4. MONITORING RESULTS

As noted in Section 2.4, a total of 92 randomly selected flights were monitored during the study; smoking was permitted on 69 (75 percent) of these flights. Characteristics of the monitored flights, including airlines, types of aircraft, and flight durations, are described in Section 4.1. For smoking flights, information on passenger counts in the smoking section and observed smoking rates is also presented. Results of environmental measurements — air exchange rates, temperatures, relative humidities, cabin pressures, ETS contaminants, and pollutants — are presented in Section 4.2.

TABLE 4-1. DISTRIBUTION BY AIRLINE FOR DOMESTIC
SMOKING, INTERNATIONAL, AND NONSMOKING FLIGHTS
THAT WERE MONITORED

	Number of Flights
Airline	Domestic Smoking	International	Nonsmoking
American (AA)	10	2	0
Braniff (BN)	0	0	1
Continental (CO)	10	0	0
Delta (DL)	8	0	6
Midway Connection (ML)	2	0	0
Northwest (NW)	0	1	5
Pan American (PA)	4	2	3
Piedmont (PI)	3	0	2
Trans World (TW)	9	1	3
United (UA)	7	2	2
U.S. Air (US)	6	0	1
Western (WN)	2	0	0
Total, All Airlines	61	8	23

4.1 CHARACTERISTICS OF MONITORED FLIGHTS

Airlines, Aircraft Types, and Flight Times

The distribution of monitored flights by airline is summarized in Table 4-1; distributions are given separately for 61 domestic smoking flights, 8 international flights, and 23 nonsmoking flights. All major airlines except Braniff, Eastern, and Northwest were represented by smoking flights. The number of smoking flights offered by these airlines was relatively small, particularly for Eastern (whose airline services were substantially curtailed during the monitoring period due to a strike) and Northwest (for which smoking flights are restricted to those between Hawaii and the continental United States). Northwest was the carrier, however, for a substantial fraction (more than 20 percent) of the nonsmoking flights. Although the number of monitored international flights was limited, most of the major U.S. carriers offering such flights were represented.

The Representativeness of monitored flights is shown more directly in Figure 4-1, in relation to all flights (more than 100,000) that were scheduled for departure from major U.S. airports during January 1989.  The comparison is restricted to domestic smoking flights, the largest subset of flights (61) that was monitored. As indicated by the comparative percentage frequency distributions in the figure, the monitored flights were proportionately representative of most airlines. The five airlines (American, Continental, Delta, Trans World, and United) that accounted for a majority of the scheduled flights were also associated with a majority of the flights that were monitored, and these five airlines were represented in nearly the same order by relative percentage. The most notable discrepancy between monitored and scheduled flights was the lack of representation of Eastern Airlines (EA); Eastern flights were deliberately avoided during the monitoring period (early April to early June 1989) because of Eastern’s curtailment of services, and associated uncertainty in flight availability, at that time.

TABLE 4-2. DISTRIBUTION BY TYPE OF AIRCRAFT FOR DOMESTIC
SMOKING, INTERNATIONAL, AND NONSMOKING FLIGHTS THAT
WERE MONITORED

	Number of Flights
Type of Aircraft	Domestic Smoking	International	Nonsmoking
Narrow Body
Boeing 727	18	0	9
Boeing 737	12	0	3
Boeing 757	3	0	1
McDonnell Douglas DC9/MD80	15	0	8
British Aerospace 111	0	0	1
Wide Body
Boeing 747	0	5	0
Boeing 767	3	1	0
McDonnell Douglas DC10	3	2	1
Lockheed L1011	6	0	0
Airbus Industries 310	1	0	0
Total, All Types	61	8	23

The distribution by type of aircraft is summarized for the three subsets of flights in Table 4-2. All international flights were on widebody aircraft and all nonsmoking flights but one were on narrow-body aircraft, consistent with the relative durations of these types of flights. Domestic smoking flights involved the greatest variety in aircraft types, with about 20 percent of these flights taken on wide-body aircraft. For all three subgroups, Boeing aircraft were most frequently represented, accounting for more than half the monitored flights, and McDonnell Douglas aircraft were next most frequently represented. As indicated in Figure 4-2, the domestic smoking flights were proportionately representative with respect to aircraft type; with the exception of Lockheed aircraft, which were over-represented, the distributions for monitored and scheduled flights differed by no more than a few percentage points for each type of aircraft.

The joint distribution by aircraft width and recirculation capability is shown for smoking flights (domestic plus international) and nonsmoking flights in Table 4-3. The smoking flights were almost equally distributed on aircraft with and without recirculation, whereas the nonsmoking flights were primarily on aircraft without recirculation.

