An ANN-based framework for mapping urban tree cooling and exposure inequality in a high-density city: A case study of Hong Kong

Abstract

As heatwaves intensify with global warming, urban trees play a critical role in cooling through shading and evapotranspiration. However, these cooling benefits are constrained by methodological limitations that trade street-level precision for citywide scalability, hindering high-resolution mapping of tree cooling effects across metropolitan areas. Additionally, exposure to these cooling benefits remains uneven across socioeconomic groups. To enable spatially precise mapping of urban tree impacts while enhancing social equity among diverse population groups, this study develops a multi-scale framework integrating microclimate field measurements, LiDAR-based tree mapping, satellite thermal imagery, and urban big data in Hong Kong, a high-density city experiencing frequent heat extremes. Implemented via a multilayer perceptron artificial neural network, the framework achieves high predictive accuracy (R² = 0.905) for estimating tree cooling effects. The cooling exposure index is computed using a population-weighted exposure model that combines citywide maps of tree cooling effects with grid-level census data. Results reveal that at the city scale, tree cooling effects can reach up to 4.49$℃$ (average: 0.68$℃$), with the Islands district showing the highest average cooling (1.27$℃$), followed by Central & Western (1.25$℃$) and Wan Chai (1.05$℃$) on Hong Kong Island. The tree cooling effect varies by species: Pterocarpus indicus demonstrates the strongest effect, reducing temperatures by 2.19$℃$ under summer conditions. However, population-weighted spatial analysis shows that marginalized communities, particularly low-income residents and solitary dwellers, experience the least exposure to these benefits, with cooling exposure indices as low as 18.0 and 18.7, respectively, compared to a citywide average of 257.7. These findings highlight the need for equity-centered greening strategies and justice-oriented urban planning that prioritizes species selection and neighborhood investment to reduce heat disparities.