Show Summary Details

Page of

PRINTED FROM the OXFORD RESEARCH ENCYCLOPEDIA, CLIMATE SCIENCE (climatescience.oxfordre.com). (c) Oxford University Press USA, 2018. All Rights Reserved. Personal use only; commercial use is strictly prohibited (for details see Privacy Policy and Legal Notice).

date: 19 November 2018

Climate and Simulation

Summary and Keywords

Climate and simulation have become interwoven concepts during the past decades because, on the one hand, climate scientists shouldn’t experiment with real climate and, on the other hand, societies want to know how climate will change in the next decades. Both in-silico experiments for a better understanding of climatic processes as well as forecasts of possible futures can be achieved only by using climate models. The article investigates possibilities and problems of model-mediated knowledge for science and societies. It explores historically how climate became a subject of science and of simulation, what kind of infrastructure is required to apply models and simulations properly, and how model-mediated knowledge can be evaluated. In addition to an overview of the diversity and variety of models in climate science, the article focuses on quasiheuristic climate models, with an emphasis on atmospheric models.

Keywords: climate research, climate modeling, dynamic meteorology, computer-based simulation, IPCC Assessment Reports, model intercomparison, community models, model thinking, model evaluation, uncertainties

Introduction

Climate simulation has become synonymous with trying to understand and forecast possible future global developments. No other scientific discipline has gained more attention and media coverage, nor posed such a challenge for politicians, as climate models have. However, neither the scientific concept of climate nor the practice of mathematical modeling and simulation is correlated with everyday experience. Climate, as defined by the World Meteorological Organization (WMO), is “the statistical description of weather variables in terms of the mean and variability of relevant quantities over a period of time” (World Meteorological Organization, 2016). The “period of time” addressed usually covers 30 years of weather observations, the “mean quantities” refer to the averaging of local data in terms of global means, and “variability” marks the changes in anomalies compared to other periods of time. For instance, global mean annual surface temperature exhibited a deviation of about 0.4°C between the 1880s and 1970s and of about 0.8°C by 2010 (GISTEMP [GISS Surface Temperature Analysis], 2016). In other words, we know that we face global warming, although none of us can ever personally experience the trend in global mean annual surface temperature. Furthermore, we know from the projections of climate models that global warming will increase. Based on model-mediated knowledge, we know that societies will have to invest billions of dollars in order to mitigate and to adapt to climate change. But what is a model? What kind of scientific knowledge does it generate? How can model-based knowledge be evaluated?

Models are common tools in science (Black, 1962; Giere, 1999; Hesse, 1963; Magnani & Nersessian, 2002; Morgan & Morrison, 1999; Toon, 2012; for an overview, see Frigg & Hartmann, 2012). In general, scientific models can be classified in two categories: material models and symbolic models. So-called material models, scale models and analog models, mimic objects and processes, respectively, in experiments. Scale models of ships, airplanes, and cars are tested in wind tunnels to help engineers investigate the aerodynamic properties of a scaled-down prototype, based on the underlying hypothesis that upscaling will not change the properties. Analog models “simulate” processes in the laboratory. For instance, hydraulic models materially simulate the behavior of tides and floods. Cloud chambers allow the study of the microphysical processes of clouds in order to gain a conceptual understanding of specific relationships.

Entirely different from material models are symbolic models, which represent objects, processes, and properties symbolically and thus provide information about the state and development of a system (Müller & von Storch, 2004). Usually expressed in mathematical terms, symbolic models can vary from conceptual models based on major simplifications to quasirealistic models containing as much information about processes as possible (von Storch, 2010). Conceptual models, sometimes formulated as only a single equation expressing a specific relationship, are easy to investigate mathematically. They can usually be solved analytically, yielding a solution that illuminates the specific aspect of interest. However, the climate system is a complex system, interweaving numerous processes that constitute its state and development. Thus, conceptual models have gradually been transformed into complex, quasirealistic models, containing information about atmospheric and oceanic processes interacting with land surface, ice, vegetation, chemistry, and anthropogenic effects. Today, comprehensive Earth System Models (ESMs) are the ultimate goal of climate science. The price for increasingly complex mathematical models is that they cannot be solved analytically. Instead, they need to be computed numerically (simulated), which requires advanced computing resources. Because mathematical models represent climate symbolically, such representations can depict either facts or fictions, not only about the climate system, but also about the projected results as well. Thus, climate models have to be thoroughly evaluated, requiring a global infrastructure to supply observational data for model comparison and evaluation.

Based on the specific circumstances of climate and simulation, this article explores how climate became a scientific object and how climate models became subject to simulation, with a focus on meteorology and the atmospheric part of climate models. The article also investigates the global infrastructure that modern climate models require in order to be applied properly. Finally, it discusses models as new “tools” of scientific knowledge production, afflicted with many problems but with the unique advantage of providing insight into complex phenomena, such as climate in the future.

Historical Development of Quasirealistic Climate Models

Advent of Dynamic Meteorology and Model Thinking in the Late 19th Century

Until the late 19th century, meteorology was a purely empirical science. Meteorologists measured the main meteorological quantities (using specific tools): temperature (thermometer), air pressure (barometer), wind direction and velocity (vane, anemometer), and humidity (hygrometer). However, the measurement data they collected gave only a description of the atmosphere, snapshots of its momentary states.

Of greater interest to scientists is understanding which atmospheric changes occur, how they are caused, and how they will develop in the future. Such a theoretical understanding can be developed either inductively or deductively (Lorenz, 1969). The inductive approach is based on measurement data and the attempt to find structures and patterns in the data. It led to the development of the statistical and climatological approaches to weather, respectively, as well as to the synoptic approach. Yet, the theoretical results of the inductive approach were meager: Some synoptic rules were developed, and a single empirical law, Buys-Ballot’s baric wind law describing the relationship between air pressure and wind direction (Buys-Ballot, 1854, 1857). Therefore, the deductive approach based on a physical and mechanical understanding of the atmosphere increasingly gained favor. The consequence was that it turned meteorology into the physics of the atmosphere—also called “dynamic meteorology”—focusing on the dynamics of atmospheric processes articulated by mathematical models.

Since the time of Isaac Newton, a physical and mechanical understanding of nature had been expressed by abstract models, and since the time of Leonard Euler, differential equations had become the lingua franca for conceptualizing the kinetics and dynamics of physical processes mathematically (Darrigol, 2005; Eckert, 2006; Euler, 1755; Newton, 1687). Thus, mechanics and mathematics became synonyms, and the study of natural phenomena was, in an increasing number of scientific fields, replaced by studying abstract mathematical models and their behavior. Celestial mechanics was the first science to use this new research style (Dear, 1995). An important reason for the primary role of astronomy was the intrinsically distant perspective of astronomers on the objects of their research, which supported an abstract model view usually expressed as the linear behavior of two bodies (e.g., planets). How were meteorologists to develop a distant perspective on the extremely complex states of the atmosphere surrounding them in order to arrive at a similar abstract model view?

The answer was the perspective on the atmosphere as a fluid with a global circulation. In 1686, Edmund Halley had realized that solar radiation differs in low and high latitudes, and that this causes a north–south circulation as heated tropical air is replaced by cooler air from polar regions (Halley, 1686/1687a, 1686/1687b). In 1735, George Hadley had pointed out that the atmospheric circulation is deflected by the Earth’s rotation (Hadley, 1735). Because the speed of rotation differs at each point on Earth, the deflection of air masses differs as well. As Dove explained in 1837, this causes a difference in rotational speed between moving air masses and the places to which the masses move (Dove, 1837). Finally, Ferrel accounted for the Coriolis effect of the Earth’s rotation in 1856 (Ferrel, 1856, 1858; cf. Fleming, 2000, 2002). From these considerations, the first three-cell model of the global circulation for each hemisphere resulted.

The three-cell model consists of the polar cells (beyond 60°N and 60°S), the middle-latitude cells (60°N−30°N and 60°S−30°S), and the tropical cells (30°N−0°N and 30°S−0°S)—the latter also called Hadley cells. In the tropical cells, the circulation of the northeasterly and southeasterly trade winds is regularly compared to the complex wind patterns of the westerlies in the middle-latitude cells. The open question in the 19th century was how the circulation in the upper atmosphere transported air from one cell to the other. In particular, James Thomson realized that, “in temperate latitudes, there are three currents at different heights” bringing air from, and back to, the poles (J. Thomson, 1857, p. 38). However, it was Ferrel, rather than Thomson (Maury, 1855; J. Thomson, 1857, 1892), who rooted the three-cell circulation model in a sound physical basis expressed by advanced mathematics (Ferrel, 1877, 1886). Therefore, he was regarded as the “standard authority” for dynamic meteorology in his day (Davis, 1887; Sprung, 1885). His approach changed the view of the atmosphere, which was now seen as a giant engine for the circulation of air masses and heat, driven by solar radiation and gravitational forces.

Methods of Computation Before the Advent of Electronic Computers

An early application of dynamic meteorology for weather forecasting was provided by the Austrian school of meteorology at the late 19th century. Max Margules tried to mathematically describe and compute a conceptual model of the “atmosphere’s tide” and its barometric fluctuations (1890)—a laborious work before the advent of electronic computers. Margules had to find analytically tractable approximations of the governing equations to calculate them by hand. Thus, Margules based his computations on Buys-Ballot’s baric wind law and Laplace’s scheme for calculating tides, and he calculated the work needed to change the state of a quantum of air from motion into equilibrium (Laplace, 1775, 1776; Margules, 1901; cf. Pichler, 2001). At the same time, his colleague Felix Exner developed a conceptual model based on hydrostatic and geostrophic approximations for the thermodynamic equation. He manually computed the advective rate of change in the potential temperature for one layer. His computations are seen as the first numerical forecast in meteorology (Exner, 1902, 1917; Fortak, 2001; Volkert, 2007).

Another way of computing weather forecasts before the advent of electronic computers was to use graphical methods in order to project into the future the course of the cyclones and anticyclones roaming over synoptic maps (Ekholm, 1904). An important contribution was made by the research program of Vilhelm Bjerknes, who developed a graphical algebra for performing computations directly upon the charts. Such a graphical method, he announced, “will be of the same importance for the progress of dynamic meteorology and hydrography as the methods of graphical statistics and of graphical dynamics have been for the progress of technical sciences” (V. Bjerknes & Sandström, 1910/1911, p. 69). As knowledge of the development of cyclones was increasingly discovered and applied—for instance, the polar front theory developed mainly by the Bergen school of meteorology (V. Bjerknes, 1919; Friedman, 1989)—synoptic maps became more reliable prognostic instruments (Fjörtoft, 1952; Scherhag, 1939).

However, synoptic forecasts depended on the subjective experience of the meteorologists and their “intuitive glance” for adjusting data (Anderson, 2005). Therefore, in addition to the conceptual models and synoptic forecasts, an objective forecasting method based directly on the hydro- and thermodynamic equations (the so-called “primitive equations” or “quasirealistic” model, respectively) became the overall aim of dynamic meteorology. As early as 1904, Vilhelm Bjerknes had outlined a fully developed circulation model consisting of seven hydro- and thermodynamic equations expressing the relationship between the seven main state variables of the atmosphere: temperature, pressure, density, humidity, and wind velocity in three directions (V. Bjerknes, 1904; Gramelsberger, 2009; Persson, 2005b). Introducing more than two variables into a model makes it nonlinear and thus, in principle, not analytically solvable, as the mathematician Henri Poincaré had already proven in 1890. A seven-variable problem like Bjerknes’ circulation model, consisting of a set of seven equations, is far beyond the possibility of ever deriving a solution algebraically. Lewis F. Richardson was the first to try to numerically compute a quasirealistic model by hand, but he failed to correctly predict a change in air pressure. However, his book Weather Prediction by Numerical Process (Richardson, 1922) anticipated the mid-20th-century style of weather and climate simulation (Lynch, 1999; Nebeker, 1995; Platzman, 1968).

Thus, in the early 20th century, meteorology was trapped between analytically tractable approximations, which were manually computable but highly idealistic, and fully developed circulation models beyond any computability by hand or by early computers. Vorticity, the core problem of every circulation model, disappeared in idealized models. Such idealized models were outlined by Hermann von Helmholtz and William Thomson (Lord Kelvin) in the middle of the 19th century (W. Thomson, 1867; von Helmholtz, 1858). Because in these simple models density depended solely on pressure, vorticity didn’t occur. However, weather is characterized by the appearance and disappearance of vorticity (storms). Therefore, V. Bjerknes and others tried to derive more complex models, but the complex models were not computable (V. Bjerknes, 1898; Schütz, 1895a, 1895b; Silberstein, 1896; Thrope, Volkert, & Ziemianski, 2003). Thus, treating vorticity in a computable way was the aim of early 20th century dynamic meteorology. In the 1930s, Carl-Gustaf Rossby articulated a linear model that conserved the vertical components of absolute vorticity in currents for the perturbations in the upper westerlies (Rossby, 1939; cf. Beyers, 1960). A little later, Ertel formulated a generalization of Bjerknes’ circulation theorem by conserving potential vorticity (Ertel, 1942a, 1942b). Ertel and Rossby were laying the foundation for today’s weather models and together they “derived another vorticity theorem for barotropic fluids, known as the Ertel-Rossby invariant” (Ertel & Rossby, 1949; Fortak, 2004; Névir, 2004, p. 485). Rossby, in particular, became the leading figure for the simulation style of dynamic meteorology in the 1940s. He introduced numerical weather prediction to Europe at the University of Stockholm as well as to the United States at the University of Chicago (cf. Allan, 2001; Harper, 2008).

