Cultures of Prediction in Atmospheric and Climate Science: Epistemic and Cultural Shifts in Computer-based Modelling and Simulation ed. by Matthias Heymann, Gabriele Gramelsberger, and Martin Mahony (review)

We live, as the editors of this valuable book argue, in a "culture of prediction." Algorithms predict how many boxes of cereal a supermarket should order, whether a suspect released on bail might commit a new crime, and what you would like to see next on YouTube. Predictions, supposedly scientific, have become a central element of society, and STS scholars ought to be in the forefront of studying it. To address this issue, a series of workshops wisely focused on case studies of a single narrow but important field: atmospheric science. Weather forecasting is deeply meaningful to most individuals and many businesses, while climate modeling (the subject of most of the chapters in this book) addresses the most urgent existential and political issue of our civilization. Fortunately, scholarship on atmospheric science is burgeoning—the workshops and resulting volume could be assembled by drawing exclusively on the subset of STS scholars and scientists working in northern Europe.

The editors offer a general analytical framework for the social practice of prediction, with no more than the usual burden of social studies jargon. Inevitably, most of the analysis restates ideas that STS students will find familiar and obvious—research is often the slave of social demands, etc. But readers with a taste for this sort of work will find it useful to have everything laid out in detail and in order, with thought-provoking phrases such as "domesticating uncertainty" and "cultures of prediction."

Most of the value in any such collection lies in the diverse facts and analysis of the individual case histories. You will get no synthesis of the entire subject but a feeling for its general nature, rather as you can learn about a country by visiting a handful of towns.

The first set of cases concentrates on the rise of computational models of the general circulation of the atmosphere. Modeling is only one of the two chief areas of modern climate science, the other being the study of past climates. Each area continues to influence scientific and popular thinking, but by the 1970s the modelers had won what the authors would call "epistemic hegemony," at least for prediction. Taking the cases in order, we begin with an overview of nineteenth-century meteorological theory, which proved incapable of representing the general circulation through plain physics equations. Next comes the birth of computational meteorology in 1945–46, as seen in the views and correspondence of key players. A chapter on H. H. Lamb features his groundbreaking but undervalued use of historical data to explore processes of climate change. The next chapter describes a pivotal transformation in the 1970s, when the main thrust of model research turned from the study of fundamental processes to policy-relevant prediction. The value and even the possibility of understanding processes in detail continued to be debated into the 2000s, as shown in a case study of a French group working on the physics of clouds.

A second set of cases comprises histories of the development of regional (and especially national) climate models; of how the Arctic moved from a peripheral to a central element in climate change studies; of approaches to geoengineering with emphasis on modeling, and of geotechnical (soil) models of a failure mode in Dutch dikes. Two final chapters tell the history of efforts by the climate modeling community to analyze and "manage" uncertainties in its predictions and to deploy graphics in communicating the results.

Each chapter focuses on a different dynamic, describing choices that were driven by scientific and political details specific to particular places and times. The most common themes are pressure from politics and tensions between different research heuristics.

Modern research in the atmospheric sciences is dominated by demands for policy-relevant prediction, demands that inevitably bring contention over rights to knowledge claims...