Wallace J.M., Hobbs P.V. Atmospheric Science. An Introductory Survey

Подождите немного. Документ загружается.

434 Climate Dynamics

in the warm pool (Fig. 10.21): it is only at these

times that convective updrafts in the atmospheric

boundary layer in the central equatorial Pacific are

warm and buoyant enough to break through the

capping inversion.

The remarkable reproducibility of ENSO-related

rainfall anomalies in different El Niño years is

demonstrated in Fig. 10.26. That the patterns of

tropospheric temperature anomalies shown in

these figures are also similar over much of the

globe suggests that tropical rainfall anomalies in

the equatorial Pacific induce a global response in

the tropospheric temperature, geopotential height,

and wind fields. This inference is supported by a

large body of theory relating to wave propagation

in planetary atmospheres and by numerous experi-

ments with global atmospheric models forced with

tropical sea-surface temperature anomalies. The

response of the global atmosphere to the shifting of

the tropical rain belts is reflected in temperature

and rainfall anomalies at the Earth’s surface at

Fig. 10.21 Sea-surface temperature (SST) and surface winds over the tropical Pacific averaged from November–April of a warm

year of the ENSO cycle (1997–1998, top) and November-April of a cold year (1998–1999, bottom). During both years the sur-

face winds converge into the region of highest sea-surface temperature, the so-called warm pool, encompassing the western Pacific

Ocean and Indonesia. During El Niño years the warm pool is shifted eastward. [SST data from the U. K. Meteorological Office

HadISST and 10-m winds from the European Remote-sensing Satellite (ERS-2). Courtesy of Todd P. Mitchell.]

the weakening of the trade winds. As a result of the

deepening of the thermocline in the eastern Pacific

during El Niño events, the upwelled water is not as

cold or as rich in nutrients as it is during normal

years. Hence, the sea surface temperatures and bio-

logical productivity in that region are controlled not

so much by the strength of the local wind-induced

upwelling as by the stronger wind anomalies in the

central Pacific, which regulate the east–west slope of

the thermocline.

The difference in the distribution of rainfall over

the tropical Pacific between warm and cold years of

the ENSO cycle is shown in Fig. 10.25. During El

Niño years the ITCZ and the belt of heavy rainfall

over the extreme western Pacific intrude into the

equatorial dry zone, bringing drenching rains to the

Pacific islands that lie close to the equator and to

the lowlands of Ecuador, while Indonesia and a

number of other areas of the tropics experience

drought. The eastward shift in the rain belt during

El Niño events is consistent with the eastward shift

30S

20S

10S

10N

20N

30N

120E 180 120W 60W

30S

20S

10S

10N

20N

30N

10 ms

–1

18 19 20 21 22 23 24 25 26 27 28 29 °C

P732951-Ch10.qxd 12/09/2005 09:12 PM Page 434

10.2 Climate Variability 435

many locations throughout the world. The regional

impacts of ENSO are summarized in Fig. 10.27.

Because the ENSO-induced anomalies in the global

circulation are due to slowly evolving tropical sea-

surface temperature anomalies, they are not subject to

the 2-week predictability limit for weather, which is

governed by the atmosphere’s capricious internal

dynamics. Tropical sea-surface temperature anomalies

tend to be highly persistent from the boreal summer

through the following winter, and an even better fore-

cast of the wintertime anomalies can be obtained by

using statistical prediction schemes or coupled atmos-

phere–ocean models, both of which exhibit significant

skill in ENSO prediction out to a year in advance.

Based on a forecast of seasonal mean sea-surface

temperature anomalies over the tropical oceans, a

variety of statistical and, in some cases, dynamical

models are used to predict the corresponding anom-

–16 –12 –8 –4 0 4 8 12 16

Height anomaly (cm)

July 97

El Niño

July 98

La Niña

Fig. 10.23 Sea level anomalies measured by the TOPEX

Poseidon altimeter. (Top) July 1997, just after the relaxation of

the trade winds following the onset of the 1997–1998 El Niño

event. (Bottom) July 1998, a month of unusually strong trade

winds. [Courtesy of the NOAA Laboratory for Satellite

Altimetry.]

