Simulating the effect of Nature-based Solutions as a mitigation tool for Urban Heat Islands
Keywords: urban heat island, susceptibility mapping, machine learning, nature-based solutions
Abstract. The Urban Heat Island (UHI) effect is a phenomenon that typically occurs in areas with dense infrastructure and limited vegetation. Nature-based solutions (NbS) have been proposed to mitigate the effects of climate change and have been proved to reduce the frequency of its hazards. Therefore, this research examines how UHIs are influenced by the simulated implementation of NbS. The area of interest (AOI) of this study is the city of Milan in Northern Italy and the purpose of this research is two-fold. First, to train machine learning (ML) models to predict Surface UHI (SUHI) susceptibility and intensity, and their corresponding SUHI maps, based on land cover, Land Surface Temperature (LST), Normalised Difference Vegetation Index (NDVI), Normalised Difference Built-up Index (NDBI) and other ancillary data. Afterwards, the produced SUHI intensity ML model was re-evaluated to analyse the expected behaviour of simulated NbS, specifically for green roofs and parks. The NDVI, NDBI, and albedo of the simulated vegetation areas were changed to the average values of the urban vegetation in the City of Milan, 0.57, −0.17, and 0.15 respectively. The SUHI statistics of the specific areas of change were analysed pre and post simulation. The results showed that green roofs have the potential to lower SUHI intensity by 5 degrees Kelvin. The proposed methodology can be extended to simulate multiple scenarios based on specific needs.