Research and analysis

Air quality onboard trains: a response to the Rail Safety and Standards Board report

Published 5 June 2023

This report is a response to the Rail Safety and Standards Board (RSSB) report on the Analysis of air quality on board trains (T1188).

The content of this report is the responsibility of the DfT Science Advisory Council alone and does not represent the views of the Department for Transport (DfT) or DfT ministers.

Background context

Poor air quality is a major environmental contributor to disease and mortality and in the UK is estimated to lead to between 15,000 and 40,000 deaths per year^{[footnote 1]}.

It creates significant economic costs through lowered productivity and health service burden.

As outdoor air quality in the UK has improved over recent decades, greater attention is being paid to indoor environments, which can represent the bulk of an individual exposure over the course of a day. Frequently, indoors can have poorer air quality than outdoors, although this is highly variable and difficult to predict.

Air pollution and transport environments have been studied for many years and are known locations where air quality can potentially be poor. This includes at the roadside where it impacts those walking and cycling, and for passengers inside cars, buses, trains, and aircraft. Transport hub buildings such as railway and bus stations are also known to be susceptible to poor air quality.

The RSSB report T1188 evaluates air pollution found in train interiors during long-distance travel and is a welcome addition to the evidence-base on current transport exposure in the UK.

It fills some important gaps in knowledge on possible exposure of passengers during longer journeys between cities (the research focus previously being mainly on commuter travel in cities).

The pollutants examined are nitrogen dioxide (NO2) and particulate matter (PM) expressed as PM2.5 and PM10 (particles smaller than 2.5 and 10 microns diameter) and as black carbon (BC). Nitrogen dioxide derives almost exclusively from diesel engine exhaust, while PM is emitted from a wider range of sources that include engine exhaust and brake wear. PM is also released from occupants themselves, via exhaled breath and from physical agitation/dust resuspension from internal surfaces such as fabrics and flooring.

Concentrations found inside trains

The report contrasts interior concentrations measured on ~100 train journeys split over 2 studies conducted by Imperial College, London. The studies used different train types/journeys and broadly compared data with concurrent air quality observations at some urban and rural fixed monitoring stations. Such comparisons are intended to provide only a guide to relative concentrations encountered since sometimes these data sources were significant distances away from the journeys being made.

A comparison is also made with literature data on air quality parameters for enclosed spaces in other transport modes. This latter comparison is an important one for long-distance train journeys. Commonly, in-cabin concentrations of air pollution are contrasted with ambient concentrations and active travel alternatives, with the latter often now resulting in lower overall exposure.

For intermediate and long-distance train travel in the UK, the most appropriate counterfactual for exposure is with the air quality that would be experienced inside cars or buses on the same journey. Since local meteorology is a major factor in determining air quality day-to-day, without matched control journeys being conducted on the same days as the train journeys, it has not been possible to quantitatively compare effects between modes.

In this study air quality observations were collected with 1-minute data averaging, a frequency that is higher than that returned from most outdoor air quality monitoring stations. A consequence is that the train interior data showed some transient high concentrations (reported at the 95th percentile).

Care should be taken to not directly compare these values with Defra or local authority outdoor ambient monitoring data, which is time-average and smoothed. There are some comparisons made of one-minute peak train values to ambient outdoor annual means. Superficially this makes the train carriage appear highly polluted compared to outside, but it is comparing very different timescales for exposure.

The most useful metric for comparison is likely to be the journey mean concentration, which in this study typically represented 1-3 hours spent inside the train. It may be valuable in future to produce a metric of ‘journey-highest’ 1-hour mean to give a concentration that can be mapped more directly onto UK and international indoor air quality guidelines, such as those from WHO and UKHSA. Short-term indoor air quality guidelines for other enclosed spaces are often expressed as 1-hour, 8-hour or 24-hour mean concentrations.

The experiments show that the type of train had an impact on the interior air quality that was experienced, although the reports do not attempt to directly apportion differences observed to individual types of aftertreatment system, or other relevant factors like braking systems. This is potentially an area to be followed up in later analyses with additional expertise.

