As we write this in the early summer of 2016, we see news stories reporting that April 2016 was the hottest month of April in the historical record. In fact, the last 12 consecutive months have set global high temperature records. All but one of the ten hottest years going back to 1880 have come in the twenty-first century, with the one exception being 1998. 2015 was the hottest year on record, having broken the previous record (2014) by the largest margin yet, but 2016 looks likely to break both of those records (it will be the hottest year ever, and it will exceed 2015 by an even larger margin than 2015 exceeded 2014). Meteorologists are now predicting that 2016 will surpass the 1.5 °C mark, meaning that it will be more than 1.5 °C higher than the pre-industrial average. 2.0 °C has long been considered a dangerous tipping point beyond which we dare not pass. It is now looking more and more unavoidable.

Every year, usually in February or March, the cap of frozen seawater floating over the North Pole in the Arctic Ocean reaches its largest size for the year before its starts to melt back for the summer. The peak in 2016 was reached on 24 March at 5.607 million square miles. That is the smallest size, in the satellite record going back to 1978, to which the Arctic cap has reached; the 13 smallest years have been the last 13 years. This is an especially worrying development, because the melting of ice is an extremely strong feedback effect in the climate system : as the temperature rises, ice melts and the melting ice reduces the amount of sunlight reflected back into space, which makes the temperature rise even more. Other potential tipping points loom on the not-so-distant horizon: the melting of the Arctic permafrost, which would release billions of tons more carbon into the atmosphere; the melting of the Thwaites glacier in Antarctica, which could destabilize enough of the Antarctic ice sheets to drive sea levels up by 16 feet; and the spread of diseases into areas where they have never been before—with dengue fever, for example, now being a significant risk in areas beyond both tropics for the first time in history.

While a fair bit of controversy concerning the cause of these phenomena remains in the body politic (especially in the United States),¹ nothing could be further from the truth when it comes to the scientific community. Multiple studies, appearing in peer-reviewed publications, all show similar findings: that roughly 97–98% of actively publishing climate scientists agree with the claim that it is extremely likely that the past century’s warming trend is due to human activities.² Eighteen major scientific associations (including the American Association for the Advancement of Science and the American Geophysical Union ) have endorsed the claim that “Observations throughout the world make it clear that climate change is occurring, and rigorous scientific research demonstrates that the greenhouse gases emitted by human activities are the primary driver.”³

Part of this confidence comes from the fact that the scientific basis for the claim of anthropogenesis (caused by human activities) rests on a wide variety of convergent evidence: the recordings of modern instruments concerning the climate going back to around 1880; observations of sea ice, glaciers, ice sheets, animal migrations, etc.; basic science in the form of energy-balance models; reconstructions of more distant climate history from “proxy data” like ice cores, tree rings, pollen samples, coral reefs, and the like; and of course the detailed study of highly complex and sophisticated computer simulation models of the climate. The same can be said about our confidence in the rather general claim that further increases in greenhouse gas concentrations are going to drive the climate further away from its pre-industrial state. That too is supported by a diverse array of evidence.

But the answers to other important questions about the climate, and its response to increases in the concentrations of greenhouse gases in the atmosphere, remain less certain: what is the correct value of the earth’s equilibrium climate sensitivity (the amount that a sustained doubling in the quantity of CO₂ in the atmosphere would raise the equilibrium global surface temperature)? What about the transient features of this response? How long does it take to reach equilibrium? What can we expect from global surface temperature in the meantime?

All these involve hypotheses about the future of a very coarse-grained variable: mean global surface temperature. We would also like to know quite a bit more about how these phenomena will play out regionally. Climate change is likely to make some regions wetter and other regions drier. But which ones, exactly? So far, global warming has been (as the models mostly predicted) concentrated around the poles. Will this continue? At what rate is the Arctic sea ice going to continue to disappear? (So far, it has disappeared faster than most models predicted.) Will the melting of the Arctic ice actually make northern Europe considerably colder? And perhaps most importantly, how likely are, and how close are we to, the kinds of climate tipping points we mentioned above: the collapse of ice sheets in Antarctica and Greenland; the cessation of the vitally important thermohaline circulation system [ocean currents driven by surface heat and freshwater flows or fluxes], or the release of massive quantities of heat-trapping gases from frozen storehouses like the Arctic permafrost?

Answers to some of these latter questions are more difficult to come by, in part because they necessarily depend on less diverse sources of evidence than the basic claim of anthropogenesis. For answering most questions about the expected future pace and tempo of climate change that would come in response to possible emissions scenarios, we are almost wholly dependent on complex simulation models.