From Weather Forecast Models to Climate Models

The situation changed in the late 1940s, with the development of electronic computers. However, due to the limited performance of early computers, the simulation of weather with numerical models started with extremely simplified barotropic models comparable to von Helmholtz’s and Thomson’s models of the 1860s. In a barotropic model, pressure is solely a function of density and fields of equal pressure running parallel to fields of equal temperature, thus reducing the effort of computation by modeling wind independent of altitude (geostrophic wind). The very first weather model ever electronically computed—Charney and his colleagues’ simulation of pressure development at a height of about 5,500 m for a 15 × 18 grid representing Northern America processed on the ENIAC computer (Charney, Fjørtoft, & von Neumann, 1950)—was a barotropic model (cf. Harper, 2008; Nebeker, 1995; Persson, 2005a). In order to give rise to cyclones and anticyclones (a quasigeostrophic model), an instability process had to be introduced (Charney & Phillips, 1953; cf. Lewis, 2000; Phillips, 1995). Thus, it was not surprising that the new approach of tackling the forecast problem with numbers was heavily criticized in the beginning. In particular, the “tricks” used to simplify the models in order to reduce computations caused major suspicions. For instance, averaging out errors by relaxation methods was seen as a major problem (Thompson, 1954, p. 320). Or, more generally, it was criticized that “500 mb geopotential is not weather” (Norbert Wiener quoted in Arakawa, 2000, p. 6). Nevertheless, the numerical approach gained ground. In 1954, Solomon Belousov computed a barotropic model on the Russian BESM computer (Blinova & Kibel, 1957; cf. Marchuk, 1974; Wiin-Nielsen, 2001). Also in 1954, a group of Scandinavian meteorologists carried out a barotropic forecast on the Swedish BESK computer (Persson, 2005a). Similar efforts took place in other countries (cf. Guillemot, 2011), and as early as 1948, the British Meteorological Office held a workshop, The Possibilities of Using Electronic Computing Machines in Meteorology (cf. Persson, 2005b; Walker, 2011), which was followed by two symposia on the development in numerical prediction, one in Stockholm in 1952 and the other in Frankfurt in 1954. However, it took years before meteorological offices could afford computers and numerical weather forecasts became operational for everyday use.

The very early barotropic models could hardly be considered weather models. The same is true for the first electronically computed “climate model.” In the mid-1950s, geophysicist Norman Phillips computed a general circulation model on von Neumann’s computer at the Institute for Advanced Studies (IAS) at Princeton (Phillips, 1956; cf. Goldstine, 1972; Lewis, 2000). His two-level quasigeostrophic model “predicted the easterly-westerly-easterly distribution of surface zonal wind, the existence of a jet, and the required net poleward transport of energy” (Phillips, 1956, p. 157). Phillips’ computations are considered to be the crucial evidence that simulations can represent large-scale dynamic patterns of the atmosphere as conceived in the three-cell-model a hundred years earlier. The pioneering work of barotropic models was later called by Akio Arakawa the “epoch-making first phase” of numerical modeling (Arakawa, 2000), although the early models were proofs of concepts rather than quasirealistic weather or climate models.

This first phase was followed by the “magnificent second phase” in the 1960s of global circulation models (GCMs), which enjoyed fewer restrictions than the barotropic models (cf. Arakawa, 2000; Edwards, 2000, 2010; Spekat, 2001). Arakawa collaborated with Yale Mintz to open up the second phase with their General Circulation Model at the University of California Los Angeles (UCLA), applying Jakob Bjerknes’ program of investigations into the general circulation of the atmosphere (J. Bjerknes & Mintz, 1955; Mintz, 1955, 1958; cf. Johnson & Arakawa, 1996; Randall, 2000). The Mintz-Arakawa model was a global, two-level model based on the primitive equations of hydro- and thermodynamics, accounting for realistic land−sea distributions, surface topography, and moisture by introducing a moisture convective adjustment. Furthermore, the model included long-wave cooling and seasonal changes in solar radiation. An initial study with the GCM was carried out in 1965 on an IBM 709 computer at UCLA (Mintz, 1965). Mintz and Arakawa used the model as an experimental test bed, changing boundary conditions artificially to study the behavior of their model by comparison across various numerical experiments. This opened up the “in silico” experimental style of using simulation models as test beds that characterizes modern computational sciences, requiring enormous computing resources.

The GCMs of the second phase increasingly included information about the oceans (swamp ocean) and finally developed into coupled atmosphere−ocean models (AOGCMs). Thus, they transformed into quasirealistic climate models because climate is influenced mainly by oceans and sea ice coverage (cf. Dahan, 2010; Heymann, 2010; Laprise, Lin, & Robert, 1997). The reason is that the top few meters of the oceans hold more heat energy than the entire atmosphere; therefore, the energy exchange between atmosphere and ocean as well as the ocean’s circulation are crucial for long-term climatic integrations. The first coupling of an atmosphere and an ocean model was carried out by Syukuro Manabe and Kirk Bryan at the Princeton Geophysical Fluid Dynamics Laboratory (GFDL) in 1969. The tradition in numerical modeling at GFDL was rooted in von Neumann’s and Charney’s meteorology project of numerical weather forecasting. However, Manabe and Bryan had climate calculations in mind, taking into account “the entire fluid envelope of the earth, consisting of the atmosphere and the hydrosphere” as well as of the cryosphere—sea ice and land ice (Manabe & Bryan, 1969, p. 786). Furthermore, water vapor, carbon dioxide, and ozone were considered for the transfer of terrestrial radiation. Simulations were carried out, using an UNIVAC 1108 computer, on nine levels for a 500-km grid resolution for the atmosphere, and on five levels for the ocean with the same horizontal resolution, with extra rows of grid points added to the western boundary. The results confirmed the hypothesis that oceanic currents have a substantial effect on the distribution of temperature, humidity, and precipitation patterns.

The development of global models during these years created, as Paul Edwards called it, a “family tree” of GCMs in the United States and Europe (Donner, Schubert, & Somerville, 2011; Edwards, 2000; Kasahara & Washington, 1967; Leith, 1964; Manabe, Smagorinsky, & Strickler, 1965; Messinger & Arakawa, 1976; Smagorinsky, 1963; Weart, 2010). The models influenced each other and sometimes one model was the direct ancestor of another as the latter model inherited parts of the software code. But not just the quasirealistic models yielded insights into the climate system. Regional and mesoscale models, too, as well as cloud-resolving models, showed promise for studying the whole spectrum of atmospheric phenomena. With the diversity of models, increasingly prominent questions concerning anthropogenic climate change could be addressed, when human carbon dioxide (CO2) production came under suspicion as having potential to irreversibly change climate. Evidence for the “vast geophysical experiment” of mankind (Revelle & Suess, 1957) was delivered by Charles D. Keeling, whose measurements at the Mauna Loa Observatory in Hawaii clearly demonstrated the increase in CO2 concentration at a site that was thought to be otherwise unpolluted (Keeling, 1958, 1978; Plass, 1956; Weart, 2003). The Keeling Curve has become the icon of human-induced climate change. It led to the core question of climate change: How much increase in the annual mean temperature of the surface would result from doubling of CO2 concentrations? However, the answer to the CO2 doubling question can be given only by climate simulations, and the first answer computed was already alarming: “Doubling the existing CO2 content of the atmosphere has the effect of increasing the surface temperature by about 2.3°C” (Archer & Pierrehumbert, 2011; Manabe & Wetherald, 1967, p. 254; Möller, 1963; Stehr & von Storch, 2010; Weart, 2003). Thus, in silico experiments were urgently needed in order to better understand the vast geophysical experiment of mankind, or to be more precise, of the highly industrialized nations in the Western world. “But such [in silico] investigations [were] in danger of becoming mere exercise, due to the lack of observations to supply the initial conditions and to check the calculations” (National Academy of Sciences, 1966, p. 4). During the 1960s, numerical simulation had outstripped the evaluation potential of meteorological and climatological observations.

Global Infrastructure for Climate Modeling and Simulation

WMO’s Global Infrastructure of Climate Science

Originally focused on weather monitoring, the WMO quickly expanded its work from the late 1960s on, due primarily to concerns about human influence on the climate (cf. Young, 1997). Understanding the climate, and climate change in particular, requires globally collected data on relevant processes within and between the components of the climate system: the atmosphere, the hydrosphere (oceans, lakes), the cryosphere (land and sea ice), the pedosphere, the terrestrial and maritime biosphere, and the anthroposphere. Intensifying observation of these processes has led to an outstanding global observation infrastructure. Along with the United Nations Environment Programme (UNEP), the WMO has installed and organized many global programs. The concerns about climate change reached a first peak with the report on the Study of Man’s Impact on Climate (SMIC) in 1971 (SMIC, 1971) and the United Nations Conference on the Human Environment (UNCHE) in 1972 (cf. Demeritt, 2001; Howe, 2014; Weart, 2003, 2014). The WMO responded to the concerns by setting up programs like the Global Atmosphere Watch (GAW), combining the Global Ozone Observing System (GO3OS) and the Background Air Pollution Monitoring Network (BAPMoN). From 1966 until 1979, the Global Atmospheric Research Program (GARP) addressed both requirements–understanding climate and climate change, respectively, with the goal of “advancing the range of deterministic weather prediction and understanding the physical basis of climate” (Barron, 1992, p. 1).

The first efforts were rooted mainly in scientific concerns and studies on ongoing developments, but at the end of the 1970s the situation changed. Attempting to answer the CO2 doubling question more robustly, Charney et al. carried out numerical studies using two GCMs that represented the state of the art at the time. They concluded that “our best estimate is that changes in global temperature on the order of 3°C will occur and that these will be accompanied by significant changes in regional climatic patterns” (Charney et al., 1979, p. 17). The highly influential “Charney Report,” as it became known, and the first World Climate Conference (WCC-1) in Geneva in 1979, marked a watershed in climate science. Climate change turned into a public policy issue, increasingly interlinking climate science and politics through expanded international conferences, research programs, working groups, and committees (cf. Depledge, 2005; Grover, 2008; Halfmann & Schuetzenmeister, 2009; Hare, Stockwell, Flachsland, & Oberthür, 2010; Jasanoff, 2011; Jasanoff & Martello, 2004). The aims of the international activities coordinated by the WMO were threefold: a better understanding of the current state of the atmosphere (observation), a framework for negotiating an adequate response to climate change (climate politics), and a better understanding of future trends (simulation; for details, see Edwards, 2010).

Ever since the first sputnik collected data from space for the International Geophysical Year from 1957 to 1958, the goal of the WMO was to gain better knowledge of the current situation by installing a global observation infrastructure (see Table 1). Through the global observation infrastructure, more than 50 different essential climate variables (ECV) have become observable.

Table 1. World Meteorological Organization Global Infrastructure of Climate Science

Since 1979

World Climate Programme (WCP)

World Climate Impact Assessment and Response Programme (WCIRP)

Since 1980

World Climate Research Programme (WCRP)

Since 1992

Global Climate Observing System (GCOS), including the Global Observing System (GOS) and the Global Atmosphere Watch (GAW)

GOS

GOS collects data from 1,000 land stations, 1,300 upper-air stations, 4,000 ships, ~1,200 drifting and 200 moored buoys, and 3,000 Advanced Research and Global Observation Satellite (ARGOS) profiling floats, as well as 3,000 commercial aircraft, five operational polar-orbiting meteorological satellites, six geostationary meteorological satellites, and several environmental research and development satellites.

GAW

GAW coordinates data from 26 global stations, 410 regional stations, and 81 contributing stations to produce high-quality data on selected variables characterizing the chemical composition of the atmosphere.

