Techniques for sensitivity analysis of simulation models: A case study of the CO2 greenhouse effect

Sensitivity analysis is needed for validation, what-if analysis, and optimization of complicated simulation models. One set of techniques for sensitivity analysis consists of least squares curve fitting, regression analysis, and statistical designs such as factorial designs. In this case study these techniques are applied to several modules of a large integrated assessment model for the greenhouse effect, developed in The Netherlands. The regression models turn out to be valid approximations to the simulation models. Some estimated effects are quite surprising for the simulation users.