Publicação
Towards an improved representation of the urban climate: An Application of Artificial Intelligence
| Resumo: | Cities are considered local “hotspots” of climate change. Urban areas concentrate a large fraction of global population, wealth, and emissions, exposing their inhabitants to climate change impacts. Therefore, the improvement of urban present climate description and future projections are paramount for designing adaptation and mitigation strategies. The Global Climate Models (GCMs) are state-of-theart tools for projecting future climate. However, most of the simulations have coarse resolutions and do not have physical urban parametrisations to adequately represent the physical properties and processes at the urban scale. The advantage of applying a machine learning approach (XGBoost) is explored for better describing Madrid’s urban climate. Namely, the ability to reproduce present and future climates: 2-m air temperature; surface temperature; urban heat island and surface urban heat island effects. The XGBoost is evaluated at monthly and daily scales for local ground temperatures and, also at hourly scale, to represent the spatial structure of land surface temperature w.r.t. remote sensing data. Firstly, for present climate, XGBoost is trained with a set of ERA5 predictors (0.25º), ground stations, and Land Surface Temperature (LST) observations. Secondly, a number of sensitivity cases are performed to assess the results dependency to predictors and their resolution. Thirdly, the learned relationships between the set of predictors and predictands, is applied to ESGCM predictors, providing historical and future climate projections for the 21st century under four emission scenarios. Overall, XGBoost results reveal a good performance and significant added value against ERA5 and the ESGCM. XGBoost greatly improves the reproduction of the present climate Tmax, Tmin, LST, and more importantly, the UHI (-0.5ºC and +3ºC for Tmax and Tmin), and the SUHI (+1ºC and +2ºC for Tmax and Tmin). For future climate, the XGBoost significantly corrects the ESGCM UHI misrepresentation but seems to underestimate the expected Madrid’s local warming (3.5ºC anomaly). |
|---|---|
| Autores principais: | Bushenkova, Angelina Vladimirovna |
| Assunto: | Clima urbano Ilha de Calor Urbana Alteração Climáticas Inteligência Artificial Extreme Gradient Boosting Teses de mestrado - 2023 |
| Ano: | 2023 |
| País: | Portugal |
| Tipo de documento: | dissertação de mestrado |
| Tipo de acesso: | acesso aberto |
| Instituição associada: | Universidade de Lisboa |
| Idioma: | inglês |
| Origem: | Repositório da Universidade de Lisboa |
| Resumo: | Cities are considered local “hotspots” of climate change. Urban areas concentrate a large fraction of global population, wealth, and emissions, exposing their inhabitants to climate change impacts. Therefore, the improvement of urban present climate description and future projections are paramount for designing adaptation and mitigation strategies. The Global Climate Models (GCMs) are state-of-theart tools for projecting future climate. However, most of the simulations have coarse resolutions and do not have physical urban parametrisations to adequately represent the physical properties and processes at the urban scale. The advantage of applying a machine learning approach (XGBoost) is explored for better describing Madrid’s urban climate. Namely, the ability to reproduce present and future climates: 2-m air temperature; surface temperature; urban heat island and surface urban heat island effects. The XGBoost is evaluated at monthly and daily scales for local ground temperatures and, also at hourly scale, to represent the spatial structure of land surface temperature w.r.t. remote sensing data. Firstly, for present climate, XGBoost is trained with a set of ERA5 predictors (0.25º), ground stations, and Land Surface Temperature (LST) observations. Secondly, a number of sensitivity cases are performed to assess the results dependency to predictors and their resolution. Thirdly, the learned relationships between the set of predictors and predictands, is applied to ESGCM predictors, providing historical and future climate projections for the 21st century under four emission scenarios. Overall, XGBoost results reveal a good performance and significant added value against ERA5 and the ESGCM. XGBoost greatly improves the reproduction of the present climate Tmax, Tmin, LST, and more importantly, the UHI (-0.5ºC and +3ºC for Tmax and Tmin), and the SUHI (+1ºC and +2ºC for Tmax and Tmin). For future climate, the XGBoost significantly corrects the ESGCM UHI misrepresentation but seems to underestimate the expected Madrid’s local warming (3.5ºC anomaly). |
|---|