An unexpected feature are some high transient and journey mean concentrations of NO2 encountered using the Class 800 bi-mode trains when working in diesel mode. These indicated significantly elevated concentrations when averaged over the full journey length.

This seems worthy of further investigation since the Class 800 is a relatively new train that includes a selective catalytic reduction (SCR) exhaust gas aftertreatment. NO2 was also consistently elevated on other train types, such as Class 755 and 230. PM was seen at elevated concentrations (relative to outdoors) on other train types, prevalent on older trains without abatement technologies such as diesel particulate filters (DPF), but also on some that included DPF. Concentrations of PM in this study were broadly in line with the types of values reported previously in literature.

To place train interior concentrations in a wider context it is useful to compare some of the observed values with existing indoor air quality guidelines for other types of environment. The most relevant guidelines to compare against are those issued from WHO (2010). These do not map directly onto the data in the RSSB study, but a close matching time-averaged standard is nonetheless still useful to compare against the journey mean (representing roughly a 1-3 hour period). This places the journey mean period as falling somewhere between a 1-hour and 24-hour WHO guideline.

Taken from the 2 study periods the highest and second highest journey-mean concentrations are shown below and compared with the nearest matching WHO indoor air quality guideline limit.

Pollutant	Highest journey mean mg m-3	Second highest journey mean mg m-3	T1188 pt1 Average of all journeys mg m-3	WHO (2010) 1-hour indoor guideline mg m-3	WHO (2010) 24-hour indoor guideline mg m-3
NO2	409 (~3.5 hr average)	207 (~3.2 hr average)	88	200
PM2.5	63 (~1.75 hr average)	55 (~3.2 hr average)	16		25
PM10	86 (~1.75 hr average)	80 (~3.2 hr average)	22		50

The data illustrate that on a worst-case train journey, it is possible that the passenger interior air quality would exceed a close-matching indoor air quality guideline value from WHO. However, it is noted that from the ~100 train journeys undertaken only a minority experienced concentration that approached or exceeded these guidelines.

Many journeys showed train carriage air quality broadly commensurate with what might be found in a typical roadside urban outdoor environment. The average concentration from ‘all journeys’ in the T1188 part 1 study was below the WHO 1- and 24-hour indoor guidelines for NO2, PM2.5 and PM10. This is not an argument for no further action to improve interior air quality, but it places the scale of effect in a suitable context. Indeed, it would be relatively unusual for an individual to routinely spend several hours outdoors, roadside, in polluted air.

Sources of pollution on trains

The scope of the study did not include directly determining the sources of air pollution that were found inside trains or make extensive links between pollution and individual locomotive or carriage types. The report rightly notes that PM can come from many different places. It may depend on the effectiveness of the DPF used at the exhaust (if present), on the type of braking system, on the efficiency of filtration used on air intakes. PM can be released from the occupants themselves, from movement and exhaled breath.

Occupancy levels and other physical factors such as air exchange rate were not reported, so it is not possible to establish the extent to which these were significant factors. The use of source apportionment was only partially successful in trying to tease these apart PM sources using black carbon as a tracer of combustion.

The source of elevated concentrations of NO2 was more straightforward to diagnose, arising primarily from entrainment of diesel engine exhaust (or possibly also from nearby trains when in stations).

The exact causes of elevated NO2 on Class 800 trains are not explicitly examined but are likely to arise from a combination of non-ideal locations for carriage air intake (relative to exhaust systems) coupled to underperformance of exhaust gas aftertreatment. In more modern engines with selected catalytic reduction (SCR) underperformance can arise during periods when exhaust systems are cool. This can include during initial start-up, and if the exhaust system is reliant on heat output from the engine to sustain catalyst temperatures.

Engines which spend extended periods under idle or low load conditions may be particularly susceptible if no ancillary exhaust heating is applied (for example direct electric heating). SCR underperformance can also occur if an aftertreatment system is disabled for engine-protection reasons (for example at low temperatures) or due to lack of urea reagent.

