influence of these mechanisms to the atmospheric dust require application of sophisticated models, such as Euler-type atmospheric models with incorporated dust concentration equations. In this kind of models, specific parametrization of dust production terms is necessary. In this article, an overview of several atmospheric dust models is presented. Particular emphasis is given to the problem of dust production modelling, parametrization of dust removal processes and modelling of dust radiation effects. Finally, problem of assimilation of available dust measurements in the atmospheric models is considered, in the light of possible routine dust prediction.

We present here an overview of our quantitative studies of African dust transport to the Mediterranean, based on Meteosat and CZCS spaceborne images of the reflected sunlight, and coincident exogeneous meteorological and chemical atmospheric or marine data. Investigations include the retrieval of the desert dust aerosol optical thickness and mass column density over seawater, of the dust deposition flux and its fate in surface seawater. The comparison between measured and Meteosat-derived dust deposition fluxes indicates that cloud cover limits short-term Meteosat survey, but that long-term averaged dust fluxes to seawater can be approximated from Meteosat data. Results from our analysis of multi-year series of Meteosat ISCCP-B2 data indicate a marked seasonal cycle of dust transport in the Mediterranean, with a high dust load in spring and summer, and a strong year to year variability. A case study of the apparent velocity and altitude of dust transport shows a primary transport of dust particles above the atmospheric boundary layer. Finally, a combined survey of desert aerosol and phytoplankton in summertime seems to provide a first observation of the impact of strong dust inputs on the phytoplankton growth.

Our off-line 3D transport model of mineral dust has been used to study a dust episode of considerable extent over the Mediterranean. As shown also by satellite observations a distinct dust plume developed over the Mediterranean Sea during the first days of July 1988. The well chosen localisation of the source region in our model — in combination with the close match of ECMWF analysed wind fields to actual synoptic conditions was found to be the reason for a remarkable reproduction of the dust plume evolution over the Mediterranean. By modification of the initialisation in the model and the omittance of selected model components we show, that together, convection and the length of the period simulated improve the correlation with satellite observations. The aerosol characteristics of the dust plume over the Mediterranean as predicted by the model are presented.

Complex model simulations of airborne Saharan dust transport were carried out in the Meteorological Synthesizing Centre — East (MSC-E) of Co-operative Programme for Monitoring and Evaluation of the Long Range Transmission of Air Pollutants in Europe (EMEP). The model presented covers the process as a whole from the moment of dust elevation caused by wind in Northern and Central parts of the Sahara to its deposition onto the Mediterranean for long time periods and with actual meteorological data. As the transport model unit the MSC-E model was used. A short description of the model is presented. One of the most complicated points in the problem is an estimation of the dust emission caused by the storms in the desert. Emission calculations are based on a published approach originally developed for the evaluation of soil erosion in the USA (Gillete et al., 1988).Preliminary model estimations of deposition and concentration of dust transported from the Sahara to the Mediterranean basin and Southern Europe for 1985–1988 and 1994 are considered. The model parametrization for the calculations was developed using published experimental data. The accuracy of the results is discussed.

We have developed an aerosol model that describes the aerosol size distribution and computes number concentrations (Schulz et al., 1996). This model is embedded in a global off-line tracer transport model described by Heimann (1995). The model is tested by a point by point comparison to satellite-derived optical depths. Here we first consider the importance of well resolving the source by reproducing both the timing and the position of dust mobilization. The improvements brought about by the formulation of a source (Marticorena and Bergametti, 1995) that depends both, on the size of the loose erodible particles and on local roughness heights are shown. We also point to the influence of the size distribution which is a major factor that affects our comparison of model and satellite observations.

Analysis of the composition of aerosol particles in the Eastern Mediterranean shows that dust particles are often coated with soluble sulphate. The role of these types of particles in the formation of clouds and rain is discussed. Numerical cloud model simulations suggest that the presence of such large particles in the atmosphere could substantially increase rain amounts from convective clouds.

