Monika Kuffer, Mila Koeva, Richard Sliuzas
Advance Image Analysis
Presently, the majority of people live in urban areas [1](United Nations, 2014). Cities suffer from various negative environmental impacts, such as negative impacts on the urban microclimate. Densely built-up urban areas are affected by the phenomenon called urban heat islands (UHIs) [2]. High temperatures are not only uncomfortable for humans but also can have severe health impacts on vulnerable groups. The accumulation of heat in an urban area depends very much on the land use composition and morphological set-up of the area, e.g., built-up areas with good vegetation mix are less prone to UHIs than high-density built-up areas. As a conceptualization of different urban typologies (classification), local climate zones (LCZs) were developed by Stewart and Oke [3]. LCZs allow the classification into areas of homogenous thermal regimes. LCZs can be extracted from remotely sensed imagery. The methodology developed by the WUDAPT group (http://www.wudapt.org/) uses moderate resolution imagery (i.e., Landsat images) and a machine learning based approach to map LCZ. Surface urban heat islands (SUHI) can be modelled from remotely sensed imagery (e.g., thermal bands of Modis). The main question will be, taking examples from cities in the Global South and North: What is the relationship between LCZs and SUHIs, making use of very high resolution (VHR) imagery and 3D models or urban areas to map LCZs. This will allow to get insights into the complex morphology of densely built-up cities. Answering the question: Can we map the details of LCZs with VHR imagery and 3D models, and are the zones relevant to analyze differences in terms of thermal patterns?
References:
1. UN Department of Economic and Social Affairs Population Division. World urbanization prospects. The 2018 revision; New York, 2019.
2. Wang, J.; Kuffer, M.; Sliuzas, R.; Kohli, D. The exposure of slums to high temperature: Morphology-based local scale thermal patterns. Science of The Total Environment 2019, 650, 1805-1817.
3. Stewart, I.D. Local climate zones: Origins, development, and application to urban heat island studies. In Annual Meeting of the American Association of Geographers, Seattle, USA, 2011.
The urban morphology can be modeled by local climate zones (LCZs) can be extracted using machine learning algorithms (e.g., random forest). The methodology has been developed for moderate-resolution imagery (see http://www.wudapt.org/), called Level 0. However, the methodology can be also transferred to VHR satellite images and air photos, using 3D modelling techniques aiming at Level 1 of the WUDAPT approach. The SUHI can be extracted from thermal bands of TIR images (e.g., Aster, MODIS). This will show variations of temperatures across the city and indicate locations of local urban heat islands (UHIs). After mapping LCZs, their relation with UHIs can be analyzed, assessing also whether the classification of LCZs is fit for describing the complex urban landscape.
1. Wang, J.; Kuffer, M.; Sliuzas, R.; Kohli, D. The exposure of slums to high temperature: Morphology-based local scale thermal patterns. Science of The Total Environment 2019, 650, 1805-1817.
2. Stewart, I.D. Local climate zones: Origins, development, and application to urban heat island studies. In Annual Meeting of the American Association of Geographers, Seattle, USA, 2011.