TABLE 4-3. DISTRIBUTION OF MONITORED FLIGHTS
BY AIRCRAFT WIDTH AND RECIRCULATION

Aircraft Width and Recirculation	Number of Flights
	Domestic Smoking and International	Nonsmoking
Narrow-body aircraft	48	22
with recirculation	21	5
without recirculation	27	17
Wide-body aircraft	21	1
with recirculation	11	1
without recirculation	10	0

The distribution by flight duration is summarized for the three subgroups of monitored flights in Table 4-4. All international flights were greater than five hours in duration, averaging 8.9 hours. Domestic smoking flights, averaging 3.2 hours in duration, also included some that were greater than five hours long, but half were less than three hours long. One of the smoking flights was slightly less than two hours in duration, due to variability from the nominal scheduled flight duration that was slightly above two hours in this case. Most of the nonsmoking flights, averaging 1.6 hours in duration, were less than two hours; the exceptions were associated with two carriers–Northwest Airlines (all flights arriving and departing within the continental United States are nonsmoking) and United Airlines (all flights of 1,000 miles or less in distance are nonsmoking). The distribution of monitored domestic smoking flights by duration closely resembled that of scheduled flights (Figure 4-3), with the exception that flight durations between two and three hours were somewhat underrepresented and durations between 3.5 and tour hours were over-represented.

TABLE 4-4. DISTRIBUTION BY FLIGHT DURATION FOR
DOMESTIC SMOKING, INTERNATIONAL, AND NONSMOKING
FLIGHTS THAT WERE MONITORED

	Number of Flights
Duration of Flight (Hours)	Domestic Smoking	International	Nonsmoking
<2.0	1	0	18
2.0 - 2.49	17	0	2
2.5 - 2.99	13	0	1
3.0 - 3.49	8	0	1
3.5 - 3.99	12	0	1
4.0 - 4.99	6	0	0
>5.0	4	8	0
Total, All Durations	61	8	23

Distributions by time of departure are shown for the three subgroups of monitored flights in Table 4-5. International flights were clustered at early morning and late afternoon/evening departure times due to the limited choice of times for direct flights to and from the international destinations. The distribution of domestic smoking flights was somewhat shifted away from morning departures toward flights that departed in the middle of the day. As shown in Table 4-6, this shift was in contrast to the nearly uniform distribution of departure times for scheduled flights. The shift away from morning departures was due in part to delays relative to scheduled times of departure, as evidenced by a comparison of scheduled versus actual departure times for the monitored flights. Clustering toward the middle of the day was due in part to the desire to conserve resources by minimizing scheduled layovers for technicians between monitored flights. These differences in the distributions were not excessive, however, and all blocks of departure times were adequately covered by the monitored flights.

4.1.2 Passengers and Smoking

TABLE 4-5. DISTRIBUTION BY TIME OF DEPARTURE
FOR DOMESTIC SMOKING, INTERNATIONAL AND
NONSMOKING FLIGHTS THAT WERE MONITORED

	Number of Flights
Time of Departure	Domestic Smoking	International	Nonsmoking
Before 9:00 a.m.	5	4	1
9:00 to 11:59 a.m.	9	0	7
Noon to 2:59 p.m.	20	0	4
3:00 to 5:59 p.m.	12	1	8
After 6:00 p.m.	15	3	3
Total, All Times	61	8	23

Information concerning passenger counts, seating capacities, and load factors (i.e., percent of seating capacity filled by passengers) is summarized for the monitored flights in Table 4-7. The information for smoking flights is segregated by narrow- versus wide-body aircraft; for these flights, passenger counts were generally higher for wide-body aircraft whereas load factors were generally higher for narrow-body aircraft. Seating capacity of wide-body aircraft averaged nearly double that of narrow-body aircraft. For the nonsmoking flights (all except one of which involved narrow-body aircraft), the average seating capacity was similar to that of narrow-body aircraft associated with smoking flights, but the average load factor was somewhat lower. with the exception of one smoking flight that had only 17 passengers, the load factor consistently ranged from 30 to 100 percent for each of the three subgroups of flights listed in the table.

TABLE 4-6. REPRESENTATIVENESS OF MONITORED DOMESTIC SMOKING FLIGHTS WITH RESPECT
TO TIME OF DEPARTURE

	Percentage of Domestic Smoking Flights
Time of Departure	Monitored Flights, as Scheduled	Monitored Flights, as Flown*	All Flights Scheduled for January 1989
Before 9:00 a.m.	13.1	8.2	19.4
9:00 to 11:59 a.m.	19.7	14.8	19.2
Noon to 2:59 p.m.	31.1	32.8	20.8
3:00 to 5:59 p.m.	21.3	19.7	17.3
After 6:00 p.m.	14.8	24.6	23.4
*Differs from monitored flights, as scheduled, due to delays in scheduled  departure times.4-10

