A Further Study of Air Pollution in Diesel Bus Garages

ABSTRACT

The concentrations of smoke, polycyclic aromatic hydrocarbons (PAHs), and some gaseous air pollutants have been measured in two London Transport diesel bus garages and compared with observations made in the same garages over 20 years earlier. The main feature of the results was a large reduction in the background concentrations of smoke and polycyclic aromatic hydrocarbons from sources such as coal fires, attributable to the implementation of the Clean Air Act. Contributions from the buses to the benzo(a) pyrene content of the air inside the garages were of the same magnitude as before, being small in relation to former coal smoke contributions.

METHODS

The intention was to repeat the procedures followed in earlier studies as closely as possible, and thus the same basic sampling instruments and analytical methods were used, even though there had been substantial developments in instrumentation and techniques since then. Similarly, the aim was to examine the distribution of pollutants for one representative day in each garage rather than to determine long term mean values. The days chosen were normal weekdays, and to ensure a minimal background from heating sources, mid-summer was selected. As before, emissions of pollution within the garages occurred mainly in the late evening and the early morning, when most buses entered and left the garages, and sampling periods were arranged to cover these. As more buses now moved in and out of the garages during the day additional sampling periods were added at the beginning and end so as to cover the full 24 hour cycle of events (table 1).

The main samples for the determination of concentrations of smoke and polycyclic aromatic hydrocarbons were collected on weighed glass fibre filter sheets measuring 8 in x 10 in (20 cm x 25 cm) overall by means of a Staplex high volume sampler, at a rate of about 1-4 m3/min.

At sites A (Merton) and P (Dalston), two such samplers were used side by side for each of the five periods to provide some check on reproducibility. A single high volume sampler was run at sites B (Merton) and Q (Dalston) during period 4 (early morning run-out) only.

Concentrations of smoke were also examined simultaneously in two other ways. At each of the five dispersed sites in each garage a low rate gravimetric sampler was set up, in which smoke was collected on to a glass fibre filter in an open filter holder with an exposed area 2 in (5 cm) diameter. The sampling rate was about 6 1/min, and each filter was run for a single 24 hour period. A standard smoke/SO2 sampler of the type used in the National Survey was also set up alongside each of these latter filters. This comprised a Whatman filter paper of 2 in (5 cm) exposed area in a closed holder, with a bubbler containing dilute H2O2 for the determination of SO2.

The sampling rate was about 1-5 I/min and each sample covered a single 24 hour period. Smoke concentrations were assessed by reflectance using the standard calibration (recognising that it may not give a true equivalent weight of smoke for the material collected in and around the garage). To provide a visual indication of peaks in smoke concentration as buses passed, a continuous sampler was also used at sites A and P, in which smoke was collected on a strip of Whatman No 1 paper.

Additional instruments were installed at the main sampling sites to assess transient peaks in gaseous pollutants. Sulphur dioxide was measured with a fast response monitor based on the conductivity principle, carbon monoxide with an electrochemical instrument, oxides of nitrogen with a chemiluminescent analyser, and the outputs were connected to a multi-
channel recorder, so that changes could be observed on a common time base. At Merton all these instruments were housed in an office alongside site A and the inlets were at breathing level a short distance from the particulate samplers. At Dalston all the instruments were grouped together at site P.

After collection, each of the (high volume) glass fibre filters was weighed and extracted with cyclohexane in a Soxhlet apparatus for eight hours. The resulting solution was reduced to small bulk, made up to a standard volume, and an aliquot of this was transferred to an alumina column. The hydrocarbons were separated by elution with cyclohexane and determined spectrophotometrically, as in our earlier studies.


RESULTS

The main results from the high volume samples are summarised in tables 2 and 3. There were no substantial differences between findings from the duplicate filters run at sites A and P, and the figures shown in the tables are the means for each pair of filters. At no time were the concentrations of smoke and polycyclic hydrocarbons as high as had been seen in the earlier series, but the distribution during the day differed, with the highest values now occurring during the morning run-out (period 4) rather than in the evening run-in (periods 1 or 2). This no doubt reflected changes in operating procedures, for the build up of buses waiting with engines running to be washed manually as they came in was now avoided. The additional sites B and Q had been set up in places considered to be particularly affected by pollution during the morning run-out. At Merton smoke concentrations were certainly higher at B than at A, though A retained the higher hydrocarbon concentrations. At Dalston Q was more polluted by both smoke and hydrocarbons than P. Pollution was more uniformly distributed during the day than before, probably as a result of the greater spread of bus movements. A notable feature was the relatively low background concentrations of smoke at the roof sites (C and R) compared with those in the earlier series, and the low concentrations of polycyclic aromatic hydrocarbons there. During the night (period 3) when no buses were moving, the concentrations of smoke and hydrocarbons in the garages were similar to those at the roof sites.

As before, pollution tended to be less in Dalston than in the much busier Merton garage.

Because of the low background pollution from coal fires now, contributions from the buses within the garage stood out more clearly than before, though this did not necessarily mean that they were any greater.

In tables 4 and 5 these contributions have been estimated by subtracting the corresponding back- ground values (sites C and R). The figures are approximate, since any variation in background during the day had to be ignored. Figures derived from the summer samples in the earlier series have been added for comparison. Broadly speaking, the contribution from the buses could be said to be of the same order of magnitude as before. At Merton, while a reduced contribution during the evening run-in (periods 1 and 2)
was evident, that at site B during the morning run-out was higher, and roughly equivalent to the evening peak before.

