Simulation of European emissions impacts on particulate matter concentrations in 2010 using CMAQ Steve Griffiths, Rob Lennard and Paul Sutton* (*RWE npower) Power Technology, Environmental Compliance Group

Overview of presentation Developments in PM regulation Relevance to the power generation sector CMAQ set-up and treatment of PM Overview of modelled scenario Results - Compliance with guidelines Sector by sector impacts Primary versus secondary particulate impacts Implications of a population exposure approach

Background to study Power Technology work on behalf of UK power generator’s Joint Environmental Programme (the JEP) Eight companies – cover majority of the UK coal and oil-fired generation Investigate environmental issues of relevance to the power industry Selected CMAQ in 1999 to address regional scale issues Air quality, acid deposition, particulate matter formation

Updates to PM guidelines Mounting evidence of link between PM exposure and health effect Fine particles primarily responsible No threshold WHO Air Quality Guidelines global update PM2.5 10 µgm-3 annual mean and 25 µgm-3 24-hour mean New EU Air Quality Directive (currently under discussion) PM10 40 µgm-3 annual mean, 50µgm-3 24-hour mean PM2.5 25 µgm-3 annual mean (binding from 2015) PM2.5 20% reduction in urban background annual mean (2020) Number of allowable exceedances of pm10 annual mean is currently being debated. MEPs suggested 55 instead of 35

Important issues for the JEP/ESI Power stations contribute to PM concentrations: 5.5% of primary UK PM10 emissions (NAEI., 2004) Secondary – 60% of SO2 and 22% of NOx emissions Mass based metric Uncertainty regarding fraction responsible for adverse health effects Toxicology studies suggest primary combustion particles have high toxic potency Other components are thought to have a lower toxic potency e.g. ammonium salts, chlorides, sulphates, nitrates Need to understand the effect of our emissions on primary and secondary PM2.5 concentrations

Models-3/CMAQ 54km European domain 3-D gridded Eulerian model Set up to run on three nested grids (54/18/6km and 108/36/12 km) 21 vertical layers (vertically 15km) Requires hourly gridded emissions and meteorology (UK Met Office) Initial conditions & boundary conditions from interpolated measurements Chemical Scheme: RADM2+aerosols+aqueous chemistry 54km European domain

Modelling Particulate in CMAQ Based on USEPA particulate model / Regional Acid Deposition Model Time-dependent size distribution & size specific chemical composition Modal approach – Coarse, accumulation & nucleation Described by particle number concentration, total surface area & total mass Species Sulphate, Nitrate, Ammonium Elemental Carbon Primary anthropogenic organic species Anthropogenic secondary organic species Biogenic secondary species Unspecified anthropogenic species Can also include aerosol water

Models-3 PM simulation validation – UK sites (12km) Comparison with measured data – 24 hour averages January 1999 July 1999

European Modelling study – 54 km resolution Compared contributions of 7 sectors to ground-level PM concentrations Model run for an entire year, month-by-month basis, 1999 met data Anthropogenic emissions scenario for the year 2010 – data from EMEP WebDab Expert Emissions Database (emep.webdab.int) SO2, NOx, PM10, PM2.5, NH3, CO and NMVOC Natural VOC emissions from GEIA (Global Emissions Inventory Activity) Temporal profiles applied using SMOKE - energy sector data provided by JEP companies VOC speciation profiles from NAEI (www.naei.org.uk) Emissions from SNAP categories 1, 3 and 4 assigned to point sources (based on assumptions from National Atmospheric Emissions Inventory)

Sectors for which impacts have been assessed Ground-level PM concentrations due to 7 different sectors compared Eight model runs performed in total Sectors: Energy SNAP 1 Residential SNAP 2 Industry SNAP 3,4,5,6,9 Shipping SNAP 8 Transport SNAP 7,8 Agriculture SNAP 10 Natural* SNAP 11 (*volcanoes, marine, forests) Ground-level concentration output: Total PM10 Primary PM10 Secondary Total PM2.5 Primary PM2.5

Results – comparison with standards

Exceedances of 25μgm-3 annual mean for PM2.5 Area Grid square Eastern Sicily (44,11) Northern Italy (34,25) Agriculture 2.3 10.9 Energy 0.7 1.9 Industry 1.2 5.0 Natural 19.4? 1.5 Residential 0.6 5.2 Transport 7.1 14.5 Shipping 1.3 1.0 All Sources 27.7 30.2

Exeedances of 12 μgm-3 annual mean for PM2.5

All source primary and secondary PM2.5 concentrations

Sector contributions to grid averages and maxima (µgm-3) Total PM2.5 Secondary PM2.5 Primary PM2.5 Average Max Agriculture 2.33 11.53 2.31 11.49 0.03 0.47 Energy 0.78 4.33 0.74 3.93 0.05 0.57 Industry 0.84 5.49 0.66 3.37 0.18 4.32 Natural 0.61 19.43 0.60 0.01 Residential 0.63 6.26 0.34 1.76 0.30 5.37 Transport 1.16 14.50 0.98 10.93 3.57 Shipping 4.64 0.59 1.54 All sources 7.19 30.17 6.40 26.71 0.79 8.88

Sector contributions to grid averages and maxima (µgm-3) Total PM2.5 Secondary PM2.5 Primary PM2.5 Average Max Agriculture 2.33 11.53 2.31 11.49 0.03 0.47 Energy 0.78 4.33 0.74 3.93 0.05 0.57 Industry 0.84 5.49 0.66 3.37 0.18 4.32 Natural 0.61 19.43 0.60 0.01 Residential 0.63 6.26 0.34 1.76 0.30 5.37 Transport 1.16 14.50 0.98 10.93 3.57 Shipping 4.64 0.59 1.54 All sources 7.19 30.17 6.40 26.71 0.79 8.88

Exposure approach – population data Based on Gridded Population of the World (GPW) dataset, available from Centre for International Earth Science Information Network (CIESIN) website Projected to Lambert Conformal Projection appropriate for CMAQ Population density in persons per km2 ■ 0 - 50 50 - 200 200 - 500 500 - 1000 > 1000

Population exposure – Secondary PM2.5 Exposure = Population x concentration (person.μgm-3km-2) Agriculture Residential Energy

Population exposure – Primary PM2.5 Exposure = Population x concentration (person.μgm-3km-2) Agriculture Residential Energy

Impact of considering population exposure – Secondary PM2.5

Impact of considering population exposure* – Primary PM2.5

Conclusions (1) No regional exceedances of 40 μgm-3 PM10 predicted in 2010 Few regional exceedances of 25 μgm-3 PM2.5 predicted in 2010 Secondary particulates dominate all source PM concentrations Emissions from agriculture dominate secondary concentrations Primary concentrations dominated by emissions from transport, residential and industrial sectors Average Energy sector contributions to concentrations are relatively low (11.2% to total PM2.5, 11.9% to secondary and 6.4% to primary)

Conclusions (2) Toxic fraction is critical to abatement strategy Current toxicological evidence strongest for primary components Greatest health benefits from tackling residential, transport & industrial sectors? Exposure approach places even greater emphasis on residential and transport sectors where exposure/tonne emitted values are highest Delivering health benefits is the key reason for understanding PM behaviour – Focus should be on delivering monitoring/modelling which can improve epidemiological and toxicological understanding

Any questions? Further information from: stephen.griffiths@eon-uk.com