A Vast Machine

Quotes

This doesn't mean that controversies are suppressed; in fact, quite the opposite is true. The IPCC brings controversy within consensus, capturing the full range of expert opinion.

- A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (Paul N. Edwards) - Highlight on Page xvii - Loc. 104-5 - Added on Friday, June 07, 2013, 08:22 AM

Yet these countless versions of the atmosphere's history have also converged. Could it be that one day some grossly different data image will emerge, in which the planet did not really warm across the period of historical records, or human activity played no significant role in climate change? Sure, it's possible; in science, never say never. But the chances of such a thing happening today are vanishingly small.

- A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (Paul N. Edwards) - Highlight on Page xviii - Loc. 119-21 - Added on Friday, June 07, 2013, 08:28 AM

In the mid 1990s, environmental conservatives and climate-change skeptics promoted the idea that "sound science" must mean "incontrovertible proof by observational data," whereas models were inherently untrustworthy. But in global climate science, at least, this is a false dichotomy. The simplistic "models vs. data" debate lingers on, but in recent years it has been largely replaced by more sophisticated approaches.

- A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (Paul N. Edwards) - Highlight on Page xviii - Loc. 127-29 - Added on Friday, June 07, 2013, 08:30 AM

Computer models hold the key to transforming these information resources into knowledge.

- A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (Paul N. Edwards) - Highlight on Page xix - Loc. 141 - Added on Friday, June 07, 2013, 08:33 AM

Between 1945 and 1965, digital computers revolutionized weather forecasting, transforming an intuitive art into the first computational science. Unlike many scientific revolutions, this one was planned.

- A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (Paul N. Edwards) - Highlight on Page 111 - Loc. 1526-27 - Added on Sunday, June 09, 2013, 05:53 PM

In a process I have called "mutual orientation," scientists and engineers oriented their military sponsors toward new techniques and technologies, while the agencies oriented their grantees toward military applications. This relationship mostly produced general directions rather than precise goals; rarely did military funders require scientists to specify exactly how their research might be used by the armed forces. Nonetheless, military funders did expect that at least some of the work they paid for would ultimately lead to weaponry or to other forms of strategic advantage, including useful practical knowledge.'

- A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (Paul N. Edwards) - Highlight on Page 112 - Loc. 1539-42 - Added on Sunday, June 09, 2013, 05:57 PM

General George C. Kennedy of the Strategic Air Command claimed in 1953 that "the nation which first learns to plot the paths of air masses accurately and learns to control the time and place of precipitation will dominate the globe."'

- A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (Paul N. Edwards) - Highlight on Page 112 - Loc. 1545-47 - Added on Sunday, June 09, 2013, 06:00 PM

In an early postwar funding request, von Neumann mentioned "high speed calculation to replace certain experimental procedures in some selected parts of mathematical physics.""

- A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (Paul N. Edwards) - Highlight on Page 115 - Loc. 1580-81 - Added on Sunday, June 09, 2013, 06:51 PM

The widely recounted D-Day story, along with many others, entrenched meteorology as a military science. For example, commanders used "applied climatology" in siting new bases, choosing transport routes, and deciding when to launch operations.17 Statistical climatology had also assisted in scheduling the Normandy invasion: analysis had revealed that May and July would probably be worse than June for operations in the English Channel.

- A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (Paul N. Edwards) - Highlight on Page 116 - Loc. 1587-89 - Added on Sunday, June 09, 2013, 06:53 PM

"The leak indicated that a comprehensive meteorological theory existed (when it most certainly did not) and emphasized the weather control aspects." In order to sell a project that could forecast, or control, the weather, the meteorologists needed to have a plausible theory to back it up."23

- A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (Paul N. Edwards) - Highlight on Page 117 - Loc. 1601-3 - Added on Sunday, June 09, 2013, 06:56 PM

In March and April of 1950, Jule Charney, Ragnar Fjortoft, George Platzman, Joseph Smagorinsky, and John Freeman spent five weeks at Aberdeen. Von Neumann's wife, Klara, taught the team to code for the ENIAC and checked the final program. Von Neumann himself rarely appeared, but called in frequently by telephone. Working around the clock for 33 days, the team at Aberdeen carried out two 12-hour and four 24-hour retrospective forecasts. A second ENIAC expedition took place a year later, but the group never published its results.29

- A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (Paul N. Edwards) - Highlight on Page 119 - Loc. 1634-37 - Added on Monday, June 10, 2013, 08:36 AM

Hurd Willett of the Massachusetts Institute of Technology, writing around 1950, bemoaned the absence of noticeable improvement: "... probably there is no other field of applied science in which so much money has been spent to effect so little real progress as in weather forecasting. . . . In spite of ... [the] great expansion of forecasting activity, there has been little or no real progress made during the past forty years in the verification skill" of basic surface forecasts.55

- A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (Paul N. Edwards) - Highlight on Page 130 - Loc. 1747-50 - Added on Monday, June 10, 2013, 07:20 PM

The history of NWP is often presented as one of continuous success and steady progress. Indeed, many scientists with whom I spoke while researching this book expressed surprise or outright skepticism when I noted that it took several decades for computer models to approach human forecast skill. Yet, as we have seen, climbing the hierarchy of models did not lead instantaneously to better forecasts; in fact, initially the opposite was true. Simple barotropic models remained extremely popular with working forecasters well into the 1970s, long after baroclinic and primitive-equation model forecasts had become routinely available.

