As identified through the risk assessment given in preceding sections, the pollutants that pose the highest risks of mortality and morbidity to airliner flight attendants and passengers are ETS contaminants and cosmic radiation. The measurement results also indicated that carbon dioxide levels on flights monitored during this study were frequently above the level thought to satisfy comfort criteria. A general framework for identifying and assessing alternative mitigation strategies for these pollutants is presented in Section 9.1. Application of this framework to strategies for reducing ETS levels in aircraft is described in Section 9.2, and application of the framework to other pollutants (cosmic radiation and carbon dioxide) is described in Section 9.3.

Approach Required

General Strategy
1. Reduction of emissions  
- ban on smoking (total or partial) X
- curtailment of smoking period X
2. Contaminant removal/confinement  
- increased ventilationX 
- local exhaust (smoking section)X 
- smoking loungeX 
- filtration/absorptionX 
3. Exposure management  
- separate smoking/nonsmoking flights X
- stationing of flight attendants X


A general framework for evaluating alternative mitigation strategies for a contaminant in an airliner cabin is depicted in Figure 9-1.

The first step in this process is to identify candidate mitigation strategies. Such strategies could include potential technological or procedural solutions to apparent problems; the technological solutions generally involve some type of change in aircraft design or equipment, whereas procedural solutions involve changes in the activities of people aboard the aircraft. Although it may be possible to identify many types of candidate strategies, only a limited number will be feasible from technological or procedural standpoints. As a simple example, addition of lead shielding could be contemplated to reduce cosmic radiation exposure, but such a procedure would be technologically impractical because of the resultant increase in aircraft mass. Some qualitative judgments obviously are required in this feasibility assessment process.

In the second step of the overall framework, strategies that survive the feasibility assessment are subjected to a more quantitative process of modeling and estimation. If performing this evaluation, it must 32 be recognized that certain options will have practical upper limits (e.g., extent to which ventilation rates or filter efficiencies can be increased). Some aspects of cost estimation will require detailed pricing or econometric models that cannot be developed within the scope of this effort. In addition to the types of costs (e.g., fuel penalties, new equipment) that can be addressed quantitatively, practical considerations such as anticipated acceptability by airline management, flight crews, or passengers need to be addressed qualitatively. The cost and practical aspects are juxtaposed with the estimated benefits of each strategy. Benefits accrue from presumed decreases in contaminant concentrations and associated health or discomfort risks. The contaminant concentrations expected to prevail when a specific strategy is applied are estimated through cabin air quality modeling, discussed later. The risk reduction associated with reduced concentrations 1s estimated through the framework used to assess risks for currently prevailing concentrations (see Section 6.0). The benefits of reduced risk can be placed in monetary terms using estimates of an individual s willingness to pay for reduced mortality or morbidity. Such estimates, which are discussed in more detail later, can be taken from other studies.

The third step of the overall framework is to determine the strategy or strategies of choice. Generally speaking, the optimal strategy would be the one with the highest net benefit (i.e., benefit minus tost), given cost and distrlbutional constraints. However, if two candidate strategies have similar estimates for risk reduction, then a cost-effectiveness analysis can be performed, with a focus on the costs and practical aspects of each alternative.


The framework described above is first applied to ETS contaminants. Alternative mitigation options that are considered, and the subset retained for further analysis, are discussed in Section 9.2.1. Modeling efforts and estimated costs and benefits for each strategy are described in Section 9.2.2. Discussion of the relative costs and benefits of the alternative strategies is provided in Section 9.2.3.

9.2.1 Identification of Options

Eight candidate options for reducing the exposure of flight attendants or passengers are identified 1n Table 9-1. Half the options require a technological approach and the other half require a procedural approach. The options are also classified according to three general types of strategies for mltignting potential exposures:

  1. Preventing or minimizing the emission of ETS contaminants from cigarettes (i.e., emissions reduction)
  2. Removing the ETS contaminants from the cabin environment after they have been introduced (i.e., contaminant removal)
  3. Reducing the exposure of cabin occupants to ETS contaminants that have been introduced (i.e., exposure arrangement).Although some of these options could obviously be used 1n combination with others, the general feasibility of each option has been assessed separately, as discussed below.

For the strategy involving emissions reduction, an obvious option is an outright ban of smoking on all flights. This procedural option would be quite feasible to implement and, in fact, has been implemented in partial form on domestic flights (i.e., smoking not allowed for flights of two hours duration or less under Public Law 100-202)1. Consideration would need to be given to possibilities such as smokers experiencing withdrawal symptoms, becoming unruly, or attempting to smoke in the lavatory, thereby creating additional hazards for other passengers.

A different type of procedural option would involve curtailment of smoking by restricting the periods when it is allowed. For example, smoking could be allowed for a period of 10 minutes after every two hours of flight time, consistent with the earlier ban for flights shorter than two hours. In this case, consideration would need to be given to the possibility of substantially elevated ETS levels during the smoking period, since most smokers would probably smoke during this time.

