University of Twente Student Theses


Development of a conceptual framework for uncertainty and sensitivity analysis: Application to forest management under climate change

Goh, Suz Suz (2009) Development of a conceptual framework for uncertainty and sensitivity analysis: Application to forest management under climate change.

[img] PDF
Abstract:Environmental model is use to predict the natural science system and provides information to decision maker for environmental management and decision making purposes. However, the important issue is not only to obtain information but is to get certain predicted output. As for decision maker, they should always question about how certain the model outputs are. At the some time, modeller should be able to answer this question. The different perspective of uncertainty from modeller and decision maker point of view has hindered the assessment of uncertainty to be effectively implemented. Thus, in order to create the common understanding of uncertainty between modellers and decision makers, an uncertainty conceptual framework is needed. To establish an uncertainty framework, one should understand the source and the propagation of uncertainty in an integrated model system. Managing forest under climate change is one of the examples which involve multiple models integration. Uncertainty and sensitivity analysis therefore is important to identify, to regconise and to assess uncertainty that occurs in the model chain. The uncertain outputs of regional climate model (RCM) in PRUDENCE are used as the site condition input (vegetation growing period, annual temperature ampltitude, mean temperature and precipitation during vegetation growing period) to SILVA to examine uncertainty propagation and conduct the sensitivity analysis. Tree input variable (tree height, tree height to crown base and crown diameter) are used to investigate uncertainty and sensitivity analysis in the SILVA modal. The sensitivity analysis revealed that the SILVA output: aggregation index, species profile index and species mingling index has very small impact from climate change. Sensitivity analysis also provides underlying information of the model and traces the uncertain inputs. Questionnaire analysis describes the uncertainty in decision making model by different experts. With all these analysis, an uncertainty conceptual framework is developed. However, the large uncertainty from different models has been “dismissed” by the classification system in habitat evaluation model. To conclude, the developed uncertainty framework is useful as a communication tool between modellers and decision makers to identify, recognise and analysis uncertainty in the model chain. However, the framework needs to be improved in terms of uncertainty assessment especially in the method of quantifying uncertainty.
Item Type:Essay (Master)
Faculty:ITC: Faculty of Geo-information Science and Earth Observation
Programme:Geoinformation Science and Earth Observation MSc (75014)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page