temperature ( c) temperature ( c) within the oceanic mixed...

26 The Earth System

from 34 to 36 g kg�1 (or parts per thousand by mass,abbreviated as o�oo). Due to the presence of thesedissolved salts, sea water is �2.4% denser than freshwater at the same temperature.

The density � of sea water (expressed as the depar-ture from 1 in g kg�1 or o�oo) typically ranges from1.02 to 1.03. It is a rather complicated function of tem-perature T, salinity s, and pressure p; i.e., � � �(T, s, p).The pressure dependence of density in liquids is muchweaker than in gases and, for purposes of this qualita-tive discussion, will be ignored.1 As in fresh water,��T is temperature dependent, but the fact that seawater is saline makes the relationship somewhat dif-ferent: in fresh water, density increases with increasingtemperature between 0 and 4 °C, whereas in seawater, density decreases monotonically with increasingtemperature.2 In both fresh water and sea water,��T is smaller near the freezing point than at highertemperatures. Hence, a salinity change of a prescribedmagnitude �s is equivalent, in terms of its effect ondensity, to a larger temperature change �T in thepolar oceans than in the tropical oceans, as illustratedin Fig. 2.1.

Over most of the world’s oceans, the density of thewater in the wind-stirred, mixed layer is smaller, by afew tenths of a percent, than the density of the waterbelow it. Most of the density gradient tends to beconcentrated within a layer called the pycnocline,which ranges in depth from a few tens of meters to afew hundred meters below the ocean surface. Thedensity gradient within the pycnocline tends toinhibit vertical mixing in the ocean in much the samemanner that the increase of temperature with heightinhibits vertical mixing in atmospheric temperatureinversions and in the stratosphere. In particular, thepycnocline strongly inhibits the exchange of heat andsalt between the mixed layer, which is in direct con-tact with the atmosphere, and the deeper layers ofthe ocean. At lower latitudes, pycnocline is synony-mous with the thermocline (i.e., the layer in whichtemperature increases with height), but in polaroceans, haloclines (layers with fresher water aboveand saltier water below) also play an important rolein inhibiting vertical mixing. The strength and depthof the thermocline vary with latitude and season, asillustrated in the idealized profiles shown in Fig. 2.2.

Within the oceanic mixed layer, temperature andsalinity (and hence density) vary in response to

Fig. 2.1 The change in temperature of a water parcelrequired to raise the density of sea water at sea level as muchas a salinity increase of 1 g kg�1, plotted as a function of thetemperature of the parcel. For example, for sea water at atemperature of 10 °C, a salinity increase of 1 g kg�1 wouldraise the density as much as a temperature decrease of �5 °C,whereas for sea water at 0 °C the same salinity increasewould be equivalent to a temperature change of �17 °C.[Adapted from data in M. Winton, Ph.D. thesis, University ofWashington, p. 124 (1993).]

0 5 10

–20

–15

–10

–5

15Te

mpe

ratu

re in

crem

ent (

°C)

Temperature (°C)

1 The small effect of pressure upon density is taken into account through the use of potential density, the density that a submergedwater parcel would exhibit if it were brought up to sea level, conserving temperature and salinity. (See Exercise 3.54.)

2 Ice floats on lakes because the density of fresh water decreases with temperature from 0 to 4 °C. In contrast, sea ice floats becausewater rejects salt as it freezes.

0 5 10 15 200

500

1000

1500

2000

2500

3000

Mid

latit

udes

Hig

h la

titud

es

SummerWinter

Tropics

Temperature (°C)

Dep

th (

m)

Fig. 2.2 Idealized profiles of the temperature plotted as afunction of depth in different regions of the world’s oceans.The layer in which the vertical temperature gradient is strongestcorresponds to the thermocline. [From J. A. Knauss, Introductionto Physical Oceanography, 2nd Edition, p. 2, © 1997. Adapted bypermission of Pearson Education, Inc., Upper Saddle River, NJ.]

P732951-Ch02.qxd 9/12/05 7:40 PM Page 26

26 The Earth System

from 34 to 36 g kg�1 (or parts per thousand by mass,abbreviated as o�oo). Due to the presence of thesedissolved salts, sea water is �2.4% denser than freshwater at the same temperature.

The density � of sea water (expressed as the depar-ture from 1 in g kg�1 or o�oo) typically ranges from1.02 to 1.03. It is a rather complicated function of tem-perature T, salinity s, and pressure p; i.e., � � �(T, s, p).The pressure dependence of density in liquids is muchweaker than in gases and, for purposes of this qualita-tive discussion, will be ignored.1 As in fresh water,��T is temperature dependent, but the fact that seawater is saline makes the relationship somewhat dif-ferent: in fresh water, density increases with increasingtemperature between 0 and 4 °C, whereas in seawater, density decreases monotonically with increasingtemperature.2 In both fresh water and sea water,��T is smaller near the freezing point than at highertemperatures. Hence, a salinity change of a prescribedmagnitude �s is equivalent, in terms of its effect ondensity, to a larger temperature change �T in thepolar oceans than in the tropical oceans, as illustratedin Fig. 2.1.

Over most of the world’s oceans, the density of thewater in the wind-stirred, mixed layer is smaller, by afew tenths of a percent, than the density of the waterbelow it. Most of the density gradient tends to beconcentrated within a layer called the pycnocline,which ranges in depth from a few tens of meters to afew hundred meters below the ocean surface. Thedensity gradient within the pycnocline tends toinhibit vertical mixing in the ocean in much the samemanner that the increase of temperature with heightinhibits vertical mixing in atmospheric temperatureinversions and in the stratosphere. In particular, thepycnocline strongly inhibits the exchange of heat andsalt between the mixed layer, which is in direct con-tact with the atmosphere, and the deeper layers ofthe ocean. At lower latitudes, pycnocline is synony-mous with the thermocline (i.e., the layer in whichtemperature increases with height), but in polaroceans, haloclines (layers with fresher water aboveand saltier water below) also play an important rolein inhibiting vertical mixing. The strength and depthof the thermocline vary with latitude and season, asillustrated in the idealized profiles shown in Fig. 2.2.

Within the oceanic mixed layer, temperature andsalinity (and hence density) vary in response to

Fig. 2.1 The change in temperature of a water parcelrequired to raise the density of sea water at sea level as muchas a salinity increase of 1 g kg�1, plotted as a function of thetemperature of the parcel. For example, for sea water at atemperature of 10 °C, a salinity increase of 1 g kg�1 wouldraise the density as much as a temperature decrease of �5 °C,whereas for sea water at 0 °C the same salinity increasewould be equivalent to a temperature change of �17 °C.[Adapted from data in M. Winton, Ph.D. thesis, University ofWashington, p. 124 (1993).]

0 5 10

–20

–15

–10

–5

15

Tem

pera

ture

incr

emen

t (°C

)

Temperature (°C)

1 The small effect of pressure upon density is taken into account through the use of potential density, the density that a submergedwater parcel would exhibit if it were brought up to sea level, conserving temperature and salinity. (See Exercise 3.54.)

2 Ice floats on lakes because the density of fresh water decreases with temperature from 0 to 4 °C. In contrast, sea ice floats becausewater rejects salt as it freezes.

0 5 10 15 200

500

1000

1500

2000

2500

3000

Mid

latit

udes

Hig

h la

titud

es

SummerWinter

Tropics

Temperature (°C)D

epth

(m

)

Fig. 2.2 Idealized profiles of the temperature plotted as afunction of depth in different regions of the world’s oceans.The layer in which the vertical temperature gradient is strongestcorresponds to the thermocline. [From J. A. Knauss, Introductionto Physical Oceanography, 2nd Edition, p. 2, © 1997. Adapted bypermission of Pearson Education, Inc., Upper Saddle River, NJ.]

P732951-Ch02.qxd 9/12/05 7:40 PM 6

2.1 Components of the Earth System 27

exchanges of heat and water with the atmosphere.Precipitation lowers the salinity by diluting the saltsthat are present in the oceanic mixed layer, and evap-oration raises the salinity by removing fresh waterand thereby concentrating the residual salts, as illus-trated in the following example.

Exercise 2.1 A heavy tropical storm dumps 20 cmof rainfall in a region of the ocean in which the salin-ity is 35.00 g kg�1 and the mixed layer depth is 50 m.Assuming that the water is well mixed, by how muchdoes the salinity decrease?

Solution: The volume of water in a column extend-ing from the surface of the ocean to the bottom ofthe mixed layer is increased by a factor

and (ignoring the small difference between the densi-ties of salt water and fresh water) the mass of thewater in the column increases by a correspondingamount. The mass of salt dissolved in the waterremains unchanged. Hence, the salinity drops to

�

Water parcels that are not in contact with theocean surface tend to conserve temperature andsalinity as they move over long distances. Hence,water masses (layers of water extending over largeareas that exhibit nearly uniform temperature andsalinity) can be tracked back to the regions of themixed layer in which they were formed by exchangesof heat and mass with the atmosphere. Among theimportant water masses in the Atlantic Ocean, inorder of increasing density, are:

• Mediterranean outflow, which is conspicuouslywarm and saline due to the excess of evaporationover precipitation in the Mediterranean Sea.

• North Atlantic deep water (NADW), formed bythe sinking of water along the ice edge in theGreenland, Iceland, and Norwegian Seas.

• Antarctic bottom water (AABW), formed bysinking along the ice edge in the Weddell Sea.

The NADW and AABW, each marked by its owndistinctive range of temperatures and salinities, are

35.00 g of salt1.004 kg of water

� 34.86 g kg�1.

0.2 m50 m

� 4 � 10�3

both clearly evident near the bottom of the tropicalsounding shown in Fig. 2.3. The AABW is slightlycolder and fresher than the NADW. When both tem-perature and salinity are taken into account, theAABW is slightly denser than the NADW, consis-tent with its placement at the bottom of the watercolumn.

b. The ocean circulation

The ocean circulation is composed of a wind-drivencomponent and a thermohaline component. Thewind-driven circulation dominates the surface cur-rents, but it is largely restricted to the topmost fewhundred meters. The circulation deeper in the oceansis dominated by the slower thermohaline circulation.

By generating ocean waves, surface winds transferhorizontal momentum from the atmosphere intothe ocean. The waves stir the uppermost layer ofthe ocean, mixing the momentum downward. Themomentum, as reflected in the distribution of surfacecurrents shown in Fig. 2.4, mirrors the pattern of sur-face winds shown in Figs. 1.18 and 1.19, with closedanticyclonic circulations (referred to as gyres) at sub-tropical latitudes and cyclonic gyres at subpolar lati-tudes. Another notable feature of the wind-driven

33.5

34.0

34.5

35.0

35.5

36.0

36.5

0

5

10

15

20T

empe

ratu

re (

°C)

Salinity (g kg–1)

5045

4035 20

1816141210

9876

54

3

2.5

1.5

2

AABWNADW

26.5

26.0

25.5

25.0

24.5

24.0

27.0

27.5

28.0

28.5

29.0

Fig. 2.3 Vertical sounding of water temperature and salinityin a vertical sounding in the subtropical Atlantic Ocean.Numbers along the sounding indicate depths in hundreds ofmeters. Potential (i.e., pressure-adjusted) density (in o�oo) isindicated by the contours. Characteristic temperature andsalinity ranges for North Atlantic deep water (NADW) andAntarctic bottom water (AABW) are indicated by shading.[Reprinted from Seawater: Its Composition, Properties and Behavior,The Open University in association with Pergamon Press, p. 48(1989), with permission from Elsevier.]

