t*i*

%*,.*

, . s$$

.:i':'

;;** , -un. . , . ,^..ri!

.*fS- ' -o] i

,$".rjl' ,__*.,1",i'

*

., I s#&"o

;:r:i .l ,*+netl-,tt +i

*+ t':'

d"ii'.,,,.{i:,. : :...- '"t .

.% &;

At f i rst g lance, i t looks l ike the

cont inents on ei ther s ide of the

At lant ic Ocean once f i t togetherqui te nicely, l ike the pieces ofa j igsaw puzzle Do cont inents

real ly move? O', to put i t anoiherway, does the map cf Earth 's

surface change over t ime?

21 INTRODUCTIONIn September 1930, fifteen explorers led by a German mereo_rologist, Alfred Wegener, set out across the endless snowfieldsof Greenland to resupply two weather observers stranded at aremote camp. The observers were planning to spend the iongpolar night recording wind speeds urrd t.-i.rntures on Green_land's polar piateau. At the time, W.g..r., was well known,not only to researchers studying climate but aiso to geologists.Some fifteen years earlier, he had published a smali book, TheOrigin of the Continents and Ocrars, in which he had dared tochallenge geologists' long-held assumption that the continentshad remained fixed in position through ali of Earth history.Wegener proposed, instead, that the coitinents once fit togetherlike pieces of a giant jig saw puzzle, to make one vasr supercon_tinent. He suggested that this supercontinent, which he namedPangaea (pronounced Pan-jee-ah; Greek for ,,a11

land,,), laterfragmented into separate continents that drifted apart, moving

lt"ytt to their present positions (Fig. 2.1). This process came ro

be known as continental drift.

. Had Wegener survived to old age, he would have seen his

hypothesis become the foundation of n scientific revolution.Toduy, geologists accept Wegener's ideas and take for grantedthat the map of the Earth constantly changes; continents waltzaround its surface, variously combining and breaking apartthrough geologic time. The revolution

"b.gnn in 1960, when

Harry Hess, of princeton University, proior.d that as con_tinents drift apart, new ocean floo, for-, tetween them by aprocess that his contemporary Robert Dietz also described andnamed sea-floor spreading. Hess and others also suggestedthat continents move toward each other when the old. oceanfloor between them sinks back down into the Earth,s interior,a process now called subduction. By 1968, geologists haddeveloped a fairly complete model d"r.ribirrg continental drift,sea-floor spreading, and subduction. In this ,iodel, trartht lith_osphere, its outer, relatively rigid shell, consists of about twentydistinct pieces, or plates, that slowly move relative to each other.Because we can confirm this model by many observations, it hasgained the status of a theory, which we now call the theory ofplate tectonics, from the Greek word. tektoin, which meansbuilder; plate movements ,,build"

regional geologic features.Geologists view plate tectonics as the irand

""ify_g theory of

and dropped off enough supplies to lasr the winter. Wegenerand one companion set out on the return trip the next dai butthey never made it home.

geology, because it can successfully explain a grear many geo_logic phenomena, as we will see.

...and later drifted apart. i';t:!:f:!iiitrf, e:f a:let:ilitloil tf:,r,r."1,1. _qlrltpil

this chapter, we introduce the observations that ledWegener to propose his continen tal drifthypothesis. Then welook at paleomagnetism, the record of B"rtt t magnetic fieldin the past, which provides a key proof of continental drift.Next, we learn how observations

""borr, the sea floor made by

geologists during the micl-twentieth century led Harry Hessto propose the concept of sea_floor spreading. We conclude bydescribing the many facets of modern plate t"ectonics theory.

CHAPTER 2 The Way the Earth Works: p late Tectonrcs

. Wegener presented many observations in favor of con_

tinental drift, but he met with strong resistance. At a widelypublicized 7926 geology conference in New york City, a.ro*dof celebrated American professors scoffed: ,,What

force couldpossibly be great enough to move the immense mass of a con-tinent?" Wegener's writings didn't provide a good answer, sodespite all the supporting observations he haJprovided, mostof the meeting's participants rejected continental drift.

. No1v, four years later, Wegener faced his greatest chal_

lenge. On October 30, 1930, W.g.rr., reached the observers

Fl* i l f l4f - t .T

Wegener,s image of pangaea, and his proposal of i tssubsequent breakup and dispersal .

.i

Alfred Wegener g:r in Greenland. $:

..iri4l!:eFait!:c*rfiiig$:::

Wegener suggested thatcontinents once fit together... fr.,:l j8lige.i;!:ryie{$itd:1tilf &Tlyq,Trlrnt-]ipry:afiir"lf ir,

(:*' vTime

Present

2.2 Weqener's Evidence for Continental Drif t

22 WEGENER'S EVIDENCE FORCONTINENTAL DRIFT

Wegener suggested that a vast supercontinent, Pangaea,existed until near the end of the Mesozoic Era (the interval ofgeologic time that lasted from 257 to 65 million years ago). Hesuggested that Pangaea then broke apart, and the land massesdrifted away from each other to form the continents we seetoday. Let's look at some of Wegener's arguments and see whyhe came to this conclusion.

The FE{ *f tFaw ffmmtFrsemtsAlmost as soon as maps of the Atlantic coastlines becameavailable in the 1500s, scholars noticed the fit of the conti-nents. The northwestern coast of Africa could tuck in againstthe eastern coast of North America, and the bulge of easternSouth America could nestle cozily into the indentation ofsouthwestern Africa. Australia, Antarctica, and India could allconnect to the southeast of Africa, while Greenland, Europe,and Asia could pack against the northeastern margin of NorthAmerica. In fact, all the continents could be joined, withremarkably few overlaps or gaps, to create Pangaea. Wegenerconcluded that the fit was too sood to be coincidence.

Lmcmtimms *f Fast #$maimt*mmsGlaciers are rivers or sheets of ice that flow across the landsurface. As a glacier flows, it carries sediment grains of allsizes (clay, silt, sand, pebbles, and boulders). Grains protrud-ing from the base of the moving ice carve scratches, calledstriations, into the substrate. When the ice melts, it leavesthe sediment in a deposit called ti l l and exposes striations.Thus, the occurrence of till and striations af a location serveas evidence that the region was covered by a glacier in thepast. By studying the age of glacial till deposits, geologistshave determined that large areas of the land were coveredby glaciers during time intervals of Earth history called iceages (discussed further in Chapter 1B). One of these ice agesoccurred from 280 to 260 mill ion years ago, near the end ofthe Paleozoic Era.

35

Wegener was an Arctic climate scientist by training, soit's no surprise that he had a strong interest in glaciers. Heknew that glaciers form mostly at high latitudes today. Sohe suspected that if he plotted a map of the locations of latePaleozoic glacial till and striations, he might gain insight intothe locations of continents atthat time. When he plotted theselocations, he found that glaciers of this time interval occurredin southern South America, southern Africa, southern India,Antarctica, and southern Australia. These places are nowwidely separated and, with the exception of Antarctica, do notcurrently lie in cold polar regions (Fig. 2.2a). To Wegener's

^mazemerrt, all late Paleozoic glaciated areas lie adjacent to

each other on his map of Pangaea (Fig.2.2b). Furthermore,when he plotted the orientation of glacial striations, theyall pointed roughly outward from a location in southeasternAfrica. In other words, Wegener determined that the distribu-tion of glaciations at the end of the PaleozoicEra could easilybe explained if the continents had been united in Pangaea, withthe southern part ofPangaea lying beneath the center ofa hugeice cap. This distribution of glaciation could not be explained ifthe continents had always been in their present positions.

The ffi$strEbuct*m$n mf #$fismatfr* ffieEts

Ifthe southern part ofPangaea had straddled the South Poleat the end of the Paleozoic Era, then during this same timeinterval, southern North America, southern Europe, andnorthwestern Africa would have straddled the eouator andwould have had tropical or subtropical climates. Wegenersearched for evidence that this was so by studying sedimen-tary rocks that were formed at this time, for the materialmaking up these rocks can reveal clues to the past climate.For example, in the swamps and jungles of tropical regions,thick deposits of plant material accumulate, and when deeplyburied, this material transforms into coal. And, in the clear,shallow seas of tropical regions, large reefs develop offshore.Finally, subtropical regions, on either side ofthe tropical belt,contain deserts, an environment in which sand dunes formand salt from evaporating seawater or salt lakes accumulates.Wegener speculated that the distribution of late Paleozoiccoal, sand-dune deposits, and salt deposits could define cli-mate belts on Pangaea.

Sure enough, in the belt of Pangaea that Wegener expectedto be equatorial, late Paleozoic sedimentary rock layers includeabundant coal and the relicts of reefs; and in the portions ofPangaea that Wegener predicted would be subtropical, latePaleozoic sedimentary rock layers include relicts of desertdunes and of salt (Fig.2.2c). On a present-day map of ourplanet, these deposits are scattered around the globe at avari-ety of latitudes. In Wegener's Pangaea, the deposits align incontinuous bands that occupy appropriate latitudes.

f f l f f iC"f F4f i ;1. ;? Wegener 's evidence for cont inental dr i f t came from analyzing the geologic record

(al The distribution of late Paleozoic glacial deposits on present-day EarthArrows indicate glacial striations.

