Towards a More Adequate Description of Uncertainty

In this chapter, we deal with the fact that the traditional description of uncertainty is only approximate. Namely, measurement uncertainty is usually represented as a combination of probabilistic (random) and interval (systematic) uncertainty. In this chapter, we explain that this approach to representing uncertainty is only an approximation.We show that a more adequate representation of uncertainty leads to a three-component model, with an additional periodic error component. The existing uncertainty propagation techniques are extended to this model. These results are intended to be applied to environmental studies, especially in the analysis of the corresponding time series.