The equiaxed investment casting process is a multi-physics problem which requires knowledge from engineers who have expertise in materials, metallurgy, fluid dynamics, thermodynamics, and heat transfer. Process modeling is a tool used by foundries to help predict casting defects such as shrinkage porosity, hot tears, and poor grain structure. The reliability of these predictions is strongly dependent on the accuracy of the thermal boundary conditions set in the model. In this work, a SGT5-2000E Vane 4 cast in Rene 80 nickel-based superalloy was modeled, using the FEA simulation package ProCAST, with two different methodologies. One methodology had very little effort invested into defining the thermal domain. The other methodology involved a thorough consideration of all heat transfer mechanisms acting on the mold. An extensive literature search was performed to define a unique natural convection heat transfer coefficient for each set of surfaces on the mold. The transient boundary layer development was also captured in the definition of the heat conditions. The shrinkage porosity predictions of the models were compared to realworld x-ray data and the transient nonuniform methodology predictions were much more representative than the low fidelity heat transfer methodology predictions. The low fidelity heat transfer model did predict some shrinkage, but not where it appeared in reality. The process modeler will be misdirected by the model results when deriving a solution to the casting process if the realworld physics are not appropriately accounted for in the model. This will be very counterproductive when the foundry is using the model to reduce developmental trials by running trials in model space. References and derived parameters are provided for material properties, emissivity of shell and insulation wraps, and external mold spatially varying heat transfer coefficients.