P732951-Ch02.qxd 9/12/05 7:40 PM Page 27

28 The Earth System

circulation is the west-to-east Antarctic circumpolarcurrent along 55 °S, the latitude of the Drake passagethat separates Antarctica and South America.Velocities in these wind-driven currents are typicallyon the order of 10 cm s�1, a few percent of thespeeds of the surface winds that drive them, but inthe narrow western boundary currents such as theGulf Stream off the east coast of the United States(Figs. 2.4 and 2.5) velocities approach 1 m s�1. Therelatively warm water transported poleward by thewestern boundary currents contributes to moderat-ing winter temperatures over high latitude coastalregions.

Over certain regions of the polar oceans, water inthe mixed layer can become sufficiently dense, byvirtue of its high salinity, to break through the pycn-ocline and sink all the way to the ocean floor tobecome what oceanographers refer to as deep wateror bottom water. In some sense, these negativelybuoyant plumes are analogous to the plumes ofwarm, moist air in low latitudes that succeedin breaking through the top of the atmosphericmixed layer and continue ascending until theyencounter the tropopause. The presence of CFCs3 inNADW and AABW indicates that these water

60N

30N

0

60S

60E 120E 180 120W 60W 0

1 ms–1

K

N

H

BSS

G

Fig. 2.4 Annual mean ocean surface currents based on the rate of drift of ships. The Gulf Stream (G) and the Kuroshio Current (K)are warm, western boundary currents. The Humboldt Current (H) is the most prominent of the cold, equatorward currents driven bythe winds along the eastern flanks of the subtropical anticyclones. The westward South Equatorial Current (S) is driven by the easter-lies along the equator and the weaker eastward North Equatorial Countercurrent (N) is a response to the winds in the vicinity of theITCZ. [Data courtesy of Philip Richardson, WHOI; graphic courtesy of Todd P. Mitchell.]

Fig. 2.5 Eddies along the landward edge of the Gulf Stream,as revealed by the pattern of sea surface temperature.Temperatures range from �20 °C in the orange regions downto �6 °C in the darkest blue regions. Note the sharpness ofthe boundary and the indications of turbulent mixing betweenthe waters of the Gulf Stream and the colder LabradorCurrent to the north of it. [Based on NASA Terra�MODISimagery. Courtesy of Otis Brown.]

3 The term chlorofluorocarbons (CFCs) refers to a family of gaseous compounds that have no natural sources; first synthesized in 1928.Atmospheric concentrations of CFCs rose rapidly during the 1960s and 1970s as these gases began to be used for a widening range of pur-poses.

P732951-Ch02.qxd 9/12/05 7:40 PM Page 28


masses were in relatively recent contact with theatmosphere.

By virtue of their distinctive chemical and isotopicsignatures, it is possible to track the flow of watermasses and to infer how long ago water in variousparts of the world’s oceans was in contact with theatmosphere. Such chemical analyses indicate theexistence of a slow overturning characterized by aspreading of deep water from the high latitude sink-ing regions, a resurfacing of the deep water, and areturn flow of surface waters toward the sinkingregions, as illustrated in Fig. 2.6. The timescale inwhich a parcel completes a circuit of this so-calledthermohaline circulation is on the order of hundredsof years.

The resurfacing of deep water in the thermohalinecirculation requires that it be ventilated (i.e., mixedwith and ultimately replaced by less dense water thathas recently been in contact with the ocean surface).Still at issue is just how this ventilation occurs in thepresence of the pycnocline. One school of thoughtattributes the ventilation to mixing along slopingisopycnal (constant density) surfaces that cutthrough the pycnocline. Another school of thoughtattributes it to irreversible mixing produced by tidalmotions propagating downward into the deep oceansalong the continental shelves, and yet another to ver-tical mixing in restricted regions characterized by

strong winds and steeply sloping isopycnal surfaces,the most important of which coincides with theAntarctic circumpolar current, which lies beneath thering of strong westerly surface winds that encirclesAntarctica.

Although most of the deep and bottom watermasses are formed in the Atlantic sector, the thermo-haline circulation involves the entire world’s oceans,as illustrated in Fig. 2.7. Within the Atlantic sectoritself, the thermohaline circulation is comprised oftwo different cells: one involving NADW and theother involving AABW, as illustrated in Fig. 2.8.

Latitude

Thermocline

Eq

0

Dep

th

Pole

ice

Fig. 2.6 Idealized schematic of the thermohaline circulationin an equatorially symmetric ocean. The domain extends fromthe sea floor to the ocean surface and from equator to pole.Pink shading indicates warmer water and blue shading indi-cates colder water. The shaded arrows represent the exchangeof energy at the air–sea interface: pink downward arrows indi-cate a heating of the ocean mixed layer and blue upwardarrows indicate a cooling. The role of salinity is not specifi-cally represented in this schematic but it is the rejection ofsalt when water freezes along the ice edge that makes thewater dense enough to enable it to sink to the bottom.

Fig. 2.7 Highly simplified schematic of the thermohalinecirculation. Shading denotes regions of downwelling, blue arrowsdenote transport of bottom water, and red arrows denote thereturn flow of surface water. [Adapted from W. J. Schmitz, Jr.,“On the interbasin-scale thermohaline circulation,” Rev. Geophys.,33, p. 166, Copyright 1995 American Geophysical Union.]

South NorthEquator

Surface Water

Intermediate Water

Increased nutrients & dissolved CO2

Warm, low nutrients, & oxygenated

AABWNADW

60° 60°30° 30°

Fig. 2.8 Idealized cross section of the thermohaline circula-tion in the Atlantic Ocean. In this diagram, Intermediate Watercomprises several different water masses formed at temperatelatitudes. Note the consistency with Fig. 2.3. [Courtesy ofSteve Hovan.]

P732951-Ch02.qxd 9/12/05 7:40 PM Page 29

30 The Earth System

c. The marine biosphere

Virtually all the sunlight that reaches the surface ofthe ocean is absorbed within the topmost hundredmeters. Within this shallow euphotic zone,4 lifeabounds wherever there are sufficient nutrients, suchas phosphorous and iron, to sustain it. In regions ofthe ocean where the marine biosphere is active, theuppermost layers are enriched in dissolved oxygen (aproduct of photosynthesis) and depleted in nutrientsand dissolved carbon, as illustrated in Fig. 2.9. Phyto-(i.e., plant) plankton are capable of consuming thenutrients in the euphotic zone within a matter ofdays. Hence, the maintenance of high primary pro-ductivity (i.e., photosynthesis) requires a continualsupply of nutrients. The most productive regions ofthe oceans tend to be concentrated in regions ofupwelling, where nutrient-rich sea water from belowthe euphotic zone is first exposed to sunlight.

Nutrients consumed within the euphotic zone byphytoplankton return to the deeper layers of theoceans when marine plants and animals that feedon them die, sink, and decompose. The continualexchange of nutrients between the euphotic zone andthe deeper layers of the ocean plays an important rolein the carbon cycle, as discussed in Section 2.3. The dis-tribution of upwelling, in turn, is controlled by the pat-tern of surface winds discussed earlier. The distributionof ocean color (Fig. 2.10) shows evidence of high bio-logical productivity and, by inference, upwelling

• beneath cyclonic circulations such as Aleutianand Icelandic lows,

• along the eastern shores of the oceans atsubtropical latitudes,

• in a narrow strip along the equator in theequatorial Atlantic and Pacific Oceans.

In contrast, the ocean regions that lie beneath thesubtropical anticyclones are biological deserts. Thedynamical basis for these relationships is discussed inSection 7.2.5. Through their effect in mediating thegeographical distribution of upwelling and the depthof the mixed layer, year-to-year changes in theatmospheric circulation, such as those that occur inassociation with El Niño, perturb the entire foodchain that supports marine mammals, seabirds, andcommercial fisheries.

d. Sea surface temperature

The global distribution of sea surface temperatureis shaped by both radiative and dynamical factorsrelating to the pattern of seasonally varying, clima-tological-mean surface wind field over the oceans(Fig. 1.18). Radiative heating is the dominant fac-tor. That incident solar radiation is so muchstronger in the tropics than in the polar regionsgives rise to a strong north–south temperature gra-dient, which dominates the annual-mean fieldshown in Fig. 2.11 (top).

The effects of the winds on the sea surface temper-ature pattern become more clearly apparent whenthe zonally averaged sea surface temperature at eachlatitude is removed from the total field, leaving just

4 Greek: eu-good and photic-light.

Carbon100

0D

epth

(m

)

Nutrients

EuphoticZone

Oxygen

Fig. 2.9 Idealized vertical profiles of dissolved carbon (left)and oxygen (right) in biologically active regions of the oceans.The intensity of sunlight is indicated by the depth of the shad-ing in the middle panel.

Fig. 2.10 Distribution of primary productivity in the marine andterrestrial biosphere, averaged over a 3-year period. Over theoceans the dark blue areas are indicative of very low productivityand the green and yellow areas are relatively more productive.Over land dark green is indicative of high productivity. [Imagerycourtesy of SeaWiFS Project, NASA/GSFC and ORBIMAGE, Inc.]

P732951-Ch02.qxd 9/12/05 7:40 PM Page 30


the departures from the zonal-mean, shown inFig. 2.11 (bottom). The coolness of the eastern oceansrelative to the western oceans at subtropical latitudesderives from circulation around the subtropical anti-cyclones (Fig. 1.16). The equatorward flow of cool airaround the eastern flanks of the anticyclones extractsa considerable quantity of heat from the ocean sur-face, as explained in Section 9.3.4, and drives cool,southward ocean currents (Fig. 2.4). In contrast, thewarm, humid poleward flow around their westernflanks extracts much less heat and drives warm west-ern boundary currents such as the Gulf Stream. Athigher latitudes the winds circulating around the sub-polar cyclones have the opposite effect, cooling thewestern sides of the oceans and warming the easternsides. The relative warmth of the eastern Atlantic atthese higher latitudes is especially striking.

Wind-driven upwelling is responsible for the rela-tive coolness of the equatorial eastern Pacific and

Atlantic, where the southeasterly trade winds pro-trude northward across the equator (Fig. 1.18). Wind-driven upwelling along the coasts of Chile, California,and continents that occupy analogous positions withrespect to the subtropical anticyclones, although notwell resolved in Fig. 2.11, also contributes to the cool-ness of the subtropical eastern oceans, as do thehighly reflective cloud layers that tend to develop atthe top of the atmospheric boundary layer over theseregions (Section 9.4.4).

