Historically, computer generated forces (CGF) have been developed using the system-level approach. That is, they have been developed as a unit with the terrain they operate on and use for behavioral reasoning tightly coupled to the rest of the system. Over the last several years, the evident trend has been towards making various CGF systems interoperable. To do this, system developers and integrators have sought to use the same Terrain Databases (TDBs) and/or live with the inconsistencies that arise when linking different systems together.

This paper will present a set of grid, polygon and task-based metrics that addresses the perceived behavioral differences of CGF when operating in TDBs of differing fidelities (by fidelity, we mean how closely does the terrain match the real world) and formats. The metrics focus on atomic behavior tasks and those that are shared across other higher level behaviors, such as planning and following a route. From the quantifiable differences in the low-level tasks, we can quantify the effect of the underlying terrain representation on the higher level behaviors. The results provide new insight into terrain interoperability issues and their effects on CGF behaviors.