As described in Section 3.0, information on the number of cigarettes smoked was collected for the smoking flights in two complementary ways: (1) through collection of cigarette butts by technicians at the end of each flight for later counting and (2) through technician observations of cigarettes smoked during one-minute intervals every 15 minutes. Technicians were unable to collect cigarette butts on five of the 69 smoking flights that were monitored. For 12 other smoking flights, ashtrays were not emptied from an immediately prior flight that also was a smoking flight. For the remaining 52 flights, the correspondence between estimates for cigarettes smoked based on technician observations versus cigarette butt counts was assessed. As illustrated in Figure 4-4, very good correspondence was obtained, with a correlation coefficient of 0.89. The regression line of best fit (R2 value of 0.8) between the two estimates was as follows:

Technician Observations = 9.07 + 0.87 x Cigarette Butt Counts

TABLE 4-7. PASSENGER COUNTS, SEATING CAPACITIES
AND LOAD FACTORS FOR FLIGHTS THAT WERE MONITORED

Type of Flight (Number)	Passenger Count	Seating Capacity	Load Factor*
Smoking Flights (69)
Narrow Body (48)
Average	105.3	138.4	75.8
Standard Deviation	35.3	19.2	21.5
Range	17-187	107-187	12-100
Wide Body (21)
Average	182.2	288.0	64.1
Standard Deviation	73.2	67.9	23.2
Range	80-347	184-431	31-100
Nonsmoking Flights (23)
Average	94.4	135.2	69.9
Standard Deviation	39.7	41.8	22.4
Range	31-181	79-284	30-100
*Percent of seating capacity filled by passengers.

The regression equation indicates that technicians observations generally yielded slightly higher estimates than butt counts when smoking levels were relatively low, whereas the reverse was true for relatively high smoking levels. Given the good correspondence between the two methods of estimation, technician observations were used as the basis for analysis in this report because such observations were taken on every monitored smoking flight.

Information on passenger counts in the smoking section and observed smoking rates are summarized for the 69 smoking flights in Table 4-8. On the average, there were 18 passengers in the smoking section smoking 68 cigarettes during the flight. The smoking rates varied from as little as one cigarette per hour to as much as one cigarette per minute for all smokers combined, averaging one cigarette every three minutes. The number of cigarettes smoked per hour per passenger in the smoking section averaged 1.5 and varied widely, ranging from 0.2 to 6.5. The estimate of 6.5 cigarettes per hour per passenger may be an artifact of the estimation procedure that was used; in this case, the estimated number of cigarettes smoked was twice as high as the number of cigarette butts collected by technicians. Discounting this case, the highest estimated smoking rate was 3.5 cigarettes per hour per passenger.

TABLE 4-8. SMOKING PASSENGERS, SMOKING QUANTITY,
AND SMOKING RATES FOR SMOKING FLIGHTS THAT WERE MONITORED

Characteristic	Average	Standard Deviation	Range
Number of Passengers in Smoking Section	18.1	12.4	2-63
Percent of Passengers in Smoking Section	13.7	6.6	1.4-41.9
Number of Cigarettes Smoked during the Flight	68.1	66.7	3-411
Number of Cigarettes Smoked per Hour	19.9	11.2	1-60
Number of Cigarettes Smoked per Passenger per Hour	1.5	1.1	0.2-6.5

Further information on the distributions underlying the summary statistics is displayed in Figure 4-5 (for passenger counts and total cigarettes smoked) and in Figure 4-6 (for cigarettes smoked per hour and cigarettes per passenger per hour). The number of smoking passengers was fairly evenly distributed about the interval 10-19, with five cases at the upper extremes (i.e., 40 or more smoking passengers). The total number of cigarettes smoked had a less symmetrical distribution about the most frequent interval (25-49 cigarettes), with a long tail due to variations in both number of smoking passengers and flight duration. With consideration of flight duration, the smoking rate (expressed as cigarettes per hour) was more symmetrical about the most frequent interval (15-20), with 11 cases at the upper extreme (30 or more cigarettes per hour). The number of cigarettes smoked per passenger per hour was also distributed fairly symmetrically about the most frequent interval (1.0 – 1.5), with 12 cases at the upper extreme (2.5 or more cigarettes per passenger per hour).

Estimated smoking rates in relation to smoking duration (length of time during which smoking was permitted) and time of departure are summarized in Table 4-9. There was no distinct pattern for cigarettes smoked per hour in relation to smoking duration, but the number of cigarettes smoked per passenger per hour was distinctly lower for flights with smoking durations of five hours or longer. This lower rate most likely reflects the tendency of passengers to sleep at times on longer flights. The number of cigarettes smoked per hour was highest for flights departing between noon and 3:00 p.m., the largest time block of monitored flights. When smoking rates were expressed per passenger per hour, however, differences were less pronounced. flights departing after 3:00 p.m. had somewhat lower rates than those departing earlier in the day.

