The process of expressing relationships among variables as a function of parameters and studying these relationships as functions of changes in the parameters.
The representation, in a dynamic model, of physical effects in terms of admittedly oversimplified parameters, rather than realistically requiring such effects to be consequences of the dynamics of the system. See Mellor-Yamada parameterization, subgrid-scale process, convective adjustment.
In numerical modeling, the method of incorporating a process by representation as a simplified function of some other fully resolved variables without explicitly considering the details of the process. The classic example is the representation of sub-grid scale turbulence as the product of a function of the velocities at the local grid points and an empirically derived eddy viscosity coefficient (in analogy to the molecular viscosity coefficient). This analogy has been known to fail. See, for example, the classic (and wonderfully titled) monograph of Starr (1968).