In the present paper an assessment of the atmospheric nutrient (nitrate, nitrite, ammonium, phosphate and silicate) and trace metal concentrations have been made by computing isentropic trajectories evolving over the Mediterranean Sea during a 20 day period of February 1989. The mass balance equation has been integrated along the trajectories in order to obtain the transport of substances within an atmospheric boundary layer along the path in sections over the Mediterranean Sea. The concentrations measured at the Blanes site have been extrapolated to the entire Western Mediterranean Sea in order to compute background concentrations at the entire region. Following this, deposition has been computed by means of trajectories passing over the points for which the background concentrations have been computed. This simple deposition model provides reasonable values of the vertical fluxes, until a more accurate data base is available. The computations carried out for the period studied show a greater deposition of all soluble substances when the air masses have a southern origin (African, Eastern Mediterranean) than when they originate in the north (N Europe and NW Europe). However, in spite of the longer duration of the NW-NE events over those originating in the SE-SW sector, the North-African air masses contribute about 2/3 of the overall deposition into the Mediterranean Sea.

Sand and dust transport over Mediterranean basin is usually associated to desert depressions. However, other meteorological systems can originate these transport processes. In the present work an event is analysed, connected to a mid-latitude disturbance deeply penetrated into western Africa and associated to a cold outbreaks in North Atlantic Ocean. Finally, a climatological analysis shows a reduction of strong Sirocco occurrences in the Mediterranean basin in last 40 years, reduction probably due to an increase of the anticyclonic activity in the western and central sectors of basin.

Results are given of a detailed study to determine the spatial and temporal (seasonal) distributions of the planetary boundary layer depths (PBLD) over the Mediterranean Basin. More then 65,000 upper-air measurements from 45 rawinsonde stations in the Basin were compiled and analysed for two consecutive years from spring 1986 through winter 1988. A methodology using both tested and newly developed criteria was applied to determine PBLD values. As expected, the values prove to be generally higher over land and are probably minimum over the eastern and western ends of the Mediterranean Sea. Factors influencing this spatial distribution are mainly topography, the distance from the shoreline, and to a lesser extent, synoptic weather systems. The most striking temporal effect on PBLD distribution over the Mediterranean Basin is caused by larger scale synoptic weather systems. The second effect is the diurnal cycle, which has its largest impact mainly during the summer months.The PBLD values are displayed as analysed seasonal maps, from which typical values are extracted and presented in a tabular form. The results can be used as input values for dynamic meteorological models in order to better assess the transport and dispersion over the Mediterranean Basin.

This manuscript assesses whether biomass burning in North Africa could affect the heat budget of the Mediterranean region. The principal biomass burning activities in North Africa are forest and maquis fires, burning of agricultural wastes, and the domestic use of firewood. In the Mediterranean zones of the five North African nations, forest fires emit 1.33 Gg yr−1 of black carbon to the atmosphere; burning of agricultural wastes, 7.27 Gg yr−1, and burning of fuelwood, 5.16 Gg ye−1. The working model calculates maximum atmospheric black carbon emissions during August, September, and October as a result of forest fires and burning of agricultural residues. The atmospheric loading of North African black carbon from the estimated burning activities ranges between 0.01 and 0.05 µg BC m−3. This estimated black carbon loading is low compared to observations of the western Mediterranean aerosol, which suggests that the European continent, rather than North Africa, is the likely source region for black carbon in the Mediterranean aerosol. While North African biomass burning may be insignificant on a global scale, the emitted particulates could potentially affect the Mediterranean climate. From estimates of Saharan dust, black carbon, and biomass burning aerosols transported to the north over the Mediterranean Sea, I have calculated the relative optical depths of these three categories of particulates. The optical depth of summed black carbon and biomass burning aerosols is only seven percent that of Saharan dust. Thus, North African biomass burning particulates cause a small impact on the Mediterranean climate and heat budget compared to Saharan dust particles.

Ice core studies indicate that rates of deposition of mineral dust at the surface of polar ice sheets were much larger during the last ice age than at present (Thompson, 1977, Hammer et al., 1985, Legrand et al., 1988). Attempts to reproduce this observation using models of the atmospheric dust cycle imbedded within general circulation models of the atmosphere ((A)GCM’s) have been unsuccessful so far (Genthon, 1992,;Joussaume, 1993). The origin of this systematic failure can be traced in a poor definition of the sources, an insufficient quality of the atmospheric cycle simulated by the GCM’s, an inadequate modeling of aerosol processes or (likely) a combination of these.We here present a new source formulation for modeling the atmospheric dust cycle and particularly the transportation of dust to the Arctic. Due to the scarce information on soil properties and vegetation for previous times we chose a formulation that uses a prescribed vegetation cover with physical restrictions inferred from the simulated climate.With this approach some features of the present dust cycle are reproduced reasonably well, but although dust production from the major deserts is somewhat enhanced for the ice age simulation it, too, fails to simulate the ice age increase of the dust in the Arctic.