All of the work described in the report preceded passage of PL101-16d, which will ban smoking on all domestic commercial flights under six hours in duration.

The next set of options to be evaluated involves the notion of removal of ETS contaminants, rather than reduction of emissions. The rate of removal could be increased, for example, by increasing the amount of fresh-air intake to the airliner cabin. The extent to which fresh air could be added has n practical upper limit, however, related to the need to maintain a prescribed cabin pressure. An added benefit of this approach would be reduction of levels of some other pollutants (e.g., carbon dioxide) having sources within the cabin environment. Potential disadvantages could include the added fuel penalty associated with increased fresh-air intake, the need for increased thermal treatment of incoming air, potential increases in ozone levels, and potential decreases in relative humidity levels. The extent of contaminant removal due to increased ventilation can be modeled, and the associated fuel penalty can be estimated. Because both the strategy and the modeling of consequences are feasible, this option can be subjected to a quantitative assessment.

Local exhaust can be thought of as a special case of increased ventilation. This strategy would be most effective 1f combined with the concept of a smoking lounge, discussed below. Under the current configuration of smoking and no-smoking sections, the notion of local exhaust would essentially be tantamount to increasing the fresh-a1r supply to the smoking section only. Although there are some uncertainties related to technological feasibility and costs, the strategy is feasible and its potential consequences can be modeled.

Another special case of increased ventilation would involve the creation of a smoking lounge, which would also serve to confine ETS emissions. In a simplified form, this option would involve creating smoking sections of fixed size with physical barriers (e.g., walls/door or curtains) separating such compartments from nonsmoking sections. This option, while technologically feasible, would be inefficient (1) because of the need to create smoking/nonsmoking sections of fixed size, as opposed to the concept of a “sliding’ boundary that is currently used to accommodate varying numbers of nonsmokers on smoking flights, and (2) because it would do little to reduce the risks of flight attendants assigned to or passing through the smoking section (as shown in Section 7.0, risks related to ETS exposures are estimated to be highest for flight attendants).

A truer version of the smoking-lounge concept would be construction of an actual lounge on one side of the plane toward the rear. This lounge could be “visited’ by smokers wishing to smoke, much in the same sense as lavatories are currently visited by cabin occupants. The size of the lounge (and maximum occupancy) would obviously need to be limited, and emissions could be effectively contained by providing an independent exhaust system for the lounge. Flight attendants would not need to enter the lounge, thereby minimizing their exposures. Some challen9es in design and financing of the lounge, however, would be likely. Some of the costs could be recovered by charging a per-visit fee for the lounge. However, this approach would add some administrative burden, and the extent of costs recovered (both the cost of building the lounge and the cost of reduced seating capacity) would be somewhat difficult to predict. Thus, while potentially attractive, concepts for emissions confinement should be dismissed at this time as impractical to implement.

A third type of contaminant-removal option involves improved filtration of particle-phase ETS constituents or sorption of gas-phase constituents. Such an option obviously would be viable only for aircraft with recirculation capabilities, but the percent of aircraft with recirculation 1s expected to steadily increase in the future. Most aircraft with recirculation are currently equipped with some type of filter in the recirculation loop, and the efficiency of these filters can presumably be improved. Some potential drawbacks of filtration are (1) that filtration of only particle-phase constituents would not remove the gas-phase constituents that can cause odor and irritation, and (2) some gas-phase9-7 constituents, following removal by sorption, could conceivably volatilize and subsequently cause odor/irritation problems throughout the aircraft. Although some uncertainties are involved, modeling can be performed with assumptions involving efficiencies of currently installed filters and the extent of improvement that may be technologically feasible. Like the option of increased fresh-air intake, filtration may also achieve some reduction of pollutants other than ETS contaminants.

The last set of options involves the notion of exposure management rather than emissions reduction or contaminant removal. An extreme example would be to have separate smoking and no-smoking flights. Although such an approach would reduce exposures for nonsmoking

passengers, it would not necessarily reduce flight attendants’ overall exposures. The model required to assess the economic consequences of separate smoking and nonsmoking flights would be difficult to construct and would involve a number of assumptions. Even without such a model, it seams unlikely that such an approach would be economically viable. Thus, it should be dismissed at this time as impractical and having questionable benefits that cannot easily be modeled.

Another approach to exposure management would involve rotating flight attendants so that each is assigned to the smoking section only for some fraction of flights. This approach, however, would merely r distribute risk; the aggregate risk for flight attendants would not be reduced, and risks for nonsmoking passengers would be unaffected. Thus, the strategy would have no apparent benefits. A variation of this theme would involve recognition rather than reduction of risk. For example, flight attendants stationed in the smoking section could be offered “hazardous duty pay.” The costs of increased risk could be estimated, translated into salary differentials, and the costs recovered through differential pricing for smoking and no-smoking seats. Such an approach, however, could affect passenger behavior (e.g., more smoking passengers opting for no-smoking seats, which would reduce ETS levels and associated risks for attendants), thereby adding a layer of assumptions and uncertainties to the assessment. Like the other options for exposure management, no benefits would accrue to nonsmoking passengers (unless ETS levels would actually decrease through this approach). Thus, approaches involving exposure management can be dismissed as having very limited benefits and posing some difficulties in econometric modeling needed to help determine the extent of any potential benefits.

