Spatial sampling for night levels estimation in urban environments.

The procedure for spatial sampling in order to find out the relation between L(day) and L(night) for urban noise is presented. From that data, other parameters like L(dn) or L(den) can be easily obtained from L(day) values. To this end, a long-term measuring campaign was carried out in eight cities of different types and sizes in northeastern Spain. The statistical treatment of the measures was based on a characterization of the streets in view of their type, including factors such as traffic and land use, which enabled a final classification of six types of streets. The results show that streets of the same type located in different cities do not have the same value for L(day)-L(night), due to socioeconomic factors, but analyzed city by city have a close to normal (or t-Student) distribution. Under this behavior, it is demonstrated that a sample of between 14 and 25 points, depending on the city characteristics, is needed in order to calculate L(day)-L(night), with a confidence level of 95% and a margin of error of +/- 1 dB(A).