North Africa serves as a huge reservoir of mineral dust. Dust sources are active all year, especially during the summer when they feed strong pulses of dust across the Mediterranean to Europe and the Middle East and across the Atlantic to the Caribbean, Central and North America. Over the North Atlantic in the low and mid latitudes, mineral dust is the major non-sea-salt aerosol component, exceeding the combined mass of all other species. Model estimates yield a dust deposition of 170 Tg yr−1 to the Atlantic, 25 Tg yr−1 to the Mediterranean, and 5 Tg yr−1 to the Caribbean. There is an extremely large seasonal variability in dust deposition and we might expect a comparable variability in consequent ocean effects. However, deposition estimates are highly uncertain because of the paucity of deposition data throughout the region. The greatest uncertainty is associated with the estimation of dry deposition fluxes which in ocean regions close to sources could constitute by far the largest fraction of the total flux. In recent time, dust mobilization may have increased markedly due to human activities; thus a substantial fraction may be regarded as an anthropogenic product.

Saharan dust inputs have a very important influence on the mineralogical composition of particulate matter in the Mediterranean atmosphere. We have characterised Saharan dusts from a mineralogical point of view with regard to their final destination in deep-sea sediments. This work is an attempt to compare mineralogical data of dusts and sediments available from literature together with our mineralogical data of dust collected at Sardinia. The mineralogical source marker for the Saharan dusts is palygorskite in the Western Mediterranean and smectite in the Eastern Mediterranean. Therefore, in Sardinia attention has been focused especially on a morphological study of palygorskite aggregates from where differences in Saharan aerosol collected in both dry and wet deposition were observed. A new question has also arisen about the alteration of calcite grains in Saharan dust collected in Sardinia.

This paper presents an up-to-date survey of the mineralogical and chemical properties and the behaviour of aeolian desert dust over Israel and the Eastern Mediterranean area.Dust storms of desert origin are classified according to their principal sources by their clay mineralogy and by their meteorology and air mass trajectories. The total dustfall in Israel is composed of a uniform local background on which are superimposed pulses due to desert dust storms. A revised estimate of the flux of settling dust due to these dust storms is given as 30–60 tonnes km−2 y−1 . The total dustfall has been divided into a number of constituent fractions which enable characterization of its various sources. The number of dust events appears to be increasing over the years with potentially important effects on climate and human health. During dusty days, the concentration of Total Suspended Particulates in the air ranges from 0.2 to 20 mg m−3.

The Saharan desert is a vast source of soil derived dust, which is transported into the surrounding regions. In the areas of deposition the dust influences the sediment, element and nutrient balances of marine and terrestrial ecosystems. This paper deals with the identification and quantification of dust additions in soils and the influence on the development and properties of soils. Three case studies are presented. The research sites are situated in SW-Niger, Lanzarote (Canary Islands, Spain) and S-Portugal. They differ concerning climate, parent material and knowledge about the local dust characteristics. Therefore different approaches have been used in order to quantify the dust additions in soils. The approaches comprise physical, mineralogical, geochemical and morphological investigations. The determined amounts represent between 4 and 66% of the recent soil material. The accretion depends mainly on the distance from the sources, climate and stability of the land surface (age of the soils). In some soils local and far transported material as well as fluvial additions are found. The impact on soil development depends on the amount and characteristics of the deposited dust in comparison to the autochthonous soil material and the stage of soil development. In general the effects decrease with distance from the dust source areas.

This paper describes similarities and differences of Australian and African dust transport systems and dust properties. On both continents, wind erosion is a major geomorphic process. Dust is entrained from arid source regions with similar characteristics and transported large distances via major pathways. The Sahara is, however, a much more prolific source of dust than is Australia, which is reflected in the measured dust concentrations. African and Australian dusts have a similar clay mineralogy, with a somewhat higher smectite percentage in African dusts. While dust collected close to source shows similar particle-size distributions on both continents, African far-travelled dust is generally finer and particle-size decreases more rapidly with transport distance than Australian dust, which is more aggregated and rarely shows a particle-size mode under 10 µm.