Based on the above discussion, the following candidate approaches to ETS mitigation have been retained for further, quantitative analysis:

  • Ban on smoking (total or partial)
  • Curtailment of smoking period
  • Increased intake of fresh air (including special case targeted at smoking section)
  • Filtration/sorption of ETS contaminants.

9.2.2 Modeling of Cabin Air Quality

Flight Characteristic Model Input123
Type of AircraftB-727MD-80MD-80
Number of Passenger Rows213338
(Number assigned to coach smoking)(2)(7)(8)
Passenger Capacity108142142
Observed Smoking Rate (cigarettes/h)3115
Measured Fresh-air Intake Rate, m3/h3,5793,1253,964
Percent Recirculation Air*02121
Chamber Airflow Rates, m3/h
- SA13238.03116.93960.9
- SA2340.8839.21056.2
- RA13058.53636.54036.6
- 2520.3319.6980.5
Per aircraft specifications

Model Description. Air quality modeling was performed to assess the potential impacts of alternative mitigation strategies on ETS concentrations in cabin environments. The focus of the modeling effort was on RSP, which was used as the ETS tracer 1n performing the risk assessment for chronic effects due to ETS exposure. A two-chamber model, depicted in Figure 9-2, was developed; this model, similar in concept to that described by Rynn et. al (1988), treats the smoking and no-smoking sections as separate compartments with communicating airflows. The model also allows contaminant emission rates to be specified for each compartment and incorporates supply airflow rates from fresh (makeup) air and recirculated air (where applicable) as well as return airflow from each compartment that is exhausted from the aircraft or recirculated.

The model can actually be thought of as a three-chamber model, with the supply airstream representing the third chamber. Under steady-state conditions (appropriate for predicting average concentrations chamber is as follows (the terms used below are defined in Figure 9-2):

Cs · SA = CD · MA + (1 – e) · (C1 RA1 + C2 RA2)/(RA1 + RA2) · (RAl + RA2 -E)

Cs · SA1 + CZ · F21 + S1 = C1 (L1 + 1 + F12)

Cs · SA2 + C1 · F12 + S2 = C2 CL2 + 2 + F21)

The above mass-balance description yields a system of three equations and three unknowns (Cs, C1, and C2) which can be obtained by solving the equations simultaneously. In solving he equations, fresh-air supply rates and interchamber airflow rates were based on PFT measurements.

Leakage rates were assumed to equal zero because no quantitative guidance was available for specifying these rates; thus, any leakage is captured in the term for exhaust flow rate, equal to the fresh-a1r intake rate by assumption. The return airflow rate incorporates recirculation airflow rates based on aircraft specifications (Lorengo and Porter, 1985). A filter efficiency of 90 percent for RSP removal was assumed for baseline modeling, based on information reported by Lorengo and Porter (1985). An emission factor of 26 mg/cigarette (NRC, 1986) was combined with technician observations of smoking rates to develop an hourly emission rate for each flight that was modeled. Supply airflow rates for each section of the aircraft were apportioned by volume, using the number of rows 1n each section as a proxy for volume. Return airflow rates were determined by flow-balance considerations, given supply and lnterchamber airflow rates.

Although PFTs were deployed to estimate airflow rates on study flights, practical limitations (i.e., the need for unobtrusive measurements) precluded obtaining meaningful measurement results in a number of cases. Ideally, PFT sources and samplers would have been distributed throughout each section (smoking and no-smoking) of the aircraft; however, logistical constraints restricted the approach to one release location and one sampling location per section. PFT measurement results were reviewed to determine cases for which results were most plausible, according to the following criteria:

Measured ventilation rates for the aircraft determined by two different PFT methods (single tracer common to both sections and trenchers unique to each section) were consistent with one another and with maximum ventilation rates indicated by aircraft specifications.

Interzonal airflow rates were positive but not excessively large.The three flights chosen for modeling involved two types of narrow body aircraft (B-727 with no recirculation and ! -80 aircraft with recirculation) that collectively accounted for more than 50 percent of the flights monitored during the study. Selected characteristics of the aircraft and the i9hts used for RSP modeling are given in Table 9-2. The flights collectively provide a tan-to-twenty-told range in smoking rates and measured ETS concentrations 1n the smoking section.

The model described previously was chosen over the one developed by Rynn et al. (1988) because of the ability to include a filtration factor for recirculated air (important to the analysis of mitigation options related to filtration/sorption). However, the software for the Ryan et al. (1988) model was obtained from the principal author and applied to the case without recirculation that was listed in Table 9-2. The published model and the model developed specifically for the mitigation assessment yielded identical results when applied to this case.

