Establishing a link between complex courtyard spaces and thermal comfort: A major advancement in evidence-based design

Amidst climate change, the importance of climate-adaptive design in architecture and landscape design has surged, particularly in residential courtyards, where optimizing the microclimate is paramount to residents' well-being. Traditional spatial indices, however, fall short in accurately characterizing complex courtyards and local spatial features. To overcome these limitations, this study introduces pixel-level spatial indicators that effectively overcome these constraints. These indicators are implemented using computational geometry algorithms such as Ray Tracing, Flood Fill, and A*, enabling simulation of various courtyard spatial indicator maps. We also utilize Graphics Processing Unit (GPU)-based rapid thermal comfort simulation technology to generate thermal comfort maps. By applying data mining methods such as Partial Least Squares Regression (PLSR), Pearson correlation, and Nearest-neighbor interpolation, we explore the relationships between spatial indicators and thermal comfort, ultimately identifying key indicators and determining the guiding thresholds and influencing trends corresponding to heat discomfort frequency. Six key indicators and th emerge: Building View Factor (BVF), indicating building coverage visibility (prefer above 0.11); Solar Beam Fraction (BEAM), illustrating Summer solstice sun shading condition (prefer below 0.78); Averaged View Factor (AVF), showing overall visibility (prefer below 0.40); Directional Sky View Factor (DSVF(W)), reflecting sky visibility in a specific orientation (prefer below 0.73); Tree View Factor (TVF), denoting tree coverage visibility (prefer above 0.18); and Plan Water Ratio (PWR), signifying water surface proportion (aim for below 0.44). These insights, integrated into design tools, contribute to evidence-based microclimate regulation strategies, thereby enhancing urban residents' thermal comfort and overall well-being.