Work Package 4

The project concerning uncertainties in transport models will build upon the Monte-Carlo method, where specific elements that contribute to uncertainties will be analyzed based upon the framework by Walker et al (2003). A promising result of Nielsen & Knudsen (2005 & 2006) is that the statistical distribution of the results may be estimated upon very limited point simulations of the input variables. This is of great practical importance, since transports models have very large calculation times. Whilst uncertainty estimation today can only be done in research environments with clusters of computers available, this may lead to a methodology that can be used in applied modelling. The Bootstrap methods described in Hugosson (2005) will be explored as an alternative to the pure simulation methods, which may improve the uncertainty estimation further.