Assessing Community Annoyance of Helicopter Noise (2017)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

14 2.1 Introduction The literature review contained in Appendix B, and described in Chapter 1, identified hypotheses about the absolute and relative annoyances of fixed- and rotary-wing aircraft and examined the published evidence in favor of and contrary to the various hypotheses. Much of the historical evidence about these hypotheses proved to be either contradictory or ambiguous. As a practical matter, the hypotheses may be expressed in terms of the ability of various factors to explain variance in the relationship between helicopter noise exposure and the prevalence of a consequential degree of annoyance in communities. The nine hypotheses described in Table 1-1 were summarized and restated in seven hypotheses that were tested in this study. The nonacoustic hypotheses (general nonacoustic, noticeability, and situational aware- ness) were combined into one, and A-weighted and cumulative hypotheses were considered in combination. Loosely stated in simplified form, the hypotheses are: 1. The prevalence of annoyance due to helicopter noise in a community is greater than that associated with comparable levels of exposure to noise produced by fixed-wing aircraft; 2. The prevalence of annoyance due to helicopter noise is most appropriately predicted in units of A-weighted cumulative exposure; 3. The prevalence of annoyance due to helicopter noise is strongly influenced by its impulsive character, and thus requires an impulsiveness “correction” to A-weighted cumulative exposure (cumulative helicopter noise exposure corrections may be different for different helicopters at different exposure points on the ground); 4. The prevalence of annoyance due to helicopter noise is strongly influenced by indoor second- ary emissions (rattle and vibration) due to its low-frequency content; 5. The prevalence of annoyance due to helicopter noise is appreciably influenced by nonacoustic factors; 6. The prevalence of annoyance due to helicopter noise is appreciably influenced by proximity to helicopter flight paths; and 7. Complaints lodged about helicopter noise are more reliable predictors of the prevalence of annoyance than measures of exposure to helicopter noise or proximity to helicopter flight paths. The following sections describe some of the factors that complicate the testing of these hypotheses. These issues are discussed next in considerable detail, including the nature and relative amounts of exposure to fixed- and rotary-wing aircraft noise, population and sample C H A P T E R 2 Development of Hypotheses

Development of Hypotheses 15 size requirements, methods for quantifying nonacoustic influences on annoyance, magni- tudes of expected effects, site selection criteria, and content and method of questionnaire administration. 2.2 Factors Complicating Hypothesis Testing Both general and site-specific factors complicate hypothesis testing and interpretations of social survey findings. For example, some of the hypotheses are not mutually exclusive. It is pos- sible that an impulsiveness correction may improve the ability of A-weighted measurements to predict the prevalence of annoyance created by helicopters, at least in flight regimes that produce conspicuous blade slap. It is also possible, however, that audible blade slap, rattle, and vibra- tion are sufficiently correlated with one another that any of these factors could provide equally plausible explanations. Likewise, simple proximity to helicopter flight paths is highly correlated with most measures of noise exposure, even if the predominant cause of annoyance (e.g., fear of a crash) is not necessarily audible airborne sound. In the abstract, the field research techniques that can produce evidence in favor or contrary to these hypotheses are clear. Opinion surveys can be conducted with representative samples of people in neighborhoods exposed to varying amounts of helicopter (and potentially fixed- wing) noise. Field measurements of aircraft noise exposure can be made prior to and during the interviewing process, in areas with large residential populations living within geographically distinct areas with well-defined boundaries with homogenous exposure to noise produced by similar amounts of rotary- and fixed-wing aircraft operations, little seasonal variability, and a wide range of aircraft types and exposure levels. Many factors can reduce the reliability and generalizability of social survey findings, com- promise the ability to make confirming field measurements of actual noise exposure, increase interviewing costs, or make it difficult to delineate geographic areas eligible for interview. The following are among the factors that complicate or even preclude conduct of a social survey of relative reactions to fixed- and rotary-wing noise exposure at any site: • Geographic disparities between areas with high helicopter noise exposure and areas with sufficiently large residential populations; • Greatly disparate amounts of noise exposure due to fixed- and rotary-wing aircraft operations; • Narrow ranges of exposure levels created by helicopter noise8; • Small numbers of operations in particular flight modes (cruise, hover, rapid ascent and/or descent, taxiing, etc.); • Insufficient numbers of respondents to yield a sample large enough to document small differences in annoyance prevalence rates; • Unavailability of reliable radar/transponder information about actual rotorcraft flight paths; • Unreliability of noise modeling due to variability, complexity, seasonality, or sketchy knowledge of operations; • Excessively high ambient neighborhood noise levels; • Unavailability of complaint records; and • Large proportions of non-English speaking residents (for reasons of cost). The consequence of all of these complications is that few sites are likely to be appropriate for testing all hypotheses. In particular, it may not be possible to test many of the other hypotheses if priority in site selection is given to a direct test of the basic hypothesis that helicopter noise is more annoying than fixed-wing aircraft noise. A major goal of site selection is to identify a set of sites that allows testing for as many hypotheses as feasible.

