3 Predicting NO2 from NOx
3.1 Introduction

Nitrogen dioxide is often described as a secondary pollutant because the majority of ambient NO2 is formed by oxidation of NO that has been emitted into the atmosphere. An understanding the oxidation processes that lead to the current measured concentrations of NO2 and how these processes are likely to be influenced by changes in future emissions are therefore essential for predicting future concentrations of NO2.

One of the main mechanisms by which NO is oxidised to NO2 is by reaction with O3. If NOx emissions are reduced and O3 concentrations remain approximately unchanged then this will increase the proportion of NOx emissions that will be rapidly converted to NO2. The trimolecular reaction of NO with O2 is thought to be one of the dominant oxidation routes on rare winter smog episode days (Bower et al, 1994). The rate of this reaction depends on the square of the NO concentration and a reduction in NOx emissions to about half of the current values on these types of days will therefore tend to significantly reduce the efficiency of this pathway. Winter smog episode concentrations of NO2 are therefore expected to be much reduced in 2005 and 2009 and this is the reason why predictions in this report are focused on annual means.

Methods for predicting future NO2 concentrations from predictions of NOx are currently the subject of considerable research, because of the importance of these predictions within the review of the NAQS. The methods that have been used to predict background concentrations of NO2 in this current report are not fully consistent with the methods that have been applied at the roadside. Projections of background NO2 for 2005 and 2009 have been derived from maps of 1996 NO2 concentrations. Projections of roadside NO2 concentrations for 2005 and 2009 have been derived from projections of roadside NOx concentrations for these years. This inconsistency will be fully addressed in future national modelling studies.


3.2 Background locations

Figure 3.1 shows a comparison of measured annual mean concentrations of NOx and NO2 for both background and roadside monitoring locations in 1997.
Estimates of NOx emissions from low levels sources have been used to derive the local source contribution to current annual mean background concentrations of both NOx and NO2 (see section 4). Background NO2 concentrations in urban areas are determined by a combination of the availability of NOx and the efficiency of the oxidation process. Our coefficient (km) for the relationship between ambient concentrations and local emissions for NO2 measurements and NOx emissions estimates for 1996. The efficiency of the oxidation process has therefore been implicitly included within the coefficient. The resulting implied linear relationship between annual mean NOx and NO2 concentrations is shown by the dotted line on figure 3.1. It is clear, however, that this relationship between NOx emissions and NO2 concentration is unlikely to remain unchanged as NOx emissions are reduced.

Table 3.1 shows the reductions in annual mean NO2 concentrations that have been used to calculate the maps of concentrations in 2005 and 2009 from the 1996 maps.

Table 3.1 Factors used to estimate annual mean background NO2 concentrations in 2005 and 2009 from 1996 values

	2005	2009
background NO2, relative to 1996	0.70	0.62

These factors have been derived from the results of a detailed modelling study of the effect of NOx emission reductions on NO2 concentrations in London (Derwent, pers comm). Hourly average NO2 concentrations for 2005 were modelled by assuming that NOx emissions will reduce to 50% of current levels. The key results of this study were that annual mean NO2 concentrations are likely to fall to 70% of current values but winter peak NO2 concentrations are likely to fall to less than 50% of current values.

In summary, background NO2 concentrations in 2005 and 2009 have been calculated by applying effective reduction factors for NO2 relative to 1996, to maps for 1996, which were calculated by assuming a linear relationship between NO2 concentrations and local NOx emissions. The solid line on Figure 3.1 shows an alternative non-linear function that has been fitted to the measurement data for background sites. This curve predicts lower NO2 concentrations at high NOx. The advantage of using this type of non-linear curve is that future NO2 concentrations can be directly predicted from NOx predictions by assuming that the curve will remain the same in future years. Further work will be carried out to validate this approach and is likely to lead to estimates of lower background concentrations in areas such as central London. The dotted line on figure 3.1 indicates that the approach we have adopted in the current work provides a worst case (or slightly worse than worst case) estimate of current NO2 concentrations, which have then been projected forwards to 2005 and 2009.


3.3 Roadside locations

Figure 3.1 shows that roadside NO2 concentration are generally lower than background concentrations for the same measured NOx concentration. This is because of the limited time that is available for NO to be oxidised to NO2 at the roadside and the limited amount of ozone that may be available in the roadside environment. The dashed line shows the non-linear function that has been fitted to current roadside measurement data. This function can then be used to estimate annual mean NO2 concentrations from annual mean NOx concentrations for 1996, 2005 and 2009. The function has been chosen to be consistent with the modelling work of Derwent (pers comm) in that a 50% reduction in NOx emissions and concentrations from 1996 values will lead to an reduction in NO2 concentrations to approximately 70% of 1996 values.