- A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (Paul N. Edwards) - Highlight on Page 132 - Loc. 1765-68 - Added on Monday, June 10, 2013, 07:29 PM

Why, then, did the national weather services of most developed countries commit themselves so rapidly and completely to this new technological paradigm? First, older techniques had reached their limits; no one put forward any competing vision for major progress in forecast quality and scope. Second, it marked a generational change. As meteorology's scientific sophistication increased, and as the field became professionalized during and after World War II, consensus had developed around the desirability of grounding forecasting in physical theory. Roger Turner has argued that a small group of theoretically savvy meteorologists, led most notably by Carl-Gustav Rossby and Francis Reichelderfer, "actively constructed" this consensus around what Turner calls "universal meteorology": "... between the 1920s and the 1940s, Rossby, Reichelderfer and their allies designed the institutions, established the curriculum, and cultivated the values that guided the weather cadets trained during World War II.i6O This new cadre of theorists stood ready, on both sides of the Atlantic, to tackle the risky physics and thorny mathematics of numerical modeling. Finally, Charney, Rossby, and von Neumann together articulated a clear research and development program for NWP. As computer power increased, researchers would climb the hierarchy of models, increasing grid resolution and adding ever more realistic physics.

- A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (Paul N. Edwards) - Highlight on Page 132 - Loc. 1768-75 - Added on Monday, June 10, 2013, 07:29 PM

With the ENIAC as proof of concept, von Neumann staked his considerable reputation on the belief that electronic digital computing would develop rapidly and inexorably.

- A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (Paul N. Edwards) - Highlight on Page 133 - Loc. 1776 - Added on Monday, June 10, 2013, 07:30 PM

From the beginning, Charney, Rossby, and von Neumann framed numerical weather prediction as a plan, not a gamble. In the end it worked, of course. But 20-20 hindsight makes it easy to miss the many ways in which it might have failed.

- A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (Paul N. Edwards) - Highlight on Page 133 - Loc. 1777-79 - Added on Monday, June 10, 2013, 07:30 PM

Second, computer modeling required a considerably different range of expertise than previous forecasting techniques. The ideal modeler had a strong background in physics and mathematics as well as in meteorology. Anticipating this requirement in 1952, John von Neumann pointed out that an "educational problem" blocked the path to operational NWP: There is an educational problem because there are practically no people available at the present time capable of supervising and operating such a program. Synoptic meteorologists who are capable of understanding the physical reasoning behind the numerical forecast are needed to evaluate the forecasts. . . . Mathematicians are needed to formulate the numerical aspects of the computations. During the first several years of the program the meteorological and mathematical aspects probably cannot be separated and personnel familiar with both aspects are needed. An intense educational program could conceivably produce enough people in about three years.12

- A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (Paul N. Edwards) - Highlight on Page 134 - Loc. 1796-1801 - Added on Monday, June 10, 2013, 07:35 PM

Von Neumann did not mention-and may not have anticipated-that veritable armies of computer technicians, programmers, and other support personnel would also be required.

- A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (Paul N. Edwards) - Highlight on Page 135 - Loc. 1801-2 - Added on Monday, June 10, 2013, 07:35 PM

And price was only one factor. Scientific programmers, skilled technical support personnel, and expertise in numerical methods all remained in short supply.

- A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (Paul N. Edwards) - Highlight on Page 136 - Loc. 1821-22 - Added on Monday, June 10, 2013, 07:38 PM

The most successful institutions paired theorists with technical wizards, many of them meteorologists who had discovered an aptitude for programming.

- A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (Paul N. Edwards) - Highlight on Page 136 - Loc. 1822-23 - Added on Thursday, June 13, 2013, 08:43 AM

Without the latter, the crucial, highly complex translation of mathematics into computer code could not proceed.

- A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (Paul N. Edwards) - Highlight on Page 136 - Loc. 1823 - Added on Thursday, June 13, 2013, 08:43 AM

For this reason, meteorology has historically been the foremost civilian consumer of supercomputer power. Only the designers of nuclear weapons have laid a greater claim to the world's most advanced computers."