16 Assessing Community Annoyance of Helicopter Noise 2.3 Some General Constraints on Hypothesis Testing 2.3.1 Geographic Disparities Between Areas with High Helicopter Noise Exposure and Areas with Sufficiently Large Residential Populations Helicopter noise exposure levels are generally greatest in geographic areas near terminal oper- ating areas and in close proximity to flight routes. Good land use and flight route planning tend to minimize residential populations in such areas. Thus, to avoid overflights of residential areas, helipads are often located near shorelines, and approach and departure routes to them often overfly bodies of water rather than residential neighborhoods. Heliports are also often located in commercial and industrial areas with relatively few residences as well as in very high-density business districts with elevated ambient noise levels and urban canyons. The net effect of good planning practice is to minimize the exposure of residential areas with low ambient noise levels to very high levels of helicopter noise exposure. This, in turn, makes it difficult to identify interviewing sites in which opinions about effects of high levels of helicopter noise can be solicited from suitably large numbers of households. 2.3.2 Disparate Exposures to Fixed- and Rotary-Wing Aircraft Operations Areas of high exposure to fixed-wing aircraft noise are concentrated around runway ends and in approach and departure corridors along extended runway centerlines. For air traffic safety reasons, these are precisely the areas from which helicopter operations are excluded. It was difficult to locate interviewing sites with high levels of exposure to both fixed- and rotary-wing aircraft noise. It may be less difficult to locate residential areas exposed to intermediate or low levels of both types of aircraft noise, but these are unlikely to be areas in which the greatest differences in the annoyance of rotary- and fixed-wing aircraft noise are likely to be observed. Smaller differences between the annoyance of the two types of aircraft noise require larger sample sizes to discern, and hence, larger residential populations from which to draw such samples. By definition, the areal extents of low-density residential areas (i.e., those with low outdoor ambient noise levels) are greater than those of high-population density areas. Aircraft noise lev- els across these greater areas are likely to vary considerably, perhaps by ± 10 dB or more.9 In turn, this implies that sub-populations in low-population density areas with similar noise exposure levels may be quite small. It may therefore be impractical to stratify samples in low-population density areas into geographic zones within narrow exposure ranges (say, ± 1.5 dB). If it is not possible to identify large enough sample strata with reasonably homogeneous noise exposure that span a wide enough exposure range, it will be necessary to model exposure levels of individual survey respondents. Because nominal integrated noise model (INM) flight tracks are often assumption-based rather than empirical, credible inferences of helicopter noise exposure levels may be limited to those at sites for which high-quality radar data are available. In practice, this may restrict interviewing sites to those near major airports with good radar coverage. INM was used because the study began before the Aviation Environmental Design Tool (AEDT) was released. INM Version 7.0d and AEDT Version 2b make identical noise predictions in any event. 2.3.3 Narrow Ranges of Exposure Levels Created by Helicopter Noise and/or Small Numbers of Operations in Particular Flight Modes A narrow range in exposure levels within a given community implies that the shape of the dosage-response curve cannot be well defined empirically, regardless of the number of