Mesosaurus lived in Africa ",and South America. $i,'i:!,1,:t:!;!.;!t!i!iif :li,l;,!:6I,f ,yy;14;qq9,;9111i11eaa:1ll

(c) The distr ibution of late Paleozoic rock types p/otssensibly in the cl imate belts of Pangaea.

Yft* ffifiwts6*xu*e$*s? m$ ffims*$$m

Today, different continents provide homes for different spe-cies. Kangaroos, for example, live only in Australia. Similarly,many kinds of plants grow only on one continent and not onothers. Why? Because land-dwelling species of animals andplants cannot swim across vast oceans, and thus evolved inde-pendently on different continents. During a period of Earthhistory when all continents were in contact, however, land ani-mals and plants could have migrated among many continents.

With this concept in mind, Wegener plotted fossil occur-rences of land-dwelling species that existed during the latePaleozoic and early Mesozoic Eras (between about 300 and

CHAPTER z The Way the Earth Works. Plate Tectonlcs

(d) Fossi ls of Mesozoic land-dwel l rng organisms occuT on mult ip le cont inents. Thiswould be hard to explarn i f continents were always separated by oceans.

210 million years ago) and found that these species had indeedexisted on several conrinents Gig.2.2d). Wegener argued thatthe distribution offossil species required the continents to havebeen adjacent to one another in the late Paleozoic and earlyMesozoic E,ras.

F.iEm€*$x6rx6 ffi m*$*g e* #sc fr gs

In the same way that an art historian can identify a Picassopainting and an architect a Victorian design, a geoscientistcan identify a distinctive group of rocks. Wegener found thatthe same distinctive Precambrian (the interval of geologictime before 542 mlllion years ago) rock assemblages occurred

Antarctica

(bl On Pangaea, areas with glacial deposits f i t together in a southern polarcap

!- [ \ Lystrosaurus lived in Afiica,'r ' - \ Antarct ica, and [ndia.

,i,4trir* lt

-

': (J

; a.(r

'o^,,,"",. ' ; ' ' , '

. -'-

,t'5*.=; lrt

, t - /n

1 ' ' { \ ,

ffi*vedin Africa r;: Glos.sopteris lived inJ;and South America. i, southern continentl

0n the eastern coast of South America and thc western coastof Africa, regions now separated by an ocean (Fig. 2.3a).Ifthe continents had been joined to create pangaea in the past,then these matching rock groups would h"rr."b..r, adjacent toeach other, and thus could have composed continuous blocksor belts. Wegener also noted that fearures of the AppalachianMountains of the United States and Canada closely resemblemountain beits in southern Greenland, Great Britain, Scandi_navia, and northwestern Africa (Fig. l.3b), regions that wouldhave lain adjacent to each other in pu,.gnenl Wegener thusdemonstrated that not only did the coallines of'continentsmatch, so did the rocks adjacent to the coastlines.

Criticicc'r"r mf W*gex,les'.w $deesWegenert model of a supercontinent that later broke apartexplained the distribution oflate paleozoicglaciers, coal, sanddunes, distinctive rock assemblages, and fosJs. Clearly, he hadcompiled a strong case for continental drift. But as noted earlier,

lr: cou-ld not adequately explain how or why conrinents drifted.

He left on his fina1 expedition to Greenland havins fhiled toconvince his peers, and he died without knowing thalhis ideas,after lying dormant for decades, would be reborn as the basis ofthe broader theory ofplate tectonics.

In effect, Wegener was ahead of his time. It would takethree more decades ofresearch before geologists obtained suflficient data to test his hypothe.., p-p..t!. Collecting thisdata required instruments and techniqrr.. ,h"t did not exist

2.3 Paleomagnetism and the proof of Contjnental Drif t

(a) Dist inct ive belts of rock jn South Amerjca would al ign wrthsimilar ones in Afr ica, without the Atlantic.

37

in Wegener's day. Of the many geologic discoveries that ulti_mately opened the door to plate tectonics, perhaps the mostimportant came from the discovery of a phenomenon calledpaleomagnetism.

? 3 PATTOMAGNETISM AI\ID THEPROOF OF COI\ITINTNTAL DRIFT

More than 1,500 years ago, Chinese sailors discovered that apiece of lodestone, when suspended from a thread, points ina northedy direction and can help guide a voyage. Ltd.rtorr.exhibits this behavior because it consists of magnetite, an iron_rich mineral that, like a compass needle, aligns with Earth,smagnetic field lines. Some rock types contain tiny crystals ofmagnetite, or other magnetic minerals, and thus behave overalllike weak magnets. In this section, we explain how the studyof such magnetic behavior 1ed to the realizitionthat rocks pre_serve paleomagnetism, a record of Earth's magnetic field

FlcusE !"n Further evidence of dr i f t : rocks on di f ferent s ides of the ocean match.

(b) l f the Atlantrc didn't exist, paieozoic mountain beltson both coasts would be adjacent.

3a

in the past. An understanding of paleomagnetism providedproof of continental drift and, as we'll see later in this chapter,contributed to the development of plate tectonics theory. As afoundation for introducing paleomagnetism, we first review thenature of the Earth's masnetic field.

Sarth's ftdlagruetic Fie{d

Circulation of liquid iron alloy in the outer core of the Earthgenerates a magnetic field. (As you may learn in a phys-ics course, a similar phenomenon happens in an electricaldynamo.) Earth's magnetic field resembles the field producedby abar magnet, in that it has two ends of opposite polarity.Thus, we can represent Earth's field by a rnagnetic dipole,an imaginary arrow (Fig. 2.a{. This dipole intersects thesurface of the planet at two points, known as the magneticpoles. By convention, the north magnetic pole is at the end ofthe Earth nearest the north geographic pole (the point wherethe northern end of the spin axis intersects the surface). Thenorth-seeking (red) end of a compass needle points to thenorth magnetic pole.

Earth's magnetic poles move constantly, so, averaged overthousands of years, they roughly coincide with Earth's geo-graphic poles (Fig. 2.4b). At present, however, the magneticpoles lie hundreds ofkilometers awayfrom the geographic poles,so the magnetic dipole tilts at about 11" relative to the Earth'sspin axis. Because of this difference, a compass today does notpoint exactly to geographic north. The angle between thedirection that a compass needle points and a line of longi-tude is the magnetic declination (Fig.2.4c).

Invisible field lines curve through space betweenmagnetic poles. In a cross-section view, these linesparallel to the surface of the Earth (that is, are horizon-tal) at the equator, tilt at an angle to the surface in mid-latitudes, and plunge perpendicular to the surface atthe magnetic poles (Fig. 2.4a). The angle between amagnetic field line and the surface of the Earth, at agiven location, is called the magnetic inclination.If you place a magnetic needle on a horizontal axis soit can pivot up and down, and then carry it from themagnetic equator to the magnetic pole, you'll see thatthe inclination varies with latitude-it is 0" at the mas-netic equator and 90" at the magnetic poles. (Note that thecompass you may carry with you on a hike does not show incli-nation, because it has been balanced to remain horizontal.)

ffm$**megffi#f$ss"!'t

In the early twentieth century, researchers developed instru-ments that could measure the weak magnetic field producedby rocks and made a surprising discovery. In rocks that formed

Si#Lf $: ! ' * , . f " ,$, Features of Earth 's maqnet ic f ie ld

:;r:$:n:s:3f,:$:ggf.f.r*tr.rf:i:1:ni:,r!ra:.:!*rlt,I:lai!:t;:aii:n:;ftr5:f:r*:!pt9*r

(a) Earth's f ield l ines curve, so the t i l t of amagnetic needle changes with lat i tudeThis t i l t is the magnet ic incl inat ion.

ffis:f [*

The dipoletilts at 11" tothe spin axis.

!:}i$rili:i:!FtiE;d;11:s;etf.!.i.1r.i!r5trl!iei!!rai!;e;s;!ili1ili:i:i!6:i:t:rtfF:eieffffrisi.i.rl!:piytgti.:{ilwlqsel*rd:l

(bl A map of the magnetic pole posit ion during the past 1800 yearsshows that the pole moves, but stays within high lat i tudes

90"w 60"w

Declination

(c) The magnet ic pole and thegeographic pole do not coincide, soin TnosL locatior-s a compass doesnot pornt exactly to geographjcnorth. The dif ference is decl ination.

I = 1O'West

millions of years ago, the orientation of the dipole representingthe magnetic field of the rocks zs not the ,oru u, that of present-day Earth (Fig.2.5a). To understand this statement, consideran example. Imagine traveling to a iocation near the coast onthe equator in South America where the inclination and decli-nation are presently 0". If you measure the weak magnetic field

The declination observed iitoday var ies wi th locat ion.