Fig. 10.22 In January 1998 (top) the 1997–1998 El Niño

event was at its height. Because of the weakness of the trade

winds at this time, the upwelling of nutrient-rich water was

suppressed in the equatorial Pacific. The absence of a green

band along the equator in this image is indicative of relatively

low chlorophyll concentrations there. By July 1998 (bottom)

the trade winds had strengthened and equatorial upwelling

had resumed, giving rise to widespread phytoplankton

blooms in the equatorial belt. [Imagery courtesy of SeaWiFS

Project, NASA/GSFC and ORBIMAGE, Inc.]

alies in rainfall, temperatures, crop yields, oil prices,

and other variables.

b. Coupling with the terrestrial biosphere

Atmosphere–ocean coupling affects northern hemi-

sphere climate mainly during its winter season and

the ocean plays an important role in it. In contrast,

drought and desertification, when they occur in

extratropical latitudes, are primarily warm season

phenomena. An extended drought popularly

known as “the dust bowl” impacted large areas of

the United States throughout most of the decade of

the 1930s. Over parts of the Great Plains and

Midwest many of the 1931–1939 summers were

P732951-Ch10.qxd 12/09/2005 09:12 PM Page 435

436 Climate Dynamics

their leaves during the daylight hours. Reduced

evapotranspiration inhibits the ability of the plants

to keep themselves and the Earth’s surface beneath

them cool during the middle of the day, when the

incoming solar radiation is strongest, favoring higher

afternoon temperatures and it lower humidity within

the boundary layer. Because boundary-layer air is

the source of roughly half the moisture that con-

denses in summer rainstorms over the central

United States, lower humidity favors reduced pre-

cipitation: a positive feedback. Higher daily maxi-

mum temperatures, lower humidity, and reduced

precipitation all place stress on the plants. If the

stress is sufficiently severe and prolonged, the

changes in plant physiology become irreversible.

Once this threshold is crossed, the earliest hope for

the restoration of normal vegetation is the next

spring growing season, which may be as much as 9

months away. Throughout the remainder of the sum-

mer and early autumn the parched land surface con-

tinues to exert a feedback on the atmosphere that

acts to perpetuate the abnormally hot, dry weather

conditions that caused it in the first place.

The wilting of plants also affects ground hydrology.

In the absence of healthy root systems, ground water

runs off more rapidly after rain storms, leaving less

behind to nurture the plants. Once the water table

drops significantly, an extended period of near or

above normal precipitation is required to restore it.

The remarkable year-to-year persistence of the 1930s

drought attests to the remarkable “memory” of the

vegetation and the ground hydrology. Hence, once

established, an arid climate regime such as the one

that prevailed during the dust bowl appears to be

capable of perpetuating itself until a well-timed series

of rainstorms makes it possible for the vegetation to

regain a foothold.

The onset and termination of the 1930s dust bowl

are examples of abrupt, but reversible regime shifts

between a climate conducive to agriculture and a

more desert-like climate. Such shifts have occurred

rather infrequently over the United States, but more

regularly in semiarid agricultural regions such as the

Sahel, northeast Brazil, and the Middle East. If the

dry regimes are sufficiently frequent or prolonged,

the cumulative loss of topsoil due to wind erosion

makes it increasingly difficult for vegetation to

thrive, and irreversible desertification occurs. The

northward expansion of the Sahara desert during

the last few centuries of the Roman empire could

have involved a series of drought episodes analogous

to the dust bowl.

Fig. 10.24 (Top) Schematic showing trade winds (red arrows)

maintaining equatorial upwelling and the slope of the equato-

rial thermocline. (Bottom) Response of equatorial sea-level,

currents, and thermocline depth to an abrupt weakening of

the trade winds. [From NOAA Reports to the Nation: El Niño

and Climate Prediction, University Corporation for Atmospheric

Research (1994) pp. 12, 14.]

anomalously hot and dry, with daily maxima often

in excess of 40 °C, as shown in Fig. 10.28. Large

quantities of topsoil were irreversibly lost—blown

away in dust storms that darkened skies from the

Great Plains as far downstream as the eastern

seaboard.

