View all publications

Appendix C: Estimation of Nitrogen Dioxide Concentrations from Modelled NOx

C1 Introduction

Nitrogen oxides are emitted mainly in the form of nitric oxide (NO), but once released into the atmosphere are oxidised to the more toxic nitrogen dioxide (NO2). The predominant short-term transformation process is the reaction of nitric oxide with ambient ozone to form nitrogen dioxide:

NO + O3 - NO2 + O2

Since the reaction is a 1 to 1 transformation that does not affect total NOx concentrations, the maximum extent of conversion of NO to NO2 that can be expected in the emission plume is directly related to the maximum ambient concentration of ozone. Photo-dissociation of NO2 to re-form NO and ozone decreases the concentrations of NO2 to some degree, but is ignored here.

Ozone concentrations in air coming off the ocean are quite reproducible and show a seasonal variation, with the highest concentrations occurring during winter, at about 35 ppb at the Baring Head site. Summer concentrations are about 20 ppb. The maximum ozone concentration is sufficient to produce 72 µg/m3 of nitrogen dioxide by oxidation of nitric oxide. This information provides the basis for a simple method for calculating generally conservative estimates of nitrogen dioxide concentrations from modelled NOx concentrations.

C2 Relationship to the US EPA Ozone Limiting Method

This method of calculating nitrogen dioxide concentrations is related to the Ozone Limiting Method (OLM) available from the US EPA SCRAM website. The OLM describes only the modelling or calculation of annual average nitrogen dioxide concentrations, presumably because the US National Ambient Air Quality Standards include only an annual average standard of 100 µg/m3 for nitrogen dioxide. However, the most sophisticated approach described under the OLM does use one-hour average modelling, together with one-hour average meteorological, ozone and nitrogen dioxide data, to calculate annual averages.

The OLM model requires simultaneously recorded one-hour average meteorological data, ozone and nitrogen dioxide concentrations extending over at least one year, with at least 90% valid data. Apart from the expense of obtaining such information at a single location, there are significant problems in locating the monitoring site relative to existing emission sources and the proposed new emission source. The US EPA guidance recommends two or more monitoring sites because of the perceived difficulty of accounting for scavenging of ozone by nitric oxide.

The method described here avoids the difficulties perceived in the OLM, and the ozone and NOx monitoring requirement, by recognising that for a small island country such as New Zealand, significant photochemical production of ozone is relatively rare at most locations and will seldom need to be taken into account when considering the local effects of NOx emissions. The combination of this situation with the stable, but seasonally varying, concentrations of ozone in air moving onto New Zealand off the oceans means that realistic, but conservative, estimates of concentrations can be based on these ozone concentrations, the total NOx concentration at the site under consideration and the percentage of nitrogen dioxide in the NOx in emissions affecting that site.

C3 The calculation methodology

For cumulative NOx concentrations less than 80 µg/m3 (expressed as nitrogen dioxide) all of the NOx is considered to be present as nitrogen dioxide. The 80 µg/m3 corresponds to 8 µg/m3 of nitrogen dioxide resulting from a default percentage of 10% of nitrogen dioxide in emitted NOx plus 72 µg/m3 of nitrogen dioxide formed by oxidation of nitric oxide by ozone.

For cumulative NOx concentrations above 80 µg/m3, the nitrogen dioxide concentrations are calculated as follows:

[NO2]cum max = 72 + [NOx]bkgrd tot x %NO2 bkgrd + [NOx]emiss x %NOx emiss

where:

  • [NO2]cum max is the maximum estimate of total cumulative NO2 from both background NOx and the additional emission under consideration
  • [NOx]bkgrd tot is the total background NOx concentration in the receiving air
  • %NO2bkgrd is the percentage of nitrogen dioxide in the NOx emitted from the sources contributing to the background levels of NOx
  • [NOx]emiss is the concentration of NOx at the receptor originating from the emission
  • %NOxemiss is the percentage of nitrogen dioxide in the NOx emitted from the source under consideration.

If either %NO2bkgrd or %NOxemiss are not known, the default percentage of 10% used in the OLM is probably the best (although often highly conservative) choice.

If the percentages of nitrogen dioxide in the emissions are not known, or happen to be 10%, the expression above simplifies to:

[NO2]cum max = 72 + [NOx]cum tot x 10%

where:

  • [NOx]cum tot is the cumulative total NOx concentration including both background NOx and the NOx concentration increment at the receptor resulting from the emission under consideration.

C4 Common situations resulting in conservative predictions

There are common situations where these calculations give substantial overestimates of nitrogen dioxide concentrations, including:

  • during the day, when the photochemical equilibrium reverses the oxidation of nitric oxide by ozone to some degree
  • during stable atmospheric conditions, particularly at night, when both nitrogen dioxide and ozone are removed by reaction with vegetation and other surfaces.

C5 Validation against monitoring data

Use of this method to calculate nitrogen dioxide concentrations from NOx concentrations measured at several monitoring sites in both Auckland and Christchurch shows that taking the percentage of nitrogen dioxide in NOx emissions contributing to the measured concentrations as 10% is highly conservative and that 5% still predicts nitrogen dioxide concentrations higher than any reliable measured concentrations. This may reflect the combined effect of the true percentage of nitrogen dioxide in NOx emissions being less than 10%, together with scavenging of both nitrogen dioxide and ozone by vegetation, plus the photochemical equilibrium.

C6 Caution re application of the method to some types of locations

The Khyber Pass site in Auckland is unique among New Zealand NOx monitoring sites in that nitrogen dioxide concentrations measured there quite frequently exceed the predictions using this method. After 7 pm and before 7 am, there are only a few hourly averages during the year when the measured nitrogen dioxide concentrations exceed those predicted using 10% as the percentage of nitrogen dioxide in NOx emissions, and essentially no measured concentrations exceed those predicted using 15%. However, during the remainder of the day, there are a considerable number of measured concentrations that exceed those predicted using 15% as the percentage of nitrogen dioxide in NOx emissions. This is in marked contrast to both the Penrose and Takapuna monitoring sites, both of which are dominated by traffic emissions, as is the Khyber Pass site, but which show few if any measured concentrations exceeding those calculated using 5% of nitrogen dioxide in NOx emissions.

The frequent, unusually high nitrogen dioxide concentrations at the Khyber Pass site may be associated with some of the following features of the site:

  • the kerbside location alongside an uphill queue for traffic lights, with uphill exits in all directions
  • the absence of any vegetation between the location of the emissions and the monitoring intake
  • the relatively confined nature of the location, which may mean that relatively high concentrations of both reactive organic compounds and NOx from vehicle emissions may move away only slowly, so that there may be the possibility of photochemical reactions (this is consistent with most of the unexpectedly high nitrogen dioxide concentrations occurring during the day)
  • the proximity of the site to the busiest section of the northern motorway, which may also contribute to relatively long residence times of high concentrations of reactive organic compounds and NOx in the area
  • the relatively high proportion of diesel vehicles among the traffic flow, which may emit high percentages of nitrogen dioxide in NOx emissions at idle, immediately adjacent to the monitoring intake, while waiting for the traffic lights.

The calculation method described here should be used only with great caution at locations where NOx concentrations originate predominantly from vehicle emissions and which share some of the other features described above for the Khyber Pass site. Ideally, some monitoring data should be obtained to determine whether unexpectedly high concentrations occur at these locations.