You are in ePLANETe >
Multiple Model Simulation
Multiple Model Simulation is a strategy to address uncertainty about model structure. Instead of doing an assessment using a single model, the assessment is carried out using different models. For instance, this can be realised by having alternative model codes with different process descriptions (Linkov and Burmistrov, 2003; Butts et al., 2004) or, in the groundwater case, by having different conceptual models based on different geological interpretations (Troldborg, 2000, Selroos et al., 2001).
The strategy of applying several alternative models based on codes with different model structures is also common in climate change modelling. Thus in its description of uncertainty related to model predictions of both present and future climates the IPCC Special Report on Emission Scenarios (IPCC, 200x) bases its evaluation on scenarios of many (up to 35) different models.
The same strategy is followed in the Dialogue Model (Visser et al. 2000). Dialogue simulates the cause effect chain of climate change, using mono-disciplinary sub-models for each step in the chain. The chain starts with scenarios for economic growth, energy demand, fuel mix etc., leading to emissions of greenhouse gasses, leading to changes in atmospheric composition, leading to radiative forcing of the climate, leading to climate change, leading to impacts of climate change on societies and ecosystems. Rather than picking one main-stream mono-disciplinary sub-model for each step, Dialogue uses multiple models for each step (for instance, three different carbon cycle models, five different GCM model-outcomes, etc.), representing the major part of the spectrum of expert opinion in each discipline. This multiple model approach facilitates the inclusion of new alternative models in each step to accommodate new scientific ideas on the structure of a given sub-model.
Refsgaard et al (2006) present a new framework for dealing with uncertainty due to model structure error, based on alternative conceptual models and assessment of their pedigree and adequacy.
Resources required
Multiple model simulation requires modelling skills to design and implement the various model structures. Varying model structure is a much more complex task than varying parameters and there are no ready to use software solutions available. The required resources will further depend on the complexity of the model and the required time for each model run. If models developed and run by different modelling groups are used, a substantial part of the resources will be required for co-ordination and project management.
Strengths and limitations
+ The effects of alternative model structures can be analysed explicitly.
+ It makes it possible to include expert knowledge on plausible model structures.
+ It substantially reduces the chances that the assessment overlooks important aspects of the problem compared to the use of single models only. In that sense it may reduce surprises.
- We cannot be sure whether we have adequately sampled the relevant space of plausible models. Important plausible model structures could be overlooked.
- Often we do not know the plausibility and even less the probability of each conceivable model structure.
- The expert knowledge on which the formulations of the alternative conceptual models have to be based has an unavoidable subjective element.
References
Butts MB, Payne JT, Kristensen M and Madsen M (2004) An evaluation of the impact of model structure on hydrological modelling uncertainty for streamflow simulation. Journal of Hydrology, 298, 242-266.
Linkov I and Burmistrov D (2003) Model uncertainty and choices made by the modelers: lessons learned from the international atomic energy agency model intercomparisons, Risk Analysis 23(6), 1297-1308.
Nakicenovic N, Alcamo J, Davis G, de Vries B, Fenhann J, Gaffin S, Gregory K, Grübler A, Jung TY, Kram T, La Rovere EL, Michaelis L, Mori S, Morita T, Pepper W, Pitcher H, Price L, Riahi K, Roehrl A, Rogner H-H, Sankovski A, Schlesinger M, Shukla P, Smith S, Swart R, van Rooijen S, Victor N, Dadi Z (2000) IPCC Special Report on Emissions Scenarios. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. 599pp.
Refsgaard JC, van der Sluijs JP, Brown J and van der Keur P (2006) A Framework For Dealing With Uncertainty Due To Model Structure Error.Advances in Water Resources, 29 (11) 1586-1597.
Selroos JO, Walker DD, Strom A, Gylling B and Follin S (2001) Comparison of alternative modelling approaches for groundwater flow in fractured rock. Journal of Hydrology, 257, 174-188.
Troldborg L (2000) Effects of geological complexity on groundwater age prediction. Poster session 62C, AGU. EOS Transactions, 81(48), F435.
Visser H, Folkert RJM, Hoekstra J and De Wolff JJ (2000) Identifying key sources of uncertainty in climate change projections. Climatic Change 45, 421-457.