Executive Summary

The first Air Quality Expert Group (AQEG) report, Nitrogen Dioxide in the United Kingdom, reported an analysis of trends in the annual mean concentrations of nitrogen dioxide (NO₂) and nitrogen oxides (NO_x)1 until 2001. Urban concentrations of NO_x showed a downward trend since the early 1990s, associated mainly with reduced emissions from road traffic. While the NO₂ concentration also declined, its decrease was not so marked as that of nitric oxide (NO), so that the ratio NO₂/NO_x showed an increase. NO_x is emitted from combustion sources, such as road vehicles, mainly as NO, which reacts with ozone (O₃) in the atmosphere to form NO₂. Some NO₂ is directly emitted, but there was evidence that this fraction was small (~5%). The partitioning in the atmosphere between NO and NO₂ changes as the concentration of NO_x falls, so that an increase in NO₂/NO_x would be expected, but the extent of the observed upward trend exceeded these expectations, indicating that additional changes were taking place. Three possible causes were identified: (a) an increase in emissions of primary NO₂, associated with increased market penetration of diesel cars, (b) an increase in primary NO₂ associated with the fitting of some mobile source pollution control devices, such as catalytically regenerative traps that had been retro-fitted to London buses, (c) an increase in the hemispheric background O₃ concentration, leading to increased oxidation of NO to NO₂ in the atmosphere. This report examines information on emissions from road vehicles and data from monitoring sites. In addition analyses using several different models have been used to try to identify the source of the increasing NO₂/NO_x concentration ratio.

Emissions measurements are difficult to interpret quantitatively, because the number of measurements on different vehicle types is relatively limited. The measurements indicate that the fraction of NO_x emitted as NO₂ (f-NO₂) is considerably in excess of 5%, with values in the range 20 – 70% for Euro III diesel cars. Heavy duty vehicles and buses show a smaller f-NO₂, but the fitting of diesel particulate filters to buses substantially increases the fraction of NO₂ in vehicle exhaust gases. In 2005, over 90% of buses in London were operating with such traps.

Estimates of emissions to 2010 have been made both for London and for other UK urban areas, using projections of the fleet mix and estimates of emissions of both NO_x and NO₂ from the different vehicle types. These calculations show a continuing fall in NO_x emissions but almost constant, or even slightly increasing, values for NO₂ emissions over the period 2002 – 2005 and slight falls in NO₂ emissions over the period 2005 – 2010.

Monitoring data in London over the period 2002 – 2005 show greater than anticipated increases in the annual mean NO₂/NO_x concentration ratio at 44 roadside and kerbside sites and at 24 out of 28 urban background sites in the London Air Quality Network. Increases were also observed at a number of Defra Automatic Urban and Rural Network (AURN) roadside and kerbside sites outside London and at a few urban background and urban centre sites. There are also, however, a significant number of sites where no clear trend was observed. It has not proved possible to rationalise these differences, because of inadequate quantitative information on vehicle emissions. Similar differences were also observed in peak hourly mean NO₂ concentrations, with some sites showing substantial increases in the number of exceedences of 200 μg m-3 and others showing no clear trend.

AQEG undertook computer modelling studies for London using five different models to try to understand the observed behaviour. The models all showed that the observed increases in the NO₂/NO_x concentration ratio could be explained by a 5% absolute increase in the percentage of NO_x emitted as NO₂ (e.g. from 10% to 15%) over the period 2002 – 2005. This increase could be due to increased penetration of Euro-III diesel vehicles fitted with oxidation catalysts and/or to the fitting of catalytically regenerative particle traps to London vehicles. It is difficult to ascribe the increases solely to increased penetration of Euro-III diesel vehicles given the observation of increases in the NO₂/NO_x concentration ratio at only some roadside and kerbside sites outside London. It may be that London is particularly sensitive to direct NO₂ emissions, because of its size and emission density, but it is difficult, on this basis, to understand why the concentration ratio behaves so differently in, for example, Glasgow. A statistical analysis of the observations at the London Marylebone Road site, and comparison with the timing of the retrofit scheme for London buses, shows that the fitting of catalytically regenerative particle traps provides a plausible explanation of the observations. Particulate matter (PM) monitoring data at Marylebone Road show strong decreases in PM_2.5, but only a slight decline in PM₁₀, indicating increases in PM_coarse. The statistical analysis shows that these changes also coincided with the timing of the retrofit scheme. AQEG has no definitive explanation of the behaviour of PMcoarse but considered that it could be associated with the introduction for more and heavier buses resulting in an increase in non exhaust particle emissions.

Other possible explanations for the observed changes in NO₂/NO_x concentration ratio were also investigated, including increases in the background concentration of O₃, and were ruled out. It was not possible to eliminate direct emissions of nitrous acid (HONO), although direct NO₂ emission provides the more plausible explanation. It is recommended that direct emissions of HONO are further investigated, especially since it is plausible that this has contributed to PM_coarse concentration changes.

Projections of future concentrations of NO₂ show that the projected increases in the primary NO₂ emission percentage between 2004 and 2010 are likely to increase the extent of exceedences of an annual mean NO₂ concentration of 40 μg m^-3 relative to projections based on no change in primary NO₂ percentage. However, reductions in total NO_X emissions are expected to result in a reduction in the extent of exceedences in 2010 relative to 2004. It is unclear exactly how this balance between NO_X emission reductions and increases in primary NO₂ percentages will influence the achievement of objectives and limit values in 2010 and beyond, because of the uncertainty associated with estimates of future emissions. In view of these difficulties, AQEG recommends a separate emissions inventory for NO₂, to be updated regularly as more data become available, within the National Atmospheric Emissions Inventory to facilitate more reliable predictions of NO₂ concentrations.