The atmospheric circulation feels the influence ofthe underlying sea surface temperature pattern, par-ticularly in the tropics. For example, from a compari-son of Figs. 1.25 and 2.11 it is evident that theintertropical convergence zones in the Atlantic andPacific sectors are located over bands of relativelywarm sea surface temperature and that the dry zoneslie over the equatorial cold tongues on the easternsides of these ocean basins.

–6 –5 –4 –3 –2 –1 0 1 2 3 4 5 6

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30

Fig. 2.11 Annual mean sea surface temperature. (Top) The total field. (Bottom) Departure of the local sea surface tempera-ture at each location from the zonally average field. [Based on data from the U.K. Meteorological Office HadISST dataset.Courtesy of Todd P. Mitchell.]

P732951-Ch02.qxd 9/12/05 7:40 PM Page 31


related: episodic surges of a few months’ to a fewyears’ duration interspersed with much longer periodsof slow retreat.

Sea ice covers a larger area of the Earth’s surfacearea than the continental ice sheets (Table 2.1) but,with typical thicknesses of only 1–3 m, is orders of

magnitude less massive. The ice is not a continuoussurface, but a fractal field comprised of ice floes(pieces) of various of shapes and sizes, as shown inFigs. 2.14 and 2.15. The individual floes are separatedby patches of open water (called leads) that open andclose as the ice pack moves, dragged by surface winds.

Seasonal limits of the northern hemisphere packice are shown in Fig. 2.12. During winter, ice coversnot only the Arctic, but also much of the Bering Seaand the Sea of Okhotsk, but during the brief polarsummer the ice retreats dramatically and large leadsare sometimes observed, even in the vicinity of theNorth Pole. Antarctic pack ice also advances andretreats with the seasons.

The annual-mean sea ice motion, shown inFig. 2.16, is dominated by the clockwise BeaufortGyre to the north of Alaska and the transpolardrift stream from Siberia toward Greenland andSpitzbergen.5 Some ice floes remain in the Arctic fora decade or more, circulating around and around theBeaufort Gyre, whereas others spend just a year ortwo in the Arctic before they exit either through theFram Strait between Greenland and Spitzbergen orthrough the Nares Strait into Baffin Bay along thewest side of Greenland. Ice floes exiting the Arcticmake a one-way trip into warmer waters, where theyare joined by much thicker icebergs that break offthe Greenland ice sheet.

New pack ice is formed during the cold season bythe freezing of water in newly formed leads and inregions where offshore winds drag the pack ice awayfrom the coastline, exposing open water. The new icethickens rapidly at first and then more gradually asit begins to insulate the water beneath it from thesubfreezing air above. Ice thicker than a meter isformed, not by a thickening of newly formed layer of

Ocean

Shelf

ice10

0

Rate of creep (m yr –1)

10 1

Fig. 2.13 Satellite image of the Antarctic ice sheet showingrate of creep of the ice (in m year�1) on a logarithmic scale.Dots show the locations of ice core sites. Vostok, the site of theice core shown in Fig. 2.31, is indicated by the solid red dot.[Adapted with permission from Bamber, J. L., D. G. Vaughanand I. Joughin, “Widespread Complex Flow in the Interior ofthe Antarctic Ice Sheet,” Science, 287, 1248–1250. Copyright2000 AAAS. Courtesy of Ignatius Rigor.]

5 The existence of a transpolar drift stream was hypothesized by Nansen6 when he learned that debris from a shipwreck north of theSiberian coast had been recovered, years later, close to the southern tip of Greenland. Motivated by this idea, he resolved to sail a researchship as far east as possible off the coast of Siberia and allow it to be frozen into the pack ice in the expectation that it would be carriedacross the North Pole along the route suggested by Fig. 2.16. He supervised the design and construction of a research vessel, the Fram(“Forward”), with a hull strong enough to withstand the pressure of the ice. The remarkable voyage of the Fram, which began in summerof 1893 and lasted for 3 years, confirmed the existence of the transpolar drift stream and provided a wealth of scientific data.

6 Fridtjof Nansen (1861–1930). Norwegian scientist, polar explorer, statesman, and humanitarian. Educated as a zoologist. Led the firsttraverse of the Greenland ice cap on skis in 1888. The drift of his research vessel the Fram across the Arctic (1893–1896) was hailed as amajor achievement in polar research and exploration. Midway through this voyage, Nansen turned over command of the Fram to HaraldSverdrup and set out with a companion on what proved to be a 132-day trek across the pack ice with dog-drawn sledges and kayaks, reach-ing 86 °N before adverse conditions forced them to turn southward.

Sacrificed his subsequent aspirations for Antarctic exploration to serve the needs of his country and to pursue humanitarian concerns.Was instrumental in peacefully resolving a political dispute between Norway and Sweden in 1905–1906 and negotiating a relaxation ofan American trade embargo that threatened Norwegian food security during World War I. Awarded the Nobel Peace Prize in 1922 inrecognition of his extensive efforts on behalf of war refugees and famine victims.

P732951-Ch02.qxd 9/12/05 7:40 PM Page 33

34 The Earth System

ice, but by mechanical processes involving collisionsof ice floes. Pressure ridges up to 5 m in thickness arecreated when floes collide, and thickening occurswhen part of one floe is pushed or rafted on top ofanother.

When sea water freezes, the ice that forms iscomposed entirely of fresh water. The concentratedsalt water known as brine that is left behind mixeswith the surrounding water, increasing its salinity.Brine rejection is instrumental in imparting enoughnegative buoyancy to parcels of water to enable themto break through the pycnocline and sink to thebottom. Hence, it is no accident that the sinkingregions in the oceanic thermohaline circulation are inhigh latitudes, where sea water freezes.

Land snow cover occupies an even larger area ofthe northern hemisphere than sea ice and it varies

Fig. 2.15 Floes in pack ice streaming southward off the eastcoast of Greenland. The white area at the upper left is landfastice that is attached to the coast, and the black channel adja-cent to it is open water, where the mobile pack ice has becomedetached from the landfast ice. [NASA MODIS imagery.]

Scale: 3 cm/s

Fig. 2.16 Wintertime Arctic sea ice motion as inferred fromthe tracks of an array of buoys dropped on ice floes by air-craft. [Courtesy of Ignatius Rigor.]

Fig. 2.14 Ice floes and leads in Antarctic pack ice. The leadin the foreground is 4–5 m across. The floe behind it consistsof multi-year ice that may have originated as an iceberg; it isunusually thick, extending from �15 m below to �1 m abovesea level. Most of the portion of the floe that extends abovesea level is snow. At the time this picture was taken, the packice in the vicinity was under lateral pressure, as evidenced bythe fact that a pressure ridge had recently developed less than100 m away. [Photograph courtesy of Miles McPhee.]

much more widely from week to week and month tomonth than does sea ice. With the warming of theland surface during spring, the snow virtually disap-pears, except in the higher mountain ranges.

Permafrost embedded in soils profoundly influ-ences terrestrial ecology and human activities overlarge areas of Siberia, Alaska, and northern Canada.If the atmosphere and the underlying land surface

P732951-Ch02.qxd 9/12/05 7:40 PM Page 34


were in thermal equilibrium, the zones of continuousand intermittent permafrost in Fig. 2.12 would strad-dle the 0 °C isotherm in annual-mean surface airtemperature. There is, in fact, a close correspondencebetween annual-mean surface air temperature andthe limit of continuous permafrost, but the criticalvalue of surface air temperature tends to be slightlyabove 0 °C due to the presence of snow cover, whichinsulates the land surface during the cold season,when it is losing heat.

Even in the zone of continuous permafrost, the top-most few meters of the soil thaw during summer inresponse to the downward diffusion of heat from thesurface, as shown in Fig. 2.17. The upward diffusionof heat from the Earth’s interior limits the verticalextent of the permafrost layer. Because the molec-ular diffusion of heat in soil is not an efficient heattransfer mechanism, hundreds of years are requiredfor the permafrost layer to adjust to changes in thetemperature of the overlying air.

2.1.3 The Terrestrial Biosphere

Much of the impact of climate upon animals andhumans is through its role in regulating the conditionand geographical distribution of forests, grasslands,

tundra, and deserts, elements of the terrestrial(land) biosphere. A simple conceptual framework forrelating climate (as represented by annual-meantemperature and precipitation) and vegetation typeis shown in Fig. 2.18. The boundary between tundraand forest corresponds closely to the limit of thepermafrost zone, which, as noted earlier, is deter-mined by annual-mean temperature. The otherboundaries in Fig. 2.18 are determined largely by thewater requirements of plants. Plants utilize waterboth as raw material in producing chlorophyll andto keep cool on hot summer days, as describedlater. Forests require more water than grasslands, andgrasslands, in turn, require more water than desertvegetation. The water demands of any specified typeof vegetation increase with temperature.

Biomes are geographical regions with climatesthat favor distinctive combinations of plant andanimal species. For example, tundra is the dominantform of vegetation in regions in which the meantemperature of the warmest month is �10 °C, andsparse, desert vegetation prevails in regions in whichpotential evaporation (proportional to the quantityof solar radiation reaching the ground) exceedsprecipitation. The global distribution of biomes isdetermined by the insolation (i.e., the incident solarradiation) at the top of the atmosphere and by theclimatic variables:

• annual-mean temperature,• the annual and diurnal temperature ranges,• annual-mean precipitation, and• the seasonal distributions of precipitation and

cloudiness.

Annual average temperature

Winter temperature

Summertemperature

Permafrost

Active layer–5 –4 –3 –2 –1 0 1 2 3

0

10

50

Dep

th (

m)

Temperature (°C)

Fig. 2.17 Schematic vertical profile of summer and wintersoil temperatures in a region of permafrost. The depth of thepermafrost layer varies from as little as a few meters in zonesof intermittent permafrost to as much as 1 km over the coldestregions of Siberia.

Ann

ual r

ainf

all

Annual mean temperature0 °C

Tundra

Forest

Grassland

Desert

Fig. 2.18 A conceptual framework for understanding howthe preferred types of land vegetation over various parts of theglobe depend on annual-mean temperature and precipitation.

P732951-Ch02.qxd 9/12/05 7:40 PM Page 35


were in thermal equilibrium, the zones of continuousand intermittent permafrost in Fig. 2.12 would strad-dle the 0 °C isotherm in annual-mean surface airtemperature. There is, in fact, a close correspondencebetween annual-mean surface air temperature andthe limit of continuous permafrost, but the criticalvalue of surface air temperature tends to be slightlyabove 0 °C due to the presence of snow cover, whichinsulates the land surface during the cold season,when it is losing heat.

Even in the zone of continuous permafrost, the top-most few meters of the soil thaw during summer inresponse to the downward diffusion of heat from thesurface, as shown in Fig. 2.17. The upward diffusionof heat from the Earth’s interior limits the verticalextent of the permafrost layer. Because the molec-ular diffusion of heat in soil is not an efficient heattransfer mechanism, hundreds of years are requiredfor the permafrost layer to adjust to changes in thetemperature of the overlying air.

