Source material
Source material:
- Chapter 1 W. Washington, C. Parkinson (2005)
- Lecture slides1:
- Drive/Slides/Lecture1CM
| Next topic >>
For centuries individuals have dreamed of being able to understand how the climate system works and from that understanding being able to forecast and perhaps even modify the future climate. This distant goal has become more approachable as a result of the invention of the modern electronic computer in the mid-twentieth century. With the computer, it is now possible to solve numerically many of the equations encapsulating the physical laws that govern climate. This provides us with exciting new possibilities, but also new dangers, and it is important to move with caution, being careful to understand the limitations as well as the potentials of our new capabilities. Improved understanding of the climate system could have substantial impact on the economic well-being of the nations of the world. For instance, better predictions of future climate states could help in the determination of more opportune times for the planting and harvesting of crops and in the appropriate warning of possible harm to the environment from various human activities, such as deforestation of the landscape and industrial insertion into the Earth’s atmosphere of carbon dioxide, other trace gases, and human-created aerosols. On time scales of months to years, many uncertainties about climate change remain which could be expected to lessen as computer modeling studies advance. For instance, still unknown are the full climatic impacts of ocean temperature anomalies occurring during El Niño and La Niña episodes and the full impacts of volcanic eruptions on ozone amounts in the stratosphere and on atmospheric cooling, the latter arising as volcanoes insert into the atmosphere particles that refiect sunlight back out of the Earth/atmosphere system. Recently, scientists have begun exploring concerns about changes in the ecology brought about by biomass burning and deforestation. These changes can be examined through the use of climate models with interactive chemistry cycles. Without computer climate models, it would be extremely difficult to give quantitative answers to many of the questions raised by the varied issues just mentioned, since the interactions are so complex. Over the past half century, computer models have become powerful investigative tools for climate research. They will almost certainly become even more so in the future, especially as the observational data base improves both in quality and in spatial coverage, with expanded satellite and conventional data networks, and as procedures are further developed for assimilating the data into the numerical models. The realization that the physical laws governing the atmosphere, oceans, and other components of the climate system could be used to determine future condi- tions was inherent in much of the philosophical reasoning in the late eighteenth and early nineteenth centuries, following the resounding early success of Newtonian me- chanics. This was stated most vividly by Pierre Simon de Laplace in 1812 when he suggested that complete knowledge of the masses, positions, and velocities of all particles at any single instant would enable precise calculation of all past and fu- ture events. At the time, the classical laws of mechanics were firmly in hand, having been derived largely from Isaac Newton’s monumental 1687 publication Principia Mathematica; but still undiscovered were the fundamental laws of thermodynamics, which would be crucial for generating scientific predictions for fluid fields such as the atmosphere and oceans. The 1840s witnessed the independent development of the concept of the conservation of energy by Robert Mayer, James Joule, and Her- mann von Helmholtz. This led Rudolf Clausius in 1850 to identify the conservation of energy as the first law of thermodynamics and to formulate the second law of ther- modynamics, that in the absence of external constraints the net flow of heat between two bodies is from the warmer to the cooler one. By the late 1800s, the fundamental laws of classical physics were known and thus the goal of accurate prediction through numerical calculation was closer to becoming realizable. Concerning forecasting events in the atmosphere or oceans, Vilhelm Bjerknes (Figure 1.1) planned a course of action that in many ways remains as valid today as when he first stated it in 1904 (Bjerknes 1904) in these words:
Quote
If it is true, as every scientist believes, that subsequent atmospheric states develop from the preceding ones according to physical law, then it is apparent that the necessary and sufficient conditions for the rational solution of forecasting problems are the following:
- A sufficiently accurate knowledge of the state of the atmosphere at the initial time.
- A sufficiently accurate knowledge of the laws according to which one state of the atmosphere develops from another.
