Sensitivity analysis of the IMAGE Greenhouse model

Abstract Sensitivity analysis is an important component of model building as it provides information about the parameters which have major influences on the model and its outputs. IMAGE (Integrated Model to Assess the Greenhouse Effect) has gained acceptance as a mid-range model for working with and predicting climate change. It is an ideal model for the application of new methods of sensitivity analysis. The sensitivity of IMAGE has been carried out by employing a new screening method which estimates first-order and second-order effects. This includes a first estimate of the non-linear interactions between the parameters of IMAGE. The new method is a development of the screening method proposed by Morris. The efficient sampling strategy in the parameter space is based on graph theory and on the solution of the “handcuffed prisoner problem”. The results of the analysis are presented and some of the important first- and second-order interactions are identified. The strengths of these interactions indicate where the modelling of processes incorporated in IMAGE has to be revised in order to properly represent reality.