2.1.3 The Terrestrial Biosphere

Much of the impact of climate upon animals andhumans is through its role in regulating the conditionand geographical distribution of forests, grasslands,

tundra, and deserts, elements of the terrestrial(land) biosphere. A simple conceptual framework forrelating climate (as represented by annual-meantemperature and precipitation) and vegetation typeis shown in Fig. 2.18. The boundary between tundraand forest corresponds closely to the limit of thepermafrost zone, which, as noted earlier, is deter-mined by annual-mean temperature. The otherboundaries in Fig. 2.18 are determined largely by thewater requirements of plants. Plants utilize waterboth as raw material in producing chlorophyll andto keep cool on hot summer days, as describedlater. Forests require more water than grasslands, andgrasslands, in turn, require more water than desertvegetation. The water demands of any specified typeof vegetation increase with temperature.

Biomes are geographical regions with climatesthat favor distinctive combinations of plant andanimal species. For example, tundra is the dominantform of vegetation in regions in which the meantemperature of the warmest month is �10 °C, andsparse, desert vegetation prevails in regions in whichpotential evaporation (proportional to the quantityof solar radiation reaching the ground) exceedsprecipitation. The global distribution of biomes isdetermined by the insolation (i.e., the incident solarradiation) at the top of the atmosphere and by theclimatic variables:

• annual-mean temperature,• the annual and diurnal temperature ranges,• annual-mean precipitation, and• the seasonal distributions of precipitation and

cloudiness.

Annual average temperature

Winter temperature

Summertemperature

Permafrost

Active layer–5 –4 –3 –2 –1 0 1 2 3

0

10

50D

epth

(m

)

Temperature (°C)

Fig. 2.17 Schematic vertical profile of summer and wintersoil temperatures in a region of permafrost. The depth of thepermafrost layer varies from as little as a few meters in zonesof intermittent permafrost to as much as 1 km over the coldestregions of Siberia.

Ann

ual r

ainf

all

Annual mean temperature0 °C

Tundra

Forest

Grassland

Desert

Fig. 2.18 A conceptual framework for understanding howthe preferred types of land vegetation over various parts of theglobe depend on annual-mean temperature and precipitation.

P732951-Ch02.qxd 9/12/05 7:40 PM Page 35

36 The Earth System

Insolation and climate at a given location, in turn, aredetermined by latitude, altitude, and position withreference to the land–sea configuration and terrain.The combined influence of altitude upon tempera-ture (Fig. 1.9), terrain upon precipitation (Fig. 1.25),and local terrain slope upon the incident solar radia-tion (Exercise 4.16) gives rise to a variegated distri-bution of biomes in mountainous regions.

Several different systems exist for assigning bio-mes, each of which consists of a comprehensive set ofcriteria that are applied to the climate statistics foreach geographical location.7 The “ground truth” forsuch classification schemes is the observed distribu-tion of land cover, as inferred from ground-basedmeasurements and high-resolution satellite imagery.An example is shown in Fig. 2.19.

The state of the terrestrial biosphere feeds backupon the climate through its effects on

• the hydrologic cycle: for example, duringintervals of hot weather, plants control theirtemperatures by evapo-transpiration (i.e., by

giving off water vapor through their leaves orneedles). Energy derived from absorbed solarradiation that would otherwise contribute toheating the land surface is used instead toevaporate liquid water extracted from the soilby the roots of the plants. In this manner, thesolar energy is transferred to the atmospherewithout warming the land surface. Hence, onhot summer days, grass-covered surfaces tendto be cooler than paved surfaces and vegetatedregions do not experience as high dailymaximum temperatures as deserts andurban areas.

• the local albedo (the fraction of the incidentsolar radiation that is reflected, without beingabsorbed): for example, snow-covered tundra ismore reflective, and therefore cooler during thedaytime, than a snow-covered forest.

• the roughness of the land surface: wind speeds inthe lowest few tens of meters above the groundtend to be higher over bare soil and tundra thanover forested surfaces.

Evergreen Needleleaf Forest

Evergreen Broadleaf Forest

Deciduous Needleleaf Forest

Deciduous Broadleaf Forest

Mixed Forest

Closed Shrublands

Open Shrublands

Woody Savannas

Savannas

Grasslands

Permanent Wetlands

Croplands

Urban and Built Up

Cropland/Natural Vegetation

Snow or Ice

Barren or Sparsely Vegetated

Water

Fig. 2.19 Global land cover characterization, as inferred from NASA AVHRR NDVI satellite imagery and ground-based datarelating to ecological regions, soils, vegetation, land use, and land cover. [From USGS Land Processes DAAC.]

7 These systems are elaborations of a scheme developed by Köppen8 a century ago.8 Wladimir Peter Köppen (1846–1940) German meteorologist, climatologist, and amateur botanist. His Ph.D. thesis (1870) explored the

effect of temperature on plant growth. His climate classification scheme, which introduced the concept of biomes, was published in 1900. Formany years, Köppen’s work was better known to physical geographers than to atmospheric scientists, but in recent years it is becoming morewidely appreciated as a conceptual basis for describing and modeling the interactions between the atmosphere and the terrestrial biosphere.

P732951-Ch02.qxd 9/12/05 7:40 PM Page 36


2.1.4 The Earth’s Crust and Mantle

The current configuration of continents, oceans, andmountain ranges is a consequence of plate tectonicsand continental drift.9 The Earth’s crust and mantlealso take part in chemical transformations that medi-ate the composition of the atmosphere on timescalesof tens to hundreds of millions of years.

The Earth’s crust is broken up into plates that floatupon the denser and much thicker layer of porousbut viscous material that makes up the Earth’s man-tle. Slow convection within the mantle moves theplates at speeds ranging up to a few centimeters peryear (tens of kilometers per million years). Platesthat lie above regions of upwelling in the mantle arespreading, whereas plates that lie above regions ofdownwelling in the mantle are being pushed together.Earthquakes tend to be concentrated along plateboundaries.

Oceanic plates are thinner, but slightly denser thancontinental plates so that when the two collide, theocean plate is subducted (i.e., drawn under the con-tinental plate) and incorporated into the Earth’smantle, as shown schematically in Fig. 2.20. Rocksin the subducted oceanic crust are subjected toincreasingly higher temperatures and pressures asthey descend, giving rise to physical and chemicaltransformations.

Collisions between plate boundaries are oftenassociated with volcanic activity and with the upliftof mountain ranges. The highest of the Earth’s moun-tain ranges, the Himalayas, was created by folding ofthe Earth’s crust following the collision of the Indianand Asian plates, and it is still going on today. TheRockies, Cascades, and Sierra ranges in westernNorth America have been created in a similarmanner by the collision of the Pacific and NorthAmerican plates. These features have all appearedwithin the past 100 million years.

Oceanic plates are continually being recycled. ThePacific plate is being subducted along much of theextent of its boundaries, while new oceanic crust isbeing formed along the mid-Atlantic ridge as magma

upwelling within the mantle rises to the surface,cools, and solidifies. As this newly formed crustdiverges away from the mid-Atlantic ridge, the floorof the Atlantic Ocean is spreading, pushing otherparts of the crust into the spaces formerly occupiedby the subducted portions of the Pacific plate. Asthe Atlantic widens and the Pacific shrinks, the con-tinents may be viewed as drifting away from theAtlantic sector on trajectories that will, in 100–200million years, converge over what is now the mid-Pacific. A similar congregation of the continentalplates is believed to have occurred about 200 millionyears ago, when they were clustered around the cur-rent position of Africa, forming a supercontinentcalled Pangaea (all Earth).

Some of the material incorporated into the mantlewhen plates are subducted contains volatile sub-stances (i.e., substances that can exist in a gaseousform, such as water in hydrated minerals). As thetemperature of these materials rises, pressure builds

9 The theory of continental drift was first proposed by Alfred Wegener10 in 1912 on the basis of the similarity between the shapes ofcoastlines, rock formations, and fossils on the two sides of the Atlantic. Wegener’s radical reinterpretation of the processes that shaped theEarth was largely rejected by the geological community and did not become widely accepted until the 1960s, with the advent of geomag-netic evidence of sea-floor spreading.

10 Alfred Wegener (1880–1930). German meteorologist, professor at University of Graz. Began his career at the small University of Marburg.First to propose that ice particles play an important role in the growth of cloud droplets. Set endurance record for time a aloft in a hot air balloon(52 h) in 1906. Played a prominent role in the first expeditions to the interior of Greenland. Died on a relief mission on the Greenland icecap.The Alfred Wegener Institute in Bremerhaven is named in his honor. Son-in-law of Vladimir Köppen and co-authored a book with him.

Mid-ocean ridge

Oceanic plate

Mantle convection

Sea floorspreading

Subduction

Fig. 2.20 Schematic showing subduction, sea floor spreading,and mountain building. [Adapted by permission of PearsonEducation, Inc., Upper Saddle River, NJ. Edward J. Tarbuck,Frederick K. Lutgens and Dennis Tasa, Earth: An Introduction toPhysical Geology, 8th Edition, © 2005, p. 426, Fig. 14.9.]

P732951-Ch02.qxd 9/12/05 7:40 PM Page 37

40 The Earth System

2. In accordance with (2.1), the geographicaldistributions of E � P and Tr in Fig. 2.21 aresimilar. The agreement is noteworthy becausethe measurements used in constructing thesetwo maps are entirely different. Thedistribution of Tr is constructed from dataon winds and atmospheric water vaporconcentrations, without reference toevaporation and precipitation.

Also of interest is the time-dependent hydrologicmass balance over land for a layer extending fromthe land surface downward to the base of the deepestaquifers. In this case

(2.2)

where is the area averaged storage of water withinsome prescribed region and the transport terminvolves the inflow or outflow of water in riversand subsurface aquifers. For the special case of a

St

dStdt

� P � E � T

land-locked basin from which there is no inflow oroutflow of surface water, the transport term vanishes,and

(2.3)

Hence the storage of water within the basin, which isreflected in the level of the lake into which the riverswithin the basin drain, increases and decreases inresponse to time variations in

Figure 2.22 shows how the level of the Great SaltLake in the reat Basin of the western United Stateshas varied in response to variations in precipitation.From the time of its historic low12 in 1963 to the timeof its high in 1987, the level of the Great Salt Lakerose by 6.65 m, the area of the lake increased by afactor of 3.5, and the volume increased by a factorof 4. The average precipitation during this 14-yearinterval was heavier than the long-term average,but there were large, year-to-year ups and downs.It is notable that the lake level rose smoothly and

P � E.

dStdt

� P � E

40N

20N

0

20S

40S

40N

20N

0

20S

40S0 60E 120E 180 120W 60W 0

0 60E 120E 180 120W 60W 0

Divergence of Moisture of Transport

Evaporation-Precipitation

–10

–5

0

5

10

mm/day

–10

–5

0

5

10

Fig. 2.21 Terms in the annual mean mass balance of atmospheric water vapor in units of mm day�1 of liquid water. (Top) Thelocal rate of change of vertically integrated water vapor due to horizontal transport by the winds. (Bottom) Difference betweenlocal evaporation and local precipitation. If the estimates were perfect, the maps would be identical. [Based on data fromNASA’s QuikSCAT and Tropical Rain Measuring Mission (TRMM). Courtesy of W. Timothy Liu and Xiaosu Xie.]

