Detailed project description

Background

Of the many different air pollutants, tiny particles in the atmosphere are thought to present one of the most significant current risks to human health. However, scientific understanding of the behaviour and health effects of this so-called particulate matter (PM) is far from comprehensive. For example, it is not known with certainty which sizes of PM cause the most serious health effects, nor is it fully clear which the most dangerous chemical components of PM are. Nevertheless, given the risks, policymaking to manage PM pollution must continue. Studies which help to close this knowledge gap and which support the policymaking are sorely needed.

Both the UK Government (via the Air Quality Strategy) and the European Union (via the Air Quality Framework Directive and Daughter Directives) have established objectives and limit values for PM10 as follows:

24-hour mean limit value of 50mg m-3, not to be exceeded more than 35 times a year (approximately equivalent to a 90th percentile of 50mgm-3) [to be achieved by 1st Jan 2005]

Where PM10 is defined as particulate matter with a diameter of 10mm or less. The European Union is presently reviewing the relevant Daughter Directive.

At this present time, the PM10 metric is established as the standard to which local and national authorities must adhere. Therefore, in order to meet these strict standards, it is important to identify and quantify all sources and pathways which contribute to total PM10 concentrations regardless of health effects.

Despite the fact that this target represents a relaxation of a previously more stringent one, its achievement remains challenging in part because of the significant particulate component which is transported long distances over regional scales. These long-range sources include pseudo-natural phenomena such as forest fires, desert dust and biomass burning, along with anthropogenic aerosols from Europe and the Eastern US. Figure 1 shows a Saharan dust episode over the UK in April 2003, demonstrating the potential for long-range transport of such material. Furthermore, inter-annual variations in weather conditions lead to unwelcome 'surprises' in terms of the number of long-range particulate pollution episodes which can occur, limiting the potential positive effects of local air quality management efforts, for example of congestion charging. The UK Department for Environment, Food and Rural Affairs publish a headline indicator each year for air quality in order to demonstrate long-term trends in UK air quality. Figure 2 below shows the headline indicator up to 2003 (provisional).

Figure 1: Saharan dust episode over the UK, 15th April 2003.

Figure 2: Number of days on which air pollution is moderate or high, 1987 - 2003.

Figure 2 shows that there were twice as many days of moderate or high air pollution during 2003 as there were in 2002. This dramatic change is thought to have been mostly caused by changes in weather conditions, promoting the import into the UK of larger quantities of air pollution than normal along with conditions of relatively poor air pollution dispersion. Figure 3 reinforces this point, focusing on PM10 measurements in London in particular.

Figure 3: PM10 concentration data recorded at the North Kensington monitoring station in London in the period January-December, 2002 and 2003.

While effective action can be taken at a local level to reduce the emission of PM, for example via traffic management (such as through the London Congestion Charging Scheme), it has become clearer in recent years that a significant fraction of atmospheric PM is 'imported' over distances of tens to thousands of kilometres, given appropriate weather conditions. Recognition of this followed the establishment of a UK national network of PM measuring instruments; synchronised elevated concentrations of atmospheric PM over wide geographical regions, especially during persistent winds from major UK and European PM source regions, were seen to occur. This same finding also highlighted a key limitation imposed by national air pollution monitoring networks, namely that what was really needed was measurement technology which was not limited by national boundaries, sea surfaces or spatial resolution. It is generally believed, but not unambiguously demonstrated, that the finer particulate fractions are disproportionately serious in terms of their potential health effects (APEG, 1998). It is these same finer fractions which are preferentially transported over long distances and which therefore need to be tracked over appropriate scales.

Over the last ten years, enormous progress has been made in satellite technology, a form of 'remote sensing'. The great benefit of satellite monitoring, over point measurements, is its continuous spatial coverage over continental scales. In the last two years, with the recent launch of American and European satellites, it has become possible to monitor the transport of atmospheric PM over both land and sea surfaces. This giant stride forward not only enables us to better understand the key sources and transport pathways for PM, it also provides us with an opportunity to validate and improve the computer models which are used to simulate PM behaviour.

The demands of the European scientific community for a global environmental monitoring system, whose technical characteristics enable the extraction of quantitative information from ocean colour data, as well as for documentation of the state and evolution of the atmosphere and land surfaces, led to the conception of MERIS (The Medium Resolution Imaging Spectrometer). The instrument is capable of detecting the low levels of radiation emerging from the ocean (linked to the water constituents by the processes of absorption and scattering), in addition to acquiring information about the atmospheric medium through which the observation is made as well as terrestrial environments (Santer et al, 1999). Since its launch in March 2002, the European Space Agency (ESA) environmental research satellite 'ENVISAT' has been returning atmospheric aerosol estimates using the on-board MERIS instrument. These estimates are returned as an aerosol optical thickness, indicating regions where atmospheric aerosol loadings are highest. In addition, an estimate of the relative particle size is returned.