The core behavior of the atmosphere can be modeled with three simple laws: Newton’s laws of motion as they apply to parcels of fluid, the law of conservation of mass, and a simple thermodynamic equation that allows us to calculate the heating effect on each parcel of air via a parameterized value of the radiation from the sun. Unfortunately, what we get out of this is a coupled set of nonlinear partial differential equations for which we have no closed form solution. We can at best hope to get a numerical approximation of how a system governed by such equations should behave. Simulation models of the climate do this by transforming the original (continuous) differential equations into discrete difference equations that approximate them, and use a computer to solve the latter step-by-step over discrete intervals of time for discrete points in space. Rather than a function that tells us values for variables like temperature and pressure for arbitrary points in time and space, the computer outputs numerical values for these variables on a space–time grid.

Modern climate models of the most advanced kind do much more than model just the circulation of the atmosphere . The atmosphere, after all, is only one part of the climate system—which consists not only of the atmosphere, but also the hydrosphere (seas), the cryosphere (ice sheets), the land surfaces, and the biosphere, and all the complex interactions between them. Not only does a climate model need to couple the circulation of the atmosphere to the circulation of the oceans, but the atmospheric component must also include representations of physical features like clouds, precipitation, and aerosols; the ocean component must include sea ice dynamics, iceberg transport of fresh water, currents, and wave dynamics; the land component will include precipitation and evaporation, streams, lakes, rivers, etc.; and the ice sheet component will include thickening and thinning and cracks and fissures.⁴ A full Earth System Model (ESM) also tracks sources and sinks of carbon into and out of the biosphere and other systems.

All of this makes a good understanding of the conceptual and philosophical foundations of these models vital. It is vital if we are going to be able to form well-informed judgments not only about what to expect in the future, but also about how we should act—both to mitigate those effects that we possibly can but also to adapt to those that might, at this point in time, be unavoidable.

Unfortunately, despite the fact that computer simulation modeling has played a prominent and ever-growing role in science since the middle of the last century, and despite the fact that it plays a starring role in one of the most socially important sets of scientific questions we have ever faced, it has received, until very recently, only a smattering of interest from philosophers of science. The first goal of this book is to improve on that situation.

The second goal is to explore the philosophical foundations of the other sources of knowledge in climate science. The central component of this goal is to get a better understanding of the relations between models of the climate system and the data that inform them. Data in climate science come from a wide variety of sources and instruments, all of which have strengths and weaknesses. The task of knitting all of those sources together into the most well-informed and responsible representation of the knowledge that is best supported by those sources is highly complex. That, in turn, makes it ripe for philosophical and foundational analysis. In the case of climate science, this kind of analysis by philosophers and foundationally inclined scientists is equally overdue.

In response to these lacunae, we offer this collection of essays by both climate scientists and philosophers writing on a broad array of issues pertaining to climate science and modeling. It is intended for both philosophical and scientific audiences. The essays range from detailed consideration of the evidence for climate models to discussions of models and values, to the robustness of models and its significance, and much more. Each part contains a mixture of pieces by both, philosophers and climate scientists, each offering unique perspectives on the topics at hand, valuable for their insight into climate-related issues and philosophical conundrums involving climate models. The book is not meant to be read from front-to-back, although the pieces in each part do benefit from being read in order. Enjoy!

Oreskes, Santer et al., Lloyd, Mann, Mearns et al.

We open Part 1 with an updated reproduction of a classic paper by Naomi Oreskes, “The Scientific Consensus on Climate Change: How Do We Know We’re Not Wrong?” Oreskes was one of the first scholars to empirically document the degree of scientific consensus regarding the anthropogenic origin of observed changes in the climate. In this paper, she presents many of her findings, supplements with several others, and then offers a philosophical account of why we should take those findings to provide us with strong reason to believe in the claims. We thought this paper would provide a nice “second introduction” to all that follows.

The piece sets the agenda for the volume by answering two central sets of questions about climate science. First: What is the scientific consensus on climate change? How do we know it exists? What exactly does it assert? And second: What should we conclude from that consensus? Might not the claims, about which the overwhelming majority of climate experts agree, nevertheless be wrong? How strong, after all, is their evidence?

An important element of Oreskes’ answer to both sets of questions is a distinction that is central to any discussion of climate science and its epistemology—the distinction between claims about the existence and anthropogenic origin of climate change in the recent past, on the one hand, and claims about the pace and mode of future changes, on the other. Oreskes concedes that there is neither consensus, nor overwhelmingly strong evidence for hypotheses about the pace and mode of future changes. What she is concerned with is claims of the...