12 The historical time series dates back to 1847.

P732951-Ch02.qxd 9/12/05 7:40 PM Page 40

2.3 The Carbon Cycle 41

monotonically despite the large year-to-year variationsin the precipitation time series. Exercises 2.11–2.13 atthe end of this chapter are designed to provide someinsight into this behavior.

2.3 The Carbon CycleMost of the exchanges between reservoirs in thehydrologic cycle considered in the previous sectioninvolve phase changes and transports of a singlechemical species, H2O. In contrast, the cycling of car-bon involves chemical transformations. The carboncycle is of interest from the point of view of climatebecause it regulates the concentrations of two of theatmosphere’s two most important greenhouse gases:carbon dioxide (CO2) and methane (CH4).

The important carbon reservoirs in the Earthsystem are listed in Table 2.3 together with theirmasses and the residence times, in the same unitsas in Table 2.2. The atmospheric CO2 reservoir isintermediate in size between the active biosphericreservoir (green plants, plankton, and the entirefood web) and the gigantic reservoirs in the Earth’scrust. The exchange rates into and out of the smallreservoirs are many orders of magnitude fasterthan those that involve the large reservoirs. Thecarbon reservoirs in the Earth’s crust have residencetimes many orders of magnitude longer than theatmospheric reservoirs, reflecting not only theirlarger sizes, but also the much slower rates at whichthey exchange carbon with the other components ofthe Earth system. Figure 2.23 provides an overviewof the cycling of carbon between the various carbonreservoirs.

Exercise 2.3 Carbon inventories are often expres-sed in terms of gigatons of carbon (Gt C), where theprefix giga indicates 109 and t indicates a metric tonor 103 kg. (Gt is equivalent to Pg in cgs units, wherethe prefix peta denotes 1015.) What is the conversionfactor between these units and the units used inTable 2.3?

1950

1955

1960

1965

1970

1975

1980

1985

1990

1995

2000

2004

60

2

4

6

8

10

8

10

12

14

Year

Pre

cipi

tatio

n (c

m m

o–1 )

Lake

Lev

el (

m)

Fig. 2.22 The black curve shows variations in the depth ofthe Great Salt Lake based on a reference level of 4170 feetabove sea level (in m). Depth scale (in m) at right. Blue barsindicate seasonal-mean precipitation at nearby Logan, Utah(in cm month�1). [Lake level data from the U.S. GeologicalSurvey. Courtesy of John D. Horel and Todd P. Mitchell.]

Table 2.3 Major carbon reservoirs in the Earth system andtheir present capacities in units of kg m�2 averaged over theEarth’s surface and their residence timesa

Reservoir Capacity Residence time

Atmospheric CO2 1.6 10 years

Atmospheric CH4 0.02 9 years

Green part of the biosphere 0.2 Days to seasons

Tree trunks and roots 1.2 Up to centuries

Soils and sediments 3 Decades to millennia

Fossil fuels 10 —

Organic C in sedimentary 20,000 2 � 108 yearsrocks

Ocean: dissolved CO2 1.5 12 years

Ocean 2.5 6,500 years

Ocean HCO� 70 200,000 years

Inorganic C in sedimentary 80,000 108 yearsrocks

a Capacities based on data in Fig. 8.3 (p. 150) of Kump, Lee R.; Kasting, James F.;Crane, Robert G., The Earth System, 2nd Edition, © 2004. Adapted by permissionof Pearson Education, Inc., Upper Saddle River, NJ.

CO2�3

Atmosphere & Oceans

Biosphere

Crust

Mantle

burial

photosynthesis

subduction

respiration & decay

weathering

volcanism

sea floor spreading

Fig. 2.23 Processes responsible for the cycling of carbonbetween the various reservoirs in the Earth system.

P732951-Ch02.qxd 9/12/05 7:40 PM Page 41

2.4 Oxygen in the Earth System 45

(CaCO3), is almost exclusively a product of themarine biosphere.

Weathering exposes organic carbon in sedimentaryrock to the atmosphere, allowing it to be oxidized,thereby completing the loop in what is sometimesreferred to as the long term inorganic carbon cycle.Currently the burning of fossil fuels is returning asmuch carbon to the atmosphere in a single year asweathering would return in hundreds of thousandsof years! The mass of carbon that exists in a formconcentrated enough to be classified as “fossil fuels”represents only a small fraction of the organic carbonstored in the Earth’s crust, but it is nearly an order ofmagnitude larger than the mass of carbon currentlyresiding in the atmosphere.

On timescales of tens to hundreds of millions ofyears, plate tectonics and volcanism play an essentialrole in renewing atmospheric CO2. This “inorganiccarbon cycle,” summarized in Fig. 2.25, involves sub-duction, metamorphism, and weathering. Limestonesediments on the sea floor are subducted into theEarth’s mantle along plate boundaries where conti-nental plates are overriding denser oceanic plates. Atthe high temperatures within the mantle, limestone istransformed into metamorphic rocks by the reaction

(2.15)

The CO2 released in this reaction eventually returnsto the atmosphere by way of volcanic eruptions. Themetamorphic rocks containing calcium in chemicalcombination with silicate are recycled in the form ofnewly formed crust that emerges in the mid-oceanridges. The metamorphism reaction (2.15), in combi-nation with weathering, and the carbonate formationreaction (2.14) form a closed loop in which carbonatoms cycle back and forth between the atmospheric

CaCO3 SiO2 : CaSiO3 CO2

CO2 reservoir and the inorganic carbon reservoir inthe Earth’s crust on a timescale of tens to hundredsof millions of years.

At times when the rate at which CO2 is injectedinto the atmosphere by volcanic eruptions exceedsthe rate at which calcium ions are made available byweathering, atmospheric CO2 concentrations increaseand vice versa. The injection rate is determined byrate of metamorphism of carbonate rocks, which, inturn, depends on the rate of plate movement alongconvergent boundaries where subduction is occur-ring. The rate of weathering, however, is proportionalto the rate of cycling of water in the atmosphericbranch of the hydrologic cycle, which increases withincreasing temperature. The fact that weatheringinvolves the chemical reaction (2.13) makes thetemperature dependence even stronger. Hence, highambient temperatures and slow plate movementsare conducive to a draw-down of atmospheric CO2and vice versa. The changes in atmospheric CO2 inresponse to imbalances between (2.14) and (2.15) ontimescales of tens of millions of years are believed tohave been quite substantial.

2.4 Oxygen in the Earth SystemEarth is unique among the planets of the solar sys-tem with respect to the abundance of atmosphericoxygen O2 and the presence of an ozone (O3) layer.Atmospheric oxygen accounts for only a very smallfraction of the “free” oxygen (i.e., oxygen not boundto hydrogen atoms in water molecules) in the Earthsystem. Much larger quantities of free oxygen arepresent in the form of oxidized minerals in sedimentsand in the crust and upper mantle. The current levelof oxidation of the Earth system as a whole is muchhigher than it was at the time when the planets firstformed.

The formation of the Earth’s molten metallic ironhad the effect of enriching oxygen concentrations inthe mantle, the source of volcanic emissions. Yet geo-logical evidence suggests that oxygen was only a traceatmospheric constituent early in the Earth’s history.Iron in sedimentary rock formations that date backmore than 2.2 billion years is almost exclusively in thepartially oxidized ferrous (FeO) form. Had substantialamounts of oxygen been present in the atmosphereand oceans at the time when these sediments formed,the iron in them would have been fully oxidized toferric oxide (Fe2O3). To account for the large concen-trations of Fe2O3 that currently reside in the Earth’s

Carbon Calcium Silicon

metamorphicrocks

W

W

limestone

limestone

ions

quartzoceans

atmosphere

M

M M

S

S

W

Fig. 2.25 Schematic of the long-term inorganic carboncycle, also referred to as the carbonate–silicate cycle. Thesymbol S denotes sedimentation, M denotes metamorphosis,and W denotes weathering.

P732951-Ch02.qxd 9/12/05 7:40 PM Page 45

2.4 Oxygen in the Earth System 47

The escape of hydrogen from the Earth’s atmos-phere is detectable (Fig. 2.26), but the rate at which itis occurring is far too small to account for the degreeto which the minerals in the Earth’s crust and mantlehave become oxidized over the lifetime of the Earth.The rate of escape is slow because the two gases thatsupply hydrogen atoms to the upper atmosphere(i.e., CH4 and H2O) are present only in concen-trations of a few parts per million by volume. Airentering the upper atmosphere from below losesmost of its water vapor as it passes through the coldequatorial tropopause (see Sections 1.3.3 and 3.5).There are indications that atmospheric methane con-centrations may have been much higher in the past,as discussed in the next section.

Photosynthesis and the escape of hydrogen oper-ate independently to liberate oxygen. The crust of alifeless planet could become highly oxidized duesolely to the action of plate tectonics. Conversely,photosynthesis could occur on a planet on which theminerals in the crust and the gases in the atmospherewere in a highly reduced state. However, of the twomechanisms, only photosynthesis is capable of pro-ducing atmospheric oxygen, and only the escape ofhydrogen from the Earth system is capable of liber-ating oxygen in the quantities required to account forthe present degree of oxidation of the minerals in thecrust and mantle.

Fig. 2.26 The corona around the Earth in this image is due tothe scattering of solar radiation by hydrogen atoms that areescaping from the atmosphere. [Image from NASA�DynamicsExplorer/Spin-Scan Auroral Imaging. Courtesy of David Catling.]

Isotopes of a given element are atoms with differ-ent numbers of neutrons in their nuclei. Unstableisotopes, such as 14C, which spontaneously changeform by radioactive decay with a known “half-life,” are used for dating ice, tree and sedimentcores, fossils, and rock samples. By comparing theabundance of the isotope in the sample with thecurrent atmospheric abundance, it is possible toinfer how long the sample has been out of contactwith the atmosphere (provided, of course, that theatmospheric abundance has not changed since thesample was deposited). Relative abundances ofstable isotopes, such as 13C, vary in accordance

with local (and in some cases regional or evenglobal) environmental conditions that prevailed atthe time when the sample was deposited. Some ofthe more widely used isotopes are the following.

• The relative abundance of deuterium (2H,or D)15 in the snow samples recovered fromice cores depends on (and hence can be usedas a proxy for) the temperature of the surfacefrom which the water vapor that condensed toform the snow was evaporated. The greaterthe relative abundance of HDO in the coresample, the warmer that evaporating surface

2.1 Isotope Abundances: Proxies for Climate Data

15 The relative abundance of D is given by

where R is a reference value. Positive (negative) values of �D are indicative of an enrichment (depletion) of D relative to the referencevalue, expressed in parts per thousand (o�oo). The same formalism applies to the isotopes in subsequent bullets.