It was noted that some of the highest NO2 concentrations were observed on journeys undertaken in January, and low ambient air temperatures may have exacerbated SCR underperformance.

Recommendations related to the reports

1. Guidelines for air quality during long-distance train travel should aim where possible to use appropriate averaging periods that reflect integrated exposure over the journey. Those tasked with developing recommendations for limits on air pollution onboard trains may wish to consider indoor air quality guidelines for other enclosed environments as a starting point, for example, those issued by UKHSA or wider NICE advice. Development of standards or guidelines would need to be accompanied with evidenced advice to operators on technical interventions and adaptations.

2. The study was designed to assess typical passenger exposure on single journeys and may not necessarily reflect the occupational exposure of train staff to air pollution. Working patterns may mean staff spend time in different interior areas of the train that were not sampled, and over longer periods than passengers. This may be an aspect of air quality on trains that could require more representative measurements to be made.

3. Concentrations of PM were frequently higher inside trains than would be experienced outdoors, sometimes exceeding the WHO 24-hour indoor guideline for both PM2.5 and PM10. The values reported were however broadly within the range of interior concentrations reported previously in literature. It may be worthwhile to follow up with additional analysis to identify links between PM and known causal factors such as type and effectiveness of engine aftertreatment system, type of braking system, interior air exchange rate and passenger occupancy level.

4. The elevated concentrations of NO2 experienced on the Class 800 trains require further investigation. The SCR system may not be functioning as anticipated and/or the air intakes to some passenger compartments may not be optimally configured, unintentionally drawing in engine exhaust. Poor performance can arise due to cooling during engine idle or prolonged periods of low loads. There are likely to be technical opportunities to rectify this, for example through additional direct heating of exhaust systems, as are typically implemented on most diesel road vehicles.

5. The chemical composition of PM released from friction and wear from train systems may be different from that found in ambient air. Further analysis of the chemical nature of airborne particulate matter found inside trains would support potentially targeted interventions that addressed the most toxicologically harmful components (such as trace metals), in addition to reducing overall mass of particles.

Broader issues arising

1. A net zero transition away from diesel powered trains would eliminate the risk of elevated NO2 inside carriages. Further indoor and outdoor air quality benefits arise from rail electrification (catenary or fuel cell) since all-electric trains also emit less PM when compared to diesel alternatives. Nonetheless, the performance of all future propulsion technologies and carriage ventilation systems (including positioning of air intakes) should be evaluated for their effects on interior air quality.

2. The use of active air filtration systems for cleaning carriage air may have a dual benefit of reducing exposure to PM pollution generated from the train itself and to other airborne particles such as respiratory viruses from the occupants. Filtration systems are however unlikely to be efficient at removing gaseous pollutants like NO2, which instead are most effectively abated at the point of exhaust.

3. Any long-term strategy which retains combustion propulsion in trains using alternative low carbon fuels such as hydrogen, ammonia or biodiesel, could potentially lead to continued emissions of NOx at the engine exhaust. This would necessitate the use of either very well-optimised combustion conditions (an approach that may work for H2, but likely not for biodiesel) or the continued use of exhaust gas aftertreatment to manage of NOx emissions.

4. A wider evidence issue for the department to consider is that there is a paucity of recent (meaning from the last 5 to 10 years, post EURO6) and UK-specific data available on interior air quality in passenger cars and buses. This makes it difficult to evaluate how present-day exposure to air pollution on trains might compare with present-day counterfactual journeys.

Lead SAC author

Professor Alastair Lewis, University of York

Contributing SAC authors

Dr Emma Taylor, Cranfield University / RazorSecure Ltd

Professor Ricardo Martinez-Botas, Imperial College London

Professor William Powrie, University of Southampton

1. Royal College of Physicians, (2016). Every breath we take: the lifelong impact of air pollution. Report of a working party. ISBN 978-1-86016-567-2 ↩