4.2 ENVIRONMENTAL MEASUREMENTS

4.2.1 Air Exchange, Temperature, Humidity and Pressure

The air exchange rate prevailing during a flight depends partly on the extent to which air can be re-circulated and the extent of control that the cockpit crew has over fresh-air intake through selective use of air-conditioning packs and recirculation capabilities. Such factors can vary with the type of aircraft. Nominal air exchange rates at a cruise altitude of 9.1 km (30,000 feet) are listed for different types of narrowbody and wide-body aircraft in Table 4-10 together with nominal values for cabin volume and extent of air recirculation. The nominal values given for air exchange rates at 9.1 km (30,000 feet) are those reported by Lorengo and Porter1 based on information collected by these researchers from equipment manufacturers and airline operators. The aircraft types with recirculation capabilities have lower nominal air exchange rates, ranging from 10/h to 15/h in most cases, than for aircraft without recirculation, for which the nominal rates vary from 23/h to 27/h in most cases.

Air exchange rates measured with PFTs are compared with nominal rates in Table 4-11. For aircraft without recirculation, the measured rates are much higher than the nominal rates” as much as four to five times as high for some aircraft. For aircraft with recirculation, however the measured rates are much closer to nominal values, albeit somewhat higher and still somewhat variable. This pattern of results indicates that there generally was insufficient mixing throughout the airliner cabin for the PFT results to be indicative of prevailing air exchange rates. (Due to the need to remain unobtrusive during sampling, PFT sources for release of tracer gas could be placed only at two locations on smoking flights and one location on nonsmoking flights.) The mixing problem affected measurement results on all types of aircraft, but particularly those without recirculation. The results for aircraft with recirculation are likely to be indicative of the prevailing air exchange rates. The frequency distribution of measured air exchange rates on aircraft with recirculation is given in Figure 4-7.

Air exchange rates on smoking and nonsmoking flights are compared in Table 4-12 for selected aircraft with recirculation. The average air exchange rates were higher on smoking flights for two of the three aircraft types. However, conclusions cannot be drawn because of the extremely limited number of measurements for nonsmoking flights. 4-21 14 The average cabin pressure was lower for smoking than for nonsmoking flights, consistent with higher altitudes that are generally attained on longer flights for which smoking is permitted. Frequency distributions for temperature and relative humidity across all study flights are given in Figure 4-8. More than one-third of the flights had temperatures in the interval from 24 to 25 C, and more than a third of the flights had humidity levels in the range from 10 to 15 percent. Humidity levels were below 25 percent on about 90 percent of the flights.

4.2.2 ETS Contaminants

Nicotine measurement results are summarized by technician seat location for both smoking and nonsmoking flights in Table 4-14. The results for smoking flights are for domestic and international flights combined, except for the remote seat; results are desegregated for this location because the remote sent on international flights was in the business class at the boundary near the business smoking section. Nicotine levels were substantially higher in the coach smoking section of smoking flights, averaging 13.4 ug/m3, than at any other locution. Measurements in the boundary section near coach smoking indicated some impact of tobacco smoking; although the average level (0.26 ug/m3) in this boundary section was much lower than in the smoking section, the level at this monitoring location was higher than the average levels in the middle seat (0.04 ug/m3) and remote seat (0.03 ug/m3) for domestic flights. For international flights, the average level for the remote location near business smoking (0.18 ug/m3) was similar to that for the boundary near coach smoking on all smoking flights. The levels in the middle and remote locations on smoking flights were similar to levels measured on nonsmoking flights, which in most cases were below minimum detection limits.

Cumulative frequency distributions for nicotine measurements on nonsmoking flights are shown in Figure 4-9. The distribution for the smoking section indicates a relatively smooth continuum of measured levels, with only the maximum value of 67.2 ug/m3 somewhat distant from 4-25 its nearest neighbor (47.4 ug/m3). The highest boundary result for domestic smoking flights (3.5 ug/m3) was quite distant from the next highest result at this location (0.6 ug/m3). The highest results for the middle location were 0.9 and 0.2 ug/m3, and the highest results for the remote location were 0.4 and 0.3 ug/m3. These maximum values for the remote site are only slightly above the minimum detection level of 0.1 ug/m3 for a two-hour flight.

The Gravimetric RSP measurements are summarized by technician seat location for smoking and nonsmoking flights in Table 4-15. Because of the relatively short sampling duration and consequent measurement uncertainty, special treatment of these data was required. In most field monitoring studies, results that are negative (after netting out values obtained for field blanks) would be assigned a value of zero; however, in this situation such a practice would have exerted a significant positive bias on the results, particularly for the nonsmoking flights, because of the relatively short sampling duration. For example, historical data from the laboratory used for Gravimetric determinations indicate a standard deviation on the order of +_7 ug for analysis of blanks. Consequently, mass determinations could easily vary from -21 to +21 ug (i.e., + three standard deviations). As a result, for a one-hour sampling duration common for nonsmoking flights, corresponding to a sample volume of 0.1 m3, the measurement result for a prevailing concentration near zero could vary from -210 to +210 ug/m3 (the lowest result obtained was -195 ug/m3).