The second goal of the WMO was to install a framework for negotiating an adequate response to climate change. In particular, the view of Earth from space—in December 1972, the crew of the Apollo 17 spacecraft sent back the famous Blue Marble image of a cloud-surrounded blue sphere embedded in the blackness of space—increased awareness about the uniqueness and vulnerability of the planet (cf. Cosgrove, 2001; Jasanoff, 2001; Poole, 2008). This view, combined with alarming reports in the 1970s and early 1980s on environmental catastrophes, such as droughts, floods, air pollution, acid rain, and depletion of the ozone layer, caused calls for action to increase dramatically. Thus, in 1985, the Villach Conference on the assessment of the role of CO2 and other greenhouse gases (GHGs) emphasized the need to institute a property rights regime for human use and modification of the carbon cycle by establishing the Advisory Group on Greenhouse Gases (AGGG). Inspired by the successful negotiations of the Montreal Protocol on Substances that Deplete the Ozone Layer in 1987 (cf. Andersen & Sarma, 2002), these activities became a role model for environmental governance of climate change. They led to the appointment of the Intergovernmental Negotiating Committee on Climate Change (INC), the United Nations Framework Convention on Climate Change (UNFCCC), and finally the adoption of the Kyoto Protocol in 1997 (cf. Agrawala, 1999; Aykut & Dahan, 2015; Boehmer-Christiansen, 1994; Bodansky, 1995; Dahan & Aykut, 2013; Elzinga & Landström, 1996; Franz, 1997; Gupta, 2014; Jasanoff & Martello, 2004; Miller & Edwards, 2001; Shackley & Wynne, 1995; Siebenhüner, 2003; Skodvin, 2000; Stehr & von Storch, 2010; van Asselt, 2014).

In particular, the Intergovernmental Panel on Climate Change (IPCC), established in 1988, and the IPCC Assessment Reports on Climate Change have become core instruments for the supranational governance of climate change. Since 1990, five IPCC Assessment Reports have been released and a sixth is in preparation (see Table 2). All of the reports have three parts of each working group (WG): the Physical Basis (WGI); the Impacts, Adaptation, and Vulnerability (WG2II); and the Mitigation of Climate Change (WG3III). Every report is the product of several hundred lead authors and contributing authors who consider tens of thousands of comments from the scientific and government community. Each report starts with a Summary for Policymakers (SPM), “reviewed at final plenary sessions, where governments have to approve the SPM text, tables and figures in detail, that is, line by line” (Petersen, 2011, p. 100). The final plenary sessions are laborious, days-long meetings attended by government officials and scientists. They document the interlinking of climate science with politics and the establishment of global climate governance (cf. Agrawala, 1998a, 1998b; Bolin, 2007; Beck, 2009; Hulme & Mahony, 2010).

Table 2. IPCC Assessment Reports on Climate Change

1990

First Assessment Report (FAR)

1995

Second Assessment Report (SAR)

2001

Third Assessment Report (TAR)

2007

Fourth Assessment Report (AR4)

2013/2014

Fifth Assessment Report (AR5)

The third goal of the WMO was to achieve a better understanding of future trends based on climate modeling and simulation. This led to a conjoint infrastructure in environmental sciences coordinating global climate modeling. Among others, the mission of the World Climate Research Programme (WCRP) was, and still is, to develop and evaluate climate system models. In accord with this mission, the Working Group on Numerical Experimentation (WGNE) was established in 1980, followed by the Working Group on Coupled Modeling (WGCM) in 1997. The working groups organize the numerical experimentation and evaluation for the IPCC Assessment Reports and have an important influence on the practice of climate modeling and simulation—an impact unique in science (cf. Mahony & Hulme, 2016). They have coordinated and synchronized model development, numerical experimentation, and evaluation for all participating modeling groups since the first IPCC Assessment Report in 1990.

Model Intercomparison and Reanalysis Data

In particular, model intercomparison has shaped the rhythm of model development for the IPCC Assessment Reports. Model intercomparison is coordinated by the Coupled Model Intercomparison Project (CMIP) under the auspices of the WGCM and is carried out at the U.S. Lawrence Livermore National Laboratory. Its mission is to provide “a community-based infrastructure in support of climate model diagnosis, evaluation, intercomparison, documentation and data access” (CMIP, 2015; Meehl et al., 2005). In past years, CMIP3 (AR4), CMIP5 (AR5), and CMIP6 (AR6) have replaced the Atmosphere Model Intercomparison Project (AMIP) of the first, second, and third IPCC Assessment Reports. Thus, “virtually the entire international climate modeling community has participated in this [intercomparison] project since its inception in 1995” (CMIP, 2015). The timeline of CMIP5 documents how the work steps are coordinated (see Table 3).

Table 3. Timeline of CMIP5 for the Physical Basis (Working Group I) Part of IPCC AR5

2006, 2007

Discussion of model improvements and a preliminary set of CMIP5 numerical experiments.

2008

Community-wide consensus about improvements and experiments.

2009

List of requested model outputs available. Final set of CMIP5 experiments approved by the WCRP Working Group on Coupled Modeling, including decadal hindcasts and prediction simulations, “long-term” simulations, and “atmosphere-only” simulations for especially computationally demanding models.

2010

Modeling groups participating in AR5 begin simulation runs with their improved models for CMIP5 and deliver their results to the Lawrence Livermore National Laboratory.

In November 2010, work on the AR5 begins with the First Lead Author Meeting of Working Group I.

2011

In February 2011, the first model output becomes available to the climate science community for analysis.

2012

By the end of July 2012, papers based on the CMIP model output have to be submitted for publication to be eligible for assessment by WG1.

2013

By March 2013, papers cited by WG1 have to be published or accepted for publication (with proof of acceptance).

The standard protocols of model simulations defined by CMIP5 combine three types of simulations: Decadal hindcasts, for evaluating how realistic the models are in simulating the recent past; forecasts, providing projections of future climate change on two time scales, near term (out to about 2035) and long term (out to 2100 and beyond); and finally, scenarios for model intercomparison in order to understand some of the factors responsible for differences in model projections (e.g., those involving clouds and the carbon cycle). In particular, the decadal hindcasts are core experiments to evaluate the quality of climate models. If a model does not represent today’s climate variables correctly, the model is deemed unable to forecast future states. Thus, decadal hindcasts are the classic test beds for climate model evaluation.

Initializing standardized simulation runs of climate models requires standardized observational data. For instance, the decadal hindcasts of CMIP5 consisted of 10-year hindcasts initialized from climate states in the years 1960, 1965, and 1970. The experiments were based on so-called reference data sets, which were used for standardized numerical experiments in order to generate comparable and reproducible results. Usually, every data product is based on its own assimilation methods, which change over the years. Therefore, reference data sets are reanalysis data “with a ‘frozen’ data assimilation system (one that would not change during the reanalysis)” (Edwards, 2010, p. 324; cf. Parker, 2011).

Data assimilation is required because measurement data exhibit great uncertainties and irregularities due to their inhomogeneous characteristics. The inhomogeneity results from the diversity of measurement platforms and methods (weather stations, buoys, radiosondes, satellites, rockets), their irregular spatial distribution, and their diverse accuracy and error characteristics. On the one hand, observational data have to be placed into a gridded model space, as climate simulations are usually performed on regular and global computing grids. For instance, satellites cover swaths of only several hundred to a few thousand kilometers of the Earth’s surface and produce distorted data at the edges of their focus. Therefore, data from various satellites have to be assimilated and composed in order to gain a global, straightened, regular data set (“making data global,” Edwards, 2010). On the other hand, reanalysis data “improve” measurement data by combining information on the actual state (measurement) with physical laws (model) accounting for observation error as well as model error.

Creating reference data sets is laborious work. In particular, two major reanalysis projects have been carried out: the NCEP/NCAR Reanalysis I and II data sets (RA-I, RA-II) of the U.S. National Centers for Environmental Prediction (NCEP) and the National Center for Atmospheric Research (NCAR), and the ERA-15 and ERA-40 data sets of the European Centre for Medium-Range Weather Forecasts (ECMWF). The NCEP/NCAR Reanalysis I and II data sets cover the period from 1948 to 2002 (Kalnay et al., 1996), while the ERA-40 data cover the period from mid-1957 to mid-2002 (Uppala et al., 2005). However, reference data sets can be afflicted with biases, as “deficiencies in the analysis method or assimilating model could introduce significant biases in the resulting analyses, and could invalidate the conclusions drawn from them” (Uppala et al., 2005, p. 2962). For instance, ERA-15 data sets showed shifts in humidity and temperature resulting from problematic assimilation of satellite data (Trenberth, Stepaniak, Hurrell, & Fiorino, 2001). Because the data sets are widely used by the modeling community, biases propagate into many simulation results. However, neither in situ measurements nor reanalysis data provide a true image of the world’s states. Reanalysis data document that in situ and in silico data are increasingly merging (cf. Feichter, 2011).

Community Models and Platforms

The IPCC Assessment Reports serve as unique documentation of the history of climate models since 1990 (Le Treut et al., 2007). They began with eleven CGMs and AOGCMs from seven countries in 1990 (IPCC, 1990, pp. 81−82), and the recent, fifth report is based on more than 40 AOGCMs and ESMs as well as 15 Earth Models of Intermediate Complexity (EMICs) from a total of 15 countries (IPCC, 2013, p. 747). The global models are supplemented by regional models, integrated assessment models, and special interest models (such as high-resolution cloud models). Modeling centers and university departments are responsible for these models (cf. Easterbrook & Johns, 2009; Krueck & Borchers, 1999; Randall, 1996). When work on climate models started, such as the UCLA Department of Meteorology’s Mintz-Arakawa model, only a few modelers were involved and not many researchers outside a given modeling center worked with the model. In fact, the Mintz-Arakawa model was one of the first distributed models used by the RAND Corporation for numerical studies of climate dynamics (Gates, Batten, Kahle, & Nelson, 1971).

Today, many modelers are collaboratively involved in generating, improving, and maintaining climate models, which have become community models. Thus, in 1983 NCAR began distributing a freely available Community Climate Model (CCM) of the atmosphere to the climate research community (Washington, 1982), followed in 1994 by a Community Climate System Model (CCSM) including models of the atmosphere, land surface, ocean, and sea ice (Kiehl, Hack, & Bonan, 1998) and the Community Earth System Model (CESM) in 2010 (Hurrell, Holland, & Gent, 2013). More than 300 researchers are involved in developing the CESM. The development requires an advanced software design and collaborative tools for community development, such as software repositories, version control systems, and procedures for introducing code into the repositories, coding standards, and testing infrastructure (Drake, Jones, & Carr, 2005). Furthermore, the community contributions have to be evaluated scientifically by the Change Review Board.

Besides community development, community-wide use has been eased by increased traffic performance, allowing online platforms and gateways to be established for downloading models and simulation data. Thus, in addition to modelers, the numbers of model users and simulation data users have increased enormously, involving more and more scientific disciplines, such as biology, agriculture, and economics. Model users apply models from others for their own experiments, while simulation data users analyze the output of the in silico experiments with models, either for scientific purposes or for policy purposes. A new type of data users are climate prediction services, which tailor simulation data and predictions for sector-specific purposes, including agriculture, environmental politics, and industry. Like daily weather forecasts by weather services, regional and seasonal temperature and precipitation projections as well as forecasts of extreme weather changes and unusual seasonal variations have become real-time products of climate prediction services all over the world (Dutton, 2002; National Research Council, 2001). Last but not least, the media and the public use simulation data (Asrar, Hurrel, & Busalacchi, 2013).

Against this backdrop, an important goal of today’s climate science is to provide a fully flexible plug-and-play modeling framework as “the community moves toward componentization and shared utility infrastructures” (Drake et al., 2005, p. 180). Such a framework has been developed by the U.S. Earth System Modeling Framework (ESMF) (Collins et al., 2005) as well as by the British Grid ENabled Integrated Earth System Model (GENIE). The vision is that a registered user can construct a model from a library of components (ocean, atmosphere, ice, etc.; Price et al., 2005). With such plug-and-play frameworks, anybody could, in principle, conduct an earth system experiment. Nevertheless, because there is no single best model or best approach, evaluating which data from which model or method to use is difficult. Furthermore, evaluation of models is a complex task and “it is well known that in many situations climate models obtain the right result for the wrong reason” (de Elía, 2014, p. 1004). Thus, the lack of a dominant model, of confirmation, and of good causality complicates the application of climate models for model users and simulation data users (Aroonruengsawat & Auffhammer, 2011; Landström et al., 2011).

Model Evaluation, Uncertainties, and the Complexity of Climate Models

Sources of Uncertainties

Meteorology is a field paradigmatic for its use of models and simulation in science. It was been among the first of the sciences to take advantage of electronic computers and today it is leading the way in building a conjoint infrastructure of model intercomparison, standardized reference data sets, community models, and data platforms. In the process, meteorology has turned back into an “experimental” science. However, the term experimental refers here to in silico experiments with only mathematical models symbolically representing the atmosphere, since experimenting with the real atmosphere is not feasible. Since mathematical models are purely representational, evaluating these models is anything but trivial.