The MODIS instrument (Moderate Resolution Imaging Spectroradiometer) aboard the NASA TERRA satellite is able to return estimates of aerosol optical depth, mass concentration and the effective particle radius of those particles in the field of view (Chu et al, 2003). MODIS can also detect dust and volcanic ash in the atmosphere in addition to the direct detection of fires, such as forest fires. The plumes from these fires can be tracked using the MODIS instrument, which will be useful in this proposed study.

Both MERIS and MODIS have, for the first time, provided estimates of aerosol over land surfaces, where this was previously only possible over the ocean, for example from the SeaWIFS instrument (Doyle and Dorling, 2002). Access to data from the ENVISAT mission has already been approved by the European Space Agency for the project described here.

The School of Environmental Science at the University of East Anglia is a top rated research department, rated 5* in the 2001 UK Research Assessment Exercise. The school has, since its inception, adopted an holistic approach to environmental research, integrating physical, biological, chemical and social disciplines together. The cross-disciplinary nature of the project proposed here is very much in keeping with this philosophy. The proposed work will form a part of and be informed by the newly launched University of East Anglia Saharan Studies Programme (http://www.uea.ac.uk/sahara/). Participants in the Programme have particular expertise concerning the varying strength of the Saharan dust emission source and this will be of great use in the study proposed here.

Hypothesis

Particulate matter in the United Kingdom is influenced by both local and more distant sources. Local air quality management measures may not be sufficient to meet strict particulate criteria enacted in National and European legislation during periods when significant long range transport of particulate matter occurs.

Objectives

The main project objective is to study how state-of-the-art satellite measurements, such as those now available from the European Space Agency ENVISAT satellite and the NASA TERRA satellite, can be used to quantify and provide advanced warning of the transport of atmospheric PM over regional scales. This PM has a diverse range of sources, for example desert dust, forest fires and burning of biomass, along with man-made PM which has either been emitted at source (so-called primary PM) or which has formed in the atmosphere through chemical reactions involving man-made gaseous pollutants (secondary PM).

A second objective is to use these satellite estimates to help validate both the US Naval Research Laboratory and UK Meteorological Office models which seek to simulate atmospheric PM transport. Finally, we wish to also test the robustness of our own classification of weather patterns which is designed to highlight those weather situations which are particularly conducive to contributing to periods of dangerously high atmospheric PM concentration in the United Kingdom.

Significance

It is only very recently that it has become possible to use satellites to reliably estimate the loading of atmospheric PM over land surfaces. This is because, until now, it has not been possible to take account of the great underlying variation in land surface characteristics (city/forest/arable/water) and the effect which this has on what the satellite perceives as being present in the overlying atmosphere. Remote sensing programmes have regularly been criticised for generating enormous volumes of 'data' which, due to their volume, are far from fully analysed. The project proposed here represents a use of state-of-the-art satellite technology in application to a very real world problem.

Policymaking which seeks to set targets for air quality needs to be demanding but also realistic. In the context of this project, policy which is over-ambitious in terms of the degree to which atmospheric PM can be locally managed may lead to disappointment in terms of the actual results. The results of this project, in helping to quantify how much atmospheric PM is imported over distances of hundreds to thousands of kilometres, will help ensure that policy targets remain both demanding and realistic.

Methods

A number of innovative and mutually supportive methods will be utilised:

(i) Now that it is possible to access both MERIS and MODIS data (which, used together, provide a daily sampling frequency), we intend to study UK particulate episodes in the context of the information which this new satellite data can provide. In particular we will address the question of vertical aerosol profile by marrying the satellite data with known estimates of boundary layer depth from the output of numerical weather prediction models. Some costs will be incurred in accessing this data. The School of Environmental Sciences at the University of East Anglia has recently purchased a corporate license for the Landcover Map 2000 from the Centre for Ecology and Hydrology (CEH). This product contains a 25m resolution land use map which will help with retrieval of aerosol parameters. Algorithms used in the retrieval of aerosol over land utilise areas of dark dense vegetation (DDV) as a suitable background for these remote measurements. Any improvement in the spatial resolution of areas of DDV will result in improvements in the spatial resolution of the retrieved aerosol.

(ii) Dr Alistair Manning has developed a novel way of running the UK Meteorological Office air pollution dispersion model (known as NAME - Redington and Derwent, 2002) in reverse, in order to model the important source regions of pollutants sampled on the extreme west coast of Ireland, at the Mace Head background monitoring station. Figure 4 shows an example of the output of this method, in which thousands of puffs are released in any six hour period and are exposed to the turbulent characteristics of the windfield, rather than the mean flow depicted by a back-trajectory. Figure 4 shows an incident where Saharan Dust was thought to have been transported over the UK (after emission from North Africa - Figure 5) and subsequently removed during a rain event. We aim to repeat this novel approach, but this time selecting a different receptor over the mainland which will be more representative of the exposure of the UK population to long-range transport of particulate. These new calculations also have a data cost associated with them, of £2500 per modelled year. We intend to look at four years of data, covering the period 2000-2003.