In view of the above consideration, sampling results with values below those of field blanks were kept as negative values in computing the summary statistics. With this treatment of the data, RSP levels for nonsmoking flights were similar to those measured in the boundary, middle, and remote locations on smoking flights. The levels in the smoking section for smoking flights, however, were considerably higher, exceeding those in other locations by more than 100 ug/m3 on the average. The considerably higher standard deviations for nonsmoking flights are a reflection of the measurement uncertainty due to short sampling duration. The counterintuitive result of higher RSP levels at the remote location for domestic than for international flights may also be an artifact of measurement uncertainty; the international results, for flights of considerably longer duration, had a much smaller standard deviation.

Cumulative frequency distributions for Gravimetric RSP measurements on domestic smoking and nonsmoking flights are shown in Figure 4-10. Negative results, shown in the graph as values of zero, were obtained in about five percent of the cases for the smoking section, in 15 to 25 percent of cases for other locations on smoking flights, and in 25 to 30 percent of cases on nonsmoking flights. The distributions for each location on smoking flights indicate a relatively smooth continuum of measured levels. For nonsmoking flights, the maximum values at each location (397 ug/m3 for the middle seat and 350 ug/m3 for the near seat) are more distant from their nearest neighbors (266 and 197 ug/m3, respectively), another possible reflection of measurement uncertainty for these shorter duration flights.

Continuous monitoring with an optical sensor afforded the opportunity to quantitate RSP levels both before and during the period when smoking was allowed on smoking flights (and prior to takeoff for nonsmoking flights). As shown in Figure 4-11, RSP levels during the base- line period (prior to smoking/takeoff) consistently averaged between 20 and 30 ug/m3 across all sent locations, both for smoking and nonsmoking flights. After the baseline period, however, RSP levels declined somewhat on nonsmoking flights whereas levels on smoking flights increased by a factor of ten in the smoking section and by a factor of two in the boundary section.

Summary statistics for optically measured RSP levels, based on averaging of the continuous results across the sampling period for each flight, are given in Table 4-16. In contrast to the Gravimetric results, the optical results indicated higher levels in all sections of smoking flights than on nonsmoking flights. The difference between the boundary section and the middle/remote locations was sampled more pronounced for 4-31 21 the optical than the Gravimetric results. The optical results for the remote section were more in line with expectations, with the international flights having slightly higher levels than Domestic flights. The optical results were internally consistent, with similar averages for nonsmoking and smoking flights during the baseline period, similar averages for the two locations on nonsmoking flights during the airborne period, and similar averages for the middle and remote location on smoking flights during the smoking period. Further analysis and discussion of the Gravimetric and optical results are provided in Section 5.0..

TABLE 4-17. MEASURED PEAK RSP (OPTICAL) CONCENTRATIONS
FOR SMOKING AND NONSMOKING FLIGHTS

	Results by Seat Location, ug/m3
Type of Flight (Number)	Smoking	Boundary	Middle	Remote (Domestic)	Remote (International)
Smoking Flights (69)	883.4	211.8	68.7	60.4	137.1
Average(during smoking)
Standard Deviation	436.7	308.6	112.8	90.6	49.7
Maximum	2076.8	2275.5	732.2	614.0	198.8
Nonsmoking Flights (23)
Average (while airborne)	18.2	--	16.4	--	--
Standard Deviation	8.9	--	5.9	--	--
Maximum	45.2	--	35.7	--	--

Cumulative frequency distributions are shown in Figure 4-12 for the time-averaged optical measurements during the smoking period. The distributions indicate a relatively smooth continuum of measured levels for smoking and boundary locations on smoking flights and for both monitoring locations on nonsmoking flights. For the middle and remote locations on smoking flights, the maximum values (118 and 103 ug/m3) were quite distant from their respective nearest neighbors (44 and 46 ug/m3).

The number of observations available for optical RSP measurements varied somewhat with technician location due to occasional instrument failures. For smoking flights, there were 65 observations for the smoking location, 63 observations for the boundary location, 62 observations for the middle location, and 58 observations for the remote location. For nonsmoking flights, there were 19 observations for each location.

