Modelling Under Risk and Uncertainty: An Introduction to Statistical, Phenomenological and Computational Methods

1.4 Modelling Under Uncertainty in Metrology, Environmental/Sanitary Assessment and Numerical Analysis

The concept of uncertainty was already an essential part of the domains reviewed previously: every type of risk analysis comprises some statement about uncertainty . . . or else the risky event becomes a certainty. It is not the purpose of the present section to discuss in epistemological terms the specificities of uncertainty as differentiated from risk: Chapters 2 and 4 will return to that point. Observe so far that, beyond the domain of risk analysis as such, the study of uncertainty affecting the results of any modelling or measurement effort has involved some dedicated approaches. Those are referred to as uncertainty and sensitivity analysis (UASA). They stand beyond the explicit risk analysis as they do not model risky events and consequences in the sense of serious accidents. Consider for instance:

• Metrology;
• quality control in industrial processes;
• qualification and calibration of numerical codes;
• environmental or sanitary impact analysis;
• quantitative project management;
• oil exploration;
• climate studies;
• and so on.

1.4.1 Uncertainty and Sensitivity Analysis (UASA)

Notwithstanding the variable terminologies in sub-domains, several authors have pointed to a generic underlying methodological structure to the UASA studies (Helton et al., 1996; de Rocquigny, Devictor, and Tarantola, 2008). Figure 1.7 sketches the conceptual framework advocated by de Rocquigny, Devictor, ...