Urban noise: Measurement duration and modelling of noise levels in three different cities

Noise maps, intended as management tools, must give information with sufficient precision (reliability), at a reasonable price (economy) with update possibilities (actualization) to accurately describe the studied region. Urban noise measurements need to be dense enough to provide a good spatial representation of noise levels in the region. They must also extend over a long enough period of time to be stable and reliable, but not extremely long in order not to increase field work expense beyond acceptable levels. Data from field measurement programs carried out to obtain a noise map might also be used to develop predictive noise models in order to update the noise map, or to evaluate noise impact of possible traffic management changes (circulation patterns, bus ways, etc.). Predictive model application in the same city for which they were developed is a guarantee of the adequacy of the tool, especially when model precision and range of applicability is specified. If no precision or applicability range is given, unpredictable errors can occur without any basis for analysis. This paper compares recommended noise measurement durations for three different cities: a European city (Valencia, Spain); a South American city (Montevideo, Uruguay); and a small town in Uruguay (Rivera). Predictive models are then used in all three cases and their outputs are compared. Differences are large enough to show the risks of using predictive models developed for cities with different characteristics (urban, social and cultural differences, including different recommended noise measurement duration), even when hourly traffic volume would suggest its applicability.