An 11 years survey (1984–1994) of the Saharan dust input was performed in Corsica, on an event basis. Dry deposition was collected separately manually since 1985. Wet deposition of dust particles is independent of the amount of rainfall. Dry deposition seems to be of minor importance in Corsica, which could be related to the high altitude of dust routes above the Mediterranean, the efficiency of dust particles acting as condensation nuclei, among other parametres. Dust fallout events are very brief (usually less than 3 days long) and irregular in intensity and temporal pattern. A seasonal pattern appears in the frequency and magnitude of events (spring and autumn maxima) but with a very high variability (monthly input σ > 100% in spring). The high magnitude events drive the variability of the dust fallout at the annual and interannual scale. The annual input of Saharan dust varies between 4.0 and 26.2 g m−2, depending on the occurrence of high magnitude events. The annual average value for the period studied is 12.5 g m−2. The annual input seems to decrease in the 1990’s. If we consider that the Saharan dust input to Corsica follows the same patterns as those recorded in the Alpine glaciers, we can assess that the maximum input during the last 40 years occured in the 1980’s.

The dust contained in red rains (rainwater with a reddish silty content) has been collected at La Castanya Biological Station in the Montseny mountains (NE Spain) over 11-years. Dust deposition was highly variable from year to year. In fact, two events accounted together for 63% of the total deposition of dust in the 11-yr record. The mean dust deposition was 5.1 g m−2 yr−1. Red rain events were associated with two main meteorological synoptic situations determining the air-mass trajectories to come from Morocco-Algeria and Tunisia-Libya. Dust of Libya-Tunisian provenance was richer in smectite and kaolinite, and poorer in quartz than dust from Morocco-Algeria. Palygorskite, an African tracer, was present in all dust samples, while chlorite, indication of local peninsular sources, was absent in all samples.

In order to characterize the Saharan aerosol end-member in the Western Mediterranean, soils samples selected as potential sources of the Saharan aerosol were collected in South Algeria. Major elements (Al, Ca, Fe, K, Mg, Mn, Na, P, Si, Ti) and trace metals (Cd, Cu, Co, Ni, Pb, Zn) were determined in fine (< 50 μm) and coarse (50–170 μm) soil sub-samples. Their matrix composition can be essentially described as a mixture of (1) a large proportion of particles with a composition of 90% SiO₂ + 10% Al₂O₃ (Fe-poor alumino-silicates with quartz); (2) a small proportion of particles with a composition 35% Al₂O₃ + 65% Fe₂O₃ (probably individual oxide particles and/or surface coatings on quartz).The overall trace metal concentrations in the fine sub-samples was rather homogeneous over the studied area, indicating a large scale homogenization. The trace metals are only associated to one half of the Al; this may biased the use of the classical Al-based enrichment factor methodology for estimating small degrees of contamination of Saharan aerosols. On the contrary, the whole trace metal content is associated to Fe, but a ‘pure’ iron phase seems also present.The fate of Al and Fe associated to the Saharan aerosol was studied via a leaching experiment. Their dissolution in pure water is extremely low (often much less than 1%), in sharp contrast to experiments undertaken with polluted aerosols and inversely related to the particulate load of the medium; this may thus provide a mechanism potentially able to regulate the dissolved Fe flux to the sea due to irregular Saharan fallout pulses.There is however some evidence that Fe dissolution properties do not remain constant when Saharan aerosols are transported far away from their source. The large increase of Fe dissolution observed with a Saharan aerosol collected at the Barbados raises the question of a relative enrichment of the aerosol with fine-gained particles presenting a larger surface area, thus more sensitive to dissolution processes, but a concomitant mixing with polluted particles during this long range transport may also contribute to this result.

At two remote coastal sites in Sardinia, the two major soluble and insoluble components of aerosols and rainwater collected were seasalt (avg. rain = 61 µg g−l; aerosol = 19 µg m−3) and minerals (avg. rain = 13 µg g−1; aerosol = 9 sg m−3). While seasalt content was more or less similar, a three to seven times higher mineral concentration was found in Saharan (SAH) as compared to background (BKG) samples. Mean wet mass flux in Sardinia for the period 1990–1995 was 3 g m−2 yr−1, and wet-to-dry deposition ratio was highly variable (0.8–2.2). Particulate scavenging ratios (SR) were calculated separately for BKG and SAH samples giving values of 600 and 1800, respectively.

