ABSTRACT
In what has gone before, this book has attempted to clarify the different notions of uncertainty associated with environmental modelling. In particular, we have looked at forward uncertainty analysis in Chapter 3, the learning process associated with having historical observations in Chapter 4, the case of real-time forecasting in Chapter 5, and how uncertainty enters into decision making in Chapter 6. In all of these cases, the predictions made will be uncertain, but will often have been made for a purpose, and some of the different ways in which uncertainties can be taken into account in the decision-making process have been outlined in Chapter 6. Throughout, there has been a recognition that achieving “the” model of an environmental system may be fraught with difficulty. Instead, there may be many different model structures, and parameter sets within model structures, that are consistent in some sense with the uncertainties in the available data, and many different ways of estimating uncertainty in the predictions.