Development of Hypotheses 17 respondents.10 While the findings of this study will be analyzed in part with respect to a fixed- shape dosage-response curve that translates laterally depending on local community tolerance to aircraft noise sources, it is highly desirable to verify the fixed-shape assumption within communities. A narrow exposure range can preclude this possibility. Furthermore, helicopter sound level emissions can differ markedly between flight modes (in addition to differences in helicopter types). These flight modes can change rapidly along a flight corridor. For example, if a helicopter is descending rapidly, then the BVI may create sig- nificant amounts of blade slap, which can affect both its A-weighted sound level as well as any impulsiveness adjustments. On the other hand, if the aircraft goes into a very shallow decent or level flight, blade slap can cease very quickly. Consistency of operation along any given flight corridor would benefit site selection, but such consistency cannot be expected from one flight to the next at sites with differing types of helicopters and modes of operation. Of greatest concern is the ability to estimate when high sound level modes of operation occur, since even a small percentage of high sound level events may control annoyance responses. 2.3.4 Unavailability of Reliable Radar Flight Performance Information About Actual Rotorcraft Flight Paths and Procedures Reliable radar information is essential for modeling noise levels over the interviewing area. Helicopters almost always operate as visual flight rules (VFR) flights, and hence do not usually file flight plans or transmit a unique transponder code. Helicopter radar tracks must therefore be distinguished from fixed-wing aircraft radar tracks based on unique level flight segments at low altitude, origin or destination at specific heliport locations, or tracks within a known and exclusive helicopter corridor. A test program is in progress in Los Angeles in which VFR helicopter flights will not use 1200 as their squawk code, but will be assigned unique helicopter codes. This simplifies identification of helicopter tracks. Radar data is a regularly acquired data set at airports with modern airport noise monitoring systems. It is also possible to obtain radar data from FAA. Radar data will be available only within reasonable distances of aircraft surveillance radar (ASR) sites that will be located near airports, and for which no terrain or building obstructions intervene between the ASR sensor and the helicopter paths. Because helicopter tracks are lower and farther from the airport than those of fixed-wing aircraft, this may limit survey sites to those near (within 20 nm and without obstructions) ASR sites. Although helicopter tracks can be distinguished from fixed-wing aircraft tracks by speed, the study sites selected all had programs in place for unique helicopter squawk codes. As noted later, LAS and DCA also assign unique call signs to helicopters. 2.3.5 Questionable Reliability of Noise Modeling Due to Operational Variability, Complexity, Seasonality, or Sketchy Knowledge of Operations INM-based noise modeling for civil airports is conventionally conducted on an “average annual day” basis. If helicopter flight activity at a potential interviewing site is concentrated in one season of the year, but interviewing is conducted in a different season, standard noise modeling contours may not work well for stratifying samples by noise exposure. Such noise modeling errors could bias observed dosage-response relationships. Likewise, as with any model, generalizations and simplifications are made regarding flight paths. Noise modeling at the block or individual residence level is preferable for estimating respondents’ noise doses. The modeling procedure can also be adjusted to reflect sound

18 Assessing Community Annoyance of Helicopter Noise level measurements made at various sites within the interviewing area. Hence, the combined uncertainty in both measurements and modeling will be reflected in the computed doses. Dose uncertainty is ultimately determined by the less reliable form of estimation, whether measure- ment or modeling. Selection of interviewing sites should be based in part on the complexity of operations to estimate the size of a difference in exposure that can be attributed to aircraft type. All of these considerations underscore the need to measure, model, and ask attitudinal ques- tions about identical time frames to maximize the strength of association between dose and response. 2.3.6 Excessively High Ambient Neighborhood Noise Levels Excessively high ambient sound levels in the vicinity of heliports pose several complications for present purposes. In extreme cases, such as heliports in very high population density areas, or in areas with high levels of highway traffic noise, extraneous noise sources may mask the noise of some helicopter operations. High ambient noise levels also complicate estimation of individual noise event levels, and thus may influence differing attitudes toward aircraft noise in urban, suburban and rural areas. Since low-frequency noise level measurements are susceptible to large pseudo-noise artifacts in windy conditions (such as wind interacting with the microphone), one criterion for survey site selection may be typical wind speeds. Areas expected to have high wind speeds and high turbulence levels were avoided. Nonetheless, two unseasonable weather fronts moved through Long Beach during the field data collection period. 2.3.7 Unavailability of Complaint Records Many airports collect detailed complaint records. This may not be true at all heliports. Avail- ability of complaint records was considered in site selection. 2.3.8 Large Proportions of Residents Ineligible or Unavailable for Interview Unless the expense of translating the survey instrument (questionnaire) into other languages is affordable, response rates may be low in areas with large proportions of non-English speakers. A highly transient population (for example, of students, as in the vicinity of a helicopter-served hospital or at a major university) can also be difficult to contact. 2.4 Discussion of Potential Tests of Hypotheses Several of the hypotheses summarized in Table 1-1 can be tested via analyses of responses to individual questionnaire items about the annoyance of aircraft noise. Several other hypoth- eses are testable by comparing responses across sites chosen for the present study, or by less direct means described below. The suggested form of closed response category annoyance items is: “While you’ve been at home during (time period of interest), have you been bothered or annoyed by (noise source)?” and if yes, “Would you say that you’ve been slightly, moderately, very, or extremely annoyed by aircraft noise while you’ve been at home during (time period of interest)?” The time period of interest can be either (or both) the week prior to interview—during which extensive empirical noise measurements were made at field sites—or the year prior

