Sensitivity analysis in linear and nonlinear models: a review

Sensitivity Analysis (SA) generally refers to an assessment of the sensitivity of the output(s) of some model with respect to changes in the input(s). The inputs may be initial state variables or model parameters, and so SA can be useful for data assimilation, model tuning, and even dimensionality reduction. There exists a wide range of approaches to SA because "sensitivity" is a user-dependent concept. This talk reviews and illustrates some of the latest approaches to SA. It is shown that the results are ambiguous in the sense that they depend on the specific approach, and that the problem is exacerbated when the model is nonlinear.