PhD thesis: Impact of vegetation on urban microclimate
- Supervisors: Prof. Dr. Thijs Defraeye, Dr. Jonas Allegrini, Dr. Dominique Derome, Prof. Dr. Jan Carmeliet
- Examiners: Prof. Dr. Jan Carmeliet, Prof. Dr. Bert Blocken, Prof. Dr. Peter Edwards
- ETH Research Collection: Direct link, DOI
Vegetation in cities is seen as an effective strategy to combat the growing UHI as it provided natural cooling. Vegetation offers natural cooling primarily by providing shading to urban structures and additionally by extracting heat from the surroundings during the photosynthesis driven transpiration process. However, the effectiveness of transpirative cooling is directly related to the water availability of the plants, and in extreme environmental conditions such as drought, the effectiveness of vegetation can be severely compromised. Furthermore, trees can obstruct ventilation which can have a negative impact on the pollution dispersion characteristics in cities. Therefore, the net impact of vegetation on pedestrian comfort and health in cities is dependent on various conditions and the most effective configuration for UHI mitigation is a non-trivial problem. Thus, an urban microclimate model that can model the airflow, radiation and the water cycle in an integrated approach is necessary for accurately assessing the impact of vegetation in a complex urban environment.
The thesis aims at establishing a more accurate and detailed pre diction of the thermal influence of vegetation in an urban environment by simultaneously taking in account of its heat, mass and momentum exchanges and the influence of the water availability. The goal is to provide better guidelines for effective mitigation strategies with vegetation for any given urban, vegetation configuration and environmental conditions. The cooling potential of vegetation such as trees on the microclimate of a street-canyon is studied using a computational fluid dynamics (CFD) model in OpenFOAM. The flow field is numerically modeled by solving the Reynolds-averaged Navier-Stokes equations (RANS) with realizable k − ε turbulence closure model. The vegetation model is integrated into the CFD solver as a porous medium, providing the necessary source terms for heat, mass and momentum fluxes, with additional closures for turbulence enhancement due to vegetation. A radiation model is developed to model the short-wave and long-wave radiative heat fluxes between the leaf surface and the surrounding. The radiation model enables to model the impact of the diurnal variation of solar intensity and direction, and the long-wave radiative fluxes between vegetation and nearby urban surfaces. Also, the water cycle driven by the transpiration process is explicitly modeled by coupling with an integrated soil heat and moisture dynamics model. A soil-plant-atmosphere continuum modeling approach is essential as the transpiration rate through the stomata is directly linked to the water availability at the roots of the plant. Therefore, the proposed method helps us understand the response of vegetation during extreme environmental conditions such as drought and provides a more accurate prediction towards the cooling performance of vegetation. The present study investigates the influence of transpirative and shaded cooling due to vegetation on pedestrian comfort inside a street canyon. The influence of various vegetation features such as size, shape, and density is studied to determine the optimal configuration for improving pedestrian comfort and health.
The thesis also employs wind tunnel experiments to provide a deeper understanding of the influence of an isolated tree on the flow. A comparative study of drag force and wake flow field of small model and natural trees shows the distinction between their responses and provide an insight into the challenges of representing trees in urban flow wind tunnel studies with model trees. Furthermore, the microclimate measurement of the small natural plant provide an understanding of the dynamic response of a plant and more a basis for the validation of the numerical model.