Multi-model ensembles in climate science: Mathematical structures and expert judgements

Abstract

Projections of future climate change cannot rely on a single model. It has become common to rely on multiple simulations generated by Multi-Model Ensembles (MMEs), especially to quantify the uncertainty about what would constitute an adequate model structure. But, as Parker points out (2018), one of the remaining philosophically interesting questions is: "How can ensemble studies be designed so that they probe uncertainty in desired ways?" This paper offers two interpretations of what General Circulation Models (GCMs) are and how MMEs made of GCMs should be designed. In the first interpretation, models are combinations of modules and parameterisations; an MME is obtained by "plugging and playing" with interchangeable modules and parameterisations. In the second interpretation, models are aggregations of expert judgements that result from a history of epistemic decisions made by scientists about the choice of representations; an MME is a sampling of expert judgements from modelling teams. We argue that, while the two interpretations involve distinct domains from philosophy of science and social epistemology, they both could be used in a complementary manner in order to explore ways of designing better MMEs.

Integration of future water scarcity and electricity supply into prospective LCA: Application to the assessment of water desalination for the steel industry

“…This expansion could provide a range of possible future environmental impacts and thus give an estimate of uncertainty for the prospective evaluation. Moreover, since several IAMs exist, another interesting research avenue is to perform multiple simulations for each SSP using multi‐model ensembles to capture uncertainty induced by modeling choices, like what has been advocated and applied in the field of climate science (Jebeilea & Crucifix, 2020).…”

Future Climatic Projections and Hydrological Responses in the Upper Beijiang River Basin of South China Using Bias‐Corrected RegCM 4.6 Data

“…Generally, the uncertainty sources in rainfall‐runoff projection include but are not limited to the GCMs, RCPs, downscaling methods, and the hydrological models, and the uncertainties of the GCMs and RCPs are more obvious than those of the other two aspects (Chen et al., 2020; Wu et al., 2015). Typical approaches to reduce uncertainties include multi‐model ensembles (MMEs) (Jebeile & Crucifix, 2020) and multi‐model weighted ensembles (Colorado‐Ruiz et al., 2018). However, due to the limitation of the number of GCMs, the use of MMEs may be unsuitable in this study.…”

Ces objets qui ne tombent pas en désuétude. Le cas des modèles de transfert de la radioactivité dans l’environnement

“…Les différentes représentations d'un même phénomène ne font pas forcément appel aux mêmes lois théoriques et aux mêmes idéalisations, peuvent servir différents objectifs ou avoir différentes fonctions épistémiques et peuvent donc être a priori incommensurables, non contradictoires ou incompatibles, et peuvent à la fois vivre dans des environnements différents et servir à instruire un même phénomène 13 . Ainsi, les sciences du climat semblent pouvoir tirer parti de la diversité de modèles qui existent en procédant par exemple à des intercomparaisons sur des points précis (grâce aux Programmes internationaux d'intercomparaison de modèles ou CMIP notamment) qui permettent in fine de quantifier les incertitudes, identifier les contraintes émergentes, tirer des prédictions robustes et calculer des moyennes statistiques [Jebeile & Crucifix 2020].…”