The ENSO cycle modulates the frequency of

drought. Over most tropical continents, drought

tends to occur more frequently during the warm

phase of the ENSO cycle, whereas in the extratropi-

cal summer hemisphere it tends to occur somewhat

more frequently during the cold phase. The ENSO

connection notwithstanding, many of the droughts

on timescales of seasons to years appear to be initi-

ated and terminated by random fluctuations in

atmospheric circulation, and sustained over long

periods of time by positive feedbacks from the bios-

phere. A few weeks of abnormally hot, dry weather

are sufficient to dry the upper layers of the soil,

reducing the water available for plants to absorb

through their root systems. The plants respond by

reducing the rate of evapotranspiration through

P732951-Ch10.qxd 12/09/2005 09:12 PM Page 436

10.2 Climate Variability 437

c. Coupling with the cryosphere

The coupling between the atmosphere and the

cryosphere exerts a feedback on surface air temper-

ature: that is to say, an increase in surface tempera-

ture leads to a decrease in the areal coverage of ice

and snow, which tends to lower the albedo and raise

the fraction of incoming solar radiation that is

absorbed at the Earth’s surface, leading to a further

increase the surface air temperature. The feedback

upon surface air temperature is in the same sense as

the temperature perturbation that caused it: hence it

is said to be a positive feedback. So-called ice-albedo

feedback contributes to the large variability of sur-

face air temperature over high latitudes, which is

evident in the annual cycle (Fig. 10.8) and on

timescales ranging from intraseasonal to the scale of

the ice ages.

Snow cover and sea ice are the dominant cryos-

pheric participants in climate variability on timescales

of centuries or shorter: the more slowly evolving conti-

nental ice sheets come into play on millennial and

longer timescales. Ice albedo feedback has operated

Fig. 10.25 As in Fig. 10.21, but for surface wind and rainfall. The ITCZ and the equatorial dry zone are clearly evident in the

lower panel and they are recognizable in a distorted form in the upper panel. In both panels, note the strong correspondence

between the regions of heavy rainfall and the regions of convergence in the surface winds. [Ten-meter wind from the European

Remote-sensing Satellite (ERs-2) and rainfall from the NCEP Climate Prediction Center Merged Analysis of Precipitation

(CMAP). Courtesy of Todd P. Mitchell.]

120E 180 120W 60W

30S

20S

10S

10N

20N

30N

30S

20S

10S

10N

20N

30N

10 m s

–1

10 15

20 25

cm/mo

Fig. 10.26 Tropical Pacific rainfall anomalies based on satel-

lite imagery and anomalies in tropospheric temperature as rep-

resented by Channel 2 of the microwave sounding unit (MSU)

during December–March 1982–1983 and 1997–1998, the

strongest El Niño events of the 20th century. Red shading indi-

cates regions of anomalously heavy rainfall, gold (green) con-

tours indicate zero and above (below) normal tropospheric

temperatures. Contour interval is 0.5 C. Note the dipole

pattern straddling the region of enhanced rainfall, where

the tropospheric temperatures are abnormally high, the preva-

lence of below normal temperatures over the North Pacific

and above normal temperatures over Canada. [Precipitation

data is the NCEP Climate Prediction Center Merged Analysis

of Precipitation (CMAP). Figure courtesy of Todd P. Mitchell.]

1982–83 1997–98

P732951-Ch10.qxd 12/09/2005 09:12 PM Page 437

438 Climate Dynamics

only during those relatively short epochs in the Earth’s

history, such as the last few million years, when the

polar regions were glaciated; within those epochs

it probably played a more important role during the

glacial intervals, when the areal coverage of ice was

more extensive, than during interglacials.

Although ice albedo feedback acts locally, its effects

are felt globally. Based on estimates of the areal cover-

age of the continental ice sheets and sea ice at the

time of the last glacial maximum 20,000 years ago,

and assuming that the cloud cover was about the same

as it is today, the Earth’s planetary albedo is estimated

to have been higher, by 0.01 than the current value

of 0.305. In Exercise 4.21 the student is asked to show

that an increase in planetary albedo from 0.305 to

0.315 would be sufficient to lower the Earth’s effective

temperature by nearly 1 °C relative to today’s value.