 RSP Concentrations, ug/m3
Flight 1 (B-727)
- no-smoking section**31.94.7
- smoking section233.5122.4
Flight 2 (MD-80)
- no-smoking section**11.05.3
- smoking section7.322.3
Flight 3 (MD-80)  
- no-smoking section**86.344.2
- smoking section302.0224.3*
* Baseline model, derived from measurements together with assumed
recirculation rate of 21 percent and filter efficiency of 90 percent for MD-80 aircraft.
** Volume-weighted average of Gravimetric and optical measurements  in
boundary, middle, and remote locations.

nl Application. Results of baseline modeling for the three study flights, to be used as a benchmark for assessing various mitigation alternatives, are compared with measured RSP concentrations 1n Table 9-3. Although the modeling results are 9enerally lower than measured values, the general patterns of results are consistent. For example, both measured and modeled values indicate somewhat greater mitigation of RSP from the smoking to the no-smoking section for the 1-80 than for the B-727 aircraft, presumably due to air recirculation.

Both the measured and modeled values also have sane uncertainties; 1n the case of modeled values, sources of uncertainty include emission, mixing, and deposition rates, fresh-a1r supply and lnterchamber airflow rates, the prevailing recirculation rate during a flight, and the filter efficiency for RSP removal.

As noted in Section 9.2.1, four alternatives for ETS mitigation were retained for further analysis:

  1. Ban on smoking (total or partial)
  2. Curtailment of the smoking period
  3. Increased ventilation (including special case targeted at smoking section)
  4. Filtration/sorption of ETS contaminants.

The total ban on snaking requires no modeling; 1f this option were exercised, then RSP levels on current smoking flights would be reduced to those prevailing on non-smoking flights, and the incremental exposure and incremental risk would be zero. 51mllnrly, modeling is not required to assess the impact of partial bans; population exposures to ETS-related RSP would be reduced essentially in proportion to the reduction 1n number of flight hours during which smoking would be permitted (the reduction would not be exactly proportional because longer flights generally have larger aircraft capacities, greater percentages of time when the no-smoking light is not illuminated, and possibly different smoking rates than shorter flights).

Flight DurationPercentage of FlightsPercentage of Flight Hours
< 1 hour17.67.4
1-1.99 hours48.737.1
2-2.99 hours21.328.1
3-3.99 hours7.213.4
4-4.99 hours3.27.6
5-5.99 hours1.54.3
> 6 hours0.62.1
Total, all durations100.0100.0

A data file supplied by DOT, containing information on a11 flights scheduled for departure from U.S. airports during January 1989, was analyzed to determine the relative frequency: for domestic flights of different durations. The analysis was based on jet flights departing from 70 airports associated with large amounts of air traffic hubs, consistent with the sampling frame used for the study (see Section 2.4). The relative frequencies of flights and flight hours represented by flights of different durations (classified into hourly duration intervals) are summarized in Table 9-4. Flights under two hours in duration account for 44.5 percent of all flight hours. Thus, under the two-hour ban enacted in April 1988 under PL 100-202, smoking would be allowed during 55.5 percent of all flight hours. (A more detailed analysis, factoring in the specific policies of Northwest Airlines and United Airlines, indicated a revised figure of 54.3 percent.) A four-hour ban would limit smoking to 14 percent of all flight hours, and a six-hour ban would restrict smoking to 2 percent of all flight hours, as illustrated in Figure 9-3.

Two hypothetical scenarios were examined for curtailment of the smoking period:

  • Restriction of smoking to a 10-minute period after every two hours of flight time
  • Restriction of smoking to a 10-minute period after every hour of flight time
Case ModeledRSP Concentration, ug/m3
No-Smoking SectionSmoking Section
Flight 1  
- no curtailment
(base case)
- ten-minute smoking period every hour
(total smoking reduced by 25 percent)
3.5 (25X)*91.8 (25X)
- ten-minute smoking period every two hours
(total smoking reduced by 70 percent)
1.4  (70X)36.7 (70X)
Flight 2  
- no curtailment
(base case)
- ten-minute smoking period every hour
(total smoking reduced by 25 percent)
 4.0 (25X) 16.8 (25X)
- ten-minute smoking period every two hours
(total smoking reduced by 70 percent)
1.6 (70X)6.7 (70X)
Flight 3
- no curtailment
(base case)
- ten-minute smoking period every hour
(total smoking reduced by 25 percent)
33.2  (25X)168.2 (25X)
- ten-minute smoking period every two hours
(total smoking reduced by 70 percent)
13.3 (70X)67.3 (70X)
* Numbers in parentheses indicate percent reduction
in concentration from the base

The impact of each scenario on the smoking rate (cigarettes per flight) was estimated for each domestic smoking flight monitored during the study by assigning that each passenger seated in the smoking section would smoke one cigarette during each period when smoking was allowed. On the average, the first scenario would lower total smoking per flight by about 70 percent (i.e., from 51.9 to 15.2 cigarettes/flight) and the second scenario would reduce total smoking by about 25 percent (from 51.9 to 39.8 cigarettes/flight). Each of the flights previously chosen for modeling was modeled with these reductions in the smoking rate. As shown in Table 9-5, the reduction in average RSP concentrations in both the no-smoking and the smoking sections was proportional to the reduction in smoking rate in all three cases. However, as noted earlier, short-term peaks in RSP and gas-phase ETS constituents could rise sharply if smoking periods were restricted, thereby increasing irritation and discomfort for flight attendants and passengers

The impact of the increased fresh-air intake rates was first examined for the flight with the highest smoking rate (flight 3).

