MEDITERRANEAN ATMOSPHERIC MERCURY CYCLE SYSTEM (MAMCS)
Funded by: European Commission-DG Research – Environment and Climate Program. It is part of the ELOISE Network
Reporting Period: 1 January 1998 – 30 June 2000
Mercury and the Mediterranean Basin
Mercury is present in the Mediterranean as cinnabar deposits (HgS, the primary mercury ore), and the Mediterranean actually contains approximately 65% of the world’s deposits. Mercury is released or transported into the Mediterranean region mostly as a result of anthropogenic emissions as a result of fossil fuel combustion, cement production and some specific industries such as chlor-alkali plants. There is also a natural contribution to Hg emissions from volcanoes, soil outgassing, and release from the surface waters of the sea.
The presence of mercury in aquatic ecosystems is of concern because of the toxicity of certain forms of mercury (particularly methylmercury), and its potential for bio-accumulation within the food web. In the Mediterranean Basin many communities depend on the sea for their livelihood and seafood is a major part of many peoples diet. The quality of the waters of the Mediterranean is therefore of great importance for all the countries which surround it.
The Project Work Packages
WP-1: Emissions Database
The compilation of the Mercury Emission Inventory (MEI) for Europe and the Mediterranean countries using the most up-to-date data available from national and international bodies where available, and using emission factor techniques to verify data and estimate emissions for those countries for which no data was available. The MEI includes both natural and anthropogenic emissions.
WP-2: Model Development
The development of models to describe the processes influential in mercury cycling, including gas and aqueous phase chemistry, gas-aqueous phase mass transfer, gas-particle partitioning and air-sea surface exchange flux.
WP-3: Atmospheric Measurements
One intercomparison campaign (Tuscany, 1998), to ensure conformity between field sampling techniques for TGM, RGM and TPM and assure acceptable QA/QC standards between laboratories. Four two-week measurement campaigns, one each in Spring, Summer, Autumn and Winter, at five Mediterranean sites, performed contemporaneously with the measurement campaigns of the Mercury Over Europe (MOE) project, which had measurement sites in Germany, Scandanavia and Ireland.
WP-4: Regional Assessment
The combination of dispersion / meteorlogical models with the MEI and process models to produce an integrated modelling system. The determination of characteristic seasonal transport and wet and dry deposition patterns for the region. Validation with data from the measurement campaigns. Determination of natural vs. antrhopogenic condtributions to Mediterranean merrcury fluxes, and the identification of the major mercury source regions on a season by season basis.
The preparation of recommendations for the control and possible reduction of atmospheric mercury input into the Mediterranean Sea region.
This three-year EU funded project studied the fate and cycling of Hg in the Mediterranean Basin, and involved:
- the compilation of an updated emission inventory for Europe and the countries around the Mediteranean;
- measurements of Total Gaseous Mercury (TGM), Reactive Gaseous Mercury (RGM) and Total Particulate Mercury (TPM), in one intercomparison and four two-week field campaigns at five sites around the Mediterranean;
- development of models to describe individual processes, such as gas-particle partitioning, air-sea exchange, which influence Hg cycling;
- the integration of meteorological / dispersion models with the emission database and the process models to provide a tool capable of predicting Hg transport and deposition in the Mediterranean Basin for use by scientists and policy makers.
The Mercury Emission Inventory
- the collection of information on past and present mercury emissions from anthropogenic sources in Europe and the countries in the Mediterranean Basin;
- the estimation, using emission factors and data on industrial production, for countries where no data were available;
- the estimation of the proportion of elemental mercury (Hg0(g)), RGM (HgII(g)) and mercury associated with particulates emitted from the various mercury source categories;
- the preparation of a gridded (50km by 50 km) database of anthropogenic emission sources, including source category and (where applicable), stack height;
- the preparation of ‘league tables’ of the major anthropogenic mercury point sources on a country by county basis.
One of the most important features of the MEI is that the emissions from source categories include not only fluxes but also the relative proportions of Hg0(g), HgII(g) and mercury associated with particulates. The figures on the left show the contributions from each source category to the total emitted mercury of each of the three types.
It is obviously difficult to be precise when estimating emissions. It is possible using emission factors combined with statistical data concerning energy production, raw material consumption and industrial output to check the available national emission data, and this has been done for those countries for which data was available and the agreement suggests that on the whole national data is reasonably accurate.
Uncertainty estimates for the various emission source categories have been made:
|staionary fossil fuel combustion||± 25%|
|non-ferrous metal production||± 30%|
|cement production||± 30%|
|iron and steel production||± 30%|
|waste incineration||up to a factor of five|
The higher uncertainty associated with waste incineration reflects the variety of materials which may be incinerated and also the dependence of emissions on both combustion regime and flue gas scrubbing technology.