Ice albedo feedback is believed to have played an

essential role in the cycling back and forth between

glacial and interglacial epochs in response to the sub-

tle, orbitally induced variations in summer insolation

over high latitudes of the northern hemisphere

described in Section 2.5.3.

d. Coupling with the Earth’s crust

On a time scale of tens of millions of years it is con-

ceivable that the carbonate-silicate cycle discussed in

Section 2.3.4, of which volcanic eruptions are a part,

could serve to regulate global surface air temperature

through the following mechanism. If, for some reason,

the temperature of the Earth were to become anom-

alously warm, weathering of CaSiO

rocks would be

accelerated, making more calcium ions available for

carbonate formation [Eq. (2.11)]. Carbonate forma-

tion takes up atmospheric carbon dioxide, reducing

the greenhouse effect, inducing a temperature

response in the opposite sense to the temperature

perturbation that initiated the changes. Such a

response, called a negative feedback, is analogous to a

restoring force in a mechanical system, as described in

Section 3.6.3.

Fig. 10.27 Regional impacts of El Niño on weather and climate during the boreal winter. [Adapted from Monthly Weather

Review, 115, p. 1625 (1987) by the NOAA PACS program.]

Dry

Wet

Warm

Dry

Wet

Dry

Warm

20S

20N

120E 180 120W 60W

60N

40N

80N

60S

40S

80S

060E

Fig. 10.28 Daily maximum temperature during June–August at Kansas City. Values for “dust bowl” years 1934–1936 are indicated

by red dots, and values for all other years by gray dots. [Courtesy of Imke Durre, NOAA National Climatic Data Center.]

AugustJulyJune

Temperature (°C)

P732951-Ch10.qxd 12/09/2005 09:12 PM Page 438

10.2 Climate Variability 439

10.2.3 Externally Forced Climate Variability

Climate variability may also be forced by changes in

the sun’s emission, by large volcanic eruptions, and by

human activities. The amplitude of such externally

forced variability depends on the amplitude of the

forcing, as well as the sensitivity of the climate system

to the forcing, as discussed quantitatively in Section

10.3. If the forcing is sufficiently gradual, like the

increase in the sun’s emission over the lifetime of the

Earth, all the components of the Earth system will

remain in equilibrium with it at all times. In contrast, if

the forcing changes instantaneously or consists of a

short-lived impulse (e.g., as in the case of volcanic

eruption), climate will exhibit a transient response with

multiple timescales. Discussion of the impact of human

activities on climate is reserved for Section 10.4.

a. Solar variability

The intensity of the radiation emitted by the sun varies

over a wide range of timescales and the amplitude of

these variations is wavelength dependent. Much of this

variability is associated with the interrelated phenom-

ena shown in Fig. 10.29 that appear intermittently in

active regions of the sun.

• Sunspots are dark (cool) patches that interrupt

the regular pattern of convective cells in the

photosphere. Sunspot groups are accompanied

by strong magnetic fields with a preferred

orientation that reverses between solar

hemispheres. Lifetimes of sunspots range from a

day up to a few months.

• Faculae are bright (hot) spots in the pattern of

convective cells that often appear in conjunction

with sunspots and are accompanied by strong

magnetic fields. Lifetimes of faculae are

comparable to those of sunspots.

• Flares are intense bursts of ultraviolet and x-ray

radiation and high-energy particles emanating

from the sun’s outer atmosphere within the

active regions. They are characterized by strong

magnetic fields and violent motions. A typical

flare lasts on the order of an hour.

Solar activity is modulated by the 11-year solar

cycle illustrated in Fig. 10.30. During the active phase

of the cycle, sunspots, faculae and flares are numer-

ous, appearing first at the higher latitudes and then at

successively lower latitudes over the course of the

next few years. During the quiet phase of the solar

cycle, few, if any, of these disturbances are observed.

The polarity of the magnetic field in sunspot groups

reverses from one cycle to the next. Between succes-

sive cycles the polarity of the magnetic field in the

sunspot groups reverses; hence a full solar cycle is

sometimes regarded as lasting around 22 years.