Hypothetical increases of 25, 50, 75 and 100 percent in fresh-air intake were modeled. The results displayed in Figure 9-4 indicate a curvilinear relationship between increase in fresh-air intake and RSP concentration in either section; however, the relationship is more direct than indicated — when the intake rate is doubled, the concentrations are halved. Thus, for example, to reduce concentrations by an order of magnitude, a tenfold increase in fresh-air intake would be required. However, such an increase is not likely achievable, and resultant airflows in the cabin would cause intolerable drafts for passengers. In addition, as noted earlier, ozone concentrations in the cabin could increase and relative humidity levels could decrease.

Case ModeledRSP Concentration, ug/m3
No-Smoking SectionSmoking Section
Flight 1
- no increase (base case)4.7122.4
- 50-percent increase3.1 (3)*81.6 (33X)*
Flight 2
- no increase (base case)5.322.3
- 50-percent increase3.6 (33X)14.8 (33X)
Flight 3
- no increase (base case)44.2224.3
- 50-percent lncrease29.5 (33X)149.5 (33X)
- 50-percent increase
for smoking section only
34.2 (23X)172.6 (2)
* Numbers in parentheses indicate percent reduction
in concentration from the base case.

The impact of a more likely achievable 50-percent lncrease in the fresh-air intake rate is shown for each of the three modeled flights in Table 9-6. In each case, concentrations 1n both the no-smoking and smoking sections are reduced by one-third; that is, the concentrations with a 50-percent increase 1n fresh-a1r intake are two-thirds of their original values, consistent with the ratio of the old-to-new intake rate (i.e., 1/1.5 or 0.67).

A special case of increased fresh air is increasing the amount supplied to the smoking section only. If the fresh air supplied to the smoking section is increased by 50 percent, the overall increase is only 10.5 percent (because the smoking section accounts for only 21 percent of the total airflow). Although the reduction in RSP concentrations (23 percent, as shown in the bottom portion of Table 9-6) is less than that achieved with a 50-percent increase in fresh air to the entire cabin, the relative effectiveness is greater; the ratio of concentrations is 0.77 (i.e., 34.2/44.2 for the no-smoking section and 172.6/224.3 for the smoking section) whereas the ratio of 1nflltrntion rates for the aircraft is 0.9 (1/1.105). An assumption 1n modeling this case was that increased air supply to the smoking section would be compensated by increased exhaust from that section; otherwise, the smoking section would be over- supplied, increasing the flow rate from the smoking to the no-smoking section. As a result, RSP levels 1n the smoking section would decrease even further, but levels in the no-smoking section would increase.

The impact of filtration was examined in greatest detail for the flight with the highest smoking rate. The MO-80 aircraft for this flight has a specified air circulation rate of 21 percent (i.e., 21 percent of the air supplied to the cabin is recirculated air). RSP concentrations were modeled with hypothetical filter efficiencies of 0 (i.e., no filter), 0.3, 0.6, 0.8, 0.9, 0.95 and 0.99 (filters currently in use on aircraft are thought to have RSP removal efficiencies in the neighborhood of 0.9). As illustrated in Figure 9-5, overall RSP reductions are less than proportional to filter efficiency, because filtration competes with fresh air for RSP removal and only a fraction of the cabin air 1s recirculated through the filter. A change in filter efficiency from 0 to 0.99 would reduce RSP concentrations for this flight by 33 percent in the no-smoking section (from 63.2 to 42.7 ug/m3) and by 8 to 9 percent in the smoking section (from 243.3 to 222.8 ug/m3).

RSP Concentration, ug/m3
Case Modeled
No-smoking SectionSmoking Section
Flight 2 (MD-80 with 21 percent recirculation)
- 90-percent filter efficiency (base case)5.922.3
- 99-percent filter efficiency5.222.2
 ( )*((IX)
Flight 3 (MP-80 with 21 percent recirculation)
- 90-percent filter efficiency (base case)44.2224.3
- 99-percent filter efficiency42.7222.8
Flight 3 (MD-80 with hypothetical recirculation of 50 percent)
- 90-percent filter efficiency46.6226.7
- 99-percent filter efficiency42.9223.0

Increased filter efficiency would provide no benefit for aircraft with no recirculation capability, such as the B-727 for flight 1. For flights 2 and 3 (MD-80 aircraft), the effect of increasing filter efficiency from 90 to 99 percent was modeled. As shown in Table 9-7, minor reductions in RSP (less than 5 percent) would be achieved with more efficient filters. Because some aircraft have higher recirculation rates (up to 50 percent), flight 3 was also modeled with an MD-80 aircraft having a hypothetical recirculation rate of 50 percent. As shown in the lower portion of Table 9-7, RSP concentrations for the base case (90-percent filter efficiency) were slightly higher with 50-percent recirculation than with 21-percent recirculation. The RSP reductions due to improved filter efficiency are projected to be somewhat greater 1f the aircraft had 50 – percent recirculation, but the reductions are still less than 10 percent.