In order to improve emission estimates and the uncertainties in emission factors it would be necessary to perform long term measurements at industrial installations in the different source categories.
Major impovements could be made to the emission database if natiaonal emission data were known for those countries which as yet furnish no estimates. As many of the countries around the Mediterranean are predicted to increase their energy consumption, and building (predominantly with cement) continues in a large part of the region it is more than probable that local emission will increase in the coming years, especially given the pressure, particularly in some countries, to increase waste incineration in order to avoid problems with landfill.
Mercury is emitted naturally from volcanoes and hydrothermal vents. Fluxes from these sources depend on how active each individual source is. Other sources usually considered ‘natural’ rather than anthropogenic include area sources such as Hg-rich soils and the sea surface. In these cases there may well be a significant contribution from re-emission of previously deposited anthropogenically emitted mercury, although this is clearly extremely difficult to quantify. Emission from these sources depends greatly on solar irradation and also in the case of soils on the soil humidity.
Measurements of TGM, RGM and TPM were performed using a number of different sampling techniques and analysis methods, see Table 1 for details.
Figure 1 shows the TGM results obtained by the different groups using different methods.The variability decreases during the sampling campaign. In the last 4 samples, all of the individual results are within one standard deviation of the overall average value. The higher variability observed during the first two samples is most likely due to start-up difficulties and disturbances caused by alterations in the sampling set-up during the initial phase.
Details of the intercomparison campaign may be found in Munthe et al., 2001 “Intercomparison of Methods for Sampling and Analysis of Atmospheric Mercury Species” (see also forthcoming Special Issue of Atmosheric Environment)
Four two-week measurement campaigns were carried out at 5 sites around the Mediterranean. The measurement periods were chosen to give data for each season of the year.
|1||November 23, 1998||December 6, 1998|
|2||February 15, 1999||March 1, 1999|
|3||May 3, 1999||May 17, 1999|
|4||July 19, 1999||August 2, 1999|
The measurement sites are shown on the map on the left
The measurement sites.
1 Mallorca (39º40’30’’N, 2º41’36’’E), 2 Calabria (39o25’N, 16o00’E), 3 Sicily (36o40’N, 15o10’E), 4 Turkey (36o28’12”N, 30o20’24”E)5 Israel (32o40’N, 34o56’E), 6 Germany (53o08’34”N, 13o02’00”E), 7 Germany (54o26’14”N,2o43’30”E),8 Sweden (57o24’48”N, 11o56’06”E), 9Sweden (58o48’00”N, 17o22’54”E), 10 Ireland (53o20’N, 9o54’W).
The map also shows the five sites where measurement campaigns were carried out as part of the MOE project. These campaigns were co-ordinated with those of the MAMCS project so that atmospheric mercury data is available from 10 sites as far apart as Galway (Ireland) and Anatalya (Turkey), for the same four two-week periods.
The results from the measurement campaigns have been the subject of a number of articles, see for instance Wangberg et al., 2001 “Atmospheric Mercury Distribution in Northern Europe and in the Mediterranean Region” (see also forthcoming Special Issue of Atmosheric Environment).
The Figures below show the average TGM, RGM and TPM from the MAMCS and MOE sites obtained for each measurement campaign:
The average TGM concentrations varied between 1.6 and 2.4 ng m-3. The relatively uniform distribution found is reasonable since TGM predominately depends on relatively stable global/hemispheric background concentrations and only occasionally shows higher values, due to influence from major sources. Except for the first campaign, the data indicates that TGM is slightly but significantly higher in the Mediterranean area than in North Europe.
The Mediterranean RGM and TPM concentrations are higher than in Northern Europe. The reason for higher average TPM and RGM concentrations in the Mediterranean region may be due to higher emission rates and/or more active atmospheric transformation processes. Another aspect influencing the atmospheric content of TPM and RGM is precipitation. The lower values in Northern Europe may be due to washout being a more efficient removal process in the north.
Atmospheric Mercury Process Modeling
Non-hygroscopic particles (and hygroscopic particles under low humidity conditions) provide a surface which elemental mercury and RGM may adsorb on to, and this will effect mercury deposition. There are many different types of aerosol particle and each will have different adsorption characteristics for elemental mercury and RGM. The low frequency of precipitation events during the Mediterranean summer prompted the evaluation of Hg adsorption and subsequent deposition as a possibly important pathway for Hg input into the Mediterranean Sea. For the purposes of this evaluation a model (GASPAR – GAS-particle PARtitioning) using a parameterised description of the ambient aerosol with four source categories and three modal values of the radius was developed, and using as far as possible measured TPM and TGM values to derive an approximate adsorption coefficient, (Pirrone et al., 2000). Because deposition to the sea surface means that particles will pass through humid air the possibility of hygroscopic material in the particles and deliquescence were introduced. Whilst investigating the aqueous phase chemistry in deliquesced aerosol particles it became clear that the concentration of chloride ions in the particles played an impportant role in determining how much HgII could be associated with the particles. It also became apparent that sea-salt particles could scavenge HgO formed by the the reaction of Hg0 and O3, and recycle it to HgCl2 which could be re-emitted from the particles. These results indicated the need for gaseous and aqueous phase chemistry (and photochemistry) in the MBL to be coupled to the simple Hg chemistry module used up to that point in GASPAR in order to fully investigate Hg chemistry over the Mediterranean Sea.