�D(o/oo) �D�H � R

R

Continued on next page

P732951-Ch02.qxd 9/12/05 7:40 PM Page 47

2.5 A Brief History of Climateand the Earth SystemThis section describes evolution of the Earth system ona logarithmically telescoping time line, as depicted inthe bottom panel of Fig. 2.27, with subsections focusingon (1) the lifetime of the Earth, (2) the past 100 millionyears, (3) the past million years, and (4) the past 20,000years. If the life (to date) of a 20-year-old student wereviewed on an analogous and proportional time line, therespective subsections would focus on his�her entire20-year life span, the past 6 months, the past 2 days, andthe past hour.

2.5.1 Formation and Evolutionof the Earth System

The sun and the planets are believed to have formed4.5 billion years ago from the gravitational collapse ofa cold cloud of interstellar gas and dust.16 The absenceof the noble gases neon, xenon, and krypton17 in the

48 The Earth System

must have been at the time the sample wasdeposited.

• Oxygen-18 (18O) abundance in marine sedimentcores containing carbonates reflects the temper-ature of the water in the euphotic zone wherethe carbonate formation took place.The lowerthe temperature of the water, the greater therelative abundance of 18C that was incorporatedinto the shells and skeletons of the marineorganisms from which the carbonates formed.

• Worldwide 18O abundances also depend onthe volume of the continental ice sheets. 16Oevaporates more readily than 18O, so adisproportionately large fraction of 16O tendsto be incorporated into the snow that falls on,and becomes incorporated into, the ice sheets.When the ice sheets grow, ocean watersthroughout the world are enriched in 18O.Hence 18O abundance in marine sedimentcores and ice cores can be used as a proxy forice volume.

• Carbon-13 (13C) abundances in depositsof organic carbon reflect the ambientCO2 concentrations at the time thatphotosynthesis occurred. Plants prefer thelighter isotope 12C, and the higher theambient CO2 level, the more strongly theyexert this preference. Hence, low relativeabundances of 13C in organic carbondeposits are indicative of high ambientCO2 concentrations, and vice versa.

• Carbon-13 is also an indicator of the sourcesand sinks of atmospheric CO2. Emissionsfrom the decay of plants, forest fires andagricultural burning, and the consumption offossil fuels tend to be low in 13C, whereas CO2outgassed from the oceans has the same 13Cabundance as atmospheric CO2. In a similarmanner, the presence of a biospheric CO2sink should tend to raise atmospheric 13Clevels, whereas the presence of an oceanicsink should not.

2.1 Continued

01234

Years before present × 109

lifet

ime

of E

arth

Had

ean

Epo

ch

orig

in o

f life

Years before present

108 107 106 105 104 103 102109

lifet

ime

of E

arth

rise

of o

xyge

n

Pan

gaea

K-T

bou

ndar

ypl

ates

col

lide,

Him

alay

as r

ise

Pan

ama

clos

es g

ap

begi

n Q

uate

rnar

y E

poch

prev

ious

inte

rgla

cial

last

gla

cial

max

imum

You

nger

Dry

as“c

limat

ic o

ptim

um”

Med

ieva

l War

m P

erio

dLi

ttle

Ice

Age

rise

of o

xyge

n

Pan

gaea

K-T

bou

ndar

yQ

uate

rnar

y E

poch

Fig. 2.27 Time line for the history of the Earth system on alinear scale (top) and on a logarithmic scale (bottom).

16 The age of the of the Earth is inferred from a comparison of the ratios of radiogenic (formed by decay of uranium) and nonradiogenicisotopes of lead in meteorites and rocks of various ages.

17 Argon in the Earth’s atmosphere is a product of the radioactive decay of the radioactive isotope 40K in the crust.

P732951-Ch02.qxd 9/12/05 7:40 PM Page 48

2.5 A Brief History of Climate and the Earth System 49

atmospheres of the Earth and the other planets, rela-tive to their cosmic abundance, is evidence that theplanets formed from the coalescence of the dust intochunks of solid materials called planetesimals thatwere drawn together by gravitation. Present within thecondensing cloud were volatile compounds (i.e., water,methane, ammonia, and other substances with lowboiling points), mainly in the form of ices. When thesun formed, the inner part of the cloud should havewarmed, driving out most of the volatiles: hence therelatively low concentrations of these substances inthe atmospheres of the inner planets.

During the first 700–800 million years of its history,referred to by geologists as the Hadean Epoch, Earthwas still under continual bombardment by smallerplanetesimals. The heating and degassing resultingfrom impacts of these collisions should have liberatedwater vapor and other volatile substances, forming aprimordial atmosphere and the oceans. The energyreleased by the larger objects may well have been suf-ficient to entirely vaporize the oceans from time totime. The formation of the moon has been attributedto one of these impacts. The bombardment graduallysubsided and, by �3.8 billion years ago, conditions onEarth had become stable enough to allow early micro-bial life forms to develop in the oceans. Cataclysmiccollisions still occur occasionally, as evidenced by theK–T meteorite impact18 that took place only 65 millionyears ago. Earth is still being bombarded by vast num-bers of much smaller objects, as illustrated in Fig. 2.28,but their impact on the Earth system is minimal.

The emission of radiation from stars increasesgradually over their lifetime due to the increasingrate of fusion within their cores, which contract andheat up as progressively more of the hydrogen isfused into helium. The luminosity of the sun isbelieved to have increased by 30% over the lifetimeof the solar system. Geological evidence indicatesthat, with the exception of a few relatively briefintervals, the oceans have been largely free of icethroughout the Earth’s history. That the Earth’ssurface was not perpetually frozen during its earlyhistory, when the sun was relatively faint, suggeststhat its early atmosphere must have contained sub-stantially higher concentrations of greenhouse gasesat that time than it does now.

Over the lifetime of the Earth, its atmosphere hascontinuously been recycled and renewed by volcanismand plate tectonics.The makeup of present-day volcanicemissions is 80–90% steam, 6–12% CO2, 1–2% SO2,and traces of H2, CO, H2S, CH4, and N2. The relativeconcentrations of the reduced gases H2, CO, and CH4could have been much higher earlier in the Earth’s his-tory when the mantle was less oxidized than it is today.

An important milestone in the evolution of theEarth system was the rise of atmospheric oxygen.Cyano-(blue-green) bacteria capable of liberatingoxygen are believed to have been present in the oceansfor at least 3.0 and perhaps as long as 3.8 billion years,yet geological evidence indicates that oxygen did notbegin to accumulate in the atmosphere until about2.4–2.2 billion years ago. Oxygen liberated by photo-synthesis early in the Earth’s history would have beenconsumed quickly by H2 and other reduced gasesemanating from the Earth’s crust, formed in reactionssuch as (2.17) or by the oxidation of minerals exposedto the atmosphere by weathering. Only after the

18 The distinctive marker of this event is the distinctive iridium-enriched layer produced by the explosion that attended the impact. Thislayer, which is evident in sediments worldwide, occurs at the boundary between Cretaceous and Tertiary sediments in the Earth’s crust (hencethe name K–T). It has been hypothesized that this event was responsible for the extinction of many species of life forms, including dinosaurs.

10–12

10–10

10–8

10–6

10–6

108

108 1010

106

106

104

(cm) (m) (km)

103104

102

102

10–4

10–4

10–2

10–2 1

1

Year

s be

twee

n im

pact

s

Impactors on the space shuttle surface(30 µ s)

(µ m)

(30 s)

(1 year)

(10,000 years)

Shooting stars

Meteorites

Arizona crater

Sudbury, Ontario,impact feature

Diameter of striking object (m)

Fig. 2.28 Frequency of occurrence of collisions of objectswith the Earth as a function of the size of the objects.[Reprinted with permission from L. W. Alvarez, “Mass extinc-tions caused by large bolide impacts,” Physics Today, 40, p. 27.Copyright 1987, American Institute of Physics.]

P732951-Ch02.qxd 9/12/05 7:40 PM Page 49

minerals in the crust became more highly oxidizeddue to the gradual escape of hydrogen from the Earthsystem and reactions such as (2.17) slowed down,would it have it possible for oxygen liberated byphotosynthesis to begin to accumulate in the atmos-phere. In this sense, atmospheric O2 can be viewed as“surplus” oxygen in the Earth system.

An array of geological evidence indicates that therise of atmospheric oxygen, once it began, was quiterapid, with concentrations rising from less than 0.01%of the present concentration 2.4 billion years ago to atleast 1–3% of the present concentration 1.9 billionyears ago. Concurrent with the rise of oxygen camethe formation of ozone layer. Photochemical modelsbased on the equations in Section 5.7.1 indicate that

atmospheric oxygen concentrations of even a few per-cent of today’s values should have been capable ofsupporting an ozone layer thick enough to protect lifeon Earth from the harmful effects of solar ultravioletradiation.

The conditions that existed on Earth prior to therise of oxygen have been the subject of considerablespeculation. Given today’s rates of production ofmethane, it is estimated that, in the absence of atmos-pheric oxygen, methane concentrations could havebeen two or three orders of magnitude greater thantheir present concentration of �1.7 ppmv, in whichcase, methane might well have been the dominantgreenhouse gas. With higher methane concentrations,the number densities of hydrogen atoms in the upper

50 The Earth System

Geological evidence suggests that majorglaciations, extending all the way into the trop-ics, occurred three times in Earth’s history:the first around 2.2–2.4 billion years ago, con-current with the rise of oxygen,19 the secondbetween 600 and 750 million years ago, and thelatest so-called Permian Glaciation, �280 millionyears ago. Until quite recently, most climateresearchers discounted this evidence on thegrounds that had the oceans ever been in acompletely ice-covered state, they would haveremained so. These skeptics argued that thealbedo of an ice-covered planet would be so lowthat very little of the incident solar radiationwould be absorbed. Hence, the surface of theEarth would have been so cold that the ice couldnot have melted. Recently, this argument hasbeen called into question by proponents of thefollowing “freeze-fry scenario.”

i. During an unusually cold period, asufficiently large fraction of the Earth’ssurface becomes ice covered so that theice-albedo feedback mechanism describedin Section 10.3 renders the climate unstable:the expanding ice cover cools the Earth’ssurface, the cooling causes the ice to expand

2.2 “Snowball Earth”

Continued on next page

19 It has been hypothesized that the decline in atmospheric methane concentrations brought about by the rise of oxygen might haveprecipitated a sudden global cooling at this time.

Fig. 2.29 Artist’s conception of the onset of a world-wide glaciation. [Courtesy of Richard Peltier.]

still farther, and the process continues untilthe entire Earth becomes ice covered, asdepicted schematically in Fig. 2.29.

ii. During the ensuing snowball Earth phase,the carbonate formation reaction (2.11)cannot occur because the oceans are frozen.