Similarly, the RSP reductions due to improved filtration will be somewhat greater if the current filter efficiency is below 90 percent; however, as shown previously in Figure 9-5, the percent reduction due to filtration is relatively insensitive to filter efficiency.

9.2.3 Cost-Benefit Analysis

A correlate cost-benefit analysis of alternative mitigation strategies would include a full accounting of all categories of costs and benefits. Mitigation costs include not only the cost of the technical approach, but the losses (1f any) to smokers required to modify their behavior, and, if appropriate, losses in profits to airlines to the extent that smokers fly less often. Economists would measure losses to smokers as their willingness to pay (WTP) to avoid having their behavior modified. This type of measure has been applied with reasonably high replicability of results in other contexts, but not to the issue of valuing smokers’ WTP.

In the analysis below, only technical costs are considered, because of lack of information on thc other cost categories. This limitation means that procedural approaches are given zero cost, clearly an underestimate.

On the benefit side, mortality, morbidity, and comfort considerations dominate. Mortality reductions and their associated economic benefits (measured in terms of the WTP for a reduction in the risk of death divided by the given risk reduction) are estimated below. The linkages between passive ETS exposure and morbidity (acute effects, such as eye irritation, exacerbation of chronic conditions, say by helping initiate an asthma attack, and increase in the probability of developing chronic conditions) are not well enough understood to include these effects (although estimates of the WTP for these effects exist in the economics literature).

Comfort effects related to odor or other effects that might be part of the WTP of nonsmokers to have their ETS exposure reduced also cannot be included because of data limitations. Because a ban on smoking has to have the largest quantifiable benefit but a zero (quantifiable) cost, it must appear as the best approach, subject to the incomplete analysis.

Benefit calculations for the mitigation analysis focused on reductions in risk of lung cancer mortality due to ETS exposure, using RSP as a tracer. To treat mortality risks in monetary terms, estimates are needed either for the w>llingness to pay to avoid specific risks of death or an assumed value of a statistical life (VSL).2 The most recent valuation and wage-risk studies provide YSL estimates in the range of 2 to 5 million (Viscusi, 1986). A value of 3.75 million was chosen for this analysis, consistent with recent EPA assessments (Fisher et al. 1987).

 418 million enplanements per year for domestic flights*
 x 1.84 (hours per flight)**
 million passenger-hours per year
 x .9375 (fraction of time smoking allowed)***
 'fl million passenger-hours per year with smoking permitted
=45 (hours per year per flying passenger used in risk assessment)
 million people flying 45 hours per year
=40 (average lifetime for flying used in risk assessment)
 million lifetimes of flying 45 per hours per year
 x 1.1 (deaths per million flying lifetimes)
 expected deaths per year due to ETS exposure
Flight Attendants
=20 (average lifetime for flying used in risk assessment) thousand lifetimes
of flying 900 hours per year
 x 0.12 (deaths per thousand flying lifetimes) expected deaths per year
due to ETS exposure *
Source: NRC (1986)
** Based on analysis of data file provided by FAA
*** Assuming no-smoking light is illuminated 6.25 percent of the time

Use of the VSL approach requires that the results of the risk assessment be translated into annual expected premature lung cancer deaths due to ETS exposure for the flying population, including both passengers and flight attendants. Based on the estimated cancer risks per 100,000 cabin occupants provided in Section 7.0, estimated annual deaths to be expected in the absence of any ban on smoking for domestic flights are 0.44 for passengers and 0.34 for flight attendants (see Table 9-8). The estimated annual deaths given here are higher than those given in Section 7.0 because the estimates in Table 9-8 assume that smoking would be allowed on all domestic flights, whereas the estimates in Section 7.0 assume that smoking would be allowed only on flights of two-hour or longer durations. Given a YSL of 53.75 million, the expected deaths in Table 9-8 translate into annual economic values of 1.65 million and 51.28 million, respectively. There are also increments in morbidity due to ETS exposure that have not been taken into account.

Projected annual benefits and costs of alternative mitigation options are given in Table 9-9. The greatest benefit ( 2.93 million) would result from a total ban on smoking; benefits other than reduced mortality risk could accrue, for example, from reduced maintenance (e.g., changing of filters) or cleaning costs in the absence of smoking. There are no direct costs of implementing such a ban, although dollar values could conceivably be attached to smokers inconvenience and discomfort.