Gas and Aqueous phase chemistry
Mercury in its elemental form is relatively unreactive in the gas phase, the reaction with ozone is slow (k=3×10-20 molecules-1 cm3 s-1), only radicals such as OH (and probably some Br containing radicals) react rapidly with Hg0(g). The tropospheric chemistry model developed for the MAMCS project required detailed photochemistry chemistry of the Marine Boundary Layer (MBL) and an accurate description of the mass transfer of soluble and partially soluble gases between the gaseous and aqueous phases, because the relative humidity of the MBL is such that sea-salt aerosol exists as solution droplets, as does the non-sea-salt (nss) sulphate aerosol. The reaction database used and the details of the photolysis routine and integration method used can be found in Hedgecock and Pirrone, 2001. (link publications).
Air-Sea Exchange of Hg
This remains one of the more elusive aspects of Hg cycling. It is known that HgII compounds in the top surface microlayer can be reduced to produce Dissolved Gaseous Mercury (DGM) which due to its low Henry’s Law constant evades from the water to the air. It also seems clear from field experiments that this process in linked to variables such as insolation and the air-sea temperature difference. An empirical model has been used so far to model this process, but it is rather unsatisfactory. Further investigations are planned for the very near future (see the MERCYMS Project page).
The Integrated Modelling System
In order to couple the new Mercury Emission Inventory and Atmospheric Process Models and obtain concentration and deposition fields for the Mediterranean Basin, knowledge of the transport of air masses across and within the Mediterranean Basin is required. Two dispersion models have been used, the Regional Atmospheric Modelling System (RAMS), and SKIRON. RAMS includes much more detailed microphysics of cloud processes and is used to calculate humidity and wet deposition fields for the periods of interest. SKIRON is less computationally demanding, but does the advantage of treating rapidly changing terrain altitude in detail, a fact of great importance given that there are numerous mountainous areas close to the Mediterranean Sea.
Once coupled to the Emission Inventory and the chemistry models the system can predict dry and wet deposition fluxes and gas phase Hg concentration fields. The shematic below gives an idea of the model architecture, and figures 1 and 2 show two of the first maps of Hg0 and TPM distributions obtained using the integrated model.
Conclusion and Recomendations
- The MAMCS project has provided the most up-to-date Inventory of Mercury emissions for Europe and the non-European countries bordering the Mediterranean.
- In conjunction with the MOE project there are four two-week atmospheric Hg measurement datasets with measurements from Galway to Haifa. Such a dataset in unprecedented in mercury research, and compatability between sampling methods and analysis were validated by a major intercomparison campaign.
- Atmospheric mercury process models have been developed from scratch or have built on and extended existing models, giving new insights into the role of atmospheric chemistry in mercury cycling and the importance particularly of the MBL and sea-salt aerosol.
- The first maps of concentration fields for Hg0, RGM and TPM have been produced by combing the emission data, process models and transport/dispersion models.
Details of the various parts of the project are now available in the literature. (link Publications) The results MAMCS and MOE projects provided a great deal of data for the discussion and the preparation of the EU Position Paper on Mercury.A number questions about the cycling mercury which remain unresolved are to be addressed in the MERCYMS project recently approved under the 5th FP of the EU DG Research.
- CNR-Institute of Atmospheric Pollution Research (CNR-IIA) – COORDINATORE
- CNR Institute of Biophisics (CNR-IB)
- Technion – Israel
- IVL – Swedish Environmental Research Institute
- NILU – Norway
- NKUA – Greece
- Universitat Autonoma de Barcelona
- 52°North Initiative for Geospatial Open Source Software GMBH
- Science and Technology B.V.
Gruppo di lavoro
Ian M. Hedgecock
Staff CNR IIA
Staff CNR IIA
Staff CNR IIA
Staff CNR IIA
Staff CNR IIA
Responsible Scientist CNR IB
Staff CNR IB
Staff CNR IIA
Responsible Scientist TECHNION
Responsible Scientist IVL
Jozef M. Pacyna
Responsible Scientist NILU
responsible Scientist NKUA