Air pollution and COVID-19: is ozone an undercover criminal?


The global COVID-19 lockdown is undoubtedly resulting in curiously low levels of air pollution. Although it might seem inappropriate to seek a silver lining during a global pandemic, the fact that the air really does seem cleaner gives my PhD topic a little more everyday credibility, which – at least for me – is quite nice.

You may have already seen satellite pictures of the effects of the lockdown in northern Italy (and other major cities) on surface nitrogen dioxide (NO2) concentrations. You may have been able to breathe some cleaner air where you live these past few weeks. It feels like we are in the middle of some sort of major air quality experiment, some kind of simulation conducted by a clueless PhD student…

What you might not notice, however, is the rise in near-surface ozone (O3) during a string of warm, sunny days. While it isn’t a primary pollutant like NO2 or particulate matter, O3 is closely associated with the amount of NOx (NO + NO2) in the air. It is also invisible to the naked eye – unless it forms photochemical smog. O3 can be harmful in short bursts of elevated concentrations to people who already suffer with asthma and other respiratory problems, which could prove to be problematic since COVID-19 is itself a respiratory disease.

Weather conditions in Reading throughout the majority of April were favourable for O3 production: lots of solar radiation and weak winds. In fact, Reading experienced its sunniest April on record, along with some of the warmest April days on record. It is therefore not surprising that peak daytime concentrations of O3 creep up over a week of warm, calm weather. For example, a measuring site located between two busy roads in Reading gives a clear indication of what the mixture of favourable conditions alongside low NOx emissions can do: Figure 1 shows that peak daytime concentrations rose everyday between 02/04 – 12/04, when the air was stagnant and thus O3 tended to accumulate within the atmospheric boundary layer. The peak concentrations between 08/04 and 12/04 are typical of “moderate” levels on DEFRA’s Daily Air Quality Index (DAQI), and are close to the WHO safe concentration exceedance guidelines.

Figure 1: From “Air Quality in Reading during COVID-19”, by Helen Dacre [4]. Hourly measured Ozone concentrations at Reading New Town from 1 March 2020 to 15 April 2020. Date of social distancing implementation on 16 March 2020 (magenta dashed) and non-essential travel restrictions on 23 March 2020 (black dashed). Data from

The DAQI is dictated by the highest concentration of any one of the five pollutants deemed harmful to human health: ozone, nitrogen dioxide (NO2), sulphur dioxide (SO2), and PM2.5 / PM10 (particulate matter). Figure 2 shows a moderate DAQI in parts of south and north-east England (and in fact, the map looked a lot more yellow and orange on Friday 04/04).

Figure 2: Daily Air Quality Index (DAQI) in the UK, for Saturday 11/04/2020. Captured from

So although there is a clear trend in unseasonably low NOx concentrations in major cities in many parts of the world (including the UK), why can O3 concentrations rise?

The answer is probably that ozone has a “love-hate” relationship with NOx, Volatile Organic Compunds (VOCs) and the weather. It skyrockets when it’s sunny and skies are blue. High pressure systems and calm weather trap much of the existing ozone within the boundary layer, near to the ground. In particular, if easterly / southerly winds prevail, they can transport both ozone and its precursors from the continent – this often happens when there is an anticyclone over the UK. Therefore, high ozone episodes tend to occur in the spring / summer, due to the frequency of such ozone-favourable conditions.

Fossil fuel combustion releases NOx, some of which is in the form of NO and goes on to oxidise to create more NO2, or it can react with VOCs, or it can directly react with ozone. The usually abundant NO and other VOCs from vehicle emissions and industry are now significantly lower than usual, so the process of ozone scavenging by NO is minimised.

On top of that, NO has a short lifetime (maybe a few minutes), and can quickly oxidise to form NO2, which has a longer lifetime and can therefore travel on to rural areas. Often, rural regions will have higher average ozone concentrations than cities (which might seem counter-intuitive!). Although the emissions in those areas are lower, they can experience net ozone production from the additional NO2 which has travelled downwind from a nearby city.  In relatively clean tropospheric air, the production / destruction of ozone is closely linked to the ratio of NO to NO2 – an equilibrium known as the photostationary state (Leighton, 1961) – and there are some studies to show a negative correlation between annual mean NOx and O3 measurements in both rural and urban areas (e.g. Bower et al., 1989). But none of this is particularly simple, because there will always be VOCs present in air, and ozone production / destruction is also highly sensitive to the ratio of NO : VOC – this was not fully understood until Greiner (1967) and several subsequent studies, which explained the role of the hydroxyl radical OH in the reaction chain to create NO2 without destroying ozone. Another phenomenon is the ‘weekend effect’, where weekday emissions tend to be quite different from emissions during the weekend because there is no morning / evening rush-hour traffic and resultant NO (Seguel et al., 2012). If VOC levels remain high, ozone production is favoured.

Figure 3: From Bower et al., 1989. Points on the scatterplot represent mean annual NOx and winter mean O3 measurements taken at 8 rural and 2 urban background sites in the UK. Pearson correlation coefficient r = -0.91.

Let us return to the present day. How might the weather conditions affect the delicate balance between NOx, VOCs and ozone? And what about other particles closely monitored throughout the pandemic, such as particulate matter (PM)?

February was unusually wet and windy in the UK. Strong winds can disperse both NO2 and PM, while rain is an efficient sink of PM by physically washing out the particles. Both pollutants have been monitored closely at a number of locations globally over the past few weeks, as they are good indicators of emissions (and ozone is not). Before the gloriously sunny weather came, I wondered about ways of distinguishing between causes of the unusually low NO2 / PM concentrations: what proportion is attributable to the lockdown, and what is attributable to a very wet and windy February / March period in this region? How might ambient ozone concentrations change as we move into the summer, as lockdown measures might begin to gradually relax and pollution returns to pre-lockdown levels? And what does this mean for people who are vulnerable to respiratory issues aggravated by ozone? All these questions – and many more – are currently being explored by air quality experts all over the world, hopefully reaching some conclusions in time for us all to act on them timely and appropriately.

Stay home, stay safe, and thank you for reading. Please leave a comment or send an email if you have any questions (I’ll be happy to answer) or corrections: I am a PhD student and there are probably still some gaps in my understanding.

For further reading: The Copernicus Atmosphere Monitoring Service (CAMS) are providing vital satellite observations of interest to COVID-19 matters, which I encourage you to check out if you are interested in the air quality aspects of the pandemic.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s