Climate Modeling
Bjorn StevensModeling of the global climate is reviewed through a delineation of different classes of models and different cultures of use. Climate models are classified on the basis of their degrees of freedom and the extent to which these are simulated—calculated based on first principles—or modeled (usually through gestalt constructs). A new class of computationally intensive models, storm-resolving Earth-system models, is introduced. Their advantages and development challenges are discussed in relation to other classes of models. Artificial intelligence provides a new tool for accessing and enabling interactivity with the information content of these models and may also help harmonize the modeling with data, that is, from impacts, more generally. Two cultures of climate model use are distinguished, one interested in understanding how the climate system works, the other interested in making predictions to inform decisions (future making).