Semantic 3D city models are a central component of Urban Digital Twins (UDTs), enabling KPI-driven, evidence-based urban planning and decision making. While significant progress has been made in reconstructing models of the existing built environment and various modeling support tools exist, the creation of hypothetical planning scenarios remains time-consuming and the modeling process, as such, insufficiently formalized. In practice, planning scenarios are manually modeled to evaluate potential future states, requiring numerous interdependent geometric and semantic modifications to the base model. Despite the substantial manual effort involved, modeling processes are rarely analyzed systematically. This work addresses this gap by formalizing scenario modeling in semantic 3D city models as a recursive aggregation of atomic edit operations and their dependencies, represented through directed dependency graphs. To distinguish between the effort required to model virtual planning scenarios and the effort required for their real-world implementation, we introduce the concepts of the ’<em>Level of Modeling Effort (LoME)</em>’ and ’<em>Level of Realization Effort (LoRE)</em>’. Using the City of Munich’s scenario modeling activities as a case study, we analyze the substantial manual effort required and highlight the need to formalize the modeling process. The proposed framework provides a foundation for comparing modeling approaches, quantifying efficiency gains, and supporting the development of automated or AI-assisted scenario-generation methods for semantic 3D city models.