- A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (Paul N. Edwards) - Highlight on Page 137 - Loc. 1828-29 - Added on Thursday, June 13, 2013, 08:44 AM

As we saw in chapter 6, John von Neumann and others understood the significance of this power "to replace certain experimental procedures" almost immediately, although the full extent of it would take some time to dawn on anyone.3

- A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (Paul N. Edwards) - Highlight on Page 140 - Loc. 1848-49 - Added on Thursday, June 13, 2013, 08:49 AM

"infinite forecast."' By this phrase, von Neumann did not intend deterministic prediction of weather over long or "infinite" periods. Instead, he had in mind the statistically "ordinary circulation pattern" that would emerge when "atmospheric conditions . . . have become, due to the lapse of very long time intervals, causally and statistically independent of whatever initial conditions may have existed." This phrase sounds remarkably like the mathematical concept of chaos, by which minute variations in initial conditions rapidly generate extreme divergences in outcomes: a butterfly flaps its wings over Brazil and causes a tornado in Texas. In fact, the theoretical meteorologist Edward Lorenz first discovered what we now call chaos theory while working with atmospheric models in the early 1960s.8 But in the mid 1950s, these results, and the idea of chaos itself, remained unknown. In 1955, von Neumann's "infinite forecast" expressed the widespread belief that global atmospheric flows might display predictable symmetry, stability, and/or periodicity. Research aimed at finding such predictable features remained active throughout the 1950s.9

- A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (Paul N. Edwards) - Highlight on Page 141 - Loc. 1856-61 - Added on Thursday, June 13, 2013, 08:51 AM

What counts are not mere tabulations of data; it is their intelligent organization according to physical laws so as to lead to physical depiction of relevant processes and schemes of motion.

- A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (Paul N. Edwards) - Highlight on Page 142 - Loc. 1874-75 - Added on Thursday, June 13, 2013, 08:54 AM

If Earth had no atmosphere or oceans, its average surface temperature would be about -19°C. Instead, the heat retained in the atmosphere and oceans maintains it at the current global average of about 15°C. At the equator, Earth receives more heat than it can re-radiate to space; at the poles, it re-radiates more heat than it receives. Thus the climate system, as a thermodynamic engine, serves to transport heat from the equator toward the poles.

- A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (Paul N. Edwards) - Highlight on Page 143 - Loc. 1890-93 - Added on Thursday, June 13, 2013, 07:08 PM

Svante Arrhenius's calculations of carbon dioxide's effect on Earth's temperature constituted one of the earliest one-dimensional (zonal) EBMs.13

- A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (Paul N. Edwards) - Highlight on Page 144 - Loc. 1899 - Added on Thursday, June 13, 2013, 07:09 PM

Parameterizing physical processes accurately is the most difficult aspect of climate modeling and is a source of considerable scientific and political controversy.

- A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (Paul N. Edwards) - Highlight on Page 146 - Loc. 1922-23 - Added on Thursday, June 13, 2013, 07:14 PM

experiment," he wrote, "contains empirical elements in that the representation of certain physical effects is based on meteorological experience with the actual atmosphere, rather than being predicted from fundamental laws ofphysics."

- A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (Paul N. Edwards) - Highlight on Page 152 - Loc. 1985-86 - Added on Friday, June 14, 2013, 07:10 PM

Phillips's model provoked enormous excitement.

- A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (Paul N. Edwards) - Highlight on Page 152 - Loc. 1989 - Added on Friday, June 14, 2013, 07:11 PM

Phillips marked out the path: start with simplifying assumptions, such as barotropy and quasi-geostrophy, then eliminate them, one by one, until nothing remained but the primary physics.

- A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (Paul N. Edwards) - Highlight on Page 152 - Loc. 1996-97 - Added on Friday, June 14, 2013, 07:14 PM

We first constructed with considerable care, and in fact programmed, the most general of a hierarchy of models in order to uncover in some detail the body of physics needed, to determine where the obvious weaknesses were, and to give us some idea of the computational limitations we could expect. The perspective thus gained was invaluable. We then laid out a program of simplified models which can be constructed as a sub-set of the most general one. The main requirements were (1) that each model represent a physically realizable state, (2) that they could be constructed computationally . . . , and (3) that they collectively would provide a step-by-step study of the behavior of new processes and their influence on the interactive system. Hence, many of the intermediate models in themselves may lack detailed similitude to the atmosphere but provide the insight necessary for careful and systematic scientific inquiry.32

- A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (Paul N. Edwards) - Highlight on Page 154 - Loc. 2021-26 - Added on Friday, June 14, 2013, 07:17 PM

This strict attention to correcting physical theory and numerical methods before seeking verisimilitude became a hallmark of the GFDL modeling approach.33 The full primitive-equation GCM ("the most general" model) served as a conceptual framework, driving work on simpler models which led to refinements of the GCM.

- A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (Paul N. Edwards) - Highlight on Page 155 - Loc. 2026-28 - Added on Friday, June 14, 2013, 07:17 PM

Arakawa persuaded Mintz to pay more attention to designing model dynamics that could sustain long-term integration.

- A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (Paul N. Edwards) - Highlight on Page 156 - Loc. 2061-62 - Added on Friday, June 14, 2013, 07:20 PM