The quasirealistic models presented in this article are deterministic models based on “first principles” of hydro- and thermodynamic theory. Although this type of model is predestinated for expressing processes, the main problem is that many important physical processes are only partly resolved in the models, either because of lack of knowledge about the processes or because of lack of computing resources to increase spatial resolution. The limit in spatial resolution divides the model in resolved and unresolved processes. For instance, the first IPCC Assessment Report was based on a 500-km grid (T21, corresponding to a 500-km grid). Such a grid does not resolve major cyclones, and small countries are represented by only a few grid points. Therefore, the image presented was extremely coarse for climate projections. The fourth IPCC Assessment Report employed a 110-km grid (T106, corresponding to a 110-km grid) and 10-min time steps, which led to 52,560 iterations to simulate a year of climate development, and more than five million iterations for a century. However, an extremely high spatial resolution (T1279, corresponding to a 16-km grid) is needed in order to simulate the circulation regime structure of the northern hemisphere (Dawson, Palmer, & Corti, 2012). Even today, such a resolution can be computed for only a limited area and short-term predictions, but not for the globe and long-term forecasts. Thus, small-scale (unresolved) processes have to be explicitly parameterized, because they have an important influence on the resolved scale of quasirealistic models. Unfortunately, subscale parameterizations are a major source of uncertainties. In particular, parameterized clouds have become prominent for afflicting climate models with uncertainties, because many cloud processes are not known and those known are often insufficiently parameterized (parameter uncertainties).

Initial and boundary conditions are another source of uncertainties. Weather models extrapolate the current state of the atmosphere into the future; therefore, they are heavily influenced by the initial condition problem of uncertain measurement data and intrinsic nonlinearities, quickly leading to unstable results after three to seven forecasted days (initial condition uncertainty; Lorenz, 1963). Different climate models, which extrapolate averaged states of the atmosphere into the future, thus average out the initial data uncertainties after a while. However, simulation results of climate models are sensitive to the boundary conditions, such as GHG concentrations forecast for the next decades by socioeconomic scenarios. Such scenarios have been developed in the course of the IPCC Assessment Reports (SA90 scenarios, IPCC, 1990, Appendix 1; IS92 scenarios for FAR, Pepper et al., 1992; SRES scenarios, Nakicenovic & Swart, 2000; and RCP scenarios, van Vuuren et al., 2011, and Meinshausen et al., 2011). They cover a wide range of possible changes in future anthropogenic GHG emissions and technological development (boundary condition uncertainty). Since the first scenarios, a “business as usual” scenario and several ecologically friendly stories have been developed, leading to temperature projections by climate models for 2100 that vary between + 0.3°C and + 4.8°C.

There are many other sources of uncertainties for climate and simulation. An uncertainty typology has been developed by Arthur Petersen (2006, 2012) that differentiates the location of uncertainty (models, input data, model implementation, and output interpretation) from the nature of uncertainty (epistemic uncertainty due to incompleteness and fallibility of knowledge, and ontic uncertainty due to the intrinsic character of a natural system) and the range of uncertainty (statistical and scenario uncertainty). The typology yields an awareness of the limits of predictability, about the adequacy or inadequacy of methods, and the value-laden choices of decisions. Nevertheless, climate simulation literature often refers to other types of uncertainties, differentiating structural uncertainty (a mix of Petersen’s epistemic and ontic uncertainty), from parameter uncertainty (including tuning) and observational uncertainty. All the sources of uncertainty contribute to the forecasting uncertainty of models, in general, compelling the IPCC to introduce a “likelihood talk” for its Assessment Reports and to suggest speaking about climate projections, rather than predictions or forecasts (Bray & von Storch, 2009; IPCC, 2005).

Evaluation of Climate Models

Nevertheless, the core topic of climate simulation is the question about the “truth” of a model: How can climate models afflicted with many sources of uncertainties be evaluated? Does a model “truly” grasp the relevant aspects of a real system for which it has been conceived? The questions are particularly important because climate models are purely representational models and because there is no single best model or best approach (cf. Hulme & Dessai, 2008; Lahsen, 2005; Parker, 2009b, 2011, 2014). Against this backdrop, Oreskes, Shrader-Frechette, and Belitz (1994) differentiated main concepts, such as the verification, validation, and confirmation of models. Verification ensures the “truth” of a statement, but this absolute certainty is only deductively achievable for, and within, first-order logical statements and linear systems. With some of the conceptual models, an exact solution can be algebraically deduced for a specific context, but none of the quasirealistic climate models discussed qualifies, so the concept of verification can’t be used. Validation, often used synonymously with verification, is similarly misleading, because the term valid “might be useful for assertions about a generic computer code but is clearly misleading if used to refer to actual model results in any particular realization” (Oreskes et al., 1994, p. 642). Such an assertion can refer to the consistency of a code, designating the correct implementation of the algorithm of a model (Sargent, 2013) or a method. Similarly, the comparison of numerical results with an analytical solution, if it exists, does not say anything about the “truth” of a model, only about the convergence of the discrete approximations to the solution. For complex (nonlinear) models, not even such a comparison is possible, because an analytical solution is missing and therefore is replaced by “semi-empirical” convergence tests. In other words, it checks the stability of numerical results of a simulation run by doubling the resolution of a control run. If both runs behave stably, it is assumed that the discrete approximations converge toward the unknown analytic solution. However, such an assumption is legitimate only for linear problems (Lax & Richtmyer, 1956).

Thus, we can only talk about confirmation of models. The crucial method for confirmation is the quantitative comparison of simulation results with observational data to assess model performance. Complex models like GCMs, AOGCMs, and ESMs are usually evaluated with observational data on the system level as well as on the component and parameter levels (Randall, Khairoutdinov, Arakawa, & Grabowski, 2003). The basic tests of the system level are decadal hindcasts (for instance, for known extreme-value statistics), as well as transient climate simulations (usually 1850 to 2100) to reproduce observed climate change. If simulation results match observational data, it is said that the model is reliable.

According to AR5, “a significant development since the AR4 is the increased use of quantitative statistical measures, referred to as performance metrics,” enabling the assessment of model improvements over time (IPCC, 2013, p. 753). Performance metrics are standardized measures of benchmark experiments established by WGNE and carried out by CMIP, which allow the strengths and weaknesses of a given model to be assessed in comparison with in situ data and other models (Gleckler, Taylor, & Doutriaux, 2008; Mearns, 1997). By comparing in silico and in situ data sets of many variables of atmospheric fields, a detailed analysis of model performance is possible, and comparing the performance of various models yields a model performance index. Such an index of a multimodel ensemble (MME) reveals that the median model outperforms every other model (Pennell & Reichler, 2011; Reichler & Kim, 2008). However, by comparing many climate models, CMIP5 does pose the question of model comparability. This has sparked a debate on model weighting (Weigel, Knutti, Liniger, & Appenzeller, 2010) versus “model democracy” (Knutti, 2010). Averaging model results has led to another ensemble method—perturbed parameter ensembles (PPEs)—which tests single models. Perturbing particular parameters of a model can yield knowledge on the behavior and uncertainty of parameters (Allen, 1999; Lambert et al., 2013; Stainforth et al., 2005). Initial condition ensembles, a variation of PPEs, change the initial conditions of a climate model slightly in order to gain knowledge about the sensitivity of a model and to average the results. Combining both methods, MME and PPE, may eventually help to better assess the structural and parameter uncertainties of models (Sexton, Murphy, Collins, & Webb, 2012), but ensemble methods are computing-intensive (cf. Parker, 2010, 2013).

Problem of Tuning

The “reality check” of climate models on the system, component, and parameter levels is laborious and beset by many problems. First, fulfilling the requirements of a specific set of numerical experiments, which is always purpose-driven, does not imply any general quality. The second problem is model tuning, which is “the idea that models need to be harmonized with observations . . . to improve the representation of some aspect of the climate system” (Mauritsen et al., 2012, p. 1). Parameter fitting and adjusting, also known as model tuning or model calibration, are issues that have been long practiced and discussed in climate modeling, although not many publications highlight this aspect of modeling (Gramelsberger, 2010; Guillemot, 2010; Mauritsen et al., 2012; Petersen, 2011; Randall & Wielicki, 1997). Furthermore, some numerical parameters lack observational data and have to be adjusted (tuned). For instance, flux corrections of the top of the atmosphere (TOA) radiative imbalance have been a well-known example of tuning for many years, although they are barely accepted in the climate modeling community today. Models without flux corrections demonstrate progress in modeling, although such models have been replaced by cloud-related parameter tuning and other methods.

Besides the fictive elements of modeling, it has to be mentioned that a good fit between a model and in situ data does not necessarily make for a “good model” (cf. Edwards, 1999; Heymann, 2012; Lenhard, 2011; Shackley et al., 1998, 1999). In situ data usually do not fit model requirements (resolution, distribution, etc.), and many interactions lack observational methods and data, most prominent among them being climate variability. Furthermore, in situ data exhibit a range of uncertainties; because they are samples of incomplete spatial and temporal coverage, they propagate the limits of instrumental capabilities. Satellite data are particularly model-laden in order to make it possible for the required indirect properties to be translated and derived into useful data products. Thus, the empirical basis contains its own data uncertainties. This has led to a new evaluation method: the inverse treatment of data and models by what is called the instrument simulator. Instead of converting satellite data into “model equivalents,” “observation equivalents” are computed and compared with satellite data by simulating a virtual satellite and its records in a model (Bodas-Salcedo & Webb, 2011). On the component and parameter levels, the particular schemes are tested in isolation against in situ data with box or column models and, afterward, within the whole model (Phillips et al., 2004).

Conclusion

Climate models are afflicted with many uncertainties. They can’t be verified. At best, they can be evaluated only more or less properly. Nevertheless, models are the only tools that can provide insights into complex phenomena and that can extrapolate past and current trends into the future. Therefore, they have become indispensable for science and, in particular, for climate science and society facing climate change. Although quasirealistic circulation models provide the dominant, deterministic view on climate, other views on climate have been developed. Generally speaking, besides the deterministic view, the statistical and probabilistic views can be employed (Lorenz, 1969). Statistical models resample observed records and project them into the future (Slingo, 2013). Stochastic models don’t represent the development of processes that deterministic models aim at, but they try to cover the influence of processes of various scales on each other (Hasselmann, 1976; for an overview, see Franzke, O’Kane, Berner, Williams, & Lucarini, 2014). Conceptually speaking, climate models can be complex and comprehensive, like GCMs, AOGCMs, and ESMs, but they can also be more simple one- or two-dimensional energy balance models (EBMs), conceptual models, or box models (Müller & von Storch, 2004; von Storch, 2010). For instance, EMICs have been developed to bridge the gap between conceptual and comprehensive models (Claussen et al., 2002). EMICs replace some interactions of GCMs by prescribing them as external forcings.

Thus, many modelers advocate a hierarchy of models as a proper strategy for dealing with climate and simulation, because “on the one hand, we try to simulate by capturing as much of the dynamics as we can in comprehensive numerical models. On the other hand, we try to understand by simplifying and capturing the essence of a phenomenon in idealized models, or even with qualitative pictures” (Held, 2005, p. 1609; Henderson-Sellars & McGuffie, 1987, 1999; Schneider & Dickinson, 1974). An increasingly important variety of climate models results from the goal of regional and seasonal climate projections. In 1981, Charney and Shukla proposed seasonal studies on the predictability of monsoons. With regional climate models, interactions can be studied in greater detail—for example, effects of lakes and mountains (high-resolution forcings). However, studies show that finer temporal (days) and spatial (local) scales do not necessarily deliver reliable results (Dickinson, Errico, Giorgi, & Bates, 1989, Giorgi, 1990; McGregor, 1997; Takle, 1995).

Whatever type of model is chosen, it is obvious that models are, in general, purpose driven. Different purposes lead to different models by taking specific views on climate and climate change into account, as integrated assessment models, paleoclimate models, data-driven models, and carbon dioxide models do (cf. Ackerman et al., 2009; Nordhaus & Boyer, 2000; Parker, 2006; Parker et al., 2002).

Although a hierarchy of models is important, it must be asked whether the diversity and variety of model approaches are in accordance with the overall aim of science to produce reliable and comparable results. Besides the variety and diversity of models, the interdependency of the models is obvious. In the case of CGMs, they share not only the same hydro- and thermodynamic equations, but also many common ideas about parameterizations and modeling techniques. In particular, the heredity of GCMs as well as the standardization for model intercomparison are accompanied by a notable unification of models, in addition to model diversity and variety. Le Queré has identified three phases in modeling: the illusion, the chaos, and the relief phase (Le Queré, 2006). During the early phase of illusion, modelers have to tackle the lack of observations in order to evaluate the models properly. Therefore, they came along with many different modeling approaches. This leads to the chaos phase, which base modeling on observational data. Thus, the diversity of approaches exhibits a lack of accordance with observational data. Nevertheless, “the chaos phase is the most creative and beneficial period in the development of models. . . . Modelers are, on the contrary, driven by the possibility of exploring new dimensions, unconstrained even by observations” (Le Queré, 2006, p. 499). Finally, the relief phase leads to good understanding of the underlying concepts and good agreement with measurements, as well as increasing unification of the formerly diverse modeling approaches.