P732951-Ch02.qxd 9/12/05 7:40 PM Page 50


atmosphere could have been orders of magnitudehigher than their current values. Under those condi-tions, large numbers of hydrogen atoms would haveescaped to space, gradually raising the level of oxida-tion of the Earth system. The abrupt transition of theatmosphere from an anoxic state, with relatively highconcentrations of CH4, CO, H2S, and other reducedgases, to a more oxidized state was marked by theextinction of anaerobic life forms that are intolerantof O2. The rise of O2 set the stage for the evolution ofmore complex life forms.

2.5.2 The Past 100 Million Years

During the Cretaceous epoch, which ended �65 mil-lion years ago, surface air temperatures were sub-stantially higher than they are today, especially at thehigher latitudes. This view is supported by the discov-ery of remains of dinosaurs and lush tropical plantsdating back to that time in Siberia, Canada, andother subarctic sites. Geological evidence indicatesthat atmospheric concentrations of CO2 were aboutan order of magnitude higher at that time than theyare today. The Cretacous epoch was followed by anextended interval of cooling and declining CO2 con-centrations, culminating in Pleistocene glaciation,20

which began around 2.5 million years ago.The cooling that set the stage for the Pleistocene

glaciation is widely attributed to the role of plate

tectonics in regulating the reactions (2.14) and (2.15)in the carbonate–silicate cycle. Geological evidenceindicates that the rate of movement of the plates hasslowed down over the past 100 million years. Areduced rate of ingestion of limestone sediments intothe mantle implies reduced metamorphism, whichwould favor reduced volcanic emissions of CO2.Meanwhile, the rise of the Himalayas following thecollision of the Indian and Asian plates (Fig. 2.30) isbelieved to have increased the rate of weathering ofCaSiO3 rocks, making more Ca2 ions available for

However, continental drift continues and,with it, metamorphism and volcanicemissions of CO2 into the atmosphere. Inthe absence of a carbon sink, atmosphericCO2 concentrations increase.

iii. Eventually, the combination of the increasinggreenhouse effect and the blackening ofparts of the ice sheets by windblown dustraise the surface temperature up to thethreshold value at which ice in the tropicaloceans begins to melt. Once this processbegins, the ice-albedo feedback exerts apowerful warming influence, which abruptlyflips the Earth system into an ice-free state.

iv. With the thawing of the oceans, carbonateproduction resumes, but, compared to thereaction time of the cryosphere in (iii), thisis a slow process, limited by the rate at whichweathering supplies Ca2 ions. Hence, forseveral million years, the Earth systemresides in the hothouse phase, with global-mean temperatures initially as high as 50 °C.

The occurrence of an extended hothouse phase issupported by the existence of rock formations(i.e., banded iron formations and cap carbonates)suggestive of a period of very high temperaturesaround the time of the Permian glaciation.

2.2 Continued

20 Pleistocene (from the Greek: pleistos: most, and ceno: new) in reference to geological sediments. The Pleistocene epoch and thesubsequent Holocene (all new) epoch comprise the Quaternary period in the Earth’s history.

Fig. 2.30 Continental configuration 65 million years ago, atthe end of the Cretaceous epoch. Note the separationbetween the Indian and Eurasian plates. [Courtesy of the U.S.Geological Survey.]

P732951-Ch02.qxd 9/12/05 7:40 PM Page 51

limestone formation in the oceans, thereby accelerat-ing the removal of CO2 from the atmosphere andoceans. A decreasing source and increasing sink ofatmospheric CO2 would account for the apparentdecline in atmospheric CO2 levels, and the conse-quent weakening of the greenhouse effect is consis-tent with the observed cooling.

Another factor that contributed to the cooling wasglaciation of the Antarctic continent 15–30 millionyears ago as it drifted into higher latitudes, whichwould have increased the fraction of the incidentsolar radiation reflected back to space. Other impor-tant milestones in the drift of the continents towardtheir present configuration were the opening of theDrake passage in the southern hemisphere 15–30million years ago, which gave rise to the formation ofthe Antarctic circumpolar current, and the joiningof the North and South American continents at theisthmus of Panama �3 million years ago. It has beensuggested that these events could have caused majorreorganizations of the oceanic thermohaline circula-tion, resulting in a reduction of the poleward heatflux in the North Atlantic, thereby accelerating thecooling of the Arctic.

2.5.3 The Past Million Years

The past 2.5 million years have been marked byclimatic swings back and forth between extended

glacial epochs in which thick ice sheets coveredlarge areas of North America, northern Europe, andSiberia and shorter interglacial epochs such as thepresent one, in which only Antarctica (and some-times Greenland) remain ice covered. Carbon andoxygen isotopes (see Box 2.1) and the remains ofliving organisms buried in marine sediments inanoxic ocean basins around the world reveal a greatdeal about the history of the pronounced climaticswings during this so-called Quaternary period ofEarth’s history.

A more accurate and detailed history of theclimate swings of the past few hundred thousandyears, as revealed by an ice core extracted fromthe dome of the Antarctic ice sheet, is shown inFig. 2.31. The temperature variations shown inFig. 2.31 are inferred from changes in concen-trations of deuterium in the ice, as explained inBox 2.1. Atmospheric concentrations of CO2 andCH4 are inferred from microscopic bubbles of airthat became trapped in the ice as the snow fromwhich it was formed was compressed and consoli-dated. Because these gases tend to be well mixed,their concentrations in the cores are indicative ofglobal conditions. The chemical signature of thedust in the cores provides a basis for identifying itwith one or more specific source regions such asthe Gobi Desert, where fine particles of soil areexposed directly to strong winds.

52 The Earth System

200

400

600

800

CH

4 (p

ptv)

150

200

250

300

CO

2 (p

pmv)

0 50 100 150 200 250 300 350 400−10

−5

0

5

Age (thousands of years)

Tem

pera

ture

cha

nge

(°C

)

Fig. 2.31 Comparison of methane, carbon dioxide, and estimated temperature (from oxygen and deuterium isotope ratios) fromthe Vostok ice core, Antarctica, over the last 440 thousand years. The location of Vostok is indicated by the red dot in Fig. 2.13.Note that the time axis runs from right to left. [Adapted from J. R. Petit et al., “Climate and atmospheric history of the past420,000 years from the Vostok ice core, Antarctica.” Nature, 399, p. 431, 1999. Courtesy of Eric Steig.]

P732951-Ch02.qxd 9/12/05 7:40 PM Page 52


Based on the analysis of ice core records likethe ones shown in Fig. 2.31 and comparisons withnumerous (but less highly resolved and accuratelydated) marine sediment cores, it is clear that ontimescales of tens of thousands of years or longer,temperature and a number of other climate-related parameters vary coherently with oneanother and that these variations are global inextent. Global temperatures cooled at an irregularrate during the extended glacial epochs and rosemuch more rapidly at the beginning of the inter-glacials, which have been recurring at intervals ofroughly 100,000 years. The last glacial maximumoccurred around 20,000 years ago. AtmosphericCO2 and CH4 concentrations have risen and fallensynchronously with temperature, and the colderepochs have been dustier, perhaps because ofhigher wind speeds or drier conditions over thesource regions.

Figure 2.32 contrasts the extent of the northernhemisphere continental ice sheets at the time of thelast glacial maximum with their current extent.Parts of Canada were covered by ice as thick as3 km. As a consequence of the large amount ofwater sequestered in the ice sheets, the global sealevel was �125 m lower than it is today. The con-centration of atmospheric CO2 was �180 ppm, ascompared with a mean value of �260 ppm duringthe current interglacial prior to the industrialrevolution. Hence, the Earth’s albedo must havebeen higher at the time of the last glacial maximum

than it is today and the greenhouse effect musthave been weaker, both of which would havefavored lower surface temperatures. Temperaturesin Greenland were �10 °C lower than they aretoday and tropical temperatures are estimated tohave been �4 °C lower.

There is evidence of a small time lag betweenfluctuations in atmospheric CO2 concentrationsand fluctuations in the volume of ice stored in thecontinental ice sheets, with ice volume leadingCO2. Hence, it would appear that the cause of thesefluctuations is intimately related to the growth andshrinkage of the continental ice sheets. The CO2fluctuations represent a positive feedback thatamplifies the temperature contrasts between glacialand interglacial epochs, as discussed further inSection 10.3.2.

The pronounced climatic swings during theQuaternary period are believed to be driven by sub-tle variations in the Earth’s orbit that affect the sum-mer insolation (i.e., the average intensity of incidentsolar radiation) at high latitudes of the northernhemisphere. During intervals when the summer inso-lation is relatively weak, snow deposited during win-ter does not completely melt, leaving a residual,which, over a time span of thousands of years, accu-mulates to form thick ice sheets. The high reflectivityof the growing ice sheets exacerbates the coolness ofthe summers, amplifying the orbital forcing. Based onthe same reasoning, it is believed that the continentalice sheets are most prone to melting during periods

Fig. 2.32 The extent of the northern hemisphere continental ice sheets at the time of the last glacial maximum 20,000 yearsago (left) as compared with their current extent (right). [Courtesy of Camille Li.]

P732951-Ch02.qxd 9/12/05 7:40 PM Page 53

when summer insolation at high northern latitudes isstrong.21

The orbital variations believed to be responsiblefor these pronounced climatic swings involve:

i. 100,000-year cycle in eccentricity (the degreeof ellipticity, defined as the distance from thecenter to either focus of the ellipse divided bythe length of the major axis), which rangesfrom 0 to 0.06 and is currently 0.017,

ii. 41,000-year cycle in the obliquity (i.e., the tiltof the Earth’s axis of rotation relative to theplane of the Earth’s orbit) which ranges from22.0° to 24.5° and is currently 23.5°, and

iii. 23,000- and 19,000-year cycles in the precessionof the Earth’s orbit. As a result of theprecession cycle, the day of the year on whichthe Earth is closest to the sun (currentlyJanuary 3) progresses through the year at arate of �1.7 calendar day per century.24

Figure 2.33 shows a schematic visual representationof these three types of orbital perturbations. Whenthe eccentricity and obliquity are both near thepeaks of their respective cycles, summertime insola-tion at 65 °N varies by up to �20% between theextremes of the precession cycle (see Exercise 4.19).

In Fig. 2.34 a time series of the rate of growth ofthe continental ice sheets, as inferred from oxygen-18 concentrations in marine sediment cores, iscompared with a time series of summertime insola-tion over high latitudes of the northern hemi-sphere, as inferred from orbital calculations. Thedegree of correspondence between the series is

quite striking. The fit can be made even betterby synchronizing the maxima and minima in theoxygen-18 record with nearby features in the inso-lation time series.25

2.5.4 The Past 20,000 Years

The transition from the last glacial to the currentinterglacial epoch was dramatic. The ice sheetsstarted shrinking around 15,000 years ago. By 12,000years ago the Laurentide ice sheet was pouring hugevolumes of melt water into newly formed lakes andrivers, setting the stage for a series of flood eventsthat shaped many of the features of today’s land-scape. Around this time the emergence from the iceage was interrupted by an �800 year relapse into iceage conditions, an event referred to by geologists asthe Younger Dryas.26 The signature of the Younger

54 The Earth System

21 Many of the elements of the orbital theory of the ice ages are embodied in works of James Croll,22 published in 1864 and 1875. In1920, Milutin Milankovitch23 published a more accurate time series of insolation over high latitudes of the northern hemisphere based onnewly available calculations of the variations in the Earth’s orbit. Wladimir Köppen and Alfred Wegener included several ofMilankovitch’s time series in their book Climates of the Geologic Past (1924). The idea that summer is the critical season in determiningthe fate of the continental ice sheets is widely attributed to Köppen. Analysis of extended sediment core records, which did not becomewidely available until the 1970s, has provided increasingly strong support for orbital theory.

