Atmospheric modelling of heavy metal pollution in Europe: Further development and evaluation of the MSCE-HM model" DRAFT

MSC-E Technical Note 1/2007

I. Ilyin, O. Rozovskaya, V. Sokovykh, O. Travnikov

ABSTRACT

The MSCE-HM chemical transport model has been developed for assessment of heavy metal airborne pollution in Europe and support of the review, extension and implementation of the Protocol on Heavy Metals. The model performance have been reviewed at the EMEP Task Force on Measurements and Modelling meeting in Zagreb and TFMM Workshop on model review in Moscow in 2005. Along with the recognition that the model is "appropriate for the operational modelling of heavy metals at the regional scale", a number of recommendations aimed at further model improvement have been formulated. Particularly, the recommendations included the following issues:

-· Extensive evaluation of input meteorological data for regional heavy metal modelling;
-· Movement to the ECMWF input data with 1°?1° spatial resolution for the meteorological fields pre-processing;
-· Development of the model parameterization for heavy metal re-suspension from soil and seawater;
-· Extension of the MSCE-HM model parameterisation for the second priority heavy metals (As, Cr, Ni, Cu, Zn and Se);
-· Improvement of the model description of atmospheric removal processes;
-· Extension of the heavy metal hemispheric model to the global scale;
-· Inclusion of a shallow lowest model layer;
-· Investigation of mercury dry deposition to forests;
-· Further research and improvement of the model description of mercury chemical transformations in the atmosphere.

Besides, the Worksop concluded that official emission data for heavy metals suffered from significant uncertainties and are of limited value in terms of model applications. Therefore further improvement of official emission data was highly recommended.

First results of MSC-E activities on implementation of the above mentioned recommendations have been reflected in the EMEP/MSC-E technical report [Gusev at al., 2006]. Particularly, an approach to validate input meteorological fields prepared for atmospheric transport modelling has been elaborated and described. A tentative parameterization of wind re-suspension of particle-bound heavy metals from soil and seawater has been developed. Pilot results of the assessment of lead and cadmium re-suspension contribution to the pollution levels in Europe have been described. Extension of the MSCE-HM model parameterisation for the second priority heavy metals (As, Cr, Ni, Cu, Zn and Se) has been reported along with modelling results of their long-rage atmospheric transport in Europe. The model performance has been evaluated using different emission inventories (official emissions data and ESPREME expert estimates).

This report continues consideration of the MSCE-HM model development and evaluation in accordance with the recommendations of the Workshop on model review. Particular attention is paid to comparison and analysis of discrepancies of different anthropogenic emissions inventories. Three available datasets of heavy metals anthropogenic emissions in Europe in 2000 are taken into consideration: the EMEP official emissions data, TNO and ESPREME non-official expert estimates. The analysis is focused on emissions data for lead and cadmium since these two heavy metals give the major concern from the modelling point of view [Ilyin and Travnikov, 2005]. The emission inventories were analyzed by intercomparison of annual emission totals from individual European countries and contributions of key source categories to national emissions.

Further development of the quality control system of input meteorological data is described and illustrated. The main goal of the system is performing quantitative quality control and verification of meteorological fields prepared by the meteorological pre-processor (MM5) for atmospheric transport modelling on the routine basis. Statistical algorithms characterizing reliability of meteorological data through comparison of their spatial and temporal variation as well as frequency distributions with reference datasets are presented. Besides, progress of MSC-E activities aimed at moving to ECMWF meteorological fields with higher spatial resolution (1? ? 1?) as input information for the data pre-processing is reported. First results of comparison of meteorological parameters obtained using this dataset with those based on data from the NCEP/DOE reanalysis are considered.

The model parameterization of wind re-suspension of particle-bound heavy metals from soil and seawater is further developed. Particularly, the effect of soil characteristics (soil texture, size distribution of soil grains etc.) on dust production and suspension are taken into account. A global dataset of soil properties is used to characterize wind erosion and re-suspension of heavy metals from different soil types in Europe. Besides, some special effects influencing dust mobilization (such as wind drag partition, inhibition by soil moisture, the Owen effect) are introduced or revised in the model. Full description of the revised dust production scheme is presented along with estimates of heavy metal re-suspension in Europe. Plans for further research are discussed.

A special attention of this report is paid to evaluation of modelling results against measurements. One of the inferences of the Workshop on model review was the conclusion that MSCE-HM model tends to underestimate measured concentrations of heavy metals in air and precipitation when using the official emissions data for modelling. Similar underestimation was demonstrated by other models participated in the intercomparison studies. An improvement of the agreement was achieved by taking into consideration the wind re-suspension process, however, some underestimation still remains [Gusev et al., 2006]. To analyse the problem the model results are compared with results of the CMAQ model - one of well-developed contemporary chemical transport models. Besides, the underestimation is analysed using measurements data with high (daily or weekly) temporal resolution. It allows evaluation of the model performance to reproduce both short-term and long-term variation of measured values and reveal possible reasons of the underestimation.