Urban energy–microclimate digital twin: Census-tract validation using mobile traverses and utility billing benchmarking

Urban Building Energy Modeling tools increasingly couple building loads with urban microclimate, yet empirical validation in hot-arid cities remains limited. We evaluate the CityDigitalTwin (CDT) platform in a neighborhood near Arizona State University's Polytechnic Campus (Mesa, AZ) during a hot July 2025 period. CDT couples a GPU-accelerated fast-fluid-dynamics microclimate solver with an archetype-based urban building energy model, driven by boundary conditions from the nearest upwind airport weather station. Mobile vehicle traverses provide 2-m air-temperature observations for model evaluation on 19 July; neighborhood-mean absolute differences across four districts range from 0.09 to 1.30 °C (mean 0.63 °C). Electricity use for 15 single-story residential buildings is simulated using DOE/ESS-DIVE archetypes and a Min-Max Archetype Sweep over eight uncertain parameters; 87% of measured daily loads fall within the model interquartile envelope and all values lie within the simulated min-max bounds. Results indicate that CDT can reproduce both outdoor thermal patterns and cooling-dominated electricity use with accuracy suitable for neighborhood-scale planning and heat-mitigation studies.