Uncertainty Analysis in Ecosystem Services Modelling

Such demands can be satisfied by a new approach of modelling, named distributed modelling. Distributed modelling defines the next generation of environmental modelling approaches. It moves from centralized, stand-alone desktop applications in which data are tightly integrated to the system towards the proposition of distributed models based on independently distributed and interoperable services. Distributed geographic information processing (DGIP) has turned more and more important in the past decade. Although much of the current research is directed towards distributed modelling, it still mainly focuses on technical issues of how to implement the new generation of models rather than on other aspects, such as the capturing of uncertainty. Uncertainty analysis techniques and uncertainty propagation tools for environmental models have been well-developed, but not yet applied in the context of distributed modelling. This PhD project aims to extend the new distributed environmental modelling approach with uncertainty assessment methodologies. It can partly build on existing methods but these will need to be extended and modified in order to be satisfactory for use in DGIP. As distributed modelling is still maturing, in this project we attempt to analyse how recent and current developments in distributed modelling can be incorporated in a methodology for capturing and visualizing uncertainty of these models. We aim at adapting existing statistical uncertainty quantification approaches and uncertainty visualisation techniques for use in distributed models that are connected via web services.

Project: Uncertainty Enabled Model Web (http://www.uncertweb.org)

More information on the progress of the project:

Check my poster entiteld "EXPERT ELICITATION OF THE VARIOGRAM" which I presented at the conference "Uncertainty in Computer Models 2010" 12-14 July 2010, Sheffield. (http://www.mucm.group.shef.ac.uk/UCM2010.html)