2.3 Paleomagnetjsm and the proof of Continental Drif t

Flsuft i l i i i i pareomagnet ism and how i t can form dur ing coor ing of lava

/)i\' ! / \ i = No net magnel iTat jon

' \ {Decause f+f = 0t

't'

*i ,1"i""t

!::

HotIavaflow

The paleomagnetic :.inclination is not 0., as Lit would be for a roci f..formed at the equator. f.lf ttiltf tiiiltitrytrt-rt,rrorSfoleryl,*:o,*i;.1

,E4.

Time'r;.

As rock cools, dlpoles ,:.align with Earthb $imagnetic field. With fmore cool ing, dipoleslock into t | r i i or i ln ia i ion. ,1.^, ,. " "..,.. .. ,rr-u"#,The pa.leomagnetic dipole is

rnotcated symbolicallv bvthe bar magnet.iel34s!Yi3gi'4!itf 4ii!ili]3,4!y ;Fls*q$rrt?i:s*l

'" f ,rr,

{a) A geologist f inds an ancient rock sample. Thethe rock's paleomagnetism is djf ferent from thaf(b) Paleomagnetism can form when iava coois and becomes sol id rock

&ppmnmmt Fea*mr Wmmdmr*& ffrmm$ qpf ffi*r$*x$.?ffiFrg#$ #r$$*

orrentatron ofof today's f ieid.

produced by, say, a 90-million-year_old rock, and representthe orientation of this fiennd,ha,ili.;;;tl?i;#i-Jl";:,'fi il",'"",T:i..I,1Inndnl

n:.r.| magnetic pol., f,.c".rs. it. dJ.ti.,ution differsfrom the declination that a compass at the location would dis_piay today and its inclination is not 0". CJ.4." have madesimilar measurements at many locations u.orrrri the globe and

]1ve found that the magnetic fields of ancient rocks indicatethe onentation of the magnetic field, relative to the rock, atthe time the rock formed. Thi. ,..ora,-fr...rrr.a in rock, ispaleomagnetism.

Why doesnt the paleomagne-tic dipoie in rocks point to thepresent-day magnetic field? When geologists firrt utt._pt.atoansw€r this question, they assumed ihat Jntinents were fixed inposition, and thus concluded.that the positio, of Eu.th,, ,rrrg_netic poles in the past were different than ttintrod'uced ,h.;;r; p.reopole to refer r" :f.iT;""lTi il;ftion of the Eartht magnetic north pole in the past. With this

::ncept.in mind, they set our to track what they thought was thechange in position of the paleopole over time. To ai' ,fri., ,fr.fmeasured the paleomagnetism i., u ,.,...rrion of rocks of diFferent ages, from the same general location on a continent, andpiotted the position of the associated succession of paleopoie

ry:1l"i. on a map (Fig.2.6a).The successive positions of datedpareopoles trace out a curving line that came to be known as anapparent polar_wander path. At first, geologists assumedthat the position of the continent from whi.h'th.y collected thesamples was fixed, and interpreted the uppu..rr, polar_wander

path to represent how the position of darth,s -"g.r.,i.

;"i.illYlj.,1,9"ugh

time. But were they in fo. u ,,r.prir.l When

, Hlwd.oespaleomagnetism developl One process happenswhen lava (molten rockl cools ,o for,n i"run ii

" 2.Sb). Lavau

I u.7-h9t liquid containing no crystals. A, it ,tu.t, to cooland solidify inro rock, tiny Jagnetiie crystals begin to grow.

ff:|, thelmal .l:.gy (see A"ppendi")'.*r., rhe magneticdtpole associated with each crystal to wobble and tumble cha_

:::jtt_:Irl at any,given insrant, the dipoles of. the mag_netrte specks are randomly oriented and the magnetic forcesthey produce cancel each. other out. e**""ffy, the rock coolssufficiently. that the dipoles slow down u.ra, fk. tiny compassneedles, align with the Earth,s

-"g".;. fi.li. As the rockcools still more, these tiny compass needles lock into perma_nent parallelism with the Eartht magnetic field at the timethe cooling takes place.

they obtained polar-wander parhs fi.; ;;;;. di#r# r##J,they found that each continent has a dffirrnt: nlporrntpolar_zaanderpath.The hypothesis that continentr"u.. fi*ed in position cannot

ffi"/,r*/ tn hot magma, the dipoles# llT^s. ?i'"ntarion 'apictiy, .so magma cannot have ipermanentmagnetization. I

;ir!f !:at!:f ti:iqi!.ei'!iiilitu,"or"orru"*,,o,","n",,_,ii

Paleomagnetic inclination and iideclination changes in a successionof rock layers at a given location.

Magnetic

paleornagnet ism in a succesionof rock layers on a continentdef ine an apparent polar,wanderpath relattve to the continent.

40

f i { i r - I f { f r : .1.{ i Apparent polar-wander paths, and their interpretat ion

CHAPTER 2 The Way rr-e Earth Worrs: p lare lecLon,cs

(c) The apparent poiar-wander path of NorthAmerlca is not the sameas that of Europe or Afr ica.

2 4 SETTING THE STAGEFOR THE DISGOVERY OFSEA.FIOOR SPREADING

ffiew $rmmges ef S*m-F$m*r ffimthyrffietryMilitary needs during World War II gave a boost to sea-floorexploration, for as submarine fleets grew, navies requireddetailed information about bathymetry, or depth variations.The invention of echo sounding (sonar) permitted such infor-mation to be gathered quickly. As a ship equipped with sonartravels, observers can obtain a continuous record ofthe deoth ofthe sea floor. The resulting cross section showing depth piottedagainst location is calied a bathymetric profile (Fig.2.7a,b).Bycruising back and forth across the ocean many times, investi-gators obtained a series of bathymetric profiles and from theseconstructed maps of the sea floor. (Geologists can now producesuch maps much more rapidly using satellite data.) Bathymerricmaps reveal several important features.

Millionyears oid

Successive layers ofrock near tocality@

explain this observation, for if the magnetic pole moved whileall the continents stayed fixed, measurements from all continentsshould produce the same apparent polar-wander paths.

Geologists suddenly realtzed that they were looking atapparent polar-wander paths in the wrong way. It\ not the polethat moves relatiae tofxed continents, but rather the continents thatmave relatirLe to afxed pole (Fig. 2.6b). Since each continent hasits own unique polar-wander path (Fig.2.6e), the continentsmust move with respect to each other. The discovery provedthat Wegener was right ai1 along-continents do driftl

(b) l f continents aref ixed, then the polemoves relat ive to thecont inent. / f the pole isf ixed, then the cont inenlmust dri f t relatrve to the pole

Th" ' t

ler'1r8tr:.

ffilry

FIGURH ?.? Bathymetry of mid-ocean r idges.

Sonar al lows a snio to map sea-f loor bathymetry easi lydetermines water oeplh by sending sound waves down

Regional bathymetrycan now be mappedby satellite.

t ,H:

The velocity of waves isknown. Distance = traveltime x velocity.

Continentalshelf

t tpd elnno i fc 1l

at flacture zones. i

Location map

Mid-ocean ridge

Etr Shallow

Trenches line the westernedge of the Pacific Ocean.

!:Y;!i;!i:i:a:s:3ii,.i3i!g!,Ii,!i:6:d3

I Deep

Continentalshelf

Coast -X'

Coast Abyssal plain Abyssal plainAxis

Abalhymetric prof i le along l ine X-X' i l lustrates how mid-ocean r idges rose above abyssal plains. Both are deeper than cont:nental shelves

FlGt, , rcf : t .* Other bathymetr ic features of the ocean f loor, and the pattern of heat f low at mid-ocean r idges

axis is l- Central America Trench Puerto Rico Trench

'" Aleu

St*wrir!l{rrisiiiFriRidop evic

lfil :

. More ,heat flow

fSan Andref

%'"t, i

FaultEast Pacific

Ridge

TongaTrench

ry'".

;;.i;t-tik;-l(a) A map i l lustrat ing the distr ibut ion ofr d-ocean r.dges, deep-ocean Lrenchesand oceanic transform faults.