TABLE 9-9: Projected annual benefits and costs of alternative mitigation options
PassengersAttendantsAnnual Costs
Total ban on smoking100X1.651.280*
Partial ban on smoking
-flights under two hours45X0.740.580
-flights under six hours98X1.621.250
Curtailment of smoking
- 10 minutes every 2 hours70X1.160.900
- 10 minutes every hour25X0.410.320
Increased fresh-air intake
- 50 percent for entire33X0.540.4230.8 to 51.5 cabin
- 50 percent for each237G0.380.296.2 to 10.3 smoking
Increased filter efficiency**
(from 90 to 99 percent)

2 The VSL can be thought of as the average willingness to pay for a given reduction in mortality risk, divided by the risk reduction. Thus, if 1,000 individuals are willing to pay an average of 2,000 for a 11,000 reduction in mortality risk, than the average VSL is 2 million.

However, there are currently no studies of smokers’ willingness to pay for the right to smoke on aircraft. There could be losses in airline ridership due to smokers opting for other modes of transportation, but such losses could not be estimated in this study. In addition to partial smoking bans, options to curtail smoking also provide significant benefits at no apparent cost, particularly the option of a 10-minute smoking period every two hours. Such an option would, however, substantially raise short-term ETS levels and thereby increase acute health responses. For example, application of a steady-state model to the third flight (MD-80 – with 25 smoking passengers) indicates that CO levels in the smoking section could be as high as 5 ppm if all passengers smoked during the 10-minute smoking period. The data from Cain et al. (1987) indicate that 10 percent of nonsmokers exposed to 5-ppm CO (due exclusively to tobacco smoking) for 10 minutes would express dissatisfaction due to eye irritation.

The other options listed in Table 9-9 either have costs that substantially exceed benefits (increased fresh-air intake) or very limited benefits (increased filter efficiency). Several manufacturers were contacted in an attempt to obtain estimates of filter costs, but the manufacturers were reluctant to divulge this information. Although the fuel penalty for increased fresh-air intake is quite small on a per-flight basis ( 10 to 20), the aggregate costs are substantial. The fuel cost penalty was estimated from the relationship shown in Figure 9-6, which was derived from data provided in an NRC report (1986). The incremental fuel cost for a 50-percent increase in fresh-air intake ranges from 0.04 per passenger-hour for OC-10-10 aircraft to 0.067 for a B-727 aircraft. Multiplication by 769 million passenger-hours per year (see Table 9-8) yields an estimated cost range of 30.8 to 51.5 million for added fuel requirements.


9.3.1 Cosmic Radiation

As noted earlier in this section, there are no practical approaches for reducing cosmic radiation levels on aircraft. Thus, the 34only potential mitigation route involves the notion of exposure management. Through this strategy, excessive exposures could be reduced by avoiding extreme northern or southern latitudes and high altitudes where possible. Exposure management could also focus on specific types of personnel facing higher risks, such as female flight attendants in different stages of pregnancy, particularly the first trimester. This type of mitigation strategy applies equally to flight crew members, cabin crew members, and passengers.

9.3.2 Carbon Dioxide

Risk assessment was not performed for carbon dioxide (C02) because health effects of C02 exposure (other than those above occupational guidelines) have not been documented. Nonetheless, C02 levels exceeding 1,000 ppm, the level recommended by ASHRAE for satisfaction of comfort criteria, were measured on a substantial fraction of the monitored flights. Consequently, alternatives for reducing C02 levels in airline cabins were investigated but no cost-benefit analysis could be performed.

There are three types of options for reducing C02 levels — emissions reduction, increased ventilation, and removal by sorption. C02 removal could be achieved, for example, by passing air through an adsorbent bed mounted on a rotating drum or revolving belt (White 1989). Regeneration of the adsorbent would permit hlgh capacity with low bed volume and weight. Continuous regeneration of the adsorbent would be accomplished by passing a small amount of purified air through a heated portion of the bed, then exhausting overboard the heated air containing high concentrations of C02. Alrcraft waste heat from the lubrication oil system or englne exhaust gas would be used as the heat source for regeneration.

Emlssions could be lowered by reduction of seating capacity, but this approach is not likely to be economically attractive to the airline industry. The potential effectiveness of remaining options, involving ventilation or removal, was investigated through modeling. Because the C02 sourccs (passengers and crew) are spread throughout the cabin, a9-32 single-chamber model can be used. Using similar terminology to that used for the two-chamber model described earlier in this section (see Figure 9-2), the model for C02 in the cabin (Cin) can be stated as follows:

TABLE 9-10: C02 concentrations related to ventilation rates
Ventilation RateFlight 1Flight 2Flight 3
Current level (base case)873.21147.8974,9
Increase of 25 percent764.6 (12.4X)*984.3 (14.2X)845.9 (13.2X)
Increase of 50 percent692.1 (20.7X)875.2 (23.7X)759,9 (22.1X)
Increase of 75 percent640.4 (26.7X)797.3 (30.5X)698.5 (28.4X)
Increase of 100 percent601.6 (31.1X)738.9 (35.6X)652.4 (33.1X)

The flights used previously for RSP modeling were also used for this modeling exercise. An emission rate of Q.3 1/min (18,000 ml/h) per passenger and an outdoor concentration of 330 ppm were assumed in making the calculations. The aircraft were assumed to be at full capacity–108 passengers for B-727 aircraft and 142 passengers for MD-80 aircraft.

