Climate-sensitive energy optimization in green buildings: Evidence from LEED-NC v4 projects

Climatic conditions shape building energy loads, yet their influence on green-building energy benchmarking is often treated implicitly or through broad climate zoning. This study examines how long-term climatic factors are associated with the achievement of the Optimize Energy Performance credit (EA7) in LEED v4 Building Design and Construction: New Construction projects (LEED-NC v4), using the United States as a climatically diverse empirical setting. A dataset of 779 certified buildings was analyzed using spatially matched climatological variables, including heating and cooling degree days, solar resource availability, precipitation, wind speed, temperature, and humidity, derived from high-resolution datasets. Multiple linear regression, Random Forest regression, and hierarchical clustering were employed. Across methods, precipitation and wind exposure were dominant negative correlates of EA7 achievement, whereas global tilted irradiance was a strong positive contributor. Cooling demand showed a modest negative association, while heating demand and mean temperature played limited roles when climatic variables were considered jointly. Random Forest outperformed linear regression, increasing explained variance from 18.4% to 44.6%, indicating non-linear interactions and threshold effects. Clustering identified two climate-performance regimes, separating solar-rich, low-load environments from wetter, wind-exposed climates with lower EA7 outcomes. The novelty of the study lies in providing a credit-specific, climate-explicit framework that links spatially matched long-term climatic indicators to EA7 achievement and triangulates evidence through interpretable regression, non-linear machine learning, and unsupervised clustering. Although calibrated on U.S. LEED projects, the framework is transferable to international green-building benchmarking where comparable certification and climate data are available.