Heat flow is greatest at iithe mid-ocean rldge axis. iii

ws.r! silsjjjjjjjiii!i,i,riiio,1.,;,41e;11;1,1;11,1;'1;5efl

. 'ADyssat

!Iat-!oppedSeamount(guyot)ain Sbamount

0^1

t4

t6

0lu12

lbl In addit ion to mjd-ocean r idges, the sea f loor displays other bathymetric features such as deep-sea trenches, oceanic islands, guyots,and seamounts, as shown in this vert ical ly exaggerated profi le. Heat f lows depends on location.

r Mid-ocean ridges: The floor beneath all major oceansincludes abyssal plains, which are broad, relativelyflat regions of the ocean that 1ie at a depth of about 4to 5 km below sea level; and mid-ocean ridges, elon-gate submarine mountain ranges whose peaks lie onlyabout 2 to 2.5 km below sea level (Frg.2.7b; Fig.2.8a).Geologists call the crest of the mid-ocean ridge theridge axis. A11 mid-ocean ridges are roughly symmetri-cal-bathymetry on one side of the axis is nearly a mir-ror image of bathymetry on the other side.

s Dee?-ocean trenches: Along much of the perimeter of thePacific Ocean, and in a few other localities as we1l, theocean floor reaches depths of B to 12 km-deep enoughto swallow Mt. Everest. These deep areas occur in elon-gate troughs that are now referred to as trenches (Fig.2.8b). tenches border volcanic arcs, curving chainsofactive volcanoes.

r Seamount chains: Numerous volcanic islands poke uofrom the ocean floor: for example, the Hawaiia" Irtntrdtlie in the middle of the Facific. In addition to islandsthat rise above sea level, sonar has detected many sea-mounts (isolated submarine mountains), which wereonce volcanoes but no longer erupt. Seamounts withflat tops, due to ancient reef growth, are called guy-ots. Volcanic islands and seamounts typically occur inchains, but in contrast to the volcanic arcs that borderdeep-ocean trenches, only one island at the end of aseamount chain remains capable of erupting today.

s Fracture zones; Surveys reveal that the ocean floor isdiced up by narrow bands ofvertical cracks and broken-up rock. These fracture zones lie roughly at rightangles to mid-ocean ridges, effectively segmenting theridges into small pieces (Fig. 2.8a).

ffi e"i*r -i3 firse e"varti* *r s m st t $': * ffi ;s{a"s rw*f #***raf+ #r#sg

By the mid-twentieth century, geoiogists had discovered manyimportant characteristics of the sea-floor crust. These discover-ies led them to realize that oceanic crust is quite different fromcontinental crust, and further that bathymetric features of theocean floor provide clues to the origin ofthe crust. Specifically:

A layer of sediment composed of clay and the tiny shellsof dead plankton covers much of the ocean floor. ThisTayer becomes progressiaely thicker away from the mid-ocean ridge axis. But even at its thickest, the sedimentlayer is too thin to have been accumulating for theentirety of Earth history.

By dredging up samples, geologists learned that oceaniccrust is fundamentally different in composition from

ct{ApTER z The Way the Earth Works: p late

ir!r i i , i i r . ; f ; , : . i i A 1953 map showing the distr ibution ollocat ions in the ocean basins. Note that earthquakes occur

80" 60. 40. 20" 0" zo' 40" 60" tio" 100" 120"

continental crust. Beneath its sed:ment cover, oceacrust bedrock consists of basalt-it does not displavgreat variety ofrock types found on continents.

Heat flow, the rate at which heat rises from the Earth'sinterior up through the crust, is not the same every-where in the oceans. Rather, more heat rises beneathmid-ocean ridges than elsewhere (Fig. 2.8b). This obsei-vation led geologists to speculate that magma might berising into the crust just below the mid-ocean ridge axis,for hot molten rock could bring heat into the crust.When maps showing the distribution of earthquakes inoceanic regions became available in the years afterWorldWar II, geologists realized that earthquakes do not occurrandomly, but rather define distinct belts (Fig. 2.9).Some belts follow trenches, some follow mid-ocean ridgeaxes, and others lie along portions of fracture zones,Since earthquakes define locations where rocks break andmove, geologists realized that these bathymetric featuresare places where movements of the crust take place.

Now 1et's see how Harry Hess used these observations to comeup with the hypothesis of sea-floor spreading.

it,

I t i Evidence for Sea,Floor Spreading

25 HARRY HESS AND HIS"ESSAY IN GEOPOETRY"

In the late 1950s, Harry Hess, after studying the observationsdescribed above, realized that because the sediment layer onthe ocean floor was thin overall, the ocean floor might be muchyounger than the continents. Also, because the sediment thick-ened progressively away ftom mid-ocean ridges, the ridges them-selves likely were younger than the deeper parts of the oceanfloor. If this was so, then somehow nev) oceaniiloor must beformingat the ridges, and thus an ocean couid be getting wider with time.But how? The association of earthquakes with mid-ocean ridgessuggested to him that the sea floor was cracking and splitting

^patt at the ridge. The discovery of high heat flow along mid-

ocean ridge axes provided the final piece ofthe puzzle,for it sug-gested the presence of molten rock beneath the ridges. In 1960,Hess suggested that molten rock rose upward beneath mid-oceanridges and that this material solidified to form oceanic crust(Fig.2.10). The new sea floor then moved away from the ridge,a process we now call sea-floor spreading. Hess realized that oldocean floor must be consumed somewhere, or the Earth wouldhave to expand. He suggested that deep-ocean trenches mightbe places where the sea floor sank back into the mantle, and thatearthquakes at trenches were evidence of this movement (Fig.2.10), but he didn't understand how the movement took place.

Hess and his contemporaries realized that the ..r-floor-spreading hypothesis instantly provided the long-sought expla-nation of how continental drift occurs. Continents passivelvmove apart as the sea floor between them spreads u,

-id-o..nnridges, and they passively move together as the sea floor betweenthem sinks back into the mantle at trenches. (Geologists nowreaTize that it is the lithosphere that moves, not just the crust.)Thus, sea-floor spreading proved to be an important srep onthe route to plate tectonics-the idea seemed so good that Hessreferred to his description of it as "an essay in geopoetry." Butother key discoveries would have to take place before the wholetheory of plate tectonics could come roqether.

2 6 EVIDENCE FOR SrA-FLOORSPREADING

For a hypothesis to become a theory (see Box p.1), it must betested-scientists must demonstrate that the idea really works.During the 1960s, geologists found that sea-floor spreadingsuccessfully explains several previously baffling observations.Here we discuss two: (1) the existence of ordedy variations inthe strength of the magnetic field over the sea floor, produc-ing a pattern of stripes called marine-magnetic anomalies; and(2) the variation in sediment thickness on the ocean crust, asmeasured by drilling.

fl#Hmri rse ffi * gc"aet$* Ac.s *s.t *ss$iesGeologists can measure the strength of Earth's magnetic fieldwith an instrument called a magnetometer. At any given loca-tion on the surface of the Earth, the magnetic field that youmeasure includes two parts: one produced by the main dipoleof the Earth, due to circulation of molten iron in the outercore, and another produced by the magnetism of near-surfacerock. A magnetic anomaly is the difference between theexpected strength of the Earth's main dipole field at a certainlocation and the actual measured strength of the magneticfield at that location. Places where the field strength is stron-ger than expected are positive anomalies, and places where thefield strength is weaker than expected are negative anomalies.

Geologists towed magnetometersback and forth across the ocean to maDvariations in magnetic f ield strength(Fig.2. l la) . As a ship cruised along i tscourse, the magnetometer's gauge mightfirst detect an interval of strong signal(a positive anomaly) and then an inter-val of weak signal (a negarive anomaly).A graph of signal strength versus dis-tance along the traverse, therefore, hasa sawtooth shape (Fig. 2.11b). Whengeologists compiled data from manycruises on a map, these marine mag-netic anornalies defined distinctive.

f ' r l# i . j * f ; l i t " l Harry Hess's basic concept of sea-f loor spreading. Hess impl ied, incorrect ly,that only the crust moved. we wi l i see that th is sketch is an oversimpl i f icatron.

CI*APTER 2 The Way the Earth Works: Plate Tectonics

i i i#Lj{ t t r #. t i The discovery of marine magnet lc anomal les

(a) A ship towing a magnetometerdetects changes in the strength of

the magnetic f ield.

ently points from the north magnetic pole to the south mag-

netic pole. When geologists measured the paleomagne-

tism of a succession of rock layers that had accumulated

over a long period of time, they found that the polarity

(which end of a magnet points north and which end points

south) of the paleomagnetic field preserved in some layers

was the same as that of Earth's present magnetic field,

whereas in other layers it was the opposite (Fig' 2'12a, b)'

At first, observations of reversed polarity wer e largely ignored,

thought to be the result of lightning strikes or of local chemi-

cal reactions between rock and water. But when repeated mea-

surements from around the world revealed a systematic pat-

tern of alternating normal and reversed polarity in rock layers,

geologists rcaltzed that reversals were a global, not a local,

fh.no*..rot . They reached the unavoidable conclusion that,

at ,ra.iorm times during Earth history' the palarity { Earth's

magneticf.eld has suddenly rer'tersed! In other words, sometimes

the Earth has normal polarity' as it does today, and sometimes

it has reversed polarity (Fig- 2.12c)' Times when the Earth's

field flips from normal to reversed polarity, or vice versa, are

called magnetic reversals. Reversals may be the result of

changes in circulation patterns in the outer core' When the

Earth has reversed polarity, the south magnetic pole lies near

the north geographic pole, and the north magnetic pole lies

near the south geographic pole. Thus' if you were to use a

compass during periods when the Earth's magnetic field was

..'o"rr"d, the north-seeking end of the needle would point to

the south geographic Pole'It the 1950s, about the same time geologists discovered

polarity reversals, they developed a technique that permitted

ih"- ,o defi.ne the age of arock in years, by measuring the decay

of radioactive elements in the rock. The technique, called iso-

topic dating, willbe discussed in detail in Chapter 10' Geolo-

gists applied the technique to determine the ages of rock layers

i" *hi.h they obtained their paleomagnetic measurements' and

thus determi ned zuhen the magnetic field of the Earth reversed.