TABLE 4-18. RATIO OF PEAK-TO-AVERAGE RSP (OPTICAL) CONCENTRATIONS FOR
SMOKING AND NONSMOKING FLIGHTS DURING PERIOD WHEN SMOKING WAS ALLOWED*

	Results by Seat Location
Type of Flight (Number)	Smoking	Boundary	Middle	Remote (Domestic)	Remote (International)
Smoking Flights (69)
Average Ratio	5.7	5.5	3.9	3.3	6.9
Standard Deviation	2.5	3.7	3.1	2.3	2.1
Maximum	13.3	18.0	17.9	14.4	9.6
Nonsmoking Flights (23)	2.5	--	2.3	--	--
Average Ratio
Standard Deviation	1.5	--	1.8	--	--
Maximum	5.4	--	6.5	--	--
* While airborne for nonsmoking flights.

One-minute peak RSP levels that were measured with optical sensors are summarized in Table 4-17. The peak levels on nonsmoking flights were not substantially greater than average levels, whereas on smoking flights the peak levels averaged near 70 ug/m3 in the remote and middle sections, above 200 ug/m3 in the boundary section, and near 900 ug/m3 in the smoking section. Peak levels at the remote site averaged substantially higher for international than for domestic smoking flights. The ratio of peak-to-average RSP concentrations (Table 4-18) was highest in the smoking and boundary sections, next highest in the middle and remote locations, and lowest on nonsmoking flights. These results collectively indicate (1) that tobacco smoking had some impacts on ETS levels in the other sections of the aircraft and (2) that the impacts were most pronounced in the boundary section.

TABLE 4-19. MEASURED AVERAGE CO CONCENTRATIONS FOR SMOKING
AND NONSMOKING FLIGHTS

	Results by Seat Location
Type of Flight (Number)	Smoking	Boundary	Middle	Remote (Domestic)	Remote (International)
Smoking Flights (69)
Baseline(before smoking)	2.0	1.7	1.9	2.0	1.9
Average (during smoking)	1.4	0.6	0.7	0.8	0.8
Standard Deviation	0.9	0.4	0.5	0.4	0.5
Maximum	4.3	1.8	2.8	2.5	1.4
Nonsmoking Flights (23)
Baseline (before takeoff)	1.9	--	1.4	--	--
Average (while airborne)	0.6	--	0.5	--	--
Standard Deviation	0.4	--	0.4	--	--
Maximum	1.3	--	1.3	--	--

Time-averaged CO levels on both smoking and nonsmoking flights were higher during the baseline period than the smoking/airborne period (Table 4-19) for both smoking and nonsmoking flights, due to intrusion of ground-level emissions outside the aircraft. During the smoking period, average CO levels were highest in the smoking section; the levels in the other sections of smoking flights were similar to but slightly higher than those for nonsmoking flights. Domestic and international smoking flights had similar average CO values for the remote location. The cumulative frequency distributions shown in Figure 4-13 indicate a relatively smooth continuum in of time-averaged CO levels for the smoking section, an isolated high value for the remote section, and several high values for the middle section.

TABLE 4-20. MEASURED PEAK CO CONCENTRATIONS FOR SMOKING AND NONSMOKING FLIGHTS

	Results by Seat Location
Type of Flight (Number)	Smoking	Boundary	Middle	Remote (Domestic)	Remote (International)
Smoking Flights (69)
Average (during smoking)	3.4	1.4	1.7	1.5	1.9
Standard Deviation	1.6	0.7	1.0	0.7	0.6
Maximum	8.0	3.3	6.6	4.5	2.6
Nonsmoking Flights (23)
Average (while airborne)	1.3	--	0.9	--	--
Standard Deviation	0.6	--	0.4	--	--
Maximum	2.4	--	1.9	--	--

The number of observations available for CO measurements varied with technician location due to occasional instrument failures. For smoking flights, there were 68 observations for the smoking location, 64 observations for the boundary location, 60 observations for the middle location, and 53 observations for the remote location. For nonsmoking flights, there were 16 observations for the location near the rear of the plane and 18 observations for the middle location.

As shown in Table 4-20, one-minute peak CO levels had a pattern similar to that of time-averaged CO levels, with the highest peaks in the smoking section and peaks in the other sections of smoking flights generally averaging somewhat higher than for nonsmoking flights. International flights had higher peak levels in the remote section, on the average, than domestic smoking flights. The ratios of peak-to-average CO levels (Table 4-21) were similar both across seats on smoking flights and for smoking versus nonsmoking flights. For nonsmoking flights, the ratios for CO were similar to those for RSP, whereas the smoking flights had higher ratios for RSP than for C0.

TABLE 4-21. RATIO OF PEAK-TO-AVERAGE CO CONCENTRATIONS FOR SMOKING AND
NONSMOKING FLIGHTS DURING PERIOD WHEN SMOKING WAS ALLOWED*

	Results by Seat Location
Type of Flight (Number)	Smoking	Boundary	Middle	Remote (Domestic)	Remote (International)
Smoking Flights (69)
Average Ratio	2.8	2.7	2.5	2.3	3.2
Standard Deviation	1.3	1.3	0.9	1.5	2.4
Maximum	9.0	7.5	6.0	7.0	7.5
Nonsmoking Flights (23)
Average Ratio	2.6	--	3.2	--	--
Standard Deviation	1.5	--	2.5	--	--
Maximum	6.0	--	11.0	--	--
* While airborne for nonsmoking flights.