From this course, the interdependency of models results. Although a lot has been written about climate and simulation—in climate science itself, but also in philosophy of science and environmental humanities—the process of modeling, how progress in model-mediated knowledge is gained, and whether model interdependency strengthens or weakens knowledge production have not been sufficiently explored. The explanation for this lack is rooted in the novelty of numerical modeling and simulation, influencing progress of scientific knowledge since the 1970s, but effectively since the 1990s (Barbrousse, Franceschelli, & Imbert, 2009; Dowling, 1999; Giere, 2009; Heymann, 2010; Morrison, 1998; Winsberg, 2001, 2010). However, it is also rooted in the increasingly abstract work of scientists who sit at computers and program models, conduct observational experiments with remote devices, and analyze empirical as well as in silico data stored in the same database. Even field studies have inevitably become dependent on computers and (data) models. Hence, climate and simulation are a good example of how science has been changed by the use of computer-based simulation (Heymann, Gramelsberger, & Mahony, 2017), and the changes are not easily understood by non-modelers (Kitcher, 2010; Oreskes & Conway, 2010; Somerville & Hassol, 2011). As Paul Edwards claimed, there are three types of models that enable research on climate: “simulation models of weather and climate; reanalysis models, which recreate climate history from historical weather data; and data models, used to combine and adjust measurements from many different sources” (Edwards, 2010). Considering this, it becomes clear that climate models are just one side of the coin, and data models are the other side. Both constitute the currency of a simulation-driven science, but in the case of climate science, this is leading to major sociopolitical debates (Edwards, 2001; Hulme, 2009, 2011; Kuik et al., 2008; Otto, Frame, Otto, & Allen, 2015; Pielke, 2007; Polley, 2010; von Storch, 2009).

Acknowledgments

The author thanks Martin Mahony, Matthias Heymann, and Johann Feichter for helpful comments and many discussions on this topic during past years. Thanks, also, to the reviewers for their beneficial comments.

References

Ackerman, F., DeCanio, S. J., Howarth, R. B., & Sheeran, K. (2009). Limitations of integrated assessment models of climate change. Climatic Change, 95, 297–315.Find this resource:

Agrawala, S. (1998a). Context and early origins of the Intergovernmental Panel on Climate Change. Climatic Change, 39, 605–620.Find this resource:

Agrawala, S. (1998b). Structural and process history of the Intergovernmental Panel on Climate Change. Climatic Change, 39(4), 621–642.Find this resource:

Agrawala, S. (1999). Early science–policy interactions in climate change: Lessons from the Advisory Group on Greenhouse Gases. Global Environmental Change, 9(2), 157–169.Find this resource:

Allan, D. R. (2001). The genesis of meteorology at the University of Chicago. Bulletin of the American Meteorological Society, 82(9), 1905–1909.Find this resource:

Allen, M. R. (1999). Do it yourself climate prediction. Nature, 401, 642.Find this resource:

Andersen, S. O., & Sarma, K. M. (2002). Protecting the ozone layer: the United Nations history. London, UK: Earthscan Press.Find this resource:

Anderson, K. (2005). Predicting the weather: Victorians and the science of meteorology. Chicago, IL: University of Chicago Press.Find this resource:

Arakawa, A. (2000). A personal perspective on the early years of General Circulation Modeling at UCLA. In D. A. Randall (Ed.), General Circulation Model development (pp. 1–65). San Diego, CA: Academic Press.Find this resource:

Archer, D., & Pierrehumbert, R. (2011). The warming papers. Chichester, UK: John Wiley & Sons.Find this resource:

Aroonruengsawat, A., & Auffhammer, M. (2011). Impacts of climate change on residential electricity consumption: Evidence from billing data. In G. Libecap & R. H. Steckel (Eds.), The economics of climate change: Adaptations past and present (pp. 311–343). Chicago, IL: University of Chicago Press.Find this resource:

Asrar, G. R., Hurrel, J. W., & Busalacchi, A. J. (2013). A need for “actionable” climate science and information. Bulletin of the American Meteorological Society, 94, ES8–ES12.Find this resource:

Aykut, S., & Dahan, A. (2015). Gouverner le climat? Vingt années de négociations internationales. Paris, France: Presses de Sciences Po.Find this resource:

Barbrousse, A., Franceschelli, S., & Imbert, C. (2009). Computer simulations as experiments. Synthese, 169(3), 557–574.Find this resource:

Barron, E. J. (1992). A decade of international climate research: The first ten years of World Climate Research Program. Washington, DC: National Academies Press.Find this resource:

Beck, S. (2009). Das Klimaexperiment und der IPCC. Marburg, Germany: Metropolis.Find this resource:

Beyers, H. B. (1960). Carl-Gustav Avrid Rossby, 1898–1957. A biographical memoir. Washington, DC: National Academy of Science.Find this resource:

Bjerknes, J., & Mintz, Y. (1955). Investigation of the general circulation. Final Report to the Air Force, Project AF 19(122)-48, Department of Meteorology. Los Angeles: University of California Press.Find this resource:

Bjerknes, V. (1898). Ueber einen hydrodynamischen Fundamentalsatz und seine Anwendung besonders auf die Mechanik der Atmosphäre und des Weltmeeres. Kungliga Svenska vetenskapsakademiens handlingar Stockholm, 31, 1–35.Find this resource:

Bjerknes, V. (1904). Das Problem der Wettervorhersage, betrachtet von Standpunkt der Mechanik und Physik. Meteorologische Zeitschrift, 21(1), 1–7. (English translation in Bjerknes, V. (2009). The problem of weather prediction, considered from the viewpoints of mechanics and physics. Meteorologische Zeitschrift, 18(6), 663–667.)Find this resource:

Bjerknes, V. (1919). The structure of the atmosphere when rain is falling. Quarterly Journal of the Royal Meteorological Society, 46, 119–140.Find this resource:

Bjerknes, V., & Sandström, J. W. (1910, 1911). Dynamic meteorology and hydrography (Vol. 1 and 2). Washington, DC: Carnegie Institution of Washington.Find this resource:

Black, M. (1962). Models and metaphors: Studies in language and philosophy. Ithaca, NY: Cornell University Press.Find this resource:

Blinova, E. N., & Kibel, I. A. (1957). Hydrodynamical methods of the short- and long-range weather forecasting in the USSR. Tellus, 9, 447–463.Find this resource:

Bodansky, D. M. (1995). The emerging climate change regime. Annual Review of Energy and the Environment, 20(1), 425–461.Find this resource:

Bodas-Salcedo, A., & Webb, M. J. (2011). COSP: Satellite simulation software for model assessment. Bulletin of the American Meteorological Society, 92, 1023–1043.Find this resource:

Boehmer-Christiansen, S. (1994). Global climate protection policy: The limits of scientific advice: Part 1. Global Environmental Change, 4(2), 140–159.Find this resource:

Bolin, B. (2007). A history of science and politics of climate change. New York, NY: Cambridge University Press.Find this resource:

Bray, D., & von Storch, H. (2009). “Prediction” or “Projection”? The nomenclature of climate science. Science Communication, 30, 534–543.Find this resource:

Bray, D., & von Storch, H. (2015). The normative orientations of climate scientists. Science and Engineering Ethics, 23(5), 1–17.Find this resource:

Buys-Ballot, C. (1854). Erläuterung einer graphischen Methode zur gleichzeitigen Darstellung der Witterungserscheinungen an vielen Orten, und Aufforderung der Beobachter das Sammeln der Beobachtungen an vielen Orten zu erleichtern. Poggendorfs Annalen, 4, 559–576.Find this resource:

Buys-Ballot, C. (1857). Note sur le rapport de l’intensité et de la direction du vent avec les écarts simultanés du barometer. Comptes rendus hebdomadaires, 45, 765–768.Find this resource:

Charney, J. G., Arakaw, A., Baker, D. J., Bolin, B., Dickinson, R. E., Goody, R. M., . . . Wunsch, C. I. (1979). Carbon dioxide and climate: A scientific assessment. Washington, DC: National Academy of Sciences Press.Find this resource:

Charney, J. G., Fjørtoft, R., & von Neumann, J. (1950). Numerical integration of the barotropic vorticity equation. Tellus, 2(4), 237–254.Find this resource:

Charney, J. G., & Phillips, N. (1953). Numerical integration of the quasi-geostrophic equations of motions for barotropic and simple baroclinic flows. Journal of Meteorology, 10, 71–99.Find this resource:

Charney, J. G., & Shukla, J. (1981). Predictability of monsoons. In J. Lighthill & R. P. Pearce (Eds.), Monsoon dynamics (pp. 99–109). Cambridge, UK: Cambridge University Press.Find this resource:

Claussen, M., Mysak, L., Weaver, A., Crucifix, M., Fichefet, T., Loutre, M.-F., . . . Wang, Z. (2002). Earth System Models of intermediate complexity: Closing the gap in the spectrum of climate system models. Climate Dynamics, 18, 579–586.Find this resource:

Coupled Model Intercomparison Project (CMIP). (2015). Homepage.

Collins, N., Theurich, G., DeLuca, C., Suarez, M., Trayanov, A., Balaji, V., . . . da Silva, A.(2005). Design and implementation of components in the Earth System Modeling Framework. International Journal of High Performance Computing Applications, 19, 341–350.Find this resource:

Cosgrove, D. (2001). Apollo’s eye: A cartographic genealogy of the Earth in the Western imagination. Baltimore, MD: Johns Hopkins University Press.Find this resource:

Dahan, A. (2010). Putting the Earth System in a numerical box? The evolution from climate modeling toward global change. Studies in History and Philosophy of Modern Physics, 41(3), 282–292.Find this resource:

Dahan, A., & Aykut, S. (2013). After Copenhagen, revisiting both the scientific and political framings of the climate change regime. In J. -B. Saulnier & M. Varella (Eds.), Global change, energy issues and regulation policies (pp. 221–237). Berlin, Germany: Springer Science & Business Media.Find this resource:

Darrigol, O. (2005). Worlds of flow: A history of hydrodynamics from the Bernoullis to Prandtl. Oxford, UK: Oxford University Press.Find this resource:

Davis, W. M. (1887). Book review: Advances in meteorology. Science, 9(226), 539–541.Find this resource:

Dawson, A., Palmer, T. N., & Corti, S. (2012). Simulating regimes structures in weather and climate prediction models. Geophysical Research Letters, 39, L21805.Find this resource:

Dear, P. (1995). Disciplines & experience: The mathematical way in the scientific revolution. Chicago, IL: University of Chicago Press.Find this resource:

De Elía, R. (2014). Specificities of climate modeling research and the challenges in communicating to users. Bulletin of the American Meteorological Society, 95, 1003–1010.Find this resource:

Demeritt, D. (2001). The construction of global warming and the politics of science. Annals of the Association of American Geographers, 91(2), 307–337.Find this resource:

Depledge, J. (2005). The organization of global negotiations: Constructing the climate change regime. London, UK: Earthscan.Find this resource:

Dickinson, R. E., Errico, R. M., Giorgi, F., & Bates, G. T. (1989). A regional climate model for the western United States. Climate Change, 15, 383–422.Find this resource:

Donner, L., Schubert, W., & Somerville, R. (Eds.). (2011). The development of atmospheric general circulation models: Complexity, synthesis and computation. New York, NY: Cambridge University Press.Find this resource:

Dove, H. W. (1837). Meteorologische Untersuchungen. Berlin, Germany: Sander.Find this resource:

Dowling, D. C. (1999). Experimenting on theories. Science in Context, 12, 261–274.Find this resource:

Drake, J. B., Jones, P. W., & Carr, G. R. (2005). Overview of the software design of the community climate system model. International Journal of High Performance Computing Applications, 19, 177–186.Find this resource:

Dutton, J. A. (2002). Opportunities and priorities in a new era for weather and climate services. Bulletin of the American Meteorological Society, 83, 1303–1311.Find this resource:

Easterbrook, S. M., & Johns, T. C. (2009). Engineering the software for understanding climate change. Computing in Science & Engineering, 11, 64–74.Find this resource:

Eckert, M. (2006). The dawn of fluid dynamics: A discipline between science and technology. Weinheim, Germany: Wiley-VCH.Find this resource:

Edwards, P. N. (1999). Global climate science, uncertainty and politics: Data‐laden models, model‐filtered data. Science as Culture, 8(4), 437–472.Find this resource:

Edwards, P. N. (2000). A brief history of atmospheric general circulation modeling. In D. A. Randall (Ed.), General circulation model development (pp. 67–90). San Diego, CA: Academic Press.Find this resource:

Edwards, P. N. (Ed.). (2001). Changing the atmosphere: Expert knowledge and environmental governance. Cambridge, MA: MIT Press.Find this resource:

Edwards, P. N. (2010). A vast machine: Computer models, climate data, and the politics of global warming. Cambridge, MA: MIT Press.Find this resource:

Ekholm, N. (1904). Wetterkarten der Luftdruckschwankungen. Meteorologische Zeitschrift, 21(8), 345–357.Find this resource:

Elzinga, A., & Landström, C. (Eds.). (1996). Internationalism and science. London, UK: Taylor Graham.Find this resource:

Ertel, H. (1942a). Ein neuer hydrodynamischer Wirbelsatz. Meteorologische Zeitschrift, 59(2), 277–281.Find this resource:

Ertel, H. (1942b). Über das Verhältnis des neuen hydrodynamischen Wirbelsatzes zum Zirkulationstheorem von V. Bjerknes. Meteorologische Zeitschrift, 59(3), 385–387.Find this resource:

Ertel, H., & Rossby, C.-G. (1949). A new conservation-theorem of hydrodynamics. Geofisica Pura e Applicata, 16, 189–195.Find this resource:

Euler, L. (1755). Principes généraux du mouvement des fluids. In L. Euler (1954), Opera omnia, I. Opera mathematica, II (Vol. 12, pp. 54–91). Basel, Switzerland: Birkhäuser.Find this resource:

Exner, F. M. (1902). Versuch einer Berechnung der Luftdruckänderung von einem Tage zum nächsten. Sitzungsberichte der Akademie der Wissenschaften zu Wien, 111, 704–725.Find this resource:

Exner, F. M. (1917). Dynamische Meteorologie. Leipzig, Germany: Teubner.Find this resource:

Feichter, J. (2011). Shaping reality with algorithms: The Earth System. In G. Gramelsberger (Ed.), From science to computational sciences (pp. 209−217). Berlin, Germany: diaphanes.Find this resource:

Ferrel, W. (1856). An essay on the winds and the currents of the ocean. Nashville Journal of Medicine and Surgery, 11, 287–301.Find this resource:

Ferrel, W. (1858). The influence of the Earth’s rotation upon the relative motion of bodies near its surface. Astronomical Journal, 5(109), 97–100.Find this resource:

Ferrel, W. (1877). Meteorological researches for the use of the coast pilot. Washington, DC: U.S. Coast Service.Find this resource:

Ferrel, W. (1886). Recent advances in meteorology: Annual Report of the Chief Signal Officer. Washington, DC: U.S. War Department.Find this resource:

Fjörtoft, R. (1952). On a numerical method of integrating the barotropic vorticity equation. Tellus, 4, 179–194.Find this resource:

Fleming, J. R. (2000). Meteorology in America, 1800–1870. Baltimore, MD: Johns Hopkins University Press.Find this resource:

Fleming, J. R. (2002). History of meteorology. In B. S. Biagre (Ed.), A History of Modern Science and Mathematics (Vol. 3, pp. 184−217). New York, NY: Scribner’s.Find this resource:

Fortak, H. (2001). Felix Maria Exner und die österreichische Schule der Meteorologie. In C. Hammerl, W. Lenhardt, R. Steinacker, & P. Steinhauser (Eds.), Die Zentralanstalt für Meteorologie und Geodynamik 1851–2001. 150 Jahre Meteorologie und Geophysik in Österreich, 1851–2001 (pp. 354–386). Graz, Austria: Leykam.Find this resource:

Fortak, H. (2004). Hans Ertel’s life and his scientific work. Meteorologische Zeitschrift, 13(6), 453–464.Find this resource:

Franz, W. E. (1997). The development of an international agenda for climate change: Connecting science to policy. Interim Report IR-97-034/August 1997. Laxenburg, Austria: International Institute for Applied Systems Analysis.Find this resource:

Franzke, C. L. E., O’Kane, T. J., Berner, J., Williams, P. D., & Lucarini, V. (2014). Stochastic climate theory and modelling. arXiv 1409.0423v1 [physics.ao-ph].Find this resource:

Friedman, M. (1989). Appropriating the weather: Vilhelm Bjerknes and the construction of a modern meteorology. Ithaca, NY: Cornell University Press.Find this resource:

Frigg, R., & Hartmann, S. (2012). Models in science. In E. N. Zalta (Ed.), Stanford encyclopedia of philosophy (Summer 2018 Edition).Find this resource:

Gates, L. W., Batten, E. S., Kahle, A. B., & Nelson, A. B. (1971). A documentation of the Mintz-Arakawa two-level atmospheric general circulation model, R-877-ARPA Report. Santa Monica, CA: RAND Corporation.Find this resource:

Grid ENabled Integrated Earth System Model (GENIE). (2010). Visions page.

Giere, R. (1999). Science without laws. Chicago, IL: University of Chicago Press.Find this resource:

Giere, R. N. (2009). Is computer simulation changing the face of experimentation? Philosophical Studies, 143, 59–62.Find this resource:

Giorgi, F. (1990). On the simulation of regional climate using a limited area model nested in a general circulation model. Journal of Climate, 3, 941–963.Find this resource:

GISTEMP. (2016). GISS Surface Temperature Analysis (GISTEMP). NASA Goddard Institute for Space Studies.

Gleckler, P., Taylor, K., & Doutriaux, C. (2008). Performance metrics for climate models. Journal of Geophysical Research: Atmospheres, 113, D06104.Find this resource:

Goldstine, H. (1972). The computer: From Pascal to von Neumann. Princeton, NJ: Princeton University Press.Find this resource:

Gramelsberger, G. (2009). Conceiving meteorology as the exact science of the atmosphere —Vilhelm Bjerknes’ revolutionary paper of 1904. Meteorologische Zeitschrift, 18(6), 663–667.Find this resource:

Gramelsberger, G. (2010). Conceiving processes in atmospheric models—General equations, subscale parameterizations, and ‘superparameterizations’. Studies in History and Philosophy of Modern Physics, 41(3), 233–241.Find this resource:

Gramelsberger, G., & Feichter, H. (Eds.). (2011). Climate change and policy: The calculability of climate change and the challenge of uncertainty. Berlin, Germany: Springer.Find this resource:

Grover, V. I. (Ed.). (2008). Global warming and climate change: Ten years after Kyoto and still counting. Enfield, NH: Science Publishers.Find this resource:

Guillemot, H. (2010). Connections between climate simulations and observation in climate computer modeling: Scientist’s practices and bottom-up epistemology lessons. Studies in History and Philosophy of Modern Physics, 41(3), 242–252.Find this resource:

Guillemot, H. (2011). Bref historique de la modélisation du climat en France. In C. Jeandel, & R. Mosseri (Eds.), Le Climat à découvert (pp. 148–149). Paris, France: CNRS Éditions.Find this resource:

Guilyardi, E. (2013). Documenting Climate Models and Their Simulations. Bulletin of the American Meteorological Society, 94(5), 623–627.Find this resource:

Gupta, J. (2014). The history of global climate governance. Cambridge, UK: Cambridge University Press.Find this resource:

Hadley, G. (1735). The cause of the general trade-wind. Philosophical Transactions of the Royal Society of London, 29, 58–62.Find this resource:

Halfmann, J., & Schuetzenmeister, F. (Eds.). (2009). Organisationen der Forschung: der Fall der Klimatologie. Wiesbaden, Germany: VS Verlag fuer Sozialwissenschaften.Find this resource:

Halley, E. (1686/1687a). A discourse of the rule of the decrease of the height of the mercury in the barometer. Philosophical Transactions of the Royal Society of London, 16, 104–116.Find this resource:

Halley, E. (1686/1687b). An historical account of the trade winds, and monsoons, observable in the seas between and near the tropicks, with an attempt to assign the phisical cause of the said wind. Philosophical Transactions of the Royal Society of London, 16, 153–168.Find this resource:

Hare, W., Stockwell, C., Flachsland, C., & Oberthür, S. (2010). The architecture of the global climate regime: A top-down perspective. Climate Policy, 10, 600–614.Find this resource:

Harper, K. (2008). Weather by the numbers: The genesis of modern meteorology. Cambridge, MA: MIT Press.Find this resource:

Hasselmann, K. (1976). Stochastic climate models: Part I. Theory. Tellus, 28, 473–484.Find this resource:

Held, I. M. (2005). The gap between simulation and understanding in climate modelling. Bulletin of the American Meteorological Society, 86, 1609–1614.Find this resource:

Hellmann, G. (1908). The dawn of meteorology. Quarterly Journal of the Royal Meteorological Society, 34(148), 221–234.Find this resource:

Henderson-Sellars, A., & McGuffie, K. (1987). A climate modelling primer. Chichester, UK: John Wiley & Sons.Find this resource:

Henderson-Sellars, A., & McGuffie, K. (1999). Concepts of good science in climate change modelling. Climatic Change, 42, 597–610.Find this resource:

Hesse, M. (1963). Models and analogies in science. London, UK: Sheed & Ward.Find this resource:

Heymann, M. (2010). The evolution of climate ideas and knowledge. WIREs: Climate Change, 1(4), 581–597.Find this resource:

Heymann, M. (2010). Understanding and misunderstanding computer simulation: The case of atmospheric and climate science—An introduction. Studies in History and Philosophy of Modern Physics, 41(3), 193–200.Find this resource:

Heymann, M. (2012). Constructing evidence and trust: How did climate scientists’ confidence in their models and simulations emerge? In H. Hastrup & M. Skrydstrup (Eds.), The social life of climate change models: Anticipating nature (pp. 203–224). London, UK: Routledge.Find this resource:

Heymann, M., Gramelsberger, G., & Mahony, M. (Eds.). (2017). Cultures of prediction: Epistemic and cultural shifts in computer-based environmental science. London, UK: Routledge.Find this resource:

Howe, J. P. (2014). Behind the curve: Science and the politics of global warming. Seattle: University of Washington Press.Find this resource:

Hulme, M. (2009). Why we disagree about climate change: Understanding controversy, inaction and opportunity. Cambridge, UK: Cambridge University Press.Find this resource:

Hulme, M. (2010). Problems with making and governing global kinds of knowledge. Global Environmental Change, 20(4), 558–564.Find this resource:

Hulme, M. (2011). Reducing the future to climate: A story of climate determinism and reductionism. Osiris, 26(1), 245–266.Find this resource:

Hulme, M., & Dessai, S. (2008). Predicting, deciding, learning: Can one evaluate the “success” of national climate scenarios? Environmental Research Letters, 3, 045013.Find this resource:

Hulme, M., & Mahony, M. (2010). Climate change: What do we know about the IPCC? Progress in Physical Geography, 34(5), 705–718.Find this resource:

Hurrell, J. W., Holland, M. M., & Gent, P. R. (2013). The Community Earth System Model: A framework for collaborative research. Bulletin of the American Meteorological Society, 94, 1339–1360.Find this resource:

Intergovernmental Panel on Climate Change (IPCC). (1990). Climate change: The IPCC Scientific Assessment (First Assessment Report [FAR]). In J. T. Houghton, G. J. Jenkins, & J. J. Ephraums (Eds.), Report prepared for IPCC by Working Group 1. Cambridge, UK: Cambridge University Press.Find this resource:

Intergovernmental Panel on Climate Change (IPCC). (2005). Guidance notes for lead authors of the IPCC Fourth Assessment Report on Addressing Uncertainties. Geneva, Switzerland: Intergovernmental Panel on Climate Change.Find this resource:

Intergovernmental Panel on Climate Change (IPCC). (2013). Climate change 2013: The physical science basis (Fifth Assessment Report [AR5]). In T. F. Stocker, D. Qin, G.-K. Plattner, Tignor, M. M. B., Allan, S. K., Boschung, J., . . . Midgley, P. M. (Eds.), Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge, UK: Cambridge University Press.Find this resource:

Jasanoff, S. (2001). Image and imagination: The formation of global environmental consciousness. In C. Miller & P. N. Edwards (Eds.), Changing the atmosphere: Expert knowledge and environmental governance (pp. 309–338). Cambridge, MA: MIT Press.Find this resource:

Jasanoff, S. (2011). Cosmopolitan knowledge: Climate science and global civic epistemology. In J. Dryzek, R. B. Norgaard, & D. Schlosberg (Eds.), Oxford handbook of climate change and society (pp. 129–143). Oxford, UK: Oxford University Press.Find this resource:

Jasanoff, S., & Martello, M. L. (2004). Earthly politics: Local and global in environmental governance. Cambridge, MA: MIT Press.Find this resource:

Johnson, D. R., & Arakawa, A. (1996). On the scientific contributions and insight of Professor Yale Mintz. Journal of Climate, 9, 3211–3224.Find this resource:

Kasahara, A., & Washington, W. M. (1967). NCAR global general circulation model of the atmosphere. Monthly Weather Review, 95, 389–402.Find this resource:

Kalnay, E., Kanamitsu, M., Kistler, R., Collins, W., Deaven, D., Gandin, L., . . . Joseph, D.(1996). The NCEP/NCAR 40-year reanalysis project. Bulletin of the American Meteorological Society, 77, 437–470.Find this resource:

Keeling, C. D. (1958). The concentration and isotopic abundances of atmospheric carbon dioxide in rural areas. Geochimica et Cosmochimica Acta, 13, 322–334.Find this resource:

Keeling, C. D. (1978). The influence of Mauna Loa Observatory on the development of atmospheric CO2 research. In J. Mill (Ed.), Mauna Loa Observatory: A 20th anniversary report (pp. 36–54). Boulder, CO: NOAA Environmental Research Laboratories.Find this resource:

Kiehl, J. T., Hack, J. J., & Bonan, G. B. (1998). The National Center for Atmospheric Research Community Climate Model: CCM3. Journal of Climate, 11, 1131–1149.Find this resource:

Kitcher, P. (2010). The climate change debates. Science, 328(5983), 1231–1232.Find this resource:

Knutti, R. (2010). The end of model democracy? An editorial comment. Climatic Change, 102, 395–404.Find this resource:

Krueck, C. P., & Borchers, J. (1999). Science in politics: A comparison of climate modeling centres. Minerva, 37(2), 105–123.Find this resource:

Kuik, O., Aerts, J., Berkhout, F., Biermann, F., Bruggink, J., Gupta, J., & Tol, R. S. J. (2008). Post-2012 climate policy dilemmas: A review of proposals. Climate Policy, 8, 317–336.Find this resource:

Kwa, C. (1994). Modelling technologies of control. Science as Culture, 4(20), 363–391.Find this resource:

Lahsen, M. (2005). Seductive simulations? Uncertainty distribution around climate models. Social Studies of Science, 35, 895–922.Find this resource:

Lambert, F. H., Harris, G. R., Collins, M., Murphy, J. M., Sexton, D. M. H., & Booth, B. B. B. (2013). Interactions between perturbations to different Earth system components simulated by a fully-coupled climate model. Climate Dynamics, 41(11), 3055–3072.Find this resource:

Landström, C., Whatmore, S. J., Lane, S. N., Odoni, N. A., Ward, N., & Bradley, S. (2011). Co-producing flood risk knowledge: Redistributing expertise in critical “participatory modelling”. Environment and Planning A, 43(7), 1617–1633.Find this resource:

Laplace, P. S. (1775). Recherches sur plusieurs points de systeme du monde. Mémoires de l’Académie royale des sciences de Paris, 88, 71–183.Find this resource:

Laplace, P. S. (1776). Recherches sur plusieurs points de systeme du monde (suite). Mémoires de l’Académie royale des sciences de Paris, 89, 187–280, 283–310.Find this resource:

Laprise, R., Lin, C. A., & Robert, A. R. (Eds.). (1997). Numerical methods in atmospheric and oceanic modelling: The André J. Robert memorial volume. Ottawa, Canada: NRC Research Press.Find this resource:

Lax, P. D., & Richtmyer, R. D. (1956). Survey of the stability of linear finite difference equations. Communications on Pure and Applied Mathematics, 9, 267–293.Find this resource:

Leith, C. E. (1964). Numerical simulation of the earth’s atmosphere. Rep. W-7405-eng-48, Lawrence Radiation Laboratories. Livermore: University of California.Find this resource:

Lenhard, J. (2011). Artificial, false, and performing well. In G. Gramelsberger (Ed.), From science to computational sciences (pp. 165–176). Berlin, Germany: diaphanes.Find this resource:

Le Treut, H., Somerville, R., Cubasch, U., Ding, Y., Mauritzen, C., Mokssit, A., Peterson, T., & Prather, M. (2007). Historical overview of climate change. In Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K. B., Tignor, M., & Miller, H. L. (Eds.) Climate change 2007: The physical science basis—Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change (pp. 93–127). Cambridge, UK: Cambridge University Press.Find this resource:

Lewis, J. M. (2000). Clarifying the dynamics of the general circulation: Phillip’s experiment. In D. A. Randall (Ed.), General circulation model development (pp. 91–164). San Diego, CA: Academic Press.Find this resource:

Le Queré, C. (2006). The unknown and the uncertain in Earth System Modeling. EOS, 87(45), 496–497.Find this resource:

Lorenz, E. N. (1963). Deterministic nonperiodic flow. Journal of the Atmospheric Sciences, 20(2), 130–141.Find this resource:

Lorenz, E. N. (1969). Three approaches to atmospheric predictability. Bulletin of the American Meteorological Society, 50, 345–351.Find this resource:

Lynch, P. (1999). Richardson’s marvelous forecast. In M. A. Shapiro & S. Gronas (Eds.), The life cycles of extratropical cyclones (pp. 61–73). Boston, MA: American Meteorological Society.Find this resource:

Magnani, L., & Nersessian, N. (Eds.). (2002). Model-based reasoning: Science, technology, values. Dordrecht, the Netherlands: Kluwer.Find this resource:

Mahony, M., & Hulme, M. (2016). Modeling and the nation: Institutionalising climate prediction in the UK, 1988−92. Minerva, 54(4), 445–470.Find this resource:

Manabe, S., & Bryan, K. (1969). Climate calculations with a combined ocean-atmosphere model. Journal of the Atmospheric Sciences, 26, 786–789.Find this resource:

Manabe, S., Smagorinsky, J., & Strickler, R. F. (1965). Simulated climatology of a general circulation model with a hydrological cycle. Monthly Weather Review, 93, 769–798.Find this resource:

Manabe, S., & Wetherald, R. T. (1967). Thermal equilibrium of the atmosphere with a given distribution of relative humidity. Journal of the Atmospheric Sciences, 24, 241–259.Find this resource:

Marchuk, G. I. (1974). Numerical methods in weather prediction. New York, NY: Academic Press.Find this resource:

Margules, M. (1890). Über die Schwingungen periodisch erwärmter Luft. Sitzungsberichte der Kaiserlichen Akademie der Wissenschaften zu Wien, mathematisch-naturwissenschaftliche Klasse IIa, 99, 204–277.Find this resource:

Margules, M. (1901). Über den Arbeitswert einer Luftdruckvertheilung und über die Erhaltung der Druckunterschiede. Denkschrift der Kaiserlichen Akademie der Wissenschaften zu Wien, 73, 329–345.Find this resource:

Mauritsen, T., Stevens, B., Roeckner, E., Traute Crueger, T., Esch, M., Giorgetta, M., . . . Tomassini, L. (2012). Tuning the climate of a global model. JAMES, 4(3).Find this resource:

Maury, M. F. (1855). The physical geography of the sea. New York, NY: Harper and Brothers.Find this resource:

McGregor, J. J. (1997). Regional climate modelling. Meteorology and Atmospheric Physics, 63, 105–117.Find this resource:

Mearns, L. O. (1997). On the use of statistics to evaluate climate model experiments. Climatic Change, 37(3), 443–448.Find this resource:

Meehl, G. A., Covey, C., McAvaney, B., Ltif, M., & Stouffer, R. J. (2005). Meeting summaries: Overview of the coupled model intercomparison. Bulletin of the American Meteorological Society, 86(1), 89–93.Find this resource:

Meehl, G. A. (2014). Decadal climate prediction: An update from the trenches. Bulletin of the American Meteorological Society, 95(2), 243–267.Find this resource:

Meinshausen, M., Meinshausen, N., Hare, W., Raper, S. C. B., Frieler, K., Knutti, R., Frame, D. J., & Allen, M. R. (2009). Greenhouse gas emission targets for limiting global warming to 2°C. Nature, 458, 1158–1162.Find this resource:

Meinshausen, M., Smith, S. J., Calvin, K., Daniel, J. S., Kainuma, M. L. T., Lamarque, J. -F., . . . van Vuuren, D. P. P. (2011). The RCP greenhouse gas concentrations and their extensions from 1765 to 2300. Climatic Change, 109(1–2), 213–241.Find this resource:

Messinger, F., & Arakawa, A. (1976). Numerical methods used in atmospheric models. Geneva, Switzerland: Global Atmospheric Research Program, World Meteorological Organization.Find this resource:

Miller, C. A., & Edwards, P. N. (2001). Changing the atmosphere: Expert knowledge and environmental governance. Cambridge, MA: MIT Press.Find this resource:

Mintz, Y. (1955). On heating models for extended- and long-range numerical weather forecasting and the study of the general circulation, Scientific Report II, Project AF 19(604)-1286, Department of Meteorology. Los Angeles: University of California Press.Find this resource:

Mintz, Y. (1958). Design of some numerical general circulation experiments. Bulletin of the Research Council of Israel, 76, 67–114Find this resource:

Mintz, Y. (1965). Very long-term global integration of the primitive equations of atmospheric motion: An experiment in climate simulation. WMO Technical Notes, 6, 141–167.Find this resource:

Möller, F. (1963). On the influence of changes in the CO2 concentration in air on the radiance balance of the earth’s surface and on the climate. Journal of Geophysical Research, 68(13), 3877–3886.Find this resource:

Morgan, M., & Morrison, M. (1999). Models as mediators: Perspectives on natural and social science. Cambridge, UK: Cambridge University Press.Find this resource:

Morrison, M. (1998). Modelling nature: Between physics and the physical world. Philosophia Naturalis, 35, 65–85.Find this resource:

Müller, P., & von Storch, H. (2004). Computer modelling in atmospheric and oceanic sciences. Berlin, Germany: Springer Science & Business Media.Find this resource:

Murphy, A. (1998). The early history of probability forecasts: Some extensions and clarifications. Weather and Forecasting, 13(3), 5–15.Find this resource:

Nakicenovic, N., & Swart, R. (Eds.). (2000). Emissions scenarios: A special report of Working Group III of the Intergovernmental Panel on Climate Change. Cambridge, UK: Cambridge University Press.Find this resource:

National Academy of Sciences. (1966). The feasibility of a global observation and analysis experiment: A report of the Panel on International Meteorological Cooperation to the Committee on Atmospheric Sciences, National Academy of Sciences, National Research Council. Washington, DC: National Academy Press.Find this resource:

National Research Council. (2001). A climate services vision: First steps toward the future. Board on Atmospheric Sciences and Climate, National Research Council, Division on Earth and Life Studies. Washington, DC: National Academy Press.Find this resource:

Nebeker, F. (1995). Calculating the weather: Meteorology in the 20th century. San Diego, CA: Academic Press.Find this resource:

Névir P. (2004). Ertel’s vorticity theorems, the particle relabelling symmetry and the energy-vorticity-theory of fluid mechanics. Meteorologische Zeitschrift, 13, 485–498.Find this resource:

Newton, I. (1687). Philosophiae Naturalis Principia Mathematica. London, UK: Royal Society.Find this resource:

Nordhaus, W. D., & Boyer, J. (2000). Warming the world: Models of global warming. Cambridge, MA: MIT Press.Find this resource:

Oreskes, N., & Conway, E. M. (2010). Merchants of doubt: How a handful of scientists obscured the truth on issues from tobacco smoke to global warming. New York, NY: Bloomsbury Press.Find this resource:

Oreskes, N., & Conway, E. M. (2014). The collapse of Western civilization: A view from the future. New York, NY: Columbia University Press.Find this resource:

Oreskes, N., Shrader-Frechette, K., & Belitz, K. (1994). Verification, validation, and confirmation of numerical models in the earth sciences. Science, 263(5147), 641–646.Find this resource:

Otto, F. E. L., Frame, D. J., Otto, A., & Allen, M. R. (2015). Embracing uncertainty in climate change policy. Nature Climate Change, 5, 917–920.Find this resource:

Parker, P., Letcherb, R., Jakeman, A., Beck, M. B., Harris, G., Argent, R. M., . . . Bin, S. (2002). Progress in integrated assessment and modelling. Environmental Modelling and Software, 17, 209–217.Find this resource:

Parker, W. S. (2006). Understanding pluralism in climate modeling. Foundations of Science, 11(4), 349–368.Find this resource:

Parker, W. S. (2009a). Does matter really matter? Computer simulations, experiments and materiality. Synthese, 169, 483–496.Find this resource:

Parker, W. S. (2009b). Confirmation and adequacy-for-purpose in climate modelling. Aristotelian Society Supplementary, 83(1), 233–249.Find this resource:

Parker, W. S. (2010). Predicting weather and climate: Uncertainty, ensembles and probability. Studies in History and Philosophy of Modern Physics, 41, 263–272.Find this resource:

Parker, W. S. (2011). When climate models agree: The significance of robust model predictions. Philosophy of Science, 78(4), 579–600.Find this resource:

Parker, W. S. (2013). Ensemble modeling, uncertainty and robust predictions. WIREs: Climate Change, 4(3), 213–223.Find this resource:

Parker, W. S. (2014). Values and uncertainties in climate prediction, revisited. Studies in Studies in History and Philosophy of Science, 46, 24–30.Find this resource:

Pennell, C., & Reichler, T. (2011). On the effective number of climate models. Journal of Climate, 24, 2358–2367.Find this resource:

Pepper, W. J., Leggett, R. J., Swart, R. J., Wasson, J., Edmonds, J., & Mintzer, I. (1992). Emission scenarios for the IPCC: An update, assumptions, methodology, and results. Washington, DC: Environmental Protection Agency.Find this resource:

Persson, A. O. (2005a). Early operational numerical weather prediction outside the USA: An historical introduction. Part I: Internationalism and engineering NWP in Sweden, 1952–69. Meteorological Application, 12, 135–159.Find this resource:

Persson, A. O. (2005b). Early operational numerical weather prediction outside the USA: An historical introduction: Part II: Twenty countries around the world. Meteorological Application, 12, 269–289.Find this resource:

