Page 55 - Earth's Climate Past and Future

CHAPTER 2 • Climate Archives, Data, and Models 31

organisms and benthic foraminifera. Cases that will be movement of distinctive chemical tracers through the
examined in detail in later chapters include isotopes of climate system.
oxygen, which record changes in the global volume of ice
and in local ocean temperatures (Chapters 6 and 9); and
isotopes of carbon, which trace movements of organic 2-6 Physical Climate Models
material among reservoirs on the continents, in the air, Most physical models are constructed to simulate the
and in the ocean (Chapter 10). operation of the climate system as it exists today.
Additional geochemical proxies gradually become The modern climatic system is described on the com-
available over the younger part of Earth’s long climatic panion Web site at www.whfreeman.com/ruddiman2e.
history. At orbital time scales, ice cores contain samples The simulation of modern climate is called the control
of air from past atmospheres, including concentrations case. Models must simulate modern climate reasonably
of the greenhouse gases carbon dioxide (CO ) and well to be trusted as a tool for exploring past climates.
2
methane (Chapter 10). Other important proxies in Simulations of past climates occur in a three-step
ice cores include changes in the thickness of snow process (Figure 2–18). The first step is to choose the
deposited (related to the temperature and moisture experiment to be run by specifying the input to the
content of the air), in the amount of dust delivered by model. One or more aspects of the model’s representa-
winds from various continents; and in isotopes of oxy- tion of the modern world are altered from their present
gen and hydrogen that measure air temperatures over form to reflect changes known to have occurred in the
the ice sheet. past. For example, the level of CO in the model atmos-
2
Cave deposits contain records of groundwater phere might be increased or decreased, the height of its
derived from atmospheric precipitation. Changes in the mountains raised or lowered, ice sheets removed or
chemical composition of this water reflect changes in added, or the position of continents moved around.
the original sources of the water vapor, in the atmos- These features that are altered to test hypotheses of cli-
pheric transport path to the site of precipitation, and in mate change are called the boundary conditions.
the groundwater environment (Chapter 10). Sedimen- The second step is the actual operation of the
tary deposits in lakes record not only changes in pollen model. Physical laws that drive the flow of heat energy
but also climatically driven fluctuations in lake levels through Earth’s climate system are incorporated in the
(Chapters 12 and 13) and other chemical tracers now internal workings of the model. When an experiment is
under active investigation. run, these laws come into play in a climate simulation.
Trees record the amount of cellulose deposited The third step is to analyze the climate data output
in each annual layer (determined from the width and that emerges from the experiment. The data from the
density of tree rings) as an index of changes in precipita- simulation can then be used to evaluate the hypotheses
tion during the rainy season in dry regions and changes being tested. For example, does a specific change in
in summer temperatures in cold regions (Chapter 16). boundary conditions cited in a hypothesis (atmospheric
Annual coral bands contain a wide range of chemical CO level, mountain elevation, or continental position)
2
information, including ratios of isotopes of oxygen affect climate in the way the hypothesis proposed?
that record changes in temperature and precipitation Often climate data output can be tested against inde-
(Chapter 16). pendent geologic data that played no part in the experi-
mental design (Figure 2–18). For example, if a model
Climate Models run simulates stronger winds in a specific region for a
particular interval of geologic time, scientists can sample
Scientists who extract records from Earth’s climate sediment cores from that area to check whether or not
archives inevitably discover new trends that were previ- larger particles of windblown dust were deposited in the
ously unknown. Usually, their proposed explanations locations indicated by the simulation.
for the trends are tested using climate models because Mismatches between geologic data and climate data
models put numbers on ideas. But models also simplify output from physical circulation models may imply sev-
some aspects of reality, and the results they provide eral possible problems: key boundary conditions were
have to be critically assessed. specified incorrectly or were omitted from the experi-
In this section we examine two kinds of numerical ment; the model does not adequately simulate some
(computer) models used by climate scientists. Physical part of the climate system; or the geologic data used for
climate models emphasize the physical operation of comparison to the model output were misinterpreted.
the climate system, particularly the circulation of the Despite this range of possible problems, the main cause
atmosphere and ocean but also interactions with vege- of data-model mismatches is often obvious enough to
tation (biology) and with atmospheric trace gases lead to useful refinements in boundary conditions, in
(chemistry). Geochemical climate models track the data interpretation, or in model construction. The