Abbasabadi, NarjesWorthy, Amanda2026-04-202026-04-202026-04-202026Worthy_washington_0250E_29213.pdfhttps://hdl.handle.net/1773/55465Thesis (Ph.D.)--University of Washington, 2026Building energy models typically reference aggregated, non-urban-specific weather and climate datasets that overlook urban microclimate conditions. This embedded shortfall leads to substantial differences between the modeled and actual performance of buildings, and thus to inaccurate energy demand-side calculations. Our research aims to close this gap by integrating microclimate information, derived from earth observational and remote sensing datasets, into urban building energy models. We evaluate the value of these products by conducting two complementary studies that examine the integration of satellite-derived microclimate information in both data-driven and physics-based building energy modeling workflows. First, we develop a bottom-up, data-driven urban building energy modeling framework that combines and integrates earth observational microclimate data, spatially interpolated Typical Meteorological Year data, and annual energy usage data, measured by the Seattle Energy Benchmarking Dataset, to capture the impacts of microclimates on urban building energy performance. Using machine learning techniques and Seattle, Washington, USA as a proof of concept, we compare predictive model performance across multiple climate input scenarios; ultimately validating the effectiveness of using earth-observational data inputs to address simulation-to-real modeled uncertainties in microclimate integration. Second, we conduct a hybrid study that augments physics-based residential building energy consumption insights with earth observational microclimate data using machine learning predictions. This approach produces a spatial heat map of residential energy consumption for a typical family home across Los Angeles County, USA, that is explicitly tailored to reflect urban microclimate variation. Here, we confirm current building EnergyPlus weather file (EPW) sampling sites to be in lower vulnerability areas with fewer streets and buildings than the city average. This result identifies a mismatch between the environmental conditions observed in dense urban areas and those normally simulated in building energy modeling protocols, further underscoring the structural misrepresentations that are embedded in current frameworks. Throughout this work, we emphasize the value in integrating satellite-derived microclimate products into data-driven building energy studies. However, we cite obstacles in integrating satellite-derived microclimate data directly into physics-based urban building energy models, suggesting future research to explore more spatially and temporally compatible datasets that measure EnergyPlus weather file parameters and respective downscaling opportunities. Despite these challenges, the Landsat thermal band 10, or land surface temperature products, show strong potential in being effective, scalable proxies for incorporating microclimate effects into both data-driven and physics-based urban building energy studies. This research advocates for the integration and validation of urban microclimate effects into building energy modeling frameworks, to enable more accurate, just, and precise energy policy and planning.application/pdfen-USCC BY-NDBuilding Energy DemandMicroclimatesSatelite DataSimulation-BasedUrban Building Energy ModelingCivil engineeringEnergyArchitectureCivil engineeringThesis