With this information, they constructed a history of magnetic

reversals for the past 4.5 million years; this history is now called

the magnetic-reversal chronology. A diagram representing

the Earth's magnetic-reversal chronology (Fig. 2'12d) shows

that reversals do not occur regularly, so the lengths ofdifferent

polarity "chrons," the time intervals between reversals, are diF

ierent. For example, we have had a normal-polarity chron for

about the last 700,000 years. Before that, there was a reversed-

polarity chron. Geologists named the youngest four polarity

.hron, (Brunhes, Matuyama, Gauss, and Gilbert) after scien-

tists who had made important contributions to the study of rock

magnetism. As more measurements became available, investiga-

tors realized that there were some short-duration reversals (less

than 200,000 years long) within the chrons, and called these

shorter durations polarity subchrons.

ion --i>4sdffi#s'

Stronger\ Positive

anomaly

i -Negative*' anomaly

i

1Average

IWeaker

n4

(b) On a paper record, intervals of st tonger magnet isn (posi l ive anomal ies)

alternate with intervals of weaker magnetism (negtive anomalles) '

(c) After making many traverses, geologists make a map showlng areas

of positive anomalies (dark) and negatlve anomalies (l ight) These define

alternating bands or striPes

alternating bands. Ifwe color positive anomalies dark and nega-

tive anomalies light, the pattern made by the anomalies resem-

bles the stripes on a candy cane (Fig. Z.lLc)' The mystery of this

marine magnetic anomaly pattern' however, remained unsolved

until geologists recognized the existence of magnetic reversals'

Magnetic Reversals' Recall that Earth's magnetic field can

be represented by an arroq representing the dipole, that pres-

The pattern of anomaliesis symmetrical, relativeto mid-ocean ridges.

2,6 Evidence for Sea-Floor Spreading

Thelnterpretation of Marine Anomalies. Whydo marinemagnetic anomalies exist? A graduate student in England,Fred Vine, working with his adviser, Drummond Mathews,and a Canadian geologist, Lawrence Morley (working inde-pendently), discovered a solution ro this riddle. Simply put, apositive anomaly occurs over areas of sea floor where basalt hasnormal polarity. In these areas, the magnetic force produced bythe basalt adds to the force produced by trartht dipole and cre-ates a stronger magnetic signal than expected, as measured by

45

the magnetometer (Fig. 2.73a). A negative anomaly occurs overregions of sea floor where basalt has.reversed polarity. Here, themagnetic force of the basalt subtracts from the force producedby the Earth's dipole and results in a weaker magnetic signal.Sea floor yielding positive anomalies formed at times when theEarth had normal polarity, whereas sea floor yielding negativeanomalies formed when the Earth had reversed polaritv.

-

The sea-floor-spreading model easily explains why themagnetic anomaly pattern on one side of a mid-ocean ridge is

FIGURH 2. ' i f ; lVagnetic reversals, and the chronology of reversals.

(al Successive layers of lava bui ld at a volcano over t ime

Brunhes(normal)

Matuyama(reversed)

Gauss(normal)

---;.:i----:-Lrl lDeIt

(reversed)

(bl Geologists discovered that somelayers nave normal po ari ty, whereassome have reversed polari ty.

s

6z

F

O

lnterpretation

(d) Observations led to the production of a reversalchronology, with named polari ty intervals.

(c) Geologists proposed that thereveTses polaf l ty every now and

Earth's magnetic f ieldthen.

Polarity ofdated samples

Jaramillo

subchron

Matuyamareversed

Olduvai chronnormal

GaussMammoth normal

reversed chron

46

Fri{} lJ l te : i .1;{ The progressive development of magnetrcanomaltes, and the long_term reversal chronology.

Earth -l"u positive

ftoor anomally

rFr ffire

Hk. ru ffiffiEarth

tr-lg$.

Negativeanomally

til= | |M

Seafloor

ffi| | lLI

(a) Positrve anomalies form when sea_floor rock has the same potarrtyas the present magnetic f ierd. Negative anomaries form when seajroorrock has polari ty that rs opposite to the present f ield

RidgeAXIS

,*; '.,1,. " :r-#, 1o

Time 2

(c) Seaj loorspreading explains thestr ipes. The f ield f l ips back andforth while the ocean basin grows wrder

The anomaly pattern representsalternating stripes of normal_polarity and reversed-polaritysea I10()r.

(d) The width ofmagnetic str ipes on thesea f loor is proporttonal to

CHAPTER e The Way the Earth Works: p late Tectonrcs

a mirror image of the anomaly pattern on the other side (Fig.2.13b). At Time 1 (sometime in the past), a time of normalpolarity, the dark stripe of sea floor forms (Fig. 2.13c). Thetiny dipoles of magnetite grains in basalt makinf up this stripealign with the Earth's field. As it forms, the roci< ;., this stripemigrates away from the ridge axis, half to the right and half tothe left. Later, at Time 2, the field has reversed, and the light_gray stripe forms with reversed polarity. As it forms, it too _Jrr.,away from the axis, and sti1l younger crust begins to developalong the axis. As the process continues over millions of v.r.r.manystripes form. A positive anomaly exists along the ridge axistodaybecause this represents sea floor that has deieloped juringthe most recent interval of time, a chron of normai polaritv. Thlmagnetism of the rock along the ridge adds to the magnetism ofthe Earth's field.

By relating the stripes on the Atlantic sea floor to magneticreversals found in basalt layers on land, geologists dated tte seafloor back to an age of 4.5 millio., y."r"r,

"r,i th"y found that

the relative widths of anomaly stripes on the sea floor exactlvcorresponded to the relative durations ofpolarity chrons in themagnetic-reversal chronology (Fig. 2.13d). The relationshipbetween anomaly-stripe width and polarity_chron durationindicates that the rate of spreading has been constant in theAtlantic for at least the last 4.5 miliion years.

ffiw$d*m*s $r*rm ffi *wg*-ffi wm fl?rx$$6*.rgIn the late 1960s, a drilling ship called the Glomar Cha/lengerset out to sail around the ocean drilling holes into the sea flo"or.This amazing ship could lower .rro.rgh drill pipe to drill in5-km-deep warer and could continue to drill until the holereached a depth of about I.7 km (1.1 miles) below the sea floor.Drillers brought up cores of rock and sediment that geoscien_tists then studied on board.

On one of its early cruises, the Glomar Challenger drilleda.series-of holes through sea-floor sediment to the blsalt layer.These.holes were spaced at progressively greater distancesfrom the axis of the Mid-At1anti.hidg.. if the model of sea_floor spreading was correct, then the ,."di-..r, layer should beprogressively thicker away from the axis, and the age of theoldest sediment just above the basalt should be protressivelyolder away from the axis. When the drilling and-the"analyseswere compiete, the predictions were confirmed.

The discovery of sea-floor spreading and the paleomagneticproofofcontinental drift set off a scientific revoiution i,r"geor-ogy in the 1960s and 70s. Geologists realized that many ofih.i,existing interpretations of global geology, based on the prem_ise that the positions of continents ancl-oceans remain fixed inposition through time, were simply wrongl Researchers droppedwhat they were doing and turned theii attention to stujyingthe broader implications of sea-floor spreading and continenltal drift. New studies clarified the meaning oi n plat., definedthe types of plate boundaries, constrained pl'ate motions, relatedplate motions to earthquakes and ,rol.urro.., showed how plate

I"? What Do We Mean by plate Tectonics?

2 7 WHAT DO WE MEAN BYPTATE TECTONICS?

Fif i iJSf j ; : . ! ,$ Nature of the l i thosphere.

A load is placed on the i;surface of the Earth. i..it1i:.rr'.i!.i:;..j!.r,r?:r.;:,:.rriirr,.r,,.t.,,,,,,,tr,,,,,,,r,,,,1

(Not to scale)

l rme I

motions generate mountain belts and seamount chains, and out_lined the history of past plate motions. From these, the moderntheory of plate tectonics evolved. Below, we describe lithosphereplates and their boundaries, and then outline the basic prirl.ipl.,of plate tectonics theory.

T$"xe #q*ffi**ffi€ #$ m Li€fum*g*fxmye ff$*€sWe learned earlier that geoscientists divide the outer part ofthe Earth into two layers. The lithosphere consists of thecr_us1 plus the top (cooler) part of the upper mantle. It behavesrelatively rigidly, meaning that when

" for.. pushes or pulls on

it, it does not flow but rather bends or breaks (trig. 2.1ja). Thelithosphere floats on a relatively soft, or ,plastic,,,iayer

called

(a) The l i thosphere is fair ly r igid, but when a glacier or volcano rs praced on i ts surface, i t bends. Thiscan nappen because the underlying "plast ic" asthenosphere can f low out of the wav.