Clearly the buses lead to substantially enhanced smoke concentrations in the garages (more notably at Merton than at Dalston) and the contributions to the polycyclics, although no greater than before, are by now quite substantial in relation to the (greatly reduced) background. It remains true, however, that the polycyclic content of the smoke as emitted by the buses is low compared with that from coal fires. This feature is brought out in tables 6 and 7, in which (as an example) the benzo(a)pyrene content of the smoke collected in each period in the recent series is shown along with corresponding figures from both the summer and the near winter (April, October) samples.

The highest figure is the one for the background site (C) at Merton in April 1956, when there were high concentrations of coal smoke from domestic chimneys nearby, and this figure is similar to that seen in a coal mining area during winter in still earlier studies of benzo(a)pyrene in urban air. The generally lower figures within the garages (as compared with outside) in the 1956-7 series indicate that the diesel smoke was relatively weak in benzo(a)pyrene, producing a diluting effect on the background material. By 1979 the tables had been turned, for the background material was then extremely weak in benzo(a)pyrene, and whereas there was still apparently little in the diesel smoke it was sufficient to enrich the background material slightly, a little more in the morning run-out (period 4, site A, at Merton and site Q at Dalston) than at other times. All these calculations may be taken a stage further, to derive "richness" figures in pg/g for the contributions from the buses (tables 4 and 5), but they then depend on the somewhat uncertain estimates by difference. Such figures, however, fall mainly in the range 10-20 pg/g, which is much weaker than the 100-300 pg/g common in coal smoke.

The additional low rate samplers were run primarily to check the distribution of smoke around the garages, and they were not analysed for hydrocarbons. A summary of results is shown in tables 8 and 9, together with weighted mean values for the
accompanying high volume samples where applicable. Comparison of smoke concentrations between sites is best based on the (low rate) gravimetric results. At Merton these indicate that the sites chosen for the main samplers (P and Q) did have the highest concentrations, when averaged over the complete 24 hours, though those in the office and dock were not far behind. The results assessed by reflectance relate particularly to the black component, and the ratio of the reflectance to the gravimetric result gives an indication of the extent to which black smoke is dominant in the sample (diesel smoke being especially black in relation to its weight). Thus the proportion of diesel smoke appeared to be higher at the dock site than anywhere else, whereas it was relatively low in the office, where cigarette smoke was an important contaminant. At the roof site there was also a substantial proportion of non-black material, from general urban sources, and the higher result obtained with the high volume sampler indicates that some relatively coarse dust, beyond the respiratory range, was present there. Similarly, there was some coarse component at the main site (A).

At Dalston the sites chosen for the main samples (P and Q) proved not to have the highest concentrations of smoke over the 24 hours as a whole, though it had been evident at the time that they were the ones most subject to intermittent periods of high pollution. Again cigarette smoke would have had some effect on the concentration as determined gravimetrically at the office site, but black smoke was an appreciable component there and the "blackest" smoke appeared to be at site Q in the corner of the garage which received emissions from buses being manoeuvred in and out of the tight parking in that area. There was, however, nowhere in Dalston garage as polluted as some of the sites at Merton, and although the concentrations of smoke, as measured by any of the three methods, were all above the (low) background levels on the day in question, they were within the range commonly found in urban air in other seasons.

Concentrations of sulphur dioxide as shown in tables 8 and 9 were not substantially above background levels (as measured at sites C and R). At Merton there was some enhancement at sites A and E that could be attributed to emissions from the buses, but in the office there was a deficit, due to absorption on clothes, walls, and other surfaces in this relatively confined space, or to some neutralisation by ammonia. All the SO2 was absorbed or neutralised in the office at Dalston, and only the dock site showed any
excess over background.

Although emissions from the buses had little effect on the mean concentrations of SO2 there were transient peaks that could be detected close to the vehicles, and the continuous instrument showed values around 500 ug/m3 as buses passed the sampling site
(A) at Merton, with occasional maxima of about 1000 pg/m3. At Dalston values were generally lower, with maxima around 350 pg/m3. Peaks of these magnitudes can occur in the general urban air as plumes from nearby chimneys serving heavy oil or coal fired heating plants blow across a sampler or in calm weather when such pollution accumulates.

Despite emissions of carbon monoxide being low from diesel (as opposed to petrol) vehicles, many of the buses that passed the sampling points produced a small momentary increase in concentration, usually up to about 10 ppm, with maxima of 20 ppm at Dalston (site P) and 70 ppm at Merton (site A). The more notable peaks in CO came, however, from private cars belonging to the staff that occasionally entered and left the garages (no cars entered during the earlier series in the 1950s but with fewer buses now, some are allowed in). In general the CO concentrations in the garage could be said to be less than those outside in a busy street.

The measurements of oxides of nitrogen proved unsatisfactory due to failure of components during transit in the initial sampling periods, but return visits were made later with a more portable instrument. Even then it was difficult to assess transient peaks of NO2 adequately, since values were obtained by difference from sample lines that were successively measuring NO and NO + NO2 (the NO2 being reduced back to NO). When sampling close to the buses as they passed, concentrations varied too rapidly for such a difference method to give true results for NO2.

Subject to these reservations the results indicated average values at sites A and P of about 0-5 ppm for NO and about 0-3 ppm for NO2 during active periods when buses were passing regularly, with occasional transient peaks up to about 4 ppm for NO and 2 5 ppm for NO2 when buses were idling nearby with exhausts directed towards the intake. Similar situations might be met in streets close to traffic (either diesel or petrol engined), and concentrations of similar orders of magnitude are liable to occur indoorswhen influed gas or kerosene applicances are in use.

