Understanding the spatial pattern of building energy consumption intensity is essential to urban energy planning and management. In this study, we attempt to explore the spatial complexity of residential energy usage intensity, with a focus on urban form and 3D geomorphometry attributes of urban ventilation, solar insolation, and vegetation. The empirical evidence from census tracts in Chicago suggest a clear role of different urban form and geomorphometry features in affecting residential electricity usage intensity: urban porosity and roughness length have consistent spillover effects on electricity usage intensity. While we find little evidence that normalized difference vegetation index (NDVI) has similar spillover effect, it could significantly reduce electricity usage intensity in its reference tract. Additionally, insolation per building volume could significantly increase electricity usage intensity, especially in summer months. Nevertheless, these effects generated by ventilation, NDVI, and insolation do not appear to exist for gas usage intensity in winter. Accounting for both spatial dependence among residuals and spatial spillover of ventilation and vegetation variables, our final spatial regression model explains up to 65% and 59% of variations in summer electricity and winter gas usage intensity at the census tract level. Our results also highlight the importance of spatially explicit policy for improving residential energy efficiency and mitigating climate change.