Domestic versus international flights are summarized in Table 4-22. RSP levels in the smoking section were lower on international than domestic flights, consistent with lower smoking rates per smoking passenger observed for longer flights (due, for example, to periods of sleeping). Average RSP levels in the other sections were similar for the two types of flights. International flights had higher peak RSP levels throughout all sections, however, most likely because of larger smoking sections with many people smoking simultaneously after takeoff or after meals. Nicotine levels and peak CO levels also were generally somewhat higher throughout the aircraft for international flights. The higher nicotine levels in the smoking section for international flights (despite lower average RSP levels) could be due to different cigarette brands used by foreign passengers, and the greater apparent migration of nicotine to the nonsmoking locations could be due either to more extensive use of recirculation or a more uniform distribution of smoking across the wide-body aircraft used for the international flights.

Carbon Dioxide and Pollutants

TABLE 4-23. MEASURED C02 CONCENTRATIONS
FOR SMOKING AND NONSMOKING FLIGHTS

	Seat Location
Type of Flight (Number)	Smoking	Middle
Smoking Flights (69)
Average, ppm	1562	1568
Standard Deviation	685	488
Minimum	711	597
Maximum	4943	3078
Nonsmoking Flights (23)
Average, ppm	--	1756
Standard Deviation	--	660
Minimum	--	765
Maximum	--	3157

Average C02 levels (Table 4-23) were somewhat lower on smoking than nonsmoking flights, indicative of generally higher air exchange rates on smoking flights. On both types of flights, however, average C02 levels exceeded 1,000 ppm 87 percent of the time and sometimes exceeded 3,000 ppm. Thus, due to the relatively high density of occupants, C02 levels in aircraft cabins often exceeded ASHRAE guidelines associated with satisfaction of comfort criteria, despite air exchange rates that are much higher than those for ground-level indoor environments. The frequency distributions provided in Figure 4-14 indicate (1) that C02 levels were typically between 1,000 and 2,000 ppm for smoking flights and between 1,000 and 2,500 ppm for nonsmoking flights, and (2) that the two locations monitored for smoking flights had similar distributions.

Average measurement results for both total bacteria and Staphylococcus (Table 4-24) were similar for smoking and nonsmoking flights; the levels were, however, slightly higher in the smoking than nonsmoking sections, possibly due to a higher proportion of passengers with respiratory conditions in the smoking section. Another possibility is that skinscales attach to settled particles and are resuspended by the movement of people, resulting in higher Staphylococcus levels in areas where particle concentrations are higher.

TABLE 4-24. MEASURED BACTERIA CONCENTRATIONS FOR SMOKING AND NONSMOKING FLIGHTS

	Total Bacteria		Staphylococcus
Type of Flight (Number)	Smoking Seat	Middle Seat	Smoking Seat	Middle Seat
Smoking Flights (69)
Average, CFU/m3	162.7	131,2	14.1	5.3
Standard Deviation	105.8	88.6	20.6	9.2
Maximum	556.4	462.1	97.8	45.0
Percent Below Minimum Detection	0.0	0.0	50.7	62.3
Nonsmoking Flights (23)
Average, CFU/m3	--	131.1	--	6.5
Standard Deviation	--	123.4	--	9.6
Maximum	--	641.6	--	30.0
Percent Below Minimum Detection	--	0.0	--	56.5

Average fungi results (Table 4-25) were very low on all flights; the levels were somewhat higher on nonsmoking flights, possibly due to slower removal (associated with lower air exchange rates) of fungi entrained at the gate and before takeoff. The most prevalent types of bacteria, measured on more than a third of the flights, were Staphylococcus aureus, Staphylococcus not aureus, Micrococcus varians, Micrococcus sedentarius, Corynebacteriun, and Arthrobacter (Table 4-26). The most prevalent types of fungi were Cladosporium and Alternaria (Table 4-27); apart from these types, only Penicillium was detected on more than 10 percent of the monitored flights.

TABLE 4-25. MEASURED FUNGI CONCENTRATIONS FOR
SMOKING AND NONSMOKING FLIGHTS

Type of Flight (Number)
	Smoking Seat	Middle Seat
Smoking Flights (69)
Average, CFU/m3	5.9	5.0
Standard Deviation	6.4	5.8
Maximum	29.2	32.0
Percent Below Minimum Detection	11.6	13.0
Nonsmoking Flights (23)
Average, CFU/m3	--	9.0
Standard Deviation	--	12.7
Maximum	--	61.1
Percent Below Minimum Detection	--	4.3

Average ozone levels on the monitored flights (Table 4-28) also were relatively low, never exceeding 0.1 ppm. Average levels were somewhat higher for nonsmoking than smoking flights; the difference could be due to flight paths, air exchange rates, cleaning equipment for aircraft, or poorer accuracy/precision for nonsmoking flights due to relatively short sample-collection intervals.

