Temporal and spatial stratification for the estimation of nocturnal long-term noise levels

Noise pollution in cities is mainly caused by the vehicular traffic but, depending on the place under assessment, it could be affected by the land use. For noise assessment and strategic noise mapping, the night period equivalent level (L_night), which evaluates sleep disturbance, is one of the requirements of the European Directive 2002/49/EC to be presented for the equivalent time of one year. This research aims to find the influence of the land use in the weekdays stratification to improve the accuracy of the long-term noise level estimation for the night period. It is found that depending on the land use of the place under assessment, the weekdays temporal and spatial stratification could be affected by leisure activities. From a statistical analysis based on a clustering procedure of L_night samples in 19 points, it is observed that both, temporal and spatial stratification depend on the intensity of the surrounding leisure activity, and not on traffic. Following these stratification criteria, a sampling method is presented that reduces by 47% the number of days needed to estimate the annual levels with respect to random sampling.