Page 12 - Winter 2009

Fig. 2. Random sampling of sound—(a) the sound is sampled 1800 times. The “A- weighted, equivalent, continuous sound level (Leq)” is 42.9 dB. (b) the sound is sampled 150 times. The Leq is 42.0 dB. (c) the sound is sampled 15 times. The Leq is 42.2 dB.
the area shaded in red is placed in the “community” bin. The failure to include the red area with the black does not signif- icantly diminish the reported airport noise. However, includ- ing the red area in the community noise bin, shaded in blue, substantially increases the apparent community noise. We termed these red areas as “tails.” These tails to the airport noise events significantly distort and increase the communi- ty noise levels. Table 2 gives the total noise levels for the eighteen daytime monitoring sites, the “community” noise levels as computed by the airport method, the true commu- nity noise levels determined with an observer present, and the amount by which the airport computational method exaggerates the true community levels. One can note that the error generated by the airport’s computational methods is on average 4 dB during daytime hours.
Another reason airport noise monitors fail to properly measure the community noise is that requiring correlation with control tower records means that any errors in record keeping propagate as airport noise being included as commu- nity noise. The airport’s records show that about four percent of the aircraft are missed by the airport’s noise analysis system. That is, the tower count of operations is about four percent higher than the noise analysis system count of operations.
Finally, not all airport noise monitoring systems are designed to measure lower noise levels, since aircraft noise near an airport is typically a relatively noisy event. So the electrical noise floor of the monitor can be a problem when trying to measure quiet community settings.
In summary, airport noise monitors do not measure community noise properly because they exclude from the air- port bin and count in the community bin events that do not reach the threshold, events that reach the threshold but not for a long enough time, they fail to include the tails of valid events, they fail to include events uncorrelated with tower records, and they fail to account for self-noise of the moni- toring system.
To show that the community noise data reported by the noise-monitoring system was biased by these various factors, two forms of proof were used. First, we compared hourly community noise levels for those hours when no aircraft were reported as present in the vicinity of a given monitor with the same hours of the day, but for days when some air- craft were recorded as present. Secondly, we graphed and compared the hourly aircraft noise level with the hourly “community” noise levels. There is no reason for the true community noise to be correlated with the airport noise—the noise from children playing, vehicles in the street, etc. should not be correlated with the noise of aircraft arriving or depart- ing MSP. Table 4 compares hours of the day for days where no aircraft were reported in the vicinity of that monitor and days where aircraft were reported to be in the vicinity. Clearly there is a large shift in the “community” level when aircraft are present versus when aircraft are not present. This is one clear indication that the reported community levels include significant amounts of airport noise. Figures 4 and 5 plot air- port reported levels versus “community” levels. As shown on these figures there is a significant correlation between the two, again indicating that the “community” levels include sig- nificant quantities of airport noise. Thus, this analysis serves
Fig. 3. Example of a nighttime aircraft fly-by noise from measured data. The black area is aircraft noise that the airport attributes to “airport noise.” The red area is aircraft noise that the airport attributes to “community noise.” The blue area is true “community noise.”
Minneapolis Sues its Airports Commission 11