Later in the book, we discuss limitations on our ability to determine the two broad conditions stated by Bjerknes. During World War I, while resting between battles during his work as a vol- unteer ambulance driver for the Red Cross, English scientist Lewis Fry Richardson (Figure 1.2) attempted to use the basic equations of atmospheric motions to develop a capability of forecasting weather, using only a mechanical calculator. By that time, it was well known that the equations were highly complex and not amenable to sim- ple solution and that the only recourse for solving the equations was through use of numerical approximations. In essence, a set of continuous equations was handled approximately by obtaining a numerical solution to a corresponding set of discrete equations. In his book Weather Prediction by Numerical Process, Richardson (1922) describes in a step-by-step manner his predictive method for one small area in Eu- rope using available observational data. Since Richardson did not know the relative importance of all the various processes in the atmosphere, he attempted to include considerable atmospheric physics. This served to encumber his calculations, and the resulting prediction was highly unrealistic. However, many of the problems he en- countered while trying to accomplish this unprecedented calculation are the same ones that climate and weather forecast modelers continue to face today. A fascinat- ing historical account is provided by Platzman (1967). Richardson’s book in many ways serves as a blueprint for constructing a numerical model. From time to time we will make reference to it in the context of present understanding, thereby illu- minating both some of the progress that has been made in this field and possible worthwhile future directions of research as more of the problem areas become less severe. Climate models now are capable of simulating almost all the major observed features of the atmosphere, oceans, and cryosphere, albeit with considerable room for improvement, especially in view of the sometimes questionable adjustments or “tuning” factors needed to obtain the desired simulations. Little was done on the direct numerical solution to the equations after the work of Richardson until the late 1940s, by which time the first electronic computers had been constructed. Amongst these were the Atanasoff-Berry Computer (ABC) at Iowa State College (now Iowa State University), developed by John V. Atanasoff with the help of graduate student Clifford E. Berry, the Electronic Numerical Integrator and Computer (ENIAC) at Aberdeen Proving Grounds, and, of particular importance for meteorology, the Institute for Advanced Studies (IAS) computer at Princeton University, constructed under the direction of mathematician John von Neumann. One of the first problems addressed by von Neumann with his new calculating tool was the forecasting of weather patterns, for which purpose he formed a team of scientists under the leadership of Jule Charney. See Charney et al. (1950), Thompson (1978, 1987), and Nebeker (1995) for interesting accounts of the history of the weather forecasting group. The computer program, or model, that Charney and others tested for their first numerical weather prediction forecast did not contain the general set of equations used by Richardson but rather a simpler set. The general equations admit a wide spectrum of motions in the atmosphere, ranging from sound waves and gravity waves to the very slow-moving large-scale meteorological waves. In the late 1930s Carl Gustav Rossby found that by making suitable approximations to the more general equations he could find simpler equations to solve for the weather patterns that in effect filtered out the unwanted, numerically troublesome, and meteorologically less important waves. It was this simpler set of equations that was used by Charney and his colleagues. A few years later, Phillips (1956) added simple forcing terms to this same set of equations and carried out a long-term integration that became independent of the initial state and showed many of the general features of the atmospheric circulation. This single experiment was a major early step in the long development of general circulation models. More details on Phillips’ effort are noted in later chapters. Later development of general circulation modeling included a shift back to a more general set of equations very close to that used by Richardson. This became possible as the reasons for Richardson’s failure, tied both to erroneous input fields and to the specific numerical scheme, were diagnosed and remedies were found. In the mid-1950s several groups of scientists dedicated to the development of numerical weather prediction and general circulation. modeling of the atmosphere were formed. In particular, in the late 1950s a major general circulation modeling effort in the United States Weather Bureau under the direction of Joseph Smagorinsky became the nucleus of what is now the Geophysical Fluid Dynamics Laboratory (GFDL) at Princeton University. Since the formation of GFDL, now part of the National Oceanic and Atmospheric Administration (NOAA); numerous other climate modeling centers have been established throughout the world: Large-scale ocean modeling was developed somewhat later; in the early 1960s; and involves distinctive problems, many of which are mentioned in the course of this book. Modeling of the important large-scale roles of seaice, snow, land processes, and the biosphere has developed since the mid- 1960s, and their incorporation in climate models is ongoing. We describe in subsequent chapters the relevant formulations, limitations, and coupling with the atmospheric and oceanic calculations. Preparatory to the presentation of the details of the various models of the climate system, the next chapter presents a basic description of some of the observed features of the atmosphere, ocean, and cryosphere. This discussion provides a basis for the later discussions of the numerical models and comparisons between the modeling results and the observed atmosphere, ocean, and sea ice distributions. Chapter 3 shows how the basic laws of physics can be made appropriate for the atmosphere , oceans , land , and sea ice , including. the important processes of cloud radiation, convection, precipitation, boundary transfers, and small-scale mixing. Chapter 3 also includes some impacts of vegetation on the climate system. Chapter 4 illustrates some of the numerical methods currently used for the solution of the model equations and provides suggestions for finding additional information on the World- Wide Web. The remainder of the book concentrates on examples of simulation results, with Chapter 5 presenting a few results of simulations that attempt to reproduce aspects of today’s climate and Chapter 6 presenting results of attempts to simulate climate under conditions altered from the present. In these two chapters we show examples of what can be learned from present computer models, and in the concluding chapter, Chapter 7, we suggest likely advances for the future. Although references are included, no effort is made to provide a complete bibliography of the field. Where possible, reference is made to suitable review papers that are readily available. Also, the appendices cover various technical details and indicate where on the Internet some simple climate models, with source codes appropriate for workstation computers, can be found. Some of the programs have the complexity of some of the models discussed in this book.