Variable amounts of Saharan dust deposit on the glaciers of the Alps, forming thin layers or mixing with local pre-existent sediments. To investigate the mineralogical composition and the source of the deposits, samples of Saharan dust, fine-grained tills and cryoconite accumulations were collected on the surface of the glaciers and/or in ice-cores. The fractionation of the samples and the bulk mineralogy allow a sharp differentiation among the different deposits. Only the clay mineralogy, however, points out the fundamental characteristics of the finest fractions of the deposited materials. Typical clay mineral assemblages were recognised in the Alpine glacial deposits, giving evidence of the different provenance of the materials. Reddish-yellow Saharan dust exhibits prevailing amounts of kaolinite and poorly-crystallised illite, with lower quantities of palygorskite and chlorite. Olive-greyish tills, connected to the local morainic deposits, are mainly composed by high percentages of well-crystallised mica associated with lesser chlorite and serpentine. Blackish cryoconite accumulations are mixture, in variable proportions, of both Saharan dust and local tills. The mineralogical data clearly differentiate distal and proximal sources of mineral particles deposited on Alpine glaciers.

Grain size analysis of aerosol particles collected by air filtration is difficult because of the low concentration and therefore the representativeness of sub-sampling. We have analysed a set of aerosol samples (European background), together with precipitations (Saharan and European background), using a Coulter Counter TA Il and a Galai Cis 1 laser system (magnetic cell mode and slide mode) to verify the reliability of those two methodologies in the particle size range 0.5–102 μm. No agreement has been found between Coulter and Galai results in European background aerosol and precipitation samples. Conversely we have found a good agreement between Coulter and Galai results in Saharan samples. The Saharan dust precipitations were not affected by the low concentration problem. The Galai magnetic cell mode is more efficient for Saharan dusts, on the other hand the slide mode is a good method to analyse background aerosols.

In July 1994, we were able to collect airborne fungal spores and pollen grains over the Adriatic Sea from the upper deck of the oceanographic ship Urania (CNR). The biological particles were collected using a modified Lanzoni VPPS 1000 sampler. Not only were the airborne concentrations of different organisms estimated, their viability was also tested with a 1% TTC solution. Results show that pollen viability and concentrations were significantly correlated with distance. Estimated time of residence in the air was never found to be significant. Relative Humidity had the strongest potential to model pollen or spores concentrations over sea. Southern winds presented the lowest pollen viability and concentrations and the highest ratio of small spores over larger ones, such as a Basidiospores/Alternaria ratio above 0.01.

Under certain, favourable conditions, a considerable amount of long-distance transported biological material can be deposited far from its place of origin. In Fennoscandia such situations occur when there is a strong southerly air circulation from the Mediterranean and/or North-Africa. Soil particles and pollen grains associated with the top soil or the present vegetation, can be lifted up by strong surface winds and brought up to higher levels, where further transportation takes place. During the last few years several of these events have been reported.The pollen analysis of coloured and uncoloured precipitation provides information about exotic pollen rain and its extent. By combining three-dimensional trajectory calculations with the actual phenological data the place of origin for the analysed material can be determined. In most cases the origin of the material has been determined to be the northern part of Sahara.

The particulate aerosol over the Mediterranean Sea consists largely of a ‘background’ of European-derived anthropogenic-rich material which has a trace metal composition similar to that over other European coastal seas, but which is subject to sporadic inputs of crustal-rich material from Saharan and Middle East desert sources which are often transported in the form of dust ‘pulses’. The input of Saharan dust results in increases in the total concentrations of crust-controlled trace metals (NEEs), such as Al and Fe, and to decreases in the values of AEEs, such as Cu, Zn and Pb, in the aerosol. Variations in the concentrations of trace metals at any one site in the Mediterranean Sea are largely controlled by: (i) the magnitude of the inputs of crust-rich dusts which are mixed with the anthropogenic-rich ‘background’ material, and (ii) precipitation scavenging, which removes both anthropogenic and crust-rich material from the air.The effects that the inputs of desert-derived crustal dusts have on the Mediterranean Sea aerosol are transmitted to the marine biogeochemical cycles via the deposition of the aerosol to the sea surface. The magnitude of the air to sea fluxes are dependent on the concentrations of the trace metals in the air, but the fates of the metals are dependent on their solid state speciation in the aerosols and the mode by which they are deposited to the sea surface. The solid state speciation signatures of some trace metals are different in the anthropogenic and crust-dominated aerosol ‘end-members’. In particular, larger fractions of the total concentrations of Cu, Zn and Pb are present in potentially mobile, i.e. sea water and rain water soluble, associations in anthropogenic-dominated aerosols. Following ‘dry’ deposition to the mixed layer the trace metals Cu, Zn and Pb are therefore considerably less soluble from crust-dominated than from anthropogenic-dominated aerosols. In the ‘wet’ deposition mode, the scavenging of desert-derived dusts affects both the pH and the trace metal composition of Mediterranean rainwaters. In general, the progressive scavenging of desert-derived dusts results in; (i) an increase in the pH of the rain water solutions, (ii) an increase in the total concentrations of crust-derived metals, such as AI, and a decrease in the EF_crust values of anthropogenic-derived metals, such as Pb and Zn, and (iii) a decrease in the extent to which both crustal and anthropogenic trace metals are soluble in the rain waters from the parent aerosols as a result of the increasing pH of the solutions.

