Laboratory versus real world: Discrepancies in NOx emissions in the EU

Last week we published the new edition of our pocketbook of European Vehicle Market Statistics. For those interested in road vehicles and their impact on our climate and air quality, the pocketbook is full of interesting stories, told by numerous charts and tables in a new design with double-page info-charts. We take it as a compliment to our work that publication of the pocketbook was quickly picked up by the media.

The online version of Germany’s weekly Die Zeit focused on two charts that demonstrate how the introduction of the mandatory EU-wide carbon dioxide (CO2) regulation in 2009 speeded up emission reductions of manufacturers, and how close most of the car makers are already today in meeting their emission targets for 2015, something that we had discussed in an earlier blog post here.

One important story that has not been picked up yet relates not so much to CO2 as to air pollutants, and in particular nitrogen oxide (NOx). As illustrated in the graphic below (which appears at the beginning of Chapter 6 of the pocketbook), NOx emissions of gasoline cars in the EU have decreased significantly since 2000, from about 0.2 grams per kilometer (g/km) to 0.05 g/km. This corresponds quite well with the Euro emission limits, which were adapted from 0.15 g/km to 0.06 g/km in the same time period. The Euro emission limits regulate how much specific pollutants, such as NOx, may be emitted by a car when it is tested under laboratory conditions and using a specific driving cycle. In the case of gasoline vehicles, the NOx emissions measured in the laboratory are fairly well in line with the level of emissions measured on-road, i.e., when driving the car under real-world conditions on a real road.

NOx emissions from gasoline and diesel cars

This, however, is not the case for diesel cars. Diesel vehicles in the EU are allowed a much higher NOx emission level than gasoline cars. In 2000, when the Euro 3 standard was introduced, the allowed level was 0.5 g/km, more than twice as much as for gasoline vehicles. Yet, as vehicle tests show, even back then the real on-road emission levels were closer to 1.0 g/km, i.e., much more than actually allowed by the standard. Still, the vehicles received their type-approval and could be sold, as the Euro emission standards have to be met under laboratory conditions only. Over time, emission limits got stricter, and the current Euro 5 emission standard sets a limit of 0.18 g/km for NOx diesel emissions. This is still more than three times as high as for gasoline vehicles, but of course much lower than back in 2000. However, recent research suggests that the on-road emissions did not really change at all during the last decade. The values measured are in the range of 0.8 g/km, only 20% lower than in 2000 and more than four times higher than allowed by the Euro 5 emission limit.

The on-road emission data comes from a study carried out recently by King’s College London and the University of Leeds for the UK government. In total, emissions data from more than 80,000 vehicles were analyzed, and the authors conclude: “In the case of light duty diesel vehicles it is found that NOx emissions have changed little over 20 years or so over a period when the proportion of directly emitted NO2 has increased substantially”. The UK study is not the only one arriving at these findings. A recently published paper by researchers from the European Commission’s Joint Research Centre (JRC) and IFEU institute summarizes on-road test results for a number of vehicles and comes to the conclusion: “The on-road NOx emissions of diesel cars, furthermore, appear to exceed substantially applicable emissions standards.” The authors even go one step further by raising doubts about whether the new Euro 6 emission standard, to be introduced in 2014, will solve the problem. It says: “Still, all tested cars, including the Euro 6 diesel car, exceed their NOx emissions standards on the road by 260 ±130%”.*

Why is it a problem if NOx levels are high? Well, not only is NO2 by itself known to cause inflammation, NOx emissions are a precursor to ground level ozone, which is known to affect lung function in some and increase asthma attacks. This is why in the EU air quality limits were set in order to protect the health of the people living in those cities. The increasing number of diesel vehicles and their failure to reduce NOx under real-world driving conditions contributes to a situation where many cities are struggling to meet their air quality targets, and where citizens are exposed to higher NOx emission levels than they actually should according to EU law.

If you are following our work, you will notice that this sounds familiar. In fact, a similar situation was discovered for fuel consumption and CO2, where the discrepancy between laboratory values and on-road values has also increased in recent years for both gasoline and diesel vehicles.

On the NOx issue, the problem has been recognized by the European Commission and Member States, and work is ongoing to reduce the discrepancy between type approval results and in-use emissions. One option currently discussed is to use so-called Portable Emissions Measurement Equipment (PEMS) in the car to measure emission levels while driving on the road, and to include this type of real-world testing in the official type approval process for vehicles that are new on the market. Another option under discussion is to introduce a test cycle that is put together randomly from a set of different driving patterns and used for the laboratory testing of the type approval process. This would, ideally, lead to a situation where it becomes more difficult to optimize the engine of a vehicle to a certain fixed test cycle and where manufacturers would be required to make sure that their vehicles meet emission limits under a broader range of driving conditions. While both of these options would be beneficial, some experts doubt they would adequately address the multiple causes of the discrepancy, and point to the system in the US as a best-practice example. There, a set of several fixed test cycles is used to ensure compliance with emission regulation under various driving conditions, and enforcement procedures exist to ensure data vehicles are representative of production vehicles.

It is expected that in 2013 the European Commission and Member States will select one or several options to help reduce the existing problems. When these fixes will be introduced is another question, and so there is a risk that we will have to live with high real-world NOx emission limits for the next few years, and that the fixes will not go far enough. In the meantime, we Europeans may look jealously at California: there, emission limits for gasoline and diesel cars are identical, and about 75% lower than what is currently allowed for Euro 5 diesel cars in Europe. Furthermore, California also requires testing under additional driving cycles, and has greater durability requirements and in-use enforcement than Europe, resulting in better real-world emission performance. How much better exactly will be in the focus of an upcoming research project comparing on-road emissions of diesel cars in the EU and California.

*Update 14 Nov 2012: This paragraph was modified to clarify that the second paper referenced is not an official JRC report, and consequently does not necessarily present the position of the European Commission.

Tracking progress