Heuristics for the design of low-carbon districts in the tropics

Abstract

Low-carbon urban design strategies suggest that energy performance of a district can be influenced by altering the urban form. Such strategies are the most effective at the early stages of low-carbon development projects. However, there exists a lack of early-stage design heuristics for urban practitioners, addressing parameters that largely influence energy performance of a district. The literature quantifying the interdependence between urban design parameters and energy performance at the district level is limited. The current study aims to fill in this knowledge gap by exploring relationships and quantifying trade-offs between district energy performance and such design parameters as typology, land use, density, district compactness and site coverage. About 10'000 energy simulations are conducted to assess the performance of various district designs in the tropical context of Singapore, considering high-density mixed-use urban setting. The simulation results are analyzed using statistical techniques (ANOVA, regression) to derive design heuristics with respect to studied design parameters, in order to inform decision-makers in the early urban design stages. The regression model shows that the variability of carbon and energy performance is mostly explained by land use mix: changing land use from mostly retail to mostly residential would decrease the operational CO2-intensity by a factor of two. The impact of density is much smaller but statistically significant: a decrease in the floor area ratio from ten to four reduces the operational CO2-intensity by 8%. The influence of district compactness as well as typology is negligible (below 2% impact on the operational CO2- intensity). Increasing the accuracy of the energy simulation and adopting more realistic district designs can be considered as immediate further extensions to this research. The framework developed in this study can serve as a basis for future energy-driven urban design studies.