Development of Hypotheses 19 to interview, over which exposure estimates may be made from modeling of annual average day exposure. Hypothesis 1: Decibel for decibel, rotary-wing aircraft noise is more annoying than fixed-wing aircraft noise. The most basic of the hypotheses holds that exposure to noise produced by rotary-wing air- craft is more annoying than exposure to an equivalent amount of noise produced by fixed-wing aircraft. The hypothesis does not specify why one type of aircraft noise may be more annoying than another—for example, because of spectral differences in emissions, indoor vibration or rattle excited by rotary-wing aircraft, greater noticeability of helicopter noise in some ambient noise environments, and so forth. Thus, even if the hypothesis can be empirically confirmed, it would not necessarily yield enough understanding to be useful for improved explanatory, regulatory, or policy purposes. As discussed in Section 1.2 in general terms, and Appendix A in greater detail, the complex and varied nature of rotary-wing operations can make it difficult to fully test this hypothesis. Helicopter noise may vary relatively little from fixed-wing aircraft noise at some locations and in some flight regimes (e.g., at off-track, long-range, sideline locations during straight and level cruise) but can vary greatly from that of fixed-wing aircraft in other flight regimes (e.g., in duration, level, audibility, predictability, and impulsiveness during low-altitude maneuvering). The most useful tests of this hypothesis must be able to characterize not just exposure levels, but also the nature of helicopter noise emissions. It may be necessary to test this hypothesis at more than one site, since no one site may include all of the helicopter flight regimes of potential interest. The most direct test of this hypothesis would compare the annoyance judgments of the same interview respondents to very similar levels of fixed- and rotary-wing aircraft noise. If it is possible to conduct interviews at sites with sufficient numbers of respondents who are exposed to comparable levels of fixed- and rotary-wing aircraft noise, the general form of questionnaire items that could test this hypothesis would be: “While you were at home last week, did helicopter noise bother or annoy you?” “Would you say you were not at all, slightly, moderately, very, or extremely annoyed by noise from helicopters while you were at home last week?” “While you were at home last week, did noise from aircraft other than helicopters bother or annoy you?” “Would you say you were not at all, slightly, moderately, very, or extremely annoyed by noise from aircraft other than helicopters while you were at home last week?” It could also be helpful to include a questionnaire item seeking a direct comparison of the annoyance of fixed- and rotary-wing aircraft noise, of the general form as follows: “While you were at home last week, were you annoyed more greatly by noise made by helicopters or noise made by other types of aircraft?” As noted earlier, it may not be possible to identify sites at which sufficient numbers of eligible respondents are exposed to similar amounts of both forms of aircraft noise. A less direct test of the hypothesis is still possible if this should prove to be the case. The opinions of respondents about helicopter noise could be compared with the opinions about fixed-wing aircraft noise of 75,000 respondents to prior surveys about the annoyance of aircraft noise (Fidell et al. 2011). Annoyance prevalence rates measured in the planned study could then be compared with previ- ously measured annoyance prevalence rates at as many as hundreds of sites with similar noise exposure levels at which respondents had been queried about their annoyance with exposure to fixed-wing aircraft noise.