Transport of dust of African origin over Northern Italy and the Alps is a well known phenomenon rather frequent in spring and summer months. In March 1991 at Stelvio National Park, during a field campaign aimed at assessing the concentration and composition of aerosol particles in the Alpine region, two episodes of African dust transport were pointed out.In these occasions the composition of particulate matter changed drastically with a significant increase of the concentration of typical crustal elements such as Al, Si, K, Ca, Ti, Fe, Mn, while the concentrations of elements such as S, Zn, Pb, Br, typically associated with industrial and vehicular traffic sources both of local and regional origin, remain constant or decrease.Titanium results a good tracer-element, the Ti/Ca, Ti/Fe and Al/Si concentration ratios are also good signatures of these events. Backward air mass trajectories, calculated on the basis of wind field provided by the European Centre for Medium Range Weather Forecast and of “3D Trajectory” computer package, indicate a direct transport from Africa. The analysis of the meteorological situation indicates that the thermodynamic instability associated with a cold front passing over the Sahara, coupled with upper strong southern winds is a decisive situation favourable to transport. The complete elemental characterization of a time series of particulate samplings supplemented by air mass trajectory reconstruction proved to be an effective method for the identification of dust transport phenomena.

A set of trace metals (Si, Al, Fe, Mn, Sr, Ba, Rb, Co, Ni, V, Zn, Cd, Cu and Pb) was analysed in rainwater from Southern Sardinia. Crust-Poor Rain (CPR) and Crust-Rich Rain (CRR) groups of events were distinguished. The CPR EF_crust value of non-marine strontium (cSr), chosen as the new “cut-off” value, allowed to discriminate crustal metals (Al, Fe, Si, Rb, Mn, Co and Ba) from anthropogenic metals (Cu, Pb, Cd and Zn). The contents of crustal metals in rainwater increase in proportion to their solubility as a consequence of mineral dissolution, while Saharan dust appears to act as a “sink” for some anthropogenic metals in rainwater.

Large amounts of mineral dust are regularly deposited over the Alps in the course of episodic Saharan air mass advections. To establish the long term chronology of this phenomenon, firn cores and one deep ice core recovered from a non-temperated glacier in the Monte Rosa summit range (Colle Gnifetti, 4500m a.s.l.) have been continuously analysed for their chemical (major ions) and isotopic (δ18O, δD) stratigraphy. The present Saharan dust impact on the ion chemistry of alpine winter snow packs is also documented by a multi-year snow pit study of high elevation snow fields in the Monte Rosa region. Saharan dust bearing snow layers are specifically marked by extremely high CaCO₃ contents as well as by increased δ18o and deuterium excess values. They contribute substantially to the overall cation and SO₄2− inventories of the seasonal snow pack, and, thus, decrease its present acid load by roughly 60 %. The Ca2+ ice core record dating back to 1600 A.D., approximately, is dominated by Saharan dust related spikes, explaining close to 70 % of the total Ca2+ inventory of the glacier, and much of the large interdecadal variability seen over the whole period. The Ca2+ record exhibits a yet unexplained increase of the background level by a factor of two, approximately, confined to the 20th century. The SO₄2− to Ca2+ ratio observed in Saharan dust layers deposited in the pre-industrial era is only slightly depleted with respect to the modem period indicating that, on the average, only 14 % of the Saharan dust related SO₄2− originates from anthropogenic sources.