The continental shelf overlies a i;passive-margin basin, where th icksediment over l ies stretchecl l i thosphere. ..!,r,rj i ; i i"r:r!:t!. j i .

i :! j i t!. j f ir,;:!t jr,.r,!:!.:!: i ,!r1r:! i::, l j :!, i t;!r:r, i!;:,!t;:!1%ri.:rri"{!!:!. j f:r l j

Continental shetf l Abyssal plain

Lithosphere is i,",relatively rigid Iand cannot flow. i

the asthenosphere, composed of warmer(>1,280"C) mantle that can flow slowlvwhen acted on by force. The asthenorph...can undergo convection, like water in a pot,though much more slowly, but the litho_sphere cannot.

Continental lithosphere ancl oceaniclithosphere differ markedly in rheir thick_ness. On average) continental lithospherehas a th ickness of 150 km, whereas ol j o. ._anic lithosphere has a thickness ofabout 100km (Fig. 2.14b). (For reasons discussed laterin this chapter, new oceanic lithosphere at amid-ocean r idge is much thinner.) Recal lthat the crustal part of continental litho_sphere ranges from 25 to 70 km thick andconsists of relatively low-density felsic andintermediate rock (see Chapter l ) . ln con_trast ,_the crusral part of oceanic l i thosphereis only 7 to 10 km thick and consists ofrelatively high-density mafic rock. Becauseof these difTerences, the continental litho_sphere "floats" at a higher level than doesthe oceanic lithosphere.

The lithosphere forms the Earth's rela_tively rigid shell. But unlike the shell of ahen's egg, the lithosphere shell contains anumber of major breaks, which separate itinto dist inct p ieces. As noted ear l ier , we cal ithe pieces lithosphere plates, or simplvpJates. The breaks between plates are kno*nas plate boundaries. Geoscientists dis_tinguish twelve major plates and severalmicroplates. Note that some plates consistentirely of oceanic l ithosphere whereas someplates consist ofboth oceanic antJ conttnen-tal lithosphere.

Lithosphere bends whi leasthenosphere flows. t

l

Time 2

"*i:-' 'c '.i s"' " "':

:$i:!;tur,iii;F*3r j.rl'ta8r r.rril

Asthenosphere is irrelatively soft and I,rLr4r lvtrry )urL dl lu i : ,able to flow. I*f ,ii:'tt.r!!i;:{::,i:!!:e:,i:r j$!iiF!,i"";r.l:'

(b) The l t thosphere consjsts of the crust plustne upperm.st mantre. rt is thrcker beneath continents than beneath oceans.

plf rU+4tr #"1f l The locat ions of p late boundar ies and the distr ibut ion of earthquakes.

CHAPTER z The Way the Earth Works: p lare Tec

oceanrc l i thosphere, whereas some consist of both continental and oceanic l i thosphere.margins do not.

(a) A map of major plates shows that some consist entirely olActive continental margins l ie along plate boundaries, passive

As illustrated in Figure2.15a, some plate boundaries followcontinental margins, the boundary between a continent andan ocean, but others do not. For this reason, we distinguishbetween two types of continental margins: active

-rr:gi..,are plate boundaries, whereas passive rrrargins are not plateboundaries. Along passive margins, continental crust is thinnerthan in continental interiors (Fig. 2.1ab). Thick (10 to 15 km)accumulations of sediment cover this thinned crust. The surfaceof this sedimentlayer is a broad, shallow (less than 500 m deep)

(b) The locations of earthquakes (red dots) mostly fal l in dist inct bands that correspond to plate boundaries. Relat ively few earthquakes occur jn thestabler plate inter iors.

region cailed the continental shelf; home to the major fisheriesof the world.

Tfum ffims$* Pr$mc6p*ew s$ ffi$**s Tes.f*ffifr*sWith the background provided above, we can restate plate tec-t-onics theory concisely as follows. The Earth's lithosphere isdivided into plates that move relative to each other and reratiyeto the underlying asthenosphere. Plate movement occurs at

?,r Divergent Plate Boundar ies and Sea_Floor Spreaorng

rates of about 1 to 15 cm per year. As a plate moves, its inter_nalarea remains largely rigid and intact, but rock along plateboundaries undergoes deformation (cracking, sliding,Gnd_ing, stretching, and squashing) as the plate grinds oi r.r"p.,against its neighbors or pulls away from its neighbors. Asplates move, so do the continents that form part oithe plates,resulting in continental drift. Because of plate tectonics, themap of Earth's surface constantly changes.

kf **txf y$r"* g ff Smtm ffi m*-c s"${$m rismHow do we recognize the location of a plate boundary? Theanswer becomes clear from looking at a map showing the loca_tions of earthquakes (Fig.2.15b). Recall from chapter 1 thatearthquakes are vibrations caused by shockwaves th;t are gen_erated where rock breaks and suddenly shears (slides) al"onga fault (a fracture on which sliding occurs). The focus of thIearthquake is the spot where the fault s1ips, and the epicentermarks the point on the surface of the Earth directly above thefocus. Earthquake epicenters do not speckle the Earth's sur_face randomly, like buckshot on a target. Rather, the major_ity occur in relatively narrow, distinct belts. These earthqu;kebelts define the position ofplate boundaries, because the frac_turing and slipping that occur along plate boundaries as platesmove generate earthquakes. Plate interiors, regions awalfromthe plate boundaries, remain relatively earthquake_free becausethey do not accommodate much movement. While earth-quakes serve as the most definitive indicator of a plate bound_ary, other prominent geologic features also de.relop along plateboundaries, as you will learn by the end of this .hupt"r.

" ^

Geologists define three types of plate boundaries, basedsimply on the relative motions of the plates on either side of theboundary (Fig.2.16a-.c). A bounciary at which two plates move

49

apart from each other is a divergent boundary. A boundaryat which two plates move toward each other so that one platlsinks benearh the orher is a convergent boundary. And aboundary at which one plate slips along the side of anotherplate is a transform boundary. Each type of boundary looksand behaves differently from the others, as we will now see.

2 B DIVERGENT PLATE BOUNDABIESAND SEA-FLOOR SPREADING

At divergent boundaries, or spreading boundaries, two oceanicplates move apart by the process of sea-floor spreading. Notethat an o?en s?ace does not clevelap between diverginiplates.Rather, as the plates move apart, new oceanic lithosphere'form,along the divergent boundary (Fig.2.IZa). This piocess takesplace at a submarine mountain range called a mij_ocean ridgethat rises 2 km above the adjacent abyssal plains ofthe ocean.Thus, geologists also commonly refer to a divergent boundaryas a mid-ocean ridge, or simply a ridge.

To characterize a divergent boundary more completely, 1et,slook at one mid-ocean ridge in more detail (Fig. i.UAi.'fti'Mid-Atlantic Ridge extends from the waters between north_ern Greenland and northern Scandinavia southward across

i r :1f i ! ."rf i r j :a. j4: The three types of

Lithosphere thickens i:away from the axis. i.r#,i j:liprry.!1f tii:.f .:gjrtt,l:*,,uu,:,r.,,,'r,",:.i:

(a) At a divergent boundary, two platesmove away from the axis of a mid_oceanr idge. New oceanrc l i thosphere forms.

piate boundar ies di f fer based on the nature of re lat ive movement

Volcanic arc

(b) At a convergent bouncary, rwo ptatesmove toward each other; the downgoingplate s inks beneath the overr id ing plate.

TJle process of consuming a ir No new plate forms, and no iplate is called subduction, '!; ,r.l -r.n-'r'^^^"...:l;.'-

,-.,r!:!i,,;:ei,ar!r!i!.ii,riiii*ii:!;r'ei!,!iF.ri:eii!,e;r!:ii.iriiii

:l3r.l*,.t"],,:,:.:.''|T;,1;..,o,,.*.;

(c) At a t ransform boundai .y, two plates s l ic iepasr eacn other on a vert ical faul t surface.

Transform fault

50

F t{3 ii !?*

Time 1

CHAPTER z The Way the Earth Works: plate Tectonics

|me z

: i . i i ' The process of sea-f loor spreading.

New ocean floor

/+

The youngest oceanic i;crust occurs along the i

,.,,;rl ridge axis. i1':' ,i'rr;.5;,;11p1;y13;.1i}:e;l!:!ie..de;!u:!:e.r!:!:';re:ii

floor floor floor(a) As sea-f loor spreading progresses, new oceanic l i thosphere forms at themid-ocean r idge axis. For s impl ic i ty, only the crust js shown.

(bl rhe bathymetry of the Mid-At lant ic Ridge in the Southern At lant ic oceanare shal lower depths

(c) Oceanic crust consists of d ist inct layers. p i l lowbasal t forms from extruded lava, dikes f rom r is ingmagma that sol id i f ied in cracks, and gabbro f rommagma that sol id i f ied at depth. Sedimentaccumulates as sea f ioor ages.

