New PCIC Science Brief: Climate Model Genealogy and its Relationship to Modelled Climate Properties

Some of the global climate models used to study the Earth's climate system and make climate projections share components and computer code, much as if they were members of the same “families.” PCIC is pleased to announce the release of a new Science Brief covering a paper in the Journal of Advances in Modeling Earth Systems that examines to what extent these family resemblances affect model results. The authors, Kuma, Bender and Jönsson, examined and quantified how this interdependence between models affects their simulated climate sensitivities, feedbacks and resulting projections of surface air temperature. They found that models with shared code tend to have greater similarity in their climate sensitivities, strengths of feedbacks, and therefore in their projected surface temperatures. They also demonstrated that weighting ensembles of models according to their family resemblance resulted in a lower equilibrium climate sensitivity than when using a simple ensemble mean, and also reduced differences in climate sensitivity between the two most recent generations of climate models.

Read the new Science Brief.