Talk (Work in Progress): Gabriel Târziu (MCMP)

Title:

The role of model hierarchies in understanding Earth’s climate system

Abstract:

A great deal of effort in climate science is dedicated to developing realistic models of Earth’s climate system with great predictive power. This objective led to the creation of highly complex models that are computationally intractable and incomprehensible to the modeller. So, unfortunately, as climate models became more realistic, they became incompatible with the goal of understanding because they started to resemble the map in Jorge Luis Borges’s story “On Exactitude in Science” (namely, a map of the country on a scale of 1 to 1). This situation determined the well-known American meteorologist Isaac Held to draw attention to “The Gap between Simulation and Understanding in Climate Modeling” (Held 2005). In Held’s opinion, the best way to bridge this gap is by “the construction and analysis of climate model hierarchies” because such hierarchies have the virtue of yielding important understanding. The problem with doing this is that there are many potential ways in which one can arrange the climate models. Is every conceivable hierarchy understanding-enhancing? If not, what exactly are the normative conditions that a model hierarchy has to satisfy in order to help with our understanding? My aim in this talk is to argue for the importance of answering these and similar questions about the role of model hierarchies in climate science and to provide the sketch of a theory about how model hierarchies can contribute to our understanding of a target system.