Modelagem estatística da ilha urbana de calor

• Main Author: Stier, Natália Pianca
• Publication Date: 2021
• Format: Master thesis
• Language: por
• Source: Repositório Institucional da UTFPR (da Universidade Tecnológica Federal do Paraná (RIUT))
• Download full: http://repositorio.utfpr.edu.br/jspui/handle/1/25093
• topic: Ilha de calor urbana; Climatologia urbana; Modelos lineares (Estatística); Temperatura atmosférica; Urban heat island; Urban climatology; Linear models (Statistics); Atmospheric temperature; CNPQ::ENGENHARIAS::ENGENHARIA SANITARIA; Engenharia/Tecnologia/Gestão

Rapid urbanization produces significant changes in the landscape of a region due to the replacement of natural surfaces with impermeable materials and human activities. Such changes decrease evapotranspiration, increase surface and air temperature, interfere with atmospheric dynamics, forming an urban heat island (IUC), which consists of the temperature difference between urban and rural areas. The IUC causes meteorological, economic, social, environmental, biological and human health impacts. Therefore, this study characterized the formation and development of the IUC in Londrina, northern Paraná, from thirteen points of data collection distributed in its urban fabric in the period from 27/08/2018 to 12/13/2018. Initially, the intensity of the urban heat island (IIUC) was calculated and later, the candidate variables were selected to apply a multiple linear regression model (RLM) to model the air temperature during the daytime study period (06:00 am-18h00) and night (18h00 - 06h00). The bootstrap methodology was applied to validate the daytime and nighttime models, separately. This approach validates a multivariate model by extracting a large number of subsamples and estimating models for each one. The results showed that the average daily air temperature varied between 12.5 ºC and 30.5 ºC during the study period. The presence of an IUC was observed, mainly in the city center and at night, with an average intensity between 1.1 ºC and 3.0 ºC, as well as the presence of a cold urban island (IUF) during the day with values between -0.3 and 0 ºC. The RLM model identified the solar irradiance, relative humidity and pressure for the day as predictive variables, as well as wind speed and direction, relative humidity and pressure for the night. The RLM presented an adjusted R² value of 0.76 and 0.60 and RMSE equal to 2.39 ºC and 2.38 ºC for day and night, respectively.