Sediment Fault scarpPillow basalt

Mid-oceanridge axis

Asthenosphere

The sea floor slopes away from theridge axis, reaching the depth of theabyssal plain (4 to 5 km) at a distance ofabout 500 to 800 km from the ridge axis(see Fig. 2.7). Roughly speaking, theMid-Atlantic Ridge is symmetrical-itseastern half looks like a mirror imageof its western half. The ridge consists,along its length, of short segments(tens to hundreds of km long) linked bybreaks called transform faults, which wewill discuss later.

F"**w ##es #***sai* #r$wgFwrm mt m ffi$d-#*emm ffi$S6*?

. The lighter colors As sea-floor spreading takes place, hotasthenosphere rises beneath the ridge(Fig. 2.17c). As this asrhenosphere rises.

it begins to melt, producing molten rock, or magma. Magmahas a lower density than solid rock, so it behaves buoyantly andrises, as oil rises above vinegar in salad dressing. It eventuallyaccumulates in the crust below the ridge axis, filling a region

Oldestoceanfloor

Olderoceanfloor

Younger Older Oldestocean ocean ocean

the equator to the latitude opposite the southern tip of SouthAmerica. Geologists have found that the formation of new seafloor takes place only along the axis (centerline) of the ridge.The axis lies at water depths of about 2 to 2.5 km.

FIGi lHl : ; : . "1t t A column of superhot water gushing from avent (known as a "black smoker") a long the mid-ocean rrdge.The cloud of "smoke" actual ly consistes of t iny mineral grains.

called a magma chamber. As the magma cools, it turns into amush of crystals. Some of the magma solidifies completely alongthe side of the chamber to make the coarse-grained, mafic igne-ous rock called gabbro. The rest rises still higher to fill verticalcracks, where it solidifies and forms wall-like sheets. or dikes. ofbasalt. Finally, some magma rises all the way to the surface ofthe sea floor at the ridge axis and spills out of small submarine

FlGUi l f , I -$tr This map of the wor ld shows the age of the sea f loor.f rom the rrdge axis. (Ma = mi l l ion years ago)

volcanoes. The resulting lava cools to form alayer of basalt. Thisbasalt occurs in small blobs called pillows. We cant easily see thesubmarine volcanoes because they occur at depths of more than2km beneath sea level, but they have been observed by the researchsubmarine Alviz. Observers in the submarine have also detectedchimneys spewing hot, mineralized water that rose throughcracks in the sea floor, after being heated by magma below thesurface. These chimneys are called black smokers because thewater they emit looks like dark smoke; the color comes from asuspension of tiny mineral grains that precipitate in the water asthe water cools (Fig. 2.18).

As soon as it forms, new oceanic crust moves away fromthe ridge axis, and when this happens, more magma rises frombelow, so still more crust forms. In other words, like a vast,continuously moving conveyor be1t, magma from the mantlerises to the Earth's surface at the ridge, solidifies to formoceanic crust, and then moves laterally away from the ridge.Because all sea floor forms at mid-ocean ridges, the youngestsea floor occurs on either side of the ridge axis, and sea floorbecomes progressively older away from the ridge. In the Atlan-tic Ocean, the oldest sea floor lies adjacent to the passive con-tinental margins on either side of the ocean (Fig. 2.19). The

Note how the sea f loor grows older wi th increasinq dtstance

Ma 0-5rPleistoceneto Pliocene

ffiMiocene

2I_38

Fi,ffiOligocene

52-65

Paleocene

65-145

FfrHHCretaceous

145-1 60

WJurassic

52

oldest ocean floor on our planet is in the

western Pacific Oceanl this crust formed

200 mill ion Years ago'

The te.tsion (stretching force) applied

to newly formed solid crust as spreading

takes place breaks this new crust, resulting

in the formation of faults' Movement (slip)

on the faults causes divergent-boundary

earthquakes and produces numerous cliffs'

o. r."rpr, that lie parallel to the ridge axis'

M*w fl)}**s 9fum &-!€h*sP&t*r**$V$ms'*€ts F*s'#t me m ffifiq$-ffi**atrxffi*dge?

So far, we've seen how oceanic crust forms

at mid-ocean ridges' What about the for-

mation of the mantle part of the oceanic

lithosphere? This part consists of the cooler

.rpp..-or, atea of the mantle, in which

t.-p"rutnr., are less than about 1'280'C'

At the ridge axis, such temperatures occur

almost at ihe base of the crust, because of

cHAPTER z The Way the Earth Works: Plate Tectonlcs

r i {1!- iHH r. ; i * changes accompanying the aging of l i thosphere'

Mid-ocean ridgeElevation

of ridge axls Sea level

iui o, ,"u f loor ages, the dense l i thospheric mantle thickens

and i ts surface gets deeper

(b) Like the bal last of a shiP, older

(thicker) l i thosphere sinks deeper into

the mant le.

the presence of rising hot asthenosphere

ur.d i.ot magma' so the lithospheric mantle .

beneath the riclge axis effecti"ely doesnt exist' But as the newly

formed oceanic crust moves away from the ridge axis'.the crust

and the uppermost mantle directly beneath it gradually cool as

,tr"y tor. tt.u, ,o the ocean above' As soon as mantle rock cools

below 1,280"C, it becomes, by definition' Part of the lithosphere'

As oceanic lithosphere continues to move away from the

ridue axis, it .or-tti"t"' to cool' so the lithospheric mantle' and

i[ill1'. "..^"itritrt"sphere as awhole' grows progressively

il;t (Fig.2.2}a,b)' Note that this process doesn't change

the thickness of the oceanic crust' for the crust formed entireiy

",1f-,. .lag. axis. The rate atwhich cooling and lithospheric

it-ri.t.,tin! occur decreases with distance from the ridge axis'

i;i".;, Uy'trt. time the lithosphere it u!i"l 80 million years old

iif-r"r;".i about reached its maximum thickness' As lithosphere

thickens and gets cooler and denser' it sinks into the astheno-

;;i.;;., like aihip taking on ballast' Thus' the ocean is deeper

orr., oid., o."nt floot than over younger ocean floor'

aries are also known as subduction zones' Because subduction i

at a convergent boundary consumes. old ocean,litT*T::* i

?g CONVERGENT PLATE BOUNDARIESAND SUBDUCTION

At convergent plate boundaries, two plates' at least one of which

is oceanic, move toward one another' But rather than butting

each other like angry rams' one oceanic plate bends and sinks

down into the asthenosphere beneath the other plate' Geologists

refer to the sinking p'ott" as subduction' so convergent bound-

thus closes (or "consume'") ottu"it basins' geologists also refer

,o .o.t r..g.r-tt boundaries as consuming boundaries' and because

,;.y "r"

ietneated by deep-ocea" '1:"*:t'

they are sometimes

,i-pft called trenches (see Fig' 2'B)' The amount of oceanic

plate consumption worldwide,"averaged over time' equals the

amount of sea-floor spreading worldwide' so the surface area of

the Earth remains constant through ti'me'

Subduction occurs for a simpie reason: oceanic lithosphere,

once it has aged at least 10 million years' is denser than asthe-

nosphere and thus can sink through the asthenosphere' When it

iies flat on the surface ofthe asthenosphere' oceanic lithosphere

doesn't sink, because the resistanct of1h" asthenosphere to flow

is too great' However, once the end of the convergent plate bends

Jo*.t'u,la sfips into the mantle, it begins to sink like an anchor

falling to the bottom of a lake (Fig' 2'21a)' As the lithosphere

I

\

a^- /o no^- D ato Rnr.nda. ies and SJbdLcLionww ,re, Yv,

: Drrr 'nn rrp nrnatrqs Of subduct ion, oceanic i thosphere

^r. t int^ +he daanor mzn+l^uou\ | rLv . r 19 uvuPUr rr 'ur lcrg,

Floating

anchor

Sinking of the^l^+^

position

tr:l

E

@

200

400

600

800

1,000

r,200

1,400

1,600

0 es tnegofan

attached to

The downgoing plate ::slowly sinks down, into i:the asthenosphere. :i

Fault belt dueto compression

Continentalvolcanic arc

(b) A belt of earthquakes (dots) defines the posit ion of thedowngorng p ate in the region above a depth of about 660 krn.

The overriding plate acts like a bulldozer,scraping sediment off the downgoing plateto bui id an accret ionary pr ism.

Marginal sea (back-arc basin)due to extension

Volcanicisland arc

TrenchRisingmagma

Partial melting

A mode of a subduct ion zone aiong theof a contrnent. Note the posi t ion of theh, accret ionarV pr ism, and volcanic arc.

Accretionarynri smr -_-___

TrenchAXlS

(d) A model of a subduct ionarong an is aro arc. Note Lheback-arc basin, here f ormed

Forearcbasin

nnci t inn nf +ho

by spreaorng.

