Pedigree matrix for evaluating the tenability of a conceptual model

The strength of the tenability of a conceptual model can be evaluated with the pedigree matrix below.

The first two criteria address the extent to which the model is supported by observational and experimentals data or experience (quality & quantity) and whether this empirical evidence is based on proxies or exact measures. Proxy refers to how good or close a measure of the quantity that we model is to the actual quantities that formed the basis of the model. Think of first order approximations, over simplifications, idealisations, gaps in aggregation levels, differences in definitions, non representativeness, and incompleteness issues.

The third criterion addresses the level of theoretical understanding. If our theoretical understanding of the modelled system is very high, we may well be able to reliably represent that system in a simulation model, even if the empirical basis is weak. On the other hand a strong empirical basis may not be sufficient to reliable simulate the system and make projections if our theoretical understanding of the mechanisms involved is absent. In that case extrapolation from past data is not warranted. This criterion aims to measure the extent and partiality of the theoretical understanding of the modelled system. Models based on well-established theory will score high on this metric, while models whose theoretical basis has the status of speculation will score low.

The next criterion addresses wheter the model has high process detail, where the equations governing the model correspond directly to the mechanism that govern the modelled system. In the worst case, the model is a black box and the model equations do not correspond to any known causal mechanism that governs the modelled system.

The plausibility criterion allows experts to epress their judgement on the plausibility of the model. Models that score high on other criteria in the matrix can still be considered implausible for other reasons and models that score low on the other criteria could still be plausible

The final colum addresses the level of expert agreement or disagreement on the model.

References
J-C. Refsgaard, J.P. van der Sluijs, J. Brown and P. van der Keur (2006), A Framework For Dealing With Uncertainty Due To Model Structure Error, Advances in Water Resources, 29 (11) 1586–1597.