Final Thesis

Data mining of weather conditions for sound propagation in the atmosphere

Key Info

Basic Information

Research Area:
Numerical Acoustics,
Noise Research
Type of Thesis:


Master Thesis of Tavoosi, Nasim

The work presented in this thesis addresses the extraction of hidden information from the climate system to reach a model of weather data for different stations. It investigates the effect of generated data based on original weather data on sound propagation in the atmosphere. Temperature, relative humidity, static pressure, wind direction and wind speed in different altitude are the main altitude dependent extracted data which are assumed as a multivariate random variable system. In contrast to the first assumption, the univariate random variables do not follow a normal distribution and they are not identically distributed. Therefore, the estimation of the multivariate probability density function seems challenging. In the course of this thesis, the dependency between univariate random variables is used to estimate the multivariate density function based on a concept named copula. By applying the copula, the multivariate joint probability distribution could be estimated independent of marginal distributions. Copula allows us to precisely model the height-dependent weather data based on collected measurements and characterizes both linear and non-linear dependency structure between single random variables. The generated model is used to evaluate the effect of weather data on sound propagation in the atmosphere resulting to find the variations of the atmospheric transfer function. The spectral contents of results could be utilized later in the context of auralization of acoustic outdoor scenario.