(iii) Dorling and Doyle (2003) reported a meteorological classification scheme, covering the period 1991-2001, which can be used to quantify the frequency of different atmospheric circulation pathways. The results have subsequently been used in the interpretation of recent Nitrogen Dioxide measurements in the UK (AQEG, 2003). As already noted, 2003 saw many more PM10 exceedances compared with 2002 (Figure 3). With data now available for 2002 and soon to be available for 2003, this classification scheme will be updated to include these years. Relationships between PM measurements and the output of the classification scheme will be created. The classification scheme will be tested using 2003 PM data in order to identify whether long range pollution events are detected using this scheme. Such events have been shown to be strongly connected to regional scale weather patterns (Stohl et al, 2003).

Figure 4: Output from the UK Met Office NAME model showing source regions of air for the Mace Head receptor on 1st and 2nd March 2000. (Courtesy of Alistair Manning, UK Met Office).

Figure 5: SeaWiFS image for February 26th 2000 showing Saharan Dust streaming off N. Africa and which was subsequently monitored in the UK.

Using these methods, an examination of how important long range transport pathways have waxed and waned in the 4 year study will be enabled. This will also allow a comparison of how this recent period differs from the 1951 - 2003 period as a whole, helping us to assess how typical the 2000 - 2003 period was when compared with this longer record. The classification scheme will also be tested for its ability to identify and possibly forewarn of such long range PM events.

(iv) Six-hourly forecasts from the US Naval Laboratory global aerosol model (NAAPS) are made freely available on the Internet (Figure 6, http://www.nrlmry.navy.mil/aerosol/globaer_world_loop.html). The aerosol is divided up into components: sulphate, smoke and dust with modelled surface concentrations also being available. This will aid in any validation exercise undertaken. Correspondence with the US Naval Research laboratory has confirmed the continued availability of this data for the proposed duration of the project. In addition to this the model will be made fully operational this year (i.e. hourly resolution as opposed to 6 hourly) (Westphal, 2004). We will monitor, archive and analyse NAAPS model output over the duration of the project.

(v) Ongoing satellite retrievals of aerosol optical thickness and images in the visible spectrum will continue to be monitored in the area off the coast of North West Africa for the duration of the project to identify periods when fluxes of Saharan dust are high. This monitoring will also be aided by the "Dust List Discussion Group" maintained by Professor Joe Prospero of the University of Miami, Florida. During periods of high dust activity around the globe, messages are sent to interested parties with links to images and further information of interest. These messages will be monitored and relevant material extracted.

The NAME model does not give any idea of the size, composition or source (biomass burning, desert dust etc) of particulate matter. NAAPS, meanwhile, does not give detailed information on source regions of the pollution experienced at defined receptor points. However, by using the two models in tandem, these issues will be resolved, resulting in the availability of data on particle size, source region and emission source over the UK and at the defined receptor points from (ii).

Figure 6: Sample output from the US Naval Research Laboratory global aerosol model (http://www.nrlmry.navy.mil/aerosol/globaer_world_loop.html)

Finally, we will assess the extent to which points (i) to (v) complement and support each other.

References

APEG (1998) Source Apportionment of Airborne Particulate Matter in the UK. A Report of the Airborne Particles Expert Group. United Kingdom Department of Environment, Trade and the Regions.

AQEG (2003) Nitrogen Dioxide in the United Kingdom. A Report of the Air Quality Expert Group. United Kingdom Department of the Environment, Food and Rural Affairs.

Chu, D. A., Y. J. Kaufman, G. Zibordi, J. D. Chern, J. Mao, C. Li, and B. N. Holben, (2003) Global Monitoring of Air Pollution over Land from EOS-Terra MODIS. J. Geophys. Res., 108(D21), 4661.

Dorling, S.R. and Doyle, M. (2003) Meteorological Classification and Aggregation Approaches in Support of Models-3 Air Quality Simulations. Proceedings of the 4th International Conference on Urban Air Quality. Prague, Czech Republic, pp424-427.

Doyle, M. and Dorling, S.R. (2002) Satellite based monitoring of aerosol plumes. Water Air and Soil Pollution, 2, pp615-629

Redington, A.L and Derwent, R.G. (2002) Calculation of sulphate and nitrate aerosol concentrations over Europe using a Lagrangian dispersion model. Atmospheric Environment, 36 (28), pp4425-4439.

Santer R., Carrère V., Dubuisson P. and Roger J.C. (1999) Atmospheric corrections over land for MERIS. Int. J. Remote Sensing, 20 (9), 1819-1840.

Stohl, A., Huntrieser, H., Richter, A., Beirle, S., Cooper, O.R.., Eckhardt, S., Forster, C., James, P., Spichtinger, N., Wenig, M., Wagner, T, Burrows, J.P. & U. Platt (2003) Rapid intercontinental air pollution transport associated with a meteorological bomb. Atmospheric Chemistry and Physics, 3, pp969-985

Westphal, D.L. (2004) personal communication.