53

Volcanic arc

New trench forms

54

sinks, asthenosphere flows out of its way, just as water flows outof the way of an anchor. Even though it is relatively soft andplastic, the asthenosphere resists flow, so oceanic lithosphere cansink only very slowly, at a rate ofless than about 15 cnlper vear.

Note that the downgoing plate, the plate that has teensubducted, tnust be composed of oceanic lithosphere. Theoverriding plate, which does not sink, can consist of either oce-anic or continental lithosphere. Continentai crust cannot besubducted because it is too buoyant; the low-density rocks ofcontinental crust act like a iife preserver keeping the continentafloat. If continental crust moves into a convergent margin, sub-duction eventually stops. Because ofsubduction, all ocean floo,on the planet is less than about 200 million years old. Becausecontinental crust cannot subduct, ,o-. .orrti.rental crust haspersisted at the surface ofthe Earth for over 3.8 billion years.

At convergent plate boundaries, the downgoing plategrinds along the base of the overriding plate, a process thargenerates large earthquakes. These earthquakes occur fairlyclose to the Eartht surface, so some of them trigger massivedestruction in coastal cities. But earthquakes also happen indowngoing plates at greater depths, deep below the overridingplate. In fact, geologists have detected earthquakes withindowngoing plates to a depth of 660 km. The telt of earth-quakes in a downgoing plate is called a Wadati-Benioffzorre, after its two discoverers (Fig.2.21b).

At depths greater than 660 km, conditions leading toearthquakes evidently do not occur. Recent observationr, ho*-ever, indicates that some downgoing plates do continue to sinkbelozo a depth of 660 km-they just do so without generat_ing earthquakes. In fact, studies suggest that the lower mantlemay be a graveyard for old subducted plates.

ffiemfi*69&c Fem€sxrew m$ m ffi*envetr##slt ffiwwndmryTo become familiar with the various geologic features thatoccur along a convergent plate boundary, let's look at an exam_ple, the boundary between the western coast of the SouthAmerican Plate and the eastern edge of the Nazca plate (aportion of the Pacific Ocean floor). A deep-ocean trench, thePeru-Chile tench, delineates this boundary (see Fig.2.I7b).Such trenches form where the plate bends as it starts to sinkinto the asthenosphere.

In the P.r.r--Chile tench, as the downgoing plate slidesunder the overriding plate, sediment (clay a"d plarrkton) thathad settled on the surface of the downgoing plate, as well assand that fell into the trench from the shores of South Amer_ica, gets scraped up and incorporated in a wedge-shaped massknown as an accretionary prism (Fig.2.27c). An accretionaryprism forms in basically the same way as a pile of snow in frontof a plow, and like the snow, the sediment tends to be squashedand contorted during the formation of the prism.

CHAPTER z The Way the Earth Works: plate Tecton cs

A chain of volcanoes known as a volcanic arc developsbehind the accretionary prism (see Geotour 2 on p. GT:6 to seea volcanic arc and other features of plate boundaries). As we willsee in Chapter 4, the magma that feeds these volcanoes forms ator just above the surface of the downgoing plate when the platereaches a depth of about 150 km below the Earth's surface. Iithevolcanic arc forms where an oceanic plate subducts beneath conti-nental lithosphere, the resulting chain ofvolcanoes grows on thecontinent and forms a continental volcanic arc (see trig.2.27c).Insome cases, the plates compress together, causing a belt of faultsto form behind the arc. If, however, the volcanic arc forms whereone oceanic plate subducts beneath another oceanic plate. theresulting volcanoes form a chain of islands known as a volcanicisland arc (Fig. 2.21d). A marginal sea, or back-arc basin, thesmall ocean basin between an island arc and the continent, formseither in cases where subduction happens to begin offshore, trap_ping ocean lithosphere behind the arc, or where stretchine ofthe lithosphere behind the arc leads to the formation of a smallspreading ridge between the arc and the continent.

21A TRANSFORM PLATE BOUNDARIESWhen researchers began to explore the bathymetry of mid_ocean ridges in detail, they discovered that mid-ocean ridses arenot smooth, uninterrupted lines, but rather consist of shoit ses_ments that appear to be offset from each other (see Fig.2.t7i).Narrow belts of broken and irregular sea floor lie roughly atright angles to the ridge segments, intersect the ends of th. ,.e_ments, and extend beyond the ends of the segments. These beliscame to be known as fracture ,or..r. Oriqinally, researchersincorrectly assumed that the entire length oieach fracture zonewas a fault, and that slip on a fracture zonehad displaced seg_ments of the mid-ocean ridge sideways, relative to each other.In other words, they imagined that a mid-ocean ridge initiatedas a continuous, fence-like line that only later was broken upby faulting' But when informarion about the distribution ofearthquakes along mid-ocean ridges became available, it wasclear that this model could not be correct. Earthquakes, andtherefore active fault slip, occur only on the segment of a frac_ture zone that lies between two ridge segments (Fig.2.22a).The portions offracture zones thatextendbeyond the edees ofridge segments, out into the abyssal plain, are not active.

2.10 Transform Plate Boundaries

FiGLifi i l i :.: ' : :.1 The concept of transform faulting

Youngerplate Older

plate

(a) Transform faul ts on the ocean f loorthe fault between the two r idge axes is

Mid-oceanridge

l ink r idge segments. Onlythe actrve t ransform.

Note that slip on adoes not offset the

transformridge axis.

(b) The sl ip direct ion on a t ransformwith the conf igurat ion of spreading.can stay constant over t ime.

Note that the transformis only between X and y.-.i:.;:s!rf i.3ilf, iilr r.i!,;!:i!f 3

The plate at point l isyounger than at point 2. ;

55

Time

Time 2

faul t must be compat ib leThe length of a t ransform

At its northern end, the SanAndreas links to the Cascade f,Trench and an oceanic transform. i1

At its southern end, the San IAndreas links to a mid-ocean Iridge in the culf of California. iW

{c) The San Andreas Faul t is a t ransform plate boundarybetween the North American and paci f ic p lates. The paci f ic ismoving northwest, re lat ive to North Amerjca.

(d) ln southern Cal i fornia, the San Andreas Faul t cutsa dry Iandscape. The faul t t race rs in the narrow val leyThe land has been pushed up sl ight ly along the faul t .

56

The distribution of movement along fracture zones remained

a mystery until a Canadian researcher, J . Tuzo Wilson, began to

think about fracture zones in the context ofthe sea-floor spread-

ing concept. Wilson proposed that fracture zones formed at the

same time as the ridge axis itself and thus the ridge consisted of

separate segments to start with. These segments were linked (not

offset) by fracture zones. With this idea in mind, he drew a sketch

map showing two ridge-axis segments linked by a fracture zone,

and he drew arrows to indicate the direction that ocean crust was

moving, relative to the ridge axis, as a result of sea-floor spread-

ing (Fig. 2.22b). Look at Wilson's arrows. Clearly, the move-

ment direction on the active portion of the fracture zone must

be opposite to the movement direction that researchers originally

thought occurred on the structure. Further, in Wilson's model,

slip occurs only along the segment of the fracture zone between

the two ridge segments. Plates on opposite sides of inactive frac-

ture zones move together, as one plate.

Wilson introduced the term transform boundary (or

transform fault) for the actively slipping segment of a fracture

zone between two ridge segments, and he pointed out that

these are a third type of plate boundary. At a transform bound-

ary, one plate slides sideways past another, but no new plate

forms and no old plate is consumed. tansform boundaries are,

therefore, defined by a vertical fault on which the slip direction

oarallels the Earth's surface.

t i i i r . , ! I t ! ; r ; t i : Examples of t r ip le junct ions. The tr ip le junct ions aremarked by dots.

Plate(-

CHItpTER 2 The Way the Earth Works: Plate Tectonrcs

So far we've discussed only transforms along mid-ocean

ridges. Not all transforms link ridge segments. Some, such asthe Alpine Fault of New Zealand, link trenches, while otherslink a trench to a ridge segment. Further, not all transform faults

occur in oceanic lithosphere; a few cut across continental litho-

sphere. The San Andreas Fault, for example, which cuts across

California, defines part of the plate boundary between the North

American Plate and the Pacific Plate-the portion of Californiathat lies to the west of the fault (including Los Angeles) is part

of the Pacific Plate, while the portion that lies to the east of the

fault is part of the North American Plate (Fig. 2.22c, d).

211 SPECIAL LOCATIONSIN THE PLATE MOSAIC

-$"*"$ p: f; e # a.p m*€**s"a w

Geologists refer to places where three plate boundaries intersect

at a point as triple iunctions, and name them after the types

of boundaries that intersect. For example, the triple junction

formed where the Southwest Indian Ocean Ridge intersects

two arms of the Mid-Indian Ocean Ridge (this is the triplejunction of the African, Antarctic, and Australian Plates) is a

ridge-ridge-ridge triple junction (Fig. 2.23a) . The triple junc-

tion north of San Francisco is a trench-transform-transform

tripie junction (Fig. 2.23b).

ffi**"ffigx**t*Most subaerial (above sea level) