Climate-responsive urban planning through generative models: Sensitivity analysis of urban planning and design parameters for urban heat island in Singapore's residential settlements.

• Planning parameters are as effective in mitigating urban heat island as urban design parameters. • Building density, podium density, and land parcel area show the highest impacts on UHI. • Planning and design parameters show complex interaction in their impact on UHI simulation results. • Prescriptive planning may prevent discovery of well-performing scenarios. • Performative planning can better support identifying dense and low UHI urban scenarios. The Urban Heat Island (UHI) effect exacerbates the sustainability and well-being challenges of extreme heat events. While city planning and design measures have been shown to mitigate UHI severity, the complex interaction among these measures has limited the ability of previous research to assess their impact holistically and across urban scales. To investigate the cross-scalar effectiveness of multiple UHI mitigation measures, this study applies sensitivity analysis (SA) to nine parameters in an urban generative model. Previously unstudied planning parameters, land parcel area and road network density, are included in the analysis. From the SA of 21,000 model solutions for a 100 ha case study site in Singapore, building density, podium density, and land parcel area are found to have greatest impacts on UHI. This finding supports a hypothesis that urban planning parameters have a high potential for UHI mitigation. Key findings include that a high green plot ratio (>50 %) combined with a low site coverage ratio (<50 %) permits even high-density model solutions (gross plot ratio >4) to maintain annual UHI below 0.89 °C. The conclusion discusses the implications of the findings for heat-resilient city planning and demonstrates that performance-based evaluation of generative urban models can improve upon prescriptive planning approaches. [ABSTRACT FROM AUTHOR]