An emerging goal of modern building design is the construction of buildings that provide comfort to occupants while consuming a minimal amount of energy. Reducing energy consumption and achieving comfort are competing factors in building operation, but with the increasing sophistication of building energy models (BEM), capable of simulating the thermodynamic behavior of a building and predicting energy consumption, these goals can be realized. Although models are a good predictor of building performance, numerous approximations must be made during their creation which may negatively affect accuracy. One such approximation is that of zoning, which is how the volume of a building is divided into regions where properties are assumed to be uniform. Zoning is practiced because it minimizes the time needed in creating a BEM and also requires less computation time to fully simulate. This process is often performed cut-and-try by the practitioner of the building model. A systematic approach for creating simplified models is presented using the Koopman Operator, a linear, infinite dimensional operator that captures nonlinear, finite dimensional dynamics without linearization. Using properties of the operator, modes of temperature oscillation can be identified using the predictions produced by a building model. These modes identify temperature behavior occurring at different time scales. From these modes, systematic approximations to the zoning can be created. An implementation of this technique is illustrated in a model of the Ocean Science Education Building at the University of California, Santa Barbara currently under construction, designed with both mechanical and natural conditioning.