The lockdown and air quality: nitrogen-dioxide pollution falls in France’s largest cities

Assessing and quantifying the impact of the lockdown measures is no easy feat, because air quality is linked not just to emissions of pollutants accumulating in the atmosphere, but also to weather conditions, which can help to disperse them to a greater or lesser degree. Therefore, pollution levels vary from day to day. Due to variability in the weather, it has not been possible to establish “benchmark values” which the observations can be compared against in order to quantify the effects of the restriction measures.
However, modelling may provide these answers. In order to do this, Ineris has conducted analysis based on the simulation tools in PREV’air’s national air quality forecasting system (www.prevair.org). This analysis has found that concentrations of nitrogen dioxide, predominately emitted by road traffic and industrial operations, were approximately 50% lower on average than their expected levels in France’s largest cities.

How has the numerical simulation led to this conclusion?
The analysis has involved comparing air quality model forecasts based on the database of average pollutant emissions (which therefore does not take into account the decreases caused by the lockdown measures), with adjusted simulations carried out a posteriori which incorporate the observation data approved by the Official Air Quality Monitoring Associations (AASQA).
The curve below displays this comparison, established as an average across France’s 100 largest cities since March 2020. It shows that the forecasts and the adjusted simulations up to 12 March 2020 strongly match. Beyond that date, the curves diverge, showing that the emissions data used in the forecasting model no longer reflect the actual situation. Based on the findings, levels of background concentrations of nitrogen dioxide (NO2) reached a maximum of 25 µg/m3 on 18 March 2020, while a level of 50 µg/m3 was forecasted by the model (not taking into account the impact of the lockdown).
In short, the lockdown has therefore resulted in nitrogen dioxide pollution in France’s 100 largest cities halving.

Figure 1: average background concentrations of nitrogen dioxide (µg/m3) across France’s 100 largest cities. The black curve indicates the forecasted concentrations, without taking into account the impact of the lockdown measures. The red curve is an adjusted simulation, taking into account the measures observed on the ground. The discrepancy after 12 March 2020 is due to the lockdown measures, which have helped to halve nitrogen dioxide pollution.

In addition, this effect is evident at a national level when mapping NO2 discrepancies (in µg/m3) between observations and forecasts, as averages between 15 and 24 March 2020.
Understandably, this map shows a more pronounced difference in urban areas, where nitrogen dioxide emissions are normally higher.
As a result, the reduced level of human activity over the lockdown period has led to background concentrations of nitrogen dioxide falling from 20 to 10 µg/m3 in many urban centres, which equates to a fall of approximately 50% in cities.
This conclusion cannot yet be extended to fine particles. This is because particle concentrations in the atmosphere are due to far more complex chemical transformations than the chemical transformations that affect concentrations of nitrogen dioxide (see info box below). They also involve other sources which will not experience a fall in their emissions during the lockdown, such as agriculture and residential heating.

Figure 2: discrepancies between background concentrations of NO2 (µg/m3) forecasted and observed between 15 and 24 March 2020. Steep falls have been detected across the entire territory, particularly in urban areas.

Sources of nitrogen dioxide
Nitrogen dioxide (NO2) is a pollutant with a short lifespan: it generally stays in the atmosphere for less than a day, before settling or interacting with other gases in the atmosphere. Therefore, concentrations of NO2 in the air are affected by local sources, such as transport, energy production, household heating and industrial operations, as far as sources of human origin are concerned. These emissions vary significantly, based on the times of day and the days of the week.