To understand the relationship between dust deposition and geochemically-linked biological processes, it is necessary to characterise the magnitude and location of dust deposition to the surface of the ocean. A simple model is presented to estimate the mineral dust deposition to the surface waters of the Atlantic Ocean. In the model, we assume that the dissolved aluminium (Al) in the surface waters is a steady state feature in which Al supply from the partial dissolution of incoming mineral dust balances its scavenging removal by biological processes. The model predictions are in good agreement with a wide range of measured mineral dust inputs at diverse locations in the Atlantic Ocean. The success of the model suggests that surface ocean Al distributions can be used to estimate the mineral deposition in remote regions of the ocean where logistics preclude direct measurements.

This paper concerns simultaneous measurements of ozone and carbon dioxide concentrations in troposphere during Saharan transport in order to highlight the effects of desert dust on gas values. Two episodes selected among those occurred at Mt. Cimone (44° 11’ N; 10° 42’ E, 2165 m asl) from March 1991 to March 1995 are presented. During these episodes a sharp depletion of ozone concentration and a change in the behaviour of carbon dioxide concentration are evident. Three hypotheses to explain these modifications registered during Saharan dust transports are proposed.

In the southeastern coast of Spain bulk deposition was collected, on a daily basis during multiple-day episodes, in order to relate precipitation composition to air mass origins. Samples were analyzed for major ions, in addition to conductivity and alkalinity measurements. The majority of rain had a neutral or alkaline character as a result of neutralization, caused mainly by calcareous soil dust. The occurrence of acidic events or red dust events simultaneously at most stations suggested the influence of transboundary sources. Air mass-trajectories were studied for each of the main events and most of the episodes fell into three main situations: Originating in the Atlantic Ocean, Saharan and Polar continental areas. Chemical rain composition was found to depend on the source area. Meteorological situations with air masses crossing parts of Europe were associated with acidic events. Events coming from the Ocean had a neutral pH and lower concentrations for all species, while those coming from the Sahara have the highest pH values and the highest concentrations for all ions. The relative importance of each source area in southeasterly Mediterranean precipitation was in terms of volume: 62% from Polar Continental, 18% from Saharan, 20% from Oceanic origin. Despite the relative importance of polar continental rainfalls, their acidity is relatively low. The importance of Saharan alkaline rains leads to a net alkaline precipitation in the area.

The chemical characteristics and the grain-size patterns of the mineral fraction in Sardinian aerosols and precipitation were highlighted by means of Q-Mode Factor Analysis (Q-FA) on major elements (Al, Ca, Fe, Mg, Na, K, Si), and trace metal data (Cd, Cr, Pb, Zn). Three factors explained more than 95% of the variability: the first two representing the background aerosols (Mediterranean) and local inputs from different source regions (Sardinian) respectively, and the third one being the Saharan factor. Grain-size modes of background and local aerosols were 1 and 2 μm respectively, whilst Saharan dust mode was in the range 2–8 μm. Mediterranean background aerosols in Sardinia showed a 3 to 25-fold decrease in Ef_crust values for Cd, Zn, Pb and Cr compared with coastal stations in the West and in Corsica. Local inputs resulted mainly enriched in Pb and Zn, which could represent Sardinian sources to the Central Mediterranean Sea.

In 1993 three strong dust episodes occurred in the Eastern Mediterranean Basin during the rainy season September and October. During dust episodes, and at normal conditions in order to compare dust and non-dust period, daily PM-10 aerosol samples and weekly six stage size separated aerosol samples were collected at Antalya, an eastern Mediterranean location. Both PM-10 and impactor samples were analysed for 40 elements by instrumental neutron activation analysis. Wind trajectories were examined and the dust episodes classified according to their patterns into two types: (a) Saharan Dust episodes, and (b) Middle East Dust episodes. Among the three studied dust episodes, similar chemical composition were found for Saharan dust and Middle Eastern dust episodes. Although similar absolute concentrations of anthropogenic elements were observed during dust and non-dust periods, for crustal elements like Al, Sc, Fe, and rare earth elements concentrations during strong dust periods up to a factor of four higher concentrations were measured. According to the enrichment factor results three broad groups were identified; group I: no size dependency, no enrichment: Sc-like elements (rare-earth’s, Mn, Fe, Ti, Sc), group II EFc’s increase with particle size (Na, Cl, Mg), and group III EFc’s decrease with particle size (As, Se, Sb, Cd, Zn).