22 James Croll (1821–1890). Largely self-educated Scottish intellectual. Variously employed as a tea merchant, manager of a temperancehotel, insurance agent, and janitor at a museum before his achievements earned him an appointment in the Geological Survey of Scotlandand substantial scientific recognition.

23 Milutin Milankovitch (1879–1958). Serbian mathematician. Professor, University of Belgrade.24 Evidence of the precession cycle dates back to the Greek astronomer, Hipparchus, who inferred, from observations made more than

a century apart, that the axis around which the heavens rotate was slowly shifting.25 The relationship between depth within the core and time depends on the rate of sedimentation, which varies from one site to another

and is not guaranteed to be linear. Hence, in assigning dates on the features in the cores, it is necessary to rely on supplementary information.The reversal in the polarity of the Earth’s magnetic field, which is known to have occurred 780,000 years ago and is detectable in the cores,provides a critical “anchor point” in dating the sediment core time series.

26 Dryas is a plant that currently grows only in Arctic and alpine tundra, fossil remains of which are found in a layer of sediments fromnorthern Europe deposited during this interval. Younger signifies the topmost (i.e., most recent layer in which dryas is present).

Sun Earth

Fig. 2.33 Schematic of the Earth’s orbital variations. Theprecession cycle in the tilt of the Earth’s axis is representedby a single cone; the cycle in the obliquity of the axis isrepresented by the presence of two concentric cones, and theextrema in the ellipticity of the orbit are represented by thepair of ellipses. The figure is not drawn to scale. [Adaptedfrom J. T. Houghton, Global Warming : The Complete Briefing,2nd Edition, Cambridge University Press, p. 55 (1997).]

P732951-Ch02.qxd 9/12/05 7:40 PM Page 54


Dryas shows up clearly in the Greenland ice core(Fig. 2.35) and in European proxy climate records,but whether it was truly global in extent, such as theclimatic swings on timescales of 10,000 years andlonger, is still under debate.

The timescale of the Younger Dryas event ismuch shorter than that of the orbital cycles dis-cussed in the previous section, so it is not clear whatcaused it. One hypothesis is that the sudden fresh-ening of the surface waters over higher latitudes ofthe North Atlantic, caused by a surge of glacial meltwater out of the St. Lawrence River, precipitated ashutdown of the oceanic thermohaline circulation,which reduced the poleward heat transport by theGulf Stream that warms Greenland and northernEurope.

High-resolution analysis of the Greenland ice coreindicates that the Younger Dryas event endedabruptly �11,700 years ago. The current interglacial,referred to in the geology literature as the Holoceneepoch, has not witnessed the large temperatureswings that characterized the previous glacial epoch.However, even the minor swings appear to have hadimportant societal impacts. For example, during therelatively cold interval from the 14th through most ofthe 19th century, popularly referred to as the Little

–700

–800

–600

–500

–400

–300

–200

–100 0

Ice

volu

me

d/dt

(ic

e vo

lum

e)

Thousands of years ago

Inso

latio

n an

omal

y (W

m–2

)

–150

–100

–50

0

50

100

Fig. 2.34 Top time axis: global ice volume (arbitrary scale) as inferred from 18O concentrations in a large collection of marinesediment cores. Bottom time axis: the time rate of change of ice volume (arbitrary scale), obtained by differentiating the curvein the top panel (blue-green curve) and summer insolation at the top of the atmosphere due to variations in the Earth’s orbit(red curve). The dating of the sediment core has been carried out without reference to the insolation time series. [The datingof the sediment core time series is based on estimates of Peter Huybers; the insolation time series is based on orbital calculationsusing the algorithm developed by Jonathan Levine. Figure courtesy of Gerard Roe.]

Tem

pera

ture

(°C

)

22

24

26

28

30

0 5 10 15 20−45

−40

−35

−30

Age (thousands of years)

δ18O

(pe

r m

il)

Fig. 2.35 Variations in sea surface temperature change inCariaco Basin (Venezuela) sediment (top) and oxygen-18 inGreenland ice cores (bottom) during the last 20,000 years. Seasurface temperature is inferred from Mg�Ca ratios. 18O valuesfrom the ice cores are indicative of air temperature over andaround Greenland. Note that the time axis runs from rightto left. [Based on data presented by Lea, D. W., D. K. Pak,L. C. Peterson and K. A. Hughen, “Synchroneity of tropical andhigh-latitude Atlantic temperatures over the last glacial termi-nation,” Science, 301, p. 1364 (2003) and Grootes, P. M.,M. Stuiver, J. W. C. White, S. Johnsen and J. Jouzel, “Comparisonof oxygen isotope records from the GISP2 and GRIP Greenlandice cores,” Nature, 366, p. 552 (1993). Courtesy of Eric Steig.]

P732951-Ch02.qxd 9/12/05 7:40 PM Page 55

Ice Age, the Viking colony in Greenland failed, thepopulation of Iceland declined substantially, andfarms were abandoned in parts of Norway and theAlps.

The global distribution of rainfall has variedsignificantly during the Holocene epoch. Fossilscollected from lake beds indicate that areas of whatis now the Sahara Desert were vegetated and, insome cases, swampy 6000 years ago. Areas of theMiddle East where grains were first cultivated havesubsequently become too dry to support extensiveagriculture, and aqueducts that the Romans builtin north Africa seem out of place in the context oftoday’s climate. Superimposed upon such long-termtrends are variations on timescales of decades tocentury, which have influenced the course of humanevents. For example, collapses of the AkkadianEmpire in the Middle East and the Mayan andAnasazi civilizations in the New World have been

attributed to the failure of those societies to adapt totrends toward drier climates.

2.6 Earth: The Habitable PlanetPhotographs of Earth and its neighbors in the solarsystem are shown in Fig. 2.36 and pertinent astro-nomical and atmospheric data are shown in Table 2.5.That Earth is the only planet in the solar system onwhich advanced life forms have evolved is due to avery special combination of circumstances.

i. The range of surface temperatures onEarth has allowed for the possibility ofoceans that have remained unfrozenthroughout most of Earth’s history. The oceanshave provided (a) an essential pathway inthe carbonate– silicate cycle that sequesterslarge amounts of carbon in the Earth system,

56 The Earth System

Fig. 2.36 Venus, Earth, Mars, and Jupiter from space. Venus and Jupiter are cloud covered. Not shown to scale. [Photographscourtesy of NASA.]

Table 2.5 Astronomical and atmospheric data for Earth and neighboring planetsa

Parameter Venus Earth Mars Jupiter

Radius (km � 103) 6,051 6,371 3390 66,911

Gravity (m s�2) 8.87 9.80 3.71 24.79

Distance from sun (AU) 0.72 1.000 1.524 5.20

Length of year (Earth years) 0.615 1.000 1.88 11.86

Length of day (Earth days) 117 1.000 1.027 0.41

Orbital eccentricity 0.0067 0.0167 0.093 0.049

Orbital obliquity 2.36 23.45 25.19 3.13

Dominant constituent (% by volume) CO2 (96.5) N2 (78.1) CO2 (95.3) H2 (90)

Secondary constituent (% by volume) N2 (3.5) O2 (21) N2 (2.7) He (10)

Surface pressure (hPa) 92,000 997 8b ��106

Surface temperature (K) 737 288 210

Diurnal temperature range (K) �0 10 40a Based on Planetary Fact Sheets on NASA Web site; Mars surface data based on records at the Viking 1 Lander site.b Varies seasonally from 7.0 hPa during the austral winter, when Mars is farthest from the sun, to 9.0 hPa during the austral summer.

P732951-Ch02.qxd 9/12/05 7:40 PM Page 56

only three layers: a troposphere, an isothermalmesosphere, and a thermosphere.27

The long-range outlook for the habitability ofEarth is not good. If stellar evolution follows itsnormal course, by a billion years from now theluminosity of the sun will have become strongenough to evaporate the oceans. The impacts ofhuman activities upon the Earth system, whilearguably not as disastrous, are a million times moreimminent.

Exercises2.7 Explain or interpret the following

(a) In the atmosphere, most of the deepconvection occurs at low latitudes, whereasin the oceans convection occurs at highlatitudes.

(b) The salinity of the oceanic mixed layer isrelatively low beneath the areas of heavyrainfall such as the ITCZ.

(c) The outflow from the Mediterranean Seadoes not rise to the surface of the NorthAtlantic, even though it is warmer than thesurface water.

(d) Variations in sea ice extent do not affectglobal sea level.

(e) Industrially produced CFCs are entirelyabsent in some regions of the oceans.

(f) North Atlantic Deep water and AntarcticBottom Water become progressivelydepleted in oxygen and enriched innutrients and CO2 as they drift away fromtheir respective high latitude sourceregions. (Note the gradations in coloringof these water masses in Fig. 2.8.)

(g) The oceanic thermohaline circulationslows the buildup of atmospheric CO2concentrations in response to the burningof fossil fuels.

(h) Equipment abandoned on a continental icesheet is eventually buried by snow, whereasequipment buried on an ice floe remainsaccessible as long as the floe remains intact(Fig. 2.37).

(i) Regions of permafrost tend to be swampyduring summer.

(j) An increase in wintertime snow cover overthe continental regions surrounding theArctic would cause the zone of continuouspermafrost to retreat.

(k) In regions of mixed forests and grasslands,forests tend to grow on the poleward-facingslopes.

(l) Many summertime high temperaturerecords have been set during periods ofextended drought.

(m) Ice samples retrieved from near the bottomsof Greenland ice cores are much older thanthe residence time listed in Table 2.2.

(n) Seas in closed drainage basins such asthe Great Salt Lake and the Caspian Seaare subject to much larger year-to-yearvariations in level than those that haveoutlets to the ocean.

(o) Under certain conditions, wet hay canundergo spontaneous combustion.

58 The Earth System

Fig. 2.37 Abandoned equipment on the Antarctic ice sheet(top) and on Arctic pack ice (bottom). [Photographs courtesyof Norbert Untersteiner.]

27 The atmosphere of Titan, Saturn’s largest moon, exhibits an intermediate temperature maximum analogous to Earth’s stratopause. Thisfeature is due to the absorption of sunlight by aerosols: organic compounds formed by the decomposition of CH4 by ultraviolet radiation.

P732951-Ch02.qxd 9/12/05 7:40 PM Page 58

temperature ( c) temperature ( c) within the oceanic mixed...

Documents