The assessment and monitoring of sound levels in urban areas is a growing phenomenon. When possible, this is done with fixed monitoring stations, sometimes combined with air pollution detectors. However, these stations cannot record the number of sources or their sound power levels, since they just measure sound levels at the sound level meter. Consequently, the outcomes of these stations cannot be used to tune, calibrate or validate predictive models, which need in input information about source, propagation and receiver positions. A possible strategy to overcome this problem is to adopt models built exclusively on past recorded sound levels with the application of the Time Series Analysis (TSA) approach. In this paper, the case study of the city of Messina (South Italy) will be presented, showing relevant predictive results of the TSA models, both with a deterministic and a stochastic approach. The ARIMA models will exhibit the best performances on a short time range, while the deterministic decomposition will show very good results on average, but with the possibility to extend the prediction to any time in the future. According to the application needed case by case, the proper technique can be selected and applied to the data under study.
On the Use of ARIMA Models to Predict Urban Sound Pressure Levels
Claudio Guarnaccia
;Gabriella Graziuso;Simona Mancini;Joseph Quartieri
2020-01-01
Abstract
The assessment and monitoring of sound levels in urban areas is a growing phenomenon. When possible, this is done with fixed monitoring stations, sometimes combined with air pollution detectors. However, these stations cannot record the number of sources or their sound power levels, since they just measure sound levels at the sound level meter. Consequently, the outcomes of these stations cannot be used to tune, calibrate or validate predictive models, which need in input information about source, propagation and receiver positions. A possible strategy to overcome this problem is to adopt models built exclusively on past recorded sound levels with the application of the Time Series Analysis (TSA) approach. In this paper, the case study of the city of Messina (South Italy) will be presented, showing relevant predictive results of the TSA models, both with a deterministic and a stochastic approach. The ARIMA models will exhibit the best performances on a short time range, while the deterministic decomposition will show very good results on average, but with the possibility to extend the prediction to any time in the future. According to the application needed case by case, the proper technique can be selected and applied to the data under study.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.