Concentrations of dissolved nutrients (PO₄, NO₃, NH₄ and Si(OH)₄) and pH have been monitored in 84 rain water samples collected during 3 winters at the northern Mediterranean Israeli coast (Haifa). The main dissolved inorganic nitrogen (DIN) compound was nitrate, with concentrations between 4.2 and 300 μM. Ammonium was dominant only in 12 out of 60 events, and its concentrations ranged between 2.2 and 96.9 μM. Phosphate showed a range of concentrations between 0.09 and 8.8 μM. High silicic acid concentrations (>2 μM) were detected only during relatively basic rains (pH>6), probably due to greater dissolution of quartz and clay particles into rain water under more basic conditions. Although about 30% of the rains are considered to be acid rains, their main H+ donor is non-seaspray sulfate and not nitrate. Nitrate and phosphate showed similar temporal variability, suggesting their common origin for most of the rain events.N/P ratios in rainwater (median of 80:1) were higher than in Levantine Waters (~27:1), while those of the river outlets along the Israeli Mediterranean coast were found to be lower, suggesting that atmospheric derived nutrients may be important in determining the abnormally high ratios in the eastern Mediterranean.The measured N flux in rainfall and the calculated total (wet + dry) N flux to the SE Mediterranean are 0.3 and 0.377 g N m−2 yr−1, respectively. These values are somewhat lower than fluxes determined previously in the western Mediterranean, probably due to reduced anthropogenic inputs from N. Europe.Total suspended particles (TSP) concentrations (referred to as dust) over the SE Mediterranean Sea were about half the concentrations recorded at the central and northern Israeli Mediterranean coast. The estimated aerosol fluxes into the coastal and the SE Mediterranean Sea were 72 and 36 g m−2 yr−1, respectively. Based on the latter, the calculated dry P flux to the SE Mediterranean Sea is about 0.0139 g P m−2 yr−1.

The results of shipboard measurements of sulphur- and nitrogen- containing compounds and trace metals in the atmosphere over the Mediterranean, conducted by the UkrSCES in 1987–91 during the fulfilment of national and international oceanographic programmes, are presented. We discuss the spatial and temporal distribution of these species and estimates of sulphur and nitrogen and trace metals deposition to the surface of the Mediterranean. The results indicate that more than 79–98% of the ions under study were of non-marine source but brought to the marine atmosphere by long-range transport from the continental-based anthropogenic sources. We estimated that during a year 770–1350, 260–390 and 450–990 thousand tons of sulphates, nitrates and ammonium, respectively, and 150–300, 4600–9200 and 20000–41000 tons of cadmium, lead and zinc, respectively, are deposited at the surface of the Mediterranean Sea. The marine and crustal enrichment factors for trace metals in Mediterranean aerosol are calculated. It can be concluded that all metals are non-marine in origin, being in part of continental nature.

Rainwater is often acidic (pH in the range 4–4.5 in many cases) in coastal, inland and mountain zones of Tuscany (central Italy). Neutralized rainfalls, i. e. exhibiting pH > 5.6, usually occur in spring and autumn; sometimes Saharan dust is visibly evident, in which cases a pH > 7 is generally found. When the contribution from Saharan dust is not evident, the neutralization (pH 5.6–7) appears to be due to local atmospheric dust or, in samples collected near to the shoreline, to marine aerosol. Collection of single rainfall, or fractions of single events, is recommended to distinguish among the various contributions to the neutralization.

This work reports preliminary composition data of marine aerosol collected under selected meteorological conditions in Leghorn (Tyrrhenian Sea -Tuscany -Italy) with two different porosity filters (1.0 μm and 8.0 μm). The samples are analyzed to identify inorganic anions (Cl−, Br−, NO₃−, H₂PO₄−, SO₄2−, F−), cations (Na+, NH₄+, K+, Mg2+, Ca2+) and some organic anions (acetate, propionate, formate, methansulphonate, pyruvate). Correlations between some analyzed components have been estimated to evaluate their sources (sea spray, secondary aerosol, crustal — such as Saharan dust — and anthropogenic imputs), transport processes and eventual fractionating phenomena.

This work reports preliminary data on the presence of methanesulphonic acid (MSA) in aerosol and rain collected at various sites and onboard ship around the Central Mediterranean Sea. Mean MSA values are 10.2 μg 1−1 in precipitation and 0.048 μg m−3 in aerosols. These values are very similar to other marine areas around the world. The MSA/nssSO₄2− ratio is lower than 1% (0.2 and 0.7% for rain and aerosol, respectively), well below North Sea and Atlantic Ocean values reported in literature (8–10%).