Page 11 - Winter 2009

Fig. 1. Site map showing Minneapolis–St. Paul Airport (MSP) monitor sites (18 daytime sites and 4 nighttime sites listed in Table 1).
for these 1,800 seconds is 42.9 dB. Figure 2b shows 150 ran- dom seconds from these 1,800 seconds and the computed Leq is 42.0 dB. Figure 2c shows just 15 random seconds selected from the 1,800 seconds and the Leq is 42.2 dB. So with this simple example, the random samples are within 1 dB of the universe of samples. (This is generally the case—if statistics are more or less Gaussian, a sample of twenty-five independent events usually gives a good estimate of the true average.) What was needed was a time- and space-average for the residential communities in the zone of interest to indicate the levels of noise generally found in these commu- nities during the day, and the eighteen 30-minute samples provided these data. The daytime measurement results are given in Table 2 and show that the typical community A- weighted Leq values, absent noise from MSP, are in a range from 44 to 54 dB with an average of 49 dB. Clearly these Leq values fall within the range defined for quietude.
For nighttime, whole-night continuous measurements at four sites were measured in an attempt to obtain two complete measurement nights per site. This approach for nighttime, using a small set of sites, was chosen for logistical reasons. Figure 1 shows the locations of four sites at which the moni- toring was performed; however only three of these sites yield- ed useful results. Table 3 shows the nighttime monitoring results at the three useful sites. At the fourth site, the student interns operating the equipment did not understand the instructions well enough during the first night, and during the second night the power to the monitor was accidentally turned off, thus data from both nights at this fourth site were lost. Again, the results in Table 3 show that these example nighttime levels are well within the levels defined for quietude.
(b) Establish why airport-noise monitors systematically overstate the actual community noise, absent the airport.
Airport noise monitoring sys- tems are installed by airports to measure the airport noise, and to ensure that the airport noise data are not corrupted by non-airport noise. That is, their event detection algorithms are designed to be fairly certain that noise included as air- port noise is truly airport noise and does not include significant amounts of non-airport noise. Thus, they divide the universe of data into two bins—(1) noise that they are certain is airport noise, and (2) noise that they are not certain is airport noise. Universally, they term the second bin as “community noise—noise that they are not cer- tain is airport noise.” In reality, there should be three bins of data
• There should be the bin that is
“almost certainly only airport
noise.”
• There should be a similar bin that is “almost certain-
ly only community noise.”
• There should be a bin of data for which it is unclear
whether it is airport noise, community noise or both. The first two of these three bins should be used to esti- mate the airport noise and the community noise, respective- ly. The third bin should be used only in computation of the total noise at the site, and not included in estimates of aircraft
noise or community noise.
However, that is not how airport noise monitors are pro-
grammed. In the case of MSP, they require that airport noise be above a threshold long enough, but not too long. Specifically, they require that an aircraft noise event be above 65 decibels for at least eight seconds, and that it correlates with control tower records for an aircraft being in the area of the monitor in question at the time in question. It is easy to list a series of reasons why this sampling strategy, or any sim- ilar sampling, cannot just divide the data into two bins: air- craft and non-aircraft.
• There are aircraft events that never reach the 65 dB threshold such as smaller planes or planes more dis- tant from the noise monitor.
• There are aircraft noise events that exceed the 65 dB threshold but not for eight consecutive seconds.
• There are valid events for which the noise below 65 dB is not included in the airport noise bin. They are included with the community noise.
Figure 3, derived from our MSP nighttime data, shows an aircraft noise event rising above the apparent 36 dB ambi- ent to 69 dB. In terms of the airport, the area of the curve shaded in black is included in the airport noise bin. However,
10 Acoustics Today, October 2009