4.3  QUALITY CONTROL SAMPLES

Samplers were deployed in duplicate on selected flights to estimate measurement precision for nicotine, RSP, C02, and ozone. The average precision for each measurement parameter is summarized in Table 4-29. With the exception of C02, the precision is poorer than would normally be expected. The poorer precision is due to the relatively short sampling duration; the typical monitoring duration for this study was several hours, whereas for most field monitoring studies the duration would be eight hours or longer.

TABLE 4-26. PERCENT OF FLIGHTS ON WHICH
DIFFERENT TYPES OF BACTERIA WERE DETECTED

Bacteria	Smoking Flights		Nonsmoking
	Middle	Smoking	Middle
Micrococcus varians	92.3%	91,0%	95.5%
Staphylococcus not aureus	78.5%	65.7%	81.8%
Corynebacterium	61.5%	53.7%	86.4%
Arthrobacter	63.1%	65.7%	40.9%
Micrococcus sedentarius	53.8%	64.2%	54.5%
Staphylococcus aureus	38.5%	49.3%	45.5%
Micrococcus nishinomiyaensls	26.2%	9.0%	45.5%
Streptococcus not pyogenes	15.4%	15.4%	4.5%
Gram,positive rod	13.8%	11.9%	22.7%
Bacillus	12.3%	9.0%	9.1%
Micrococcus ylae	6.2%	3.0%	0.0%
Micrococcus roseus	4.6%	3.0%	13.6%
Micrococcus kristinae	3.1%	0.0%	0.0%
Micrococcus luteus	1.5%	13.4%	0.0%
Gram,negative rod	0.0%	3.0%	0.0%
Gram,negative cocci	1.5%	0.0%	0.0%
Gram,variable cocci	1.5%	1.5%	0.0%
Gram,variable rod	1.5%	3,0%	0.0%
Stomatococcus	0.0%	1.5%	0.0%
Streptococcus pyogenes	0.0%	0.0%	0.0%

TABLE 4-27. PERCENT OF FLIGHTS ON WHICH DIFFERENT
TYPES OF FUNGI WERE DETECTED

Fungi	Smoking Flights		Nonsmoking
Middle	Smoking	Middle
Cladosporium	72.3%	70.1%	90.9%
Alternarea	46.2%	43.3%	31.8%
Aspergillus,niger	9.2%	1.5%	9.1%
Penicillium	7.7%	10.4%	18.2%
Epicoccum	7.7%	6.0%	9.1%
Black yeast	1.5%	6.0%	9.1%
Aspergillus	6.2%	4.5%	0.0%
Curvalaria	4.6%	3.0%	4.5%
Arthrinium	4.6%	1.5%	4.5%
Mucor	4.6%	4.5%	4.5%
Pithomyces	4.6%	1.5%	0.0%
Drechslera	0.0%	1.5%	4.5%
Nigrospora	3.1%	3.0%	0.0%
Monilia	0.0%	0.0%	4.5%
Aspergillus glaucus	0.0%	0.0%	4.5%
Sporotrichum	0.0%	3.0%	0.0%
White yeast	1.5%	1.5%	0.0%
Aspergillus fumigatus	1.5%	1.5%	0.0%
Phialophora	0.0%	1.5%	0.0%
Erysiphe	1.5%	0.0%	0.0%
Scopularlopsis	1.5%	0.0%	0.0%
Yeast	0.0%	1.5%	0.0%
Botrytis	0.0%	1.5%	0.0%
Unidentified fungi	1.5%	0.0%	0.0%

TABLE 4-28. MEASURED OZONE CONCENTRATIONS
FOR SMOKING AND NONSMOKING FLIGHTS

	Seat Location
Type of Flight (Number)	Boundary*	Remote
Smoking Flights (69)
Average, ppm	0.010	0.010
Standard Deviation	0.011	0.010
Maximum	0.054	0.044
Percent Below Minimum Detection	22.0	24.5
Nonsmoking Flights (23)
Average, ppm	0.022	--
Standard Deviation	0.023	--
Maximum	--	0.078
Percent Below Minimum Detection	0.0	--
*Middle seat on nonsmoking flights.

TABLE 4-29. MEASUREMENT PRECISION FOR
SELECTED PARAMETERS

Measurement Parameter	Average Precision*
Nicotine	+/- 27%
RSP	+/- 33%
C02	+/- 8%
Ozone	+/- 37%
* Precision for a set of duplicate samplers is the standard deviation for the two results expressed as a percent of the average result.