Petersen, A. (2000). Philosophy of climate science. Bulletin of the American Meteorological Society, 81, 265–271.Find this resource:

Petersen, A. (2006). Simulating nature: A philosophical study of computer-simulation uncertainties and their role in climate science and policy advice. Apeldoorn, the Netherlands: Het Spinhuis Publishers.Find this resource:

Petersen, A. (2011). Climate simulation, uncertainty, and policy advice─The case of the IPCC. In G. Gramelsberger & H. Feichter (Eds.), Climate change and policy: The calculability of climate change and the challenge of uncertainty (pp. 91–111). Berlin, Germany: Springer.Find this resource:

Petersen, A. (2012). Simulating nature: A philosophical study of computer-model uncertainties and their role in climate science and policy advice. Boca Raton, FL: CRC Press.Find this resource:

Phillips, N. (1956). The general circulation of the atmosphere: A numerical experiment. Quarterly Journal of the Royal Meteorological Society, 82, 132–164.Find this resource:

Phillips, N. (1995). Jule Gregory Charney 1917—1981. A biographical memoir. Washington, DC: National Academies Press.Find this resource:

Phillips, T. J., Potter, G. L., Williamson, D. L., Cederwall, R. T., Boyle, J. S., Fiorino, M., . . . Yio, J. J. (2004). Evaluating parameterizations in general circulation models: Climate simulation meets weather prediction. Bulletin of the American Meteorological Society, 85, 1903–1915.Find this resource:

Pichler, H. (2001). Von Margules zu Lorenz. In C. Hammerl, W. Lenhardt, & R. Steinacker, et al. (Eds.), Die Zentralanstalt für Meteorologie und Geodynamik 1851–2001. 150 Jahre Meteorologie und Geophysik in Österreich, 1851–2001 (pp. 387–397). Graz, Austria: Leykam.Find this resource:

Pielke, R. A. J. (2007). The honest broker: Making sense of science in policy and politics. Cambridge, UK: Cambridge University Press.Find this resource:

Plass, G. N. (1956). The carbon dioxide theory on climate change. Tellus, 8, 140–154.Find this resource:

Platzman, G. (1968). Richardson’s weather prediction. Bulletin of the American Meteorological Society, 60, 302–312.Find this resource:

Poincaré, H. (1890). Sur le problème des trois corps et les équations de la dynamique. Acta Mathematica, 13, 1–270.Find this resource:

Poole, R. (2008). Earthrise: How man first saw the Earth. New Haven, CT: Yale University Press.Find this resource:

Pooley, E. (2010). The climate war: True believers, power brokers, and the fight to save the planet. New York, NY: Hyperion.Find this resource:

Price, A., Lenton, T., Cox, S., Valdes, P., Shepherd, J., & the GENIE team (2005). GENIE: Grid ENabled Integrated Earth System Model. ERCIM 61.

Rahmstorf, S. (2008). Anthropogenic climate change: Revisiting the facts. In E. Zedillo (Ed.), Global warming: Looking beyond Kyoto (pp. 34–53). Washington, DC: Brookings Institution Press.Find this resource:

Randall, D. A. (1996). A university perspective on global climate modeling. Bulletin of the American Meteorological Society, 77(11), 2685–2690.Find this resource:

Randall, D. A. (Ed.). (2000). General circulation model development: Past, present, and future. San Diego, CA: Academic Press.Find this resource:

Randall, D. A., Khairoutdinov, M. F., Arakawa, A., & Grabowski, W. (2003). Breaking the cloud parameterization deadlock. Bulletin of the American Meteorological Society, 84, 1547–1564.Find this resource:

Randall, D. A., & Wielicki, B. A. (1997). Measurements, models and hypotheses in the atmospheric sciences. Bulletin of the American Meteorological Society, 82, 283–294.Find this resource:

Randalls, S. (2010). History of the 2°C climate target. WIREs Climate Change, 1(4), 598–602.Find this resource:

Reichler, T., & Kim, T. (2008). How well do coupled models simulate today’s climate? Bulletin of the American Meteorological Society, 89, 303–311.Find this resource:

Revelle, R., & Suess, H. E. (1957). Carbon dioxide exchange between atmosphere and ocean and the question of an increase of atmospheric CO2 during the past decades. Tellus, 9, 18–27.Find this resource:

Richardson, L. F. (1922). Weather prediction by numerical process. Cambridge, UK: Cambridge University Press.Find this resource:

Rossby, C.-G. (1939). Relation between variations in the intensity of the zonal circulation of the atmosphere and the displacements of the semi-permanent pressure systems. Journal of Marine Research, 2, 38–55.Find this resource:

Sargent, R. G. (2013). Verification and validation of simulation models. Journal of Simulation, 7, 12–24.Find this resource:

Scherhag, R. (1939). Verbesserung der Wettervorhersage durch Berechnung der Druckverteilung des Folgetages mit Hilfe der Höhenwetterkarte. Analen der Hydrographie und Maritimen Meteorologie, 67(9), 462–465.Find this resource:

Schneider, S. H., & Dickinson, R. E. (1974). Climate modeling. Reviews of Geophysics and Space Physics, 12, 447–493.Find this resource:

Schütz, J. R. (1895a). Über die Herstellung von Wirbelbewegungen in idealen Flüssigkeiten. Annalen der Physik, 292(9), 144–147.Find this resource:

Schütz, J. R. (1895b). Über eine bei der theoretischen Einführung incompressibler Flüssigkeiten gebotene Vorsicht. Annalen der Physik, 292(9), 148–150.Find this resource:

Sexton, D. M. H., Murphy, J. M., Collins, M., & Webb, M. J. (2012). Multivariate probabilistic projections using imperfect climate models part I: Outline of methodology. Climate Dynamics, 38, 2513–2542.Find this resource:

Shackley, S., Risbey, J., Stone, P. & Wynne, B. (1999). Adjusting to policy expectations in climate change modeling. Climatic Change, 43(2), 413–454.Find this resource:

Shackley, S., & Wynne, B. (1995). Global climate change: The mutual construction of an emergent science-policy domain. Science and Public Policy, 22, 218–230.Find this resource:

Shackley, S., Young, P., Parkinson, S., & Wynne, B. (1998). Uncertainty, complexity and concepts of good science in climate change modelling: Are GCMs the best tools? Climatic Change 38(2), 159–205.Find this resource:

Siebenhüner, B. (2003). The changing role of nation states in international environmental assessments—The case of the IPCC. Global Environmental Change, 13(2), 113–123.Find this resource:

Silberstein, L. (1896). Über die Entstehung von Wirbelbewegungen in einer reibungslosen Flüssigkeit. Bulletin international de l‘Académie des Sciences de Cracovie, Comptes rendus des séances de l’année, 280–290.Find this resource:

Skodvin, T. (2000). Structure and agent in the scientific diplomacy of climate change, Berlin. Heidelberg, Germany: Springer.Find this resource:

Slingo, J. (2013). Statistical models and the global temperature record. Report of the Met Office.

Smagorinsky, J. (1963). General circulation experiments with the primitive equations. I. The basic experiment. Monthly Weather Review, 91, 99–164.Find this resource:

Study of Man’s Impact on the Climate. (1971). Man’s impact on the climate. Cambridge, MA: MIT Press.Find this resource:

Somerville, R. C. J., & Hassol, S. J. (2011). Communicating the science of climate change. Physics Today, 64, 48–53.Find this resource:

Spekat, A. (Ed.). (2001). 50th anniversary of numerical weather prediction. Commemorative Symposium Potsdam, March 9–10, 2000. Berlin, Germany: German Meteorological Society.Find this resource:

Sprung, A. W. (1885). Lehrbuch der Meteorologie. Hamburg, Germany: Hoffman und Campe.Find this resource:

Stainforth, D. A., Aina, T., Christensen, C., Collins, M., Faull, N., Frame, D. J., . . . Allen, M. R. (2005). Uncertainty in predictions of the climate response to rising levels of greenhouse gases. Nature, 433, 403–406.Find this resource:

Stehr, N., & von Storch, H. (Eds.). (2000). Eduard Brückner─The sources and consequences of climate change and climate variability in historical times. Dordrecht, the Netherlands: Kluwer Academic Publishers.Find this resource:

Stehr, N., & von Storch, H. (2010). Climate and society: Climate as a resource, climate as a risk. Singapore: World Scientific.Find this resource:

Von Helmholtz, H. (1858). Ueber Integrale der hydrodynamischen Gleichungen, welche den Wirbelbewegungen entsprechen. Journal für die reine und angewandte Mathematik, 55, 25–55.Find this resource:

Von Storch, H. (2010). Climate models and modeling: An editorial essay. WIREs Climate Change, 1, 305–310.Find this resource:

Von Storch, H. (2009). Climate research and policy advice: Scientific and cultural constructions of knowledge. Environmental Science & Policy, 12, 741–747.Find this resource:

Von Storch, H., & Flöser, G. (Eds.). (1999). Anthropogenic climate change. Proceedings of the First GKSS School on Environmental Research. Heidelberg, Germany: Springer.Find this resource:

Takle, E. S. (1995). Project to intercompare regional climate simulations (PIRCS), preliminary workshop, November 17−18, 1994. Bulletin of the American Meteorological Society, 76, 1625–1626.Find this resource:

Thomson, J. (1857). On the grand currents of atmospheric circulation. Report of the British Association Meeting, Dublin, pp. 38–39.

Thomson, J. (1892). Bakerian lecture: On the grand currents of atmospheric circulation. Philosophical Transactions of the Royal Society of London A, 183, 653–684.Find this resource:

Thomson, W. (Lord Kelvin) (1867). On vortex atoms. Proceedings of the Royal Society of Edinburgh, 6, 94–105.Find this resource:

Thompson, P. D. (1954). Note on integration of vorticity for quasi-geostrophic flow. Tellus, 6(4), 319–325.Find this resource:

Toon, A. (2012). Models as make-believe: Imagination, fiction and scientific representation. Basingstoke, UK: Palgrave-Macmillan.Find this resource:

Thrope, A., Volkert, H., & Ziemianski, M. J. (2003). The Bjerknes’ circulation theorem: A historical perspective. Bulletin of the American Meteorological Society, 84, 471–480.Find this resource:

Trenberth, K. E., Stepaniak, D. P., Hurrell, J. W., & Fiorino, M. (2001). Quality of reanalyses in the tropics. Journal of Climate, 14, 1499–1510.Find this resource:

Uppala, S. M., Kallberg, P. W., Simmons, A. J., Andrae, U., Da Costa Bechtold, V., Fiorino, M, . . . Woollen, J. (2005). The ERA-40 re-analysis. Quarterly Journal of the Royal Meteorological Society, 131(612), 2961–3012.Find this resource:

Van Asselt, H. (2014). The fragmentation of global climate governance: Consequences and management of regime interactions. Cheltenham, UK: Edward Elgar.Find this resource:

Van Vuuren, D. P., Edmonds, J., Kainuma, M., Riahi, K., Thomson, A., Hibbard, K., . . . Rose, S. K. (2011). The representative concentration pathways: An overview. Climatic Change, 109(1), 5–31.Find this resource:

Volkert, H. (2007). Felix Maria Exner-Ewarten. In N. Koertge (Ed.), New dictionary of scientific biography (pp. 425–427). New York, NY: Charles Scribner’s Sons.Find this resource:

Walker, M. (2011). History of the Meteorological Office. Cambridge, UK: Cambridge University Press.Find this resource:

Weart, S. R. (2003). The discovery of global warming. Cambridge, MA: Harvard University Press.Find this resource:

Weart, S. R. (2010). The development of general circulation models of climate. Studies in History and Philosophy of Modern Physics, 41, 208–217.Find this resource:

Weart, S. R. (2014). General circulation models of climate. In S. R. Weart (Ed.), The discovery of global warming.Find this resource:

Washington, W. M. (1982). Documentation for the Community Climate Model (CCM), Technical Report NTIS No. PB82 194192. Boulder, CO: National Center for Atmospheric Research.Find this resource:

Weigel, A. P., Knutti, K., Liniger, M. A., & Appenzeller, C. (2010). Risks of model weighting in multimodel climate projections. Journal of Climate, 23, 4175–4191.Find this resource:

Wiin-Nielsen, A. (2001). Numerical weather prediction: The early development with emphasis on Europe. In A. Spekat (Ed.), 50th anniversary of numerical weather prediction. Commemorative Symposium, Potsdam, March 9–10, 2000. Berlin, Germany: German Meteorological Society.Find this resource:

Winsberg, E. (2001). Simulations, models and theories: Complex physical systems and their representations. Philosophy of Science, 68, 442–454.Find this resource:

Winsberg, E. (2010). Science in the age of computer simulation. Chicago, IL: University of Chicago Press.Find this resource:

World Meteorological Organization. (2016). Understanding climate.

Young, O. R. (1997). Global governance: Drawing insights from the environmental experience. Cambridge, MA: The MIT Press.Find this resource: