A carregar...
Publicação
A multi-sensor approach to the study of the urban heat island effect in cities : a comparative study
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 23.96 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
Com o aumento da urbanização, as cidades tiveram de se adaptar para acomodar o aumento da população resultante da industrialização. As superfícies naturais, tais como a relva, as árvores e vários outros tipos de vegetação, foram substituídas por materiais artificiais, como o cimento, o asfalto e o metal. Estes dois tipos de materiais apresentam características físicas distintas, como o albedo, a capacidade e a condutividade térmica que resultam em diferentes taxas de aquecimento/arrefecimento. Adicionalmente, a estrutura vertical das cidades provoca alterações na circulação do ar e no sombreamento, que também influenciam as taxas de aquecimento/arrefecimento no seu interior. As ilhas de calor urbanas traduzem a diferença de temperatura entre as zonas urbanas e as zonas rurais, e são há muito estudadas pelo seu impacto na vida das populações. No século XIX, as diferenças de temperatura entre as duas zonas começaram a ser notadas através de medições da temperatura do ar em estações meteorológicas. A dependência de estações meteorológicas constituía uma barreira ao estudo deste fenómeno, devido à escassez de dados. O advento de satélites em órbita terrestre possibilitou o estudo global de várias características da superfície, nomeadamente a temperatura da superfície. A disponibilidade global desta variável aumentou consideravelmente o interesse e o estudo do ambiente térmico urbano através das chamadas ilhas de calor urbano de superfície. Apesar da variedade de sensores de deteção remota, os dados podem ter uma boa resolução temporal ou uma boa resolução espacial, dependendo da respetiva altitude orbital. Está situação é problemática no estudo de fenómenos com variabilidade diurna e em regiões altamente heterogéneas, como é o caso das cidades, e consiste numa das desvantagens dos dados provenientes de sensores de deteção remota. O objetivo desta tese é abordar algumas das questões em aberto relativas ao estudo das ilhas de calor urbanas, nomeadamente as características espácio-temporais necessárias para a sua identificação, as implicações que diferentes referências rurais e a utilização de diferentes produtos de temperatura da superfície têm na sua intensidade e a possibilidade de redução da escala espacial de modo a melhor distinguir as heterogeneidades da superfície, mantendo uma alta frequência de observações ao longo do dia.
The increased availability of remote sensing observations has led to a deeper understanding of the urban thermal environment. Substantial progress has been made in understanding the phenomenon of urban regions having higher temperatures than their surroundings. Usually referred to as the surface urban heat island (SUHI), it represents the difference in land surface temperature (LST) between a selected urban area and its neighboring rural one. Although many studies have characterized SUHI in various climates and based on several data sources, remotely sensed observations have the peculiarity of having either a high temporal resolution or high spatial resolution. This means that highly heterogeneous areas, such as urban ones, cannot be fully characterized based on a single sensor. This limitation can be mitigated by considering multiple data sources/sensors, however we found that considerable differences can arise from changes in LST retrieval algorithms, sensor calibration and viewing angle/time, affecting the resulting SUHI intensity. To overcome these limitations, data from different sensors, with different spatiotemporal characteristics, were used to compute LSTs based on the same algorithm and input information. This allows for comparisons between different source LST and were consequently used within a multi-layer perceptron model to obtain an improved spatiotemporal dataset based on multisensor data. The new LST dataset was used to compute the SUHI and has been shown to be useful in characterizing SUHI’s diurnal variability and its dependence on both the rural and urban land cover types. In addition, it allowed for the analysis of SUHI interactions with abnormally cold and hot days. The proposed methodology is used to characterize SUHI for the city of Madrid, acting as a basis for more comprehensive studies where a unified methodology could facilitate the comparison and characterization of SUHI for different cities.
The increased availability of remote sensing observations has led to a deeper understanding of the urban thermal environment. Substantial progress has been made in understanding the phenomenon of urban regions having higher temperatures than their surroundings. Usually referred to as the surface urban heat island (SUHI), it represents the difference in land surface temperature (LST) between a selected urban area and its neighboring rural one. Although many studies have characterized SUHI in various climates and based on several data sources, remotely sensed observations have the peculiarity of having either a high temporal resolution or high spatial resolution. This means that highly heterogeneous areas, such as urban ones, cannot be fully characterized based on a single sensor. This limitation can be mitigated by considering multiple data sources/sensors, however we found that considerable differences can arise from changes in LST retrieval algorithms, sensor calibration and viewing angle/time, affecting the resulting SUHI intensity. To overcome these limitations, data from different sensors, with different spatiotemporal characteristics, were used to compute LSTs based on the same algorithm and input information. This allows for comparisons between different source LST and were consequently used within a multi-layer perceptron model to obtain an improved spatiotemporal dataset based on multisensor data. The new LST dataset was used to compute the SUHI and has been shown to be useful in characterizing SUHI’s diurnal variability and its dependence on both the rural and urban land cover types. In addition, it allowed for the analysis of SUHI interactions with abnormally cold and hot days. The proposed methodology is used to characterize SUHI for the city of Madrid, acting as a basis for more comprehensive studies where a unified methodology could facilitate the comparison and characterization of SUHI for different cities.
Descrição
Tese de doutoramento em Ciências Geofísicas e da Geoinformação (Deteção Remota), Universidade de Lisboa, Faculdade de Ciências, 2025.
Palavras-chave
surface urban heat island land surface temperature downscaling multi-layer perceptron ilha de calor urbana de superfície temperatura da superfície do solo redução de escala espacial perceptrão multicamada