Uncertainty and sensitivity studies of models of environmental systems

ABSTRACT

Traditionally, sensitivity and uncertainty analyses in environmental modelling contribute quantitative descriptions of the relative importances of individual parameters and processes, highlighting areas of significant contributors to the overall uncertainties of model predictions and giving markers for areas requiring substantial improvement (perhaps through directed experimental work). Both sensitivity and uncertainty analysis require extensive re-sampling from input data and simulation of model response and there is a large and growing literature concerning their use in environmental modelling. There is however growing interest in the sensitivity of predictions to model structure, and in evaluating the contribution of model structural uncertainty to the overall uncertainty within the general framework of sensitivity analysis (Draper, 1995, Beven, 1993, Beven and Binley, 1992). Sensitivity and uncertainty analysis contribute to all stages of model development, testing and assessment and their impact on model reliability and validity will be described. After a general introduction to sensitivity and uncertainty analysis and discussion of their contributions, some applications of sensitivity analysis to several environmental modelling studies will be presented.

REFERENCES

• 1: Andres T H, 1987. Statistical sampling strategies. Procs NEA Workshop on Uncertainty analysis for Performance Assessments of Radioactive Waste Disposal Systems, Seattle 1987.
• 2: Andres T H, 1997. Sampling methods and sensitivity analysis for large parameter sets. Journal of Statistical Computation and Simulation, to appear.
• 3: Beven K, Binley A, 1992. The future of distributed models: Model calibration and uncertainty prediction. Hydrological Processes, 6, 279-298.
• 4: Beven K, 1993. Prophecy, reality and uncertainty in distributed hydrological modelling. Advances in Water Research 16, 41-51.
• 5: BIOMOVS, 1993. Final report, BIOMOVS Technical report 15, Swedish Radiation Protection Institute, Stockholm.
• 6: Cooke R, 1994a. Parameter fitting for uncertain models: modelling uncertainty in small models. Reliability Engineering and System Safety 44, 89-102.
• 7: Cooke R, 1994b. Uncertainty in dispersion and deposition accident consequence modelling assessed with performance-based expert judgement. Reliability Engineering and System Safety, 45 35-46.
• 8: Draper D, 1995. Assessment and propagation of model uncertainty. Journal of the Royal Statistical Society, Series B
• 9: EC, 1995. Review on development of methodologies for modelling with uncertainty and variability: Munvar project. European Commission report, EUR 16174 EN.
• 10: Efron B, Tibshirani R, 1993. An introduction to the Bootstrap. Chapman & Hall, New York.
• 11: Hamby D M, 1995. A comparison of sensitivity analysis techniques. Health Physics, 68(2), 195-204.
• 12: Hoffman F O, Hammonds J S, 1994. Propagation of uncertainty in risk assessments: the need to distinguish between uncertainty due to lack of knowledge and uncertainty due to variability. Risk Analysis 14, 707- 712.
• 13: Hoffman F O, Thiessen K M, 1995. Use of Chemobyl data to test predictions and uncertainty estimates from exposure assessment models. Proceedings of 'Environmental impact or radioactive releases', IAEA- SM-339/20, 325-336.
• 14: Hoffirmn F O, Thiessen K M, Watkins B, 1996. Opportunities for the testing of environmental transport models using data obtained following the Chemobyl accident. Health Physics 70, 5-7.
• 15: Iman R L, Conover W J, 1980. Small sample sensitivity analysis techniques for computer models, with an application to risk assessment. Communications in Statistical Theory and Methods, A9(17), 1749-1842.
• 16: lman R L, Helton J C, 1988. An investigation of uncertainty and sensitivity analysis techniques for computer models. Risk Analysis 8, 71-90.
• 17: International Atomic Energy Agency, 1989. Evaluating the reliability of predictions made using environmental transfer models. Safety Series, 100.
• 18: International Atomic Energy Agency, 1995. Validation of models using Chemobyl fallout data from the central Bohemia region of the Czech republic. IAEA- Tecdoc 795.
• 19: McKay M D, 1995. Evaluating prediction uncertainty. Los Alamos National Lab Technical report, NUREG/CR-631 I.
• 20: A. Saltelli , T. Homma, Sensitivity analysis for model output: performance of black box techniques on three international benchmark exercises, Computational Statistics & Data Analysis, v.13 n.1, p.73-94, Jan. 1992 [doi>10.1016/0167-9473(92)90155-9]
• 21: A. Saltelli , T. H. Andres , T. Homma, Sensitivity analysis of model output: an investigation of new techniques, Computational Statistics & Data Analysis, v.15 n.2, p.211-238, Feb. 1993 [doi>10.1016/0167-9473(93)90193-W]
• 22: Scott E M, Gurbutt P, Harms I, Heling R, Nielsen S P, Preller R, Osvath I, Sazykina T, Wada A, Sjoeblom K L, 1997. Benchmarking of numerical models describing the dispersion of radionuclides in the Arctic Seas. Science of the Total Environment, to appear.