To understand how disturbances on the sun affect

Earth’s atmosphere it is necessary to consider in

more detail the vertical structure of the sun’s outer-

most layers. Above the photosphere lies the chro-

mosphere, a layer 2500 km thick in which the

temperature increases from a minimum of 4300 K

near the bottom to 10

K at the top. Emanating from

the chromosphere in all directions is a stream of

extremely hot, ionized particles known as the solar

wind. In Earth’s orbit, particle velocities in the solar

Fig. 10.29 An intense flare in the upper left quadrant of the solar disk that occurred June 6, 2000. The image on the left was

taken at the Big Bear Solar Observatory using radiation in the hydrogen alpha line; the second from left is a soft x-ray image

obtained from the Yohkoh satellite (“sun beam” in Japanese), the third is another soft x-ray image from the SOHO (Solar and

Heliospheric Observatory) satellite, and the panel on the right shows a visible image of the sun taken from the SOHO satellite in

which the light coming directly from the sun is blocked (the position of the solar disk is indicated in the image by the white cir-

cle). [Assembled by Syun-Ichi Akasofu.]

P732951-Ch10.qxd 12/09/2005 09:12 PM Page 439

440 Climate Dynamics

wind are 500 km s

1

and temperatures sometimes

approach 10

Photospheric radiation scattered by

the free electrons in the dense inner portion of the

solar wind produces a diffuse luminosity called the

corona (Fig. 10.29, right), which is visible to the naked

eye during solar eclipses. In accordance with Wien’s

displacement law [Eq. (4.11)], gases in the outer

chromosphere and corona emit radiation in the far

ultraviolet and x-ray regions of the electromagnetic

spectrum. These superheated gases account for virtu-

ally all of the sun’s emission at wavelengths 0.1



In contrast to the steadiness of the emission from

the photosphere, the radiation from the outer chro-

mosphere and corona increases in response to solar

activity and is enhanced greatly at times when flares

are in progress. The enhancement is particularly

strong in the x-ray region of the spectrum: radiation

with wavelengths 0.02



m is more than an order of

magnitude stronger during flares than during quiet

periods. However, radiation emitted by the outer

chromosphere and corona accounts for less than

1 part in 10

of the total solar output. Hence even

the strongest flares have no detectable influence on

the total flux density of solar radiation reaching the

Earth.

In ground-based measurements, such as those

described in Box 4.1, temporal variations in the sun’s

emission are masked by variations in the optical

properties of the Earth’s atmosphere, which are

virtually impossible to quantify. Direct measurements

from space-based instruments, which began in 1979,

indicate that the sun’s total emission varies in syn-

chrony with the solar cycle, with higher values during

active intervals than during quiet intervals, as shown

in Fig. 10.31. However, the peak-to-peak amplitude of

these variations is only 1W m

2

, less than 0.1% of

the mean value of the emission, too small to produce

even a 0.1 K change in the Earth’s equivalent black-

body temperature.

Upon close inspection of Fig. 10.30 it is evident

that the areal coverage of sunspots varies from one

solar cycle to the next and that the cycles from 1880

to 1940 were less pronounced than those later in the

record. The sun was even less active during the first

few decades of the 19th century, and there is com-

pelling historical evidence of an almost complete

1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

Year

90N

30N

30S

90S

0.5

0.4

0.3

0.2

0.1

0.0

Fig. 10.30 (Top) Fractional area of the photosphere that is covered by sunspots as a function of latitude and time. Yellow shading

indicates areal coverage in excess of 1%, red between 0.1 and 1%, and unshaded less than 0.1%. (Bottom) The fractional area of the

surface of the entire solar photosphere (in %) covered by sunspots as a function of time. [Courtesy of David Hathaway, NASA

Marshall Space Flight Center.]

High-energy particles emitted by the sun in flares reach the Earth a few days later. The resulting enhancements of the solar wind

often produce auroral displays and interfere with radio and electrical power transmission.

P732951-Ch10.qxd 12/09/2005 09:12 PM Page 440

10.2 Climate Variability 441

absence of sunspots from 1645 to 1715, a period

referred to as the Maunder

minimum in solar activ-

ity, as shown in Fig. 10.32. Solar physicists believe

that these extended intervals of reduced solar activ-

ity may have been marked by reductions in total

emission that were larger, by a factor of 10–30, than

those observed in recent decades during the quiet

years of the solar cycle. If this interpretation is cor-

rect, the consequent decrease in Earth’s equivalent

blackbody temperature could have been large

enough to account for the relatively cold conditions

that prevailed during the so-called Little Ice Age.

b. Volcanic eruptions

The impact of geothermal energy released in volcan-

ism upon the global heat balance is negligible. The

climatic impacts of volcanic eruptions are mainly due

to the radiative effects of sulfate aerosols formed from

emissions. Scavenging by cloud droplets cleanses

the troposphere of such volcanic debris within a matter

of a few weeks. Hence, it is only major eruptions whose

plumes penetrate upward into the lower stratosphere

that are capable of significantly perturbing the Earth’s

climate. A recent example of a volcanic eruption that

exerted a descernible influence on global climate is the

eruption of Mt. Pinatubo in the Philippines in June

1991 (Fig. 10.33).

