The present work can be generically divided into two main areas. In the first one, Chapters 2, 3 and 4, under a qualitative point of view the conceptual framework of the studied phenomenology is presented. In the second one, Chapters 5 and 6, essentially devoted to quantitative aspects, the formulation and the development of the interpretative model of the considered climate variables (temperature, rainfall, wind speed and wind direction is decribed.
The applicability of different statistical techniques in the treatment of climatic data, namely those characterising the urban climate of the city of Porto, is assessed in the second area. This exploratory work aims at clarifying the spatial-temporal variability associated to climatic data samples resulting from mobile acquisition campaigns and from automatic meteorological stations situated in the urban perimeter of Porto.
In the field of the spatial characterisation of the studied system and defined a series of data treatments are presented based in nonparametric statistics, multivariate statistics and geostatistical methodologies, leading to mapping the spatial variability of its considered climatic parameters.
The attempt to interpret the temporal variability of this phenomenon is achieved by employing classical methods used in this research such as, spectral analysis, directional satistics, Markov chain and auto-correlation models.