Local UHI mitigation and utilization: Urban building energy modeling, simulation, and urban design responses based on localized weather data

Urban population growth and the expansion of built-up areas are placing increasing pressure on energy systems in large cities. Anthropogenic heat from dense buildings further intensifies local climate differences and impacts building energy use. However, existing urban building energy models (UBEMs) lack the ability to differentiate micro-scale environments or capture energy variations across local climate zones (LCZs). To address this, we developed a UBEM tool driven by localized weather data (LWD), which flexibly defines weather grid resolution and simulates building energy use in specific urban contexts. Focusing on a 3.34 km² campus in Beijing with 880 buildings and 170 LCZs at 200 m resolution, the study analyzes the impact of summer and winter urban heat island (UHI) effects on local weather and energy demand. Machine learning models explore how urban morphology influences local weather and how building form affects energy use. Results show that UHI significantly increases cooling degree days (CDD) in dense urban areas and exacerbates climate disparities between different local environments. The CDD in the hotspot is four times higher than in suburban areas, while heating degree days (HDD) are reduced by more than half. Street aspect ratio and floor area ratio are key at the cluster scale; building shape coefficient (BSC) and height dominate at the building scale. Heating demand is especially sensitive to BSC. Ignoring UHI, the bias in total energy use intensity (EUI) results for poorly insulated buildings can be 3 to 10 times higher than that for other typical buildings. Except for large offices and hospitals, in Beijing, UHI tends to reduce total annual energy use for most building types. Using the integrated urban canopy model (UCM) and UBEM simulation tool and the XGBoost-SHAP computational design method enables more accurate prediction and response to climate–building interactions, offering a quantitative basis for integrated urban design strategies. This supports improved regulation of building spaces and envelopes in response to seasonal UHI variations.