20 Assessing Community Annoyance of Helicopter Noise Hypothesis 2: The prevalence of annoyance due to rotary-wing noise is most appropriately predicted in units of A-weighted cumulative exposure. No specific questionnaire items are required to test this hypothesis. The utility of the A-weighting network for predicting the annoyance of helicopter noise can be gauged instead via simple calculations of variance accounted for in the relationship between various measures of noise exposure and the prevalence of a consequential degree of annoyance at interviewing sites. All that is required is that noise measurements accompanying interviewing be conducted in such a manner that alternative frequency weightings and other adjustments can be calculated. This can be accomplished by capturing raw acoustic waveforms and post-processing them with reference to radar-confirmed helicopter flight operations. As in the case of testing Hypothesis 1, a fully generalizable test of Hypothesis 2 requires both social and acoustic measurements of helicopter noise produced in varying flight regimes. Hypotheses 3 and 4: Main rotor impulsive noise controls the annoyance of helicopter noise (and hence requires an impulsive noise “correction” to A-weighted measurements); the prevalence of annoyance due to helicopter noise is strongly influenced by indoor secondary emissions (rattle and vibration) due to its low-frequency content. Hypotheses 3 and 4 are most appropriately tested at sites exposed to considerable amounts of BVI (or “blade slap”) noise. Due to the highly directional nature of blade slap noise, this con- straint may limit testing of these hypotheses to sites exposed to landing noise in the immediate vicinity of helipads, or to cruise noise in the direction of flight and directly beneath helicopter flight paths. Questionnaire items of interest for testing Hypothesis 3 require a “yes” response to a prior question about annoyance with helicopter noise.11 Respondents who report some degree of annoyance with helicopter noise can then be asked questions of the form: “Have you been not at all, slightly, moderately, very, or extremely annoyed by repeated pounding or slapping noises made by helicopter rotors?” “Have you been not at all, slightly, moderately, very, or extremely annoyed by droning noises made by helicopters?” “Have you been slightly, moderately, very, or extremely annoyed by whining noises made by helicopters?” “What sort of helicopter noise annoys you most?” Questionnaire items of interest for testing Hypothesis 4 also require a “yes” response to a prior question about annoyance with helicopter noise. Respondents who express some degree of annoyance with helicopter noise can then be asked previously tested (Fidell et al., 1999, 2002a) questions of the form: “Do helicopters make vibrations or rattling sounds in your home?” “Are you bothered or annoyed by these vibrations or rattling sounds in your home?” “Would you say that you are slightly annoyed, moderately annoyed, very annoyed, or extremely annoyed by vibrations or rattling sounds in your home?” “About how often do you notice vibrations or rattling sounds in your home made by helicopters?”

Development of Hypotheses 21 Hypothesis 5: The prevalence of annoyance due to helicopter noise is heavily influenced by nonacoustic factors. The most direct test of this hypothesis would require soliciting annoyance judgments from respondents in two or more communities with very similar helicopter noise exposure but very different tolerances for helicopter noise. It is not yet apparent whether such pairs of communi- ties can be found. An alternative test of this hypothesis could be conducted, however, with reference to the database of observations of annoyance prevalence rates for fixed-wing aircraft in more than 500 communities worldwide. The survey instrument itself would not need any items other than the customary ones described in the discussion of Hypothesis 1. Hypothesis 6: The prevalence of annoyance due to helicopter noise is heavily influenced by proximity to helicopter flight paths. This hypothesis is most readily tested at sites along well-defined and heavily trafficked helicopter routes. Geographic information system (GIS) methods can be used to estimate how long helicop- ters flew within varying distances of respondents’ homes over the course of the week prior to inter- view. Since proximity to flight paths and noise exposure levels are highly correlated, it would be necessary to conduct ancillary GIS-based analyses of complaint rates to distinguish between expo- sure and proximity as determinants of annoyance and complaints, such as those described below. Figure 2-1 and Figure 2-2 show spatial complaint densities in the vicinity of Seattle-Tacoma International Airport (SEA) before and after the opening of a new runway. Both the numbers and westward shift of complaints are consistent with a small but abrupt shift in aircraft noise Figure 2-1. Three-dimensional spatial density representation (viewed obliquely) of complaints in 12 months prior to the start of operations on third runway.

22 Assessing Community Annoyance of Helicopter Noise exposure levels in the immediate vicinity of the airport. Actual changes in the geographic distri- bution of complaints were closely contained in the vicinity of changes in flight paths associated with the runway opening. The actual change in DNL was minor. Even though the change received widespread media coverage, the pattern of changes in complaints could not be attributed to the media coverage per se. Rather than reflecting a community-wide response to media coverage, the changes in spatial density of complaints were limited to the vicinity of changed flight tracks. Hypothesis 7: Complaints lodged about helicopter noise are more reliable predictors of the prevalence of annoyance than measures of exposure to helicopter noise or proximity to helicopter flight paths. One or more questionnaire items inquiring whether social survey respondents had lodged single or multiple complaints about helicopter noise might be a useful predictor of the prevalence of annoyance with helicopter noise. It is conceivable that responses to such items might predict actual annoyance prevalence rates as well as measures of exposure, per se, or measures of proxim- ity to helicopter flight paths. If access is available to helicopter noise complaints at airports with appreciable numbers of helicopter operations, it might be possible to compare empirical measurements of annoyance prevalence rates with total numbers of complaints and numbers of complaints per complain- ant, in the manner described by Fidell et al. (2012). The latter reference demonstrated that the number of complaints per complainant at half a dozen airports closely followed a power law relationship known as Zipf ’s Law. Figure 2-2. Three-dimensional spatial density representation (viewed obliquely) of complaints in 12 months following the start of operations on third runway.