Predicting Long-term Urban Overheating and Their Mitigations from Nature Based Solutions Using Machine Learning and Field Measurements
Authors:
Jiwei Zou,
Lin Wang,
Senwen Yang,
Michael Lacasse,
Liangzhu,
Wang
Abstract:
Urban overheating, exacerbated by climate change, threatens public health and urban sustainability. Traditional approaches, such as numerical simulations and field measurements, face challenges due to uncertainties in input data. This study integrates field measurements with machine learning models to predict the duration and severity of future urban overheating events, focusing on the role of urb…
▽ More
Urban overheating, exacerbated by climate change, threatens public health and urban sustainability. Traditional approaches, such as numerical simulations and field measurements, face challenges due to uncertainties in input data. This study integrates field measurements with machine learning models to predict the duration and severity of future urban overheating events, focusing on the role of urban greening under different global warming (GW) scenarios. Field measurements were conducted in summer 2024 at an office campus in Ottawa, a cold-climate city. Microclimate data were collected from four locations with varying levels of greenery: a large lawn without trees (Lawn), a parking lot without greenery (Parking), an area with sparsely distributed trees (Tree), and a fully covered forested area (Forest). Machine learning models, including Artificial Neural Networks (ANN), Recurrent Neural Networks (RNN), and Long Short-Term Memory (LSTM) networks, were trained on local microclimate data, with LSTM achieving the best predictions. Four GW scenarios were analyzed, corresponding to different Shared Socioeconomic Pathways (SSP) for 2050 and 2090. Results show that the Universal Thermal Climate Index (UTCI) at the "Parking" location rises from about 27,\textdegree C under GW1.0 to 31,\textdegree C under GW3.5. Moreover, low health risk conditions (UTCI > 26,\textdegree C) increase across all locations due to climate change, regardless of greenery levels. However, tree-covered areas such as "Tree" and "Forest" effectively prevent extreme heat conditions (UTCI > 38.9,\textdegree C). These findings highlight the crucial role of urban greening in mitigating severe thermal stress and enhancing thermal comfort under future climate scenarios.
△ Less
Submitted 25 February, 2025;
originally announced February 2025.