C02 concentrations related to ventilation rates (currently measured levels and hypothetical increases up to 100 percent) are shown in Table 9-10. Concentrations are projected to decrease by about a third if the fresh-air intake rate were to be doubled. Thus, some flights with C02 levels above 1,000 ppm would likely remain under this scenario. As discussed earlier, this mitigation option would carry a fuel penalty and could also increase ozone levels and decrease humidity levels.

TABLE 9-11: C02 concentrations related to filter removal efficiencies
Filter EfficiencyFlight 2
(21X recirc.)
Flight 3
(21X recirc.)
Flight 3
(50X recirc.)
*Zero (base case)1147.0974.91348.9
25 percent1076.3 (6.2X)**914.1 (6.2X)1079:1 (2O.OX)
50 percent1013.2 (11.7X)860.5 (11.7X)899,3 (33.3X)
75 percent957.0 (16.6X)812.8 (16.6X)770.8 (42.9X)

C02 concentrations related to filter removal efficiencies (zero assumed as current efficiency) are given in able 9-11. (Discussions with a filter manufacturer indicated that removal efficiencies in the neighborhood of 50 to 75 percent may be attainable.) At a 50 percent removal efficiency, C02 levels could be reduced by 12 percent for current MD-80 aircraft (21 percent recirculation) air by 33 percent 1f the recirculation rate were as high as 50 percent).


  • Cain, W.S., T. Tarik, L. See, and B. Leaderer. 1987. “Environmental Tobacco Smoke: Sensory Reactions of Occupants.” Atmos. Environ. 21(2):347-353.
  • Fisher, A., D. Violette, and L. Chestnut. 1987. The Value of Reducing Risks of Death: A Note on New Evidence. Draft report, U. . Environmental Protection Agency.
  • Lorengo, D.G., and A. Porter. 1985. Aircraft Ventilation 5 stems Stud . Final Report DTFA-03-84-C-0084. Atlantic City, NJ: U. . Federal Aviation Administration Technical Center. National Research Council. 1986. Environmental Tobacco Smoke: Measuring Exposures and Assessing Health Effects. National academy Press. Washington, D.C.
  • Ryan, P.B., J. Spengler, and P. Halfpenny. 1988. “Sequential Box Models for Indoor Air Quality: Application to Airliner Cabin Air Quality”. Atmos. Environ. 22(6):1031-1038.
  • U.S. Department of Transportation, Federal Aviation Administration. 1987. Airport Activity Statistics of Certificated Route Air Carriers: Twelve Months Ending December 1, 19 7. U. . department of transportation, Washington, 0. .
  • Viscusi, W.K. 1986. “The Valuation of Risks to Life and Health.” In Bentkover et al., Benefit Assessment: The State of the Art. Dordrecht, Holland, Reidel Publishers.
  • Whlte, O.H. 1989. “Modern Aircraft Cabin Air Purification Including Carbon Dioxide Removal.” Pall Corporation, Glen Cove, NY.

Download poster


  • "Es ist schwieriger, eine vorgefaßte Meinung zu zertrümmern als ein Atom."
    (Het is moeilijker een vooroordeel aan flarden te schieten dan een atoom.)
    Albert Einstein

  • "Als je alles zou laten dat slecht is voor je gezondheid, dan ging je kapot"
    Anonieme arts

  • "The effects of other people smoking in my presence is so small it doesn't worry me."
    Sir Richard Doll, 2001

  • "Een leugen wordt de waarheid als hij maar vaak genoeg wordt herhaald"
    Joseph Goebbels, Minister van Propaganda, Nazi Duitsland

  • "First they ignore you, then they laugh at you, then they fight you, then you win."
    Mahatma Gandhi

  • "There''s no such thing as perfect air. If there was, God wouldn''t have put bristles in our noses"
    Coun. Bill Clement

  • "Better a smoking freedom than a non-smoking tyranny"
    Antonio Martino, Italiaanse Minister van Defensie

  • "If smoking cigars is not permitted in heaven, I won't go."
    Mark Twain

  • I've alllllllways said that asking smokers "do you want to quit?" and reporting the results of that question, as is, is horribly misleading. It's a TWO part question. After asking if one wants to quit it must be followed up with "Why?" Ask why and the majority of the answers will be "because I'm supposed to" (victims of guilt and propaganda), not "because I want to."
    Audrey Silk, NYCCLASH