Edward Walter Maunder (1851–1928) English astronomer. After working briefly in a bank he became a photographic and spec-

troscopic assistant at the Royal Greenwich Observatory, where he carried out his studies of the sun. Also involved in the debate on the

canals of Mars.

100−day smoothed

daily data

1980 1985 1990 1995 2000 2005

Year

Solar flux density (W m

–2

)

1368

1367

1366

1365

1364

1363

0.1%

Fig. 10.31 Time variations in the flux density of solar

radiation incident on the Earth as inferred from space-based

measurements from a variety of instruments. [Courtesy of

Judith Lean.]

Sunspot Number

Maunder

Minimim

Dalton

Minimim

Modern

Minimim

1600 1700 20001800 1900

Year

100

150

200

Fig. 10.32 Extended time series of sunspot number (defined

here as the number of sunspots on the sun’s (visible light)

disk plus 10 times the number of sunspot groups) showing

intervals of low solar activity. [Courtesy of Judith Lean.]

Fig. 10.33 The plume from one of a series of eruptive events

of Mt. Pinatubo that preceded the massive June 15, 1991

eruption. The event pictured here began at 0841 LT June 13.

The photograph was taken 7 min after the start of the erup-

tion. The plume of hot gases is just beginning to spread out at

the top as it approaches the tropopause. The cloud from this

eruption eventually reached altitudes in excess of 24 km, the

maximum range of the weather radar. [Courtesy of Richard

Hoblitt, U.S. Geological Survey.]

P732951-Ch10.qxd 12/09/2005 09:12 PM Page 441

442 Climate Dynamics

of the aerosol cloud from the tropics into higher

latitudes took place much more slowly. A year later

(August 1992, Fig. 10.34, bottom panel), concen-

trations of aerosols in the tropics have decreased

markedly relative to the previous year, but over the

globe as a whole, concentrations remain higher than

the background levels indicated by the top panel.

By this time the stratospheric aerosols are well

mixed globally and the gradients in concentration

are mainly a reflection of regional sources of tropos-

pheric aerosols.

For a period of several months following the

Pinatubo eruption, scattering by the aerosol cloud

reduced the flux density of direct solar radiation

incident on the Earth’s surface by 30%. Because

most of the scattering was in the forward direction,

this reduction was compensated by an increase in

Figure 10.34 shows the dispersion of the sulfate

aerosols that formed in the plume of the Pinatubo

eruption. The top panel shows the global distrribu-

tion of aerosol concentrations in August 1990, a year

before the eruption. Concentrations are enhanced

immediately downstream of local source regions

such as the Sahara Desert, but they are low else-

where. This pattern, which is representative of back-

ground conditions, is dominated by tropospheric

aerosols. The middle panel shows concentrations in

August 1991, 2 months after the eruption. Compared

to the previous August, concentrations throughout

the tropics are strongly enhanced due to the pres-

ence of sulfate aerosols in the lower stratosphere.

The rapid dispersion of stratospheric aerosols in the

east–west direction is due to the strongly sheared

zonal flow in the lower stratosphere. The dispersion

Fig. 10.34 Total concentrations of atmospheric aerosols based on NOAA/AVHRR satellite imagery. The colors represent the

monthly averages of optical depth at wavelength of 0.63



m for cloud-free pixels over the oceans, as inferred from the monochro-

matic intensity of the reflected radiation at that wavelength. [Data from J. Geophys. Res., 102(D14), 16923–16934, 1997. Copyright

1997 American Geophysical Union. Reproduced/modified by permission of American Geophysical Union. Courtesy of Alan Robock.]

