Air Quality in Reading during COVID-19

By Helen Dacre

I’ve been working from home for exactly a month now and like everyone else have been adapting to a new routine.  This involves taking a daily shuffle around Caversham to get some exercise. There’s been a noticeable lack of traffic on the roads which makes my daily shuffle a lot more enjoyable.  This got me thinking about the effect of the current travel restrictions on air pollution. If there are fewer cars on the road emitting pollutants, then perhaps air quality may have changed?

Air pollution measurements are taken routinely at a network of over 100 monitoring sites across the UK. One of these sites is located in the centre of Reading at Cemetery Junction which sits between 2 busy roads, Wokingham Road and London Road.  It’s been taking measurements since 2003.  The data from this station (Reading New Town) is freely available from the Defra website.

Figure 1: 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

So, my first port of call was to take a look at the data from the Reading New Town monitoring station.  Figure 1 shows the hourly ozone measurements between 1 March and 15 April 2020.   The magenta dashed line shows the date on which social distancing was recommended (16 March) and the black dashed line shows the date on which non-essential travel restrictions were enforced, the so-called “lockdown” (23 March).  As you can see, there’s lots of variability in the data including a strong diurnal cycle.  There does appear to be an increase in ozone towards the end of the timeseries but it’s difficult to say whether this is due to changes in emissions of ozone precursors or due to changes in the meteorology.

The amount of ozone formed depends on the concentrations of other substances present in the air, such nitrogen oxides (NOx) and hydrocarbons. The concentration of these substances tends to be higher in polluted air, so we expect ozone concentrations to be lower when NOx is higher. However, NOx concentrations also tend to be higher when meteorological conditions are such that atmospheric dispersion is less efficient. These conditions are often associated with sunny high pressure, such as that we experienced last week. Therefore, meteorology plays a large role in determining the concentration of ozone, which is one of the research topics that I’m interested in.

So, if we want to find out what’s driving the increase in ozone, we need to work out how to remove the variations that are due to changes in the weather.  The best way to do this is to build a statistical model to predict the ozone concentrations using meteorological variables and other inputs. So that’s what I’ve done.  As inputs to my model, I used 12 years of data (2008-2020) including wind speed, wind direction, temperature, time of day, day of the week, Julian day and the date. My model predicts what we might expect ozone concentrations to be during the current meteorological conditions.

Figure 2: Hourly measured Ozone concentrations (grey), 24-hour moving averaged measured Ozone concentrations (blue) and predicted Ozone concentrations (red) at Reading New Town from 1 March 2020 to 15 April 2020.

Figure 2 shows the observations, with a 24-hour moving average applied (blue) and my model predictions for the same period (red).  Surprisingly, my simple model captures the ozone variability in the period prior to lockdown quite well. After lockdown it appears that the measured ozone is higher than my model predictions, possibly indicating the effect of reduced NOx emissions?

Figure 3: Accumulated difference between measured concentrations and predicted concentrations from 1 March 2020. Ozone (left) and NOx (right)

To emphasise the differences, I also plotted the accumulated difference between my model prediction and the observed ozone concentrations from the 1 March, shown in Figure 3 (left).  The accumulated difference is initially small since the over or underestimations predicted by my model (which is far from perfect) cancel each other out.  However, after lockdown there’s a steep increase in the difference indicating that ozone is above that expected, possibly due to a reduction in NOx.

To see if this increase in ozone is due to a reduction in NOx emissions, I also built a statistical model to predict NOx concentrations. My model for NOx isn’t quite as good as that for ozone because NOx has much larger extremes.  But the accumulated difference plot shown in Figure 3 (right), does show the opposite behaviour to that for ozone; i.e. that my model overpredicts the observed NOx after the 16 March 2020, when social distancing was introduced.

NOx contains NO2 which is bad for human health, particularly for those with asthma. It increases the likelihood of respiratory problems and can cause wheezing, coughing and bronchitis.  So, the evidence suggests that NOx emissions have been decreasing during the COVID-19 travel restriction period which is good news for my daily shuffle. There’s plenty of analysis still to be done to see if these results are robust across other sites in the UK, but for now enjoy the peace and quiet on the roads and stay safe and well.

This entry was posted in Air quality, Boundary layer, Climate. Bookmark the permalink.