60 °N

°N

°S

120

°W

°E

120

°E

180

°E

60 °N

40°N

°N

°S

°N

°S

0.0 0.1 0.2 0.3 0.4 0.5 0.6 >0.6

August 1990

August 1992

August 1991

Optical Depth

P732951-Ch10.qxd 12/09/2005 09:12 PM Page 442

10.3 Climate Equilibria, Sensitivity, and Feedbacks 443

–0.1

Years Years

Temperature (°C)

20–2 4–4 20–2 4–4

Fig. 10.35 (Left) Composite of global-mean surface air tem-

perature around the time of seven major volcanic eruptions.

The time of the eruption corresponds to Year 0 on the x axis

and positive values on the x axis denote time after the erup-

tion. The 2- to 3-year-long cool interval beginning at the time

of the eruption shows that the eruptions affected global-mean

temperature. (Right) The same temperature data, adjusted

to remove the temperature variability that is attributable to

El Niño, the annular mode and the PNA pattern, so as to

more clearly reveal the radiatively induced temperature change

following these eruptions. The temperature scales for the two

panels are the same. [Courtesy of David W. J. Thompson.]

the flux density of diffuse solar radiation reaching

the surface. The absorption and back-scattering by

the aerosol layer reduced the total (direct plus

diffuse) solar radiation reaching the surface by

3%. In response to the reduction in total incident

solar radiation, global-mean surface air tempera-

ture tends to be significantly lower in the wake of

major volcanic eruptions than in the long-term

average, as shown in Fig. 10.35.

The absorption of incident solar radiation by sul-

fate aerosols results in a pronounced warming of the

lower stratosphere in response to volcanic eruptions,

as shown in Fig. 10.36. The widths of the spikes in the

temperature time series suggests that the residence

time of the aerosols is on the order of 1–2 years,

consistent with Fig. 10.34 and the discussion in

Section 5.7.3.

The duration of the cooling at the Earth’s surface in

the wake of volcanic eruptions, as indicated by

Fig. 10.35, appears to be somewhat longer than the 1-

to 2-year residence time of the stratospheric aerosols.

The difference is attributable to the involvement of

the ocean mixed layer, with its large heat capacity.

Throughout the time interval that the aerosols remain

The seven largest volcanic eruptions in the modern record: Krakatoa (August 1883), Tarawera (June 1886), PeleeSoufriereSanta

Maria (May–October 1902), Katmai (June 1912), Agung (March 1963), El Chichon (April 1982), and Pinatubo (June 1991).

in the stratosphere, the Earth’s surface and the air

immediately above it tends to be anomalously cool.

Latent and sensible heat fluxes through the ocean sur-

face tend to be enhanced in response to the enhanced

air–sea temperature difference. The enhanced fluxes

reduce the amplitude of the atmospheric cooling by

extracting heat from the ocean mixed layer and

depositing it in the atmosphere. As the ocean mixed

layer loses heat it becomes anomalously cool. With the

removal of the aerosols the insolation incident upon

the troposphere returns to normal and the anomalous

fluxes at the air–sea interface reverse direction. An

additional year or two is required for the ocean mixed

layer to regain the heat that it lost while the aerosols

were present in the atmosphere. Not until the ocean

comes back into equilibrium does global-mean surface

air temperature return to its preeruption level. In a

similar manner, the exchange of heat between the

atmosphere and the ocean mixed layer damps and

delays the response of local sea-surface temperature

to the annual cycle and to random month-to-month

and year-to-year atmospheric variability.

10.3 Climate Equilibria,

Sensitivity, and Feedbacks

To gain a more quantitative understanding of the

climate response to a prescribed forcing, it will be

useful to place the concepts of climate feedbacks and

climate sensitivity into a mathematical framework.

Fig. 10.36 Time series of the globally-averaged temperature

in the lower stratosphere. The blue curve is based on the

lower stratospheric channel of the microwave sounding unit,

which is representative of the layer 15–20 km above the

Earth’s surface and the black curve is based on radiosonde

data for the same layer. [Adapted from Intergovernmental

panel on climate change, Climate Change 2001: The Scientific

Basis, Cambridge University Press, p. 121 (2001).]

Global anomaly (°C)

Ye a r

200019701960 1980 1990

El ChichonAgung Pinatubo

–2

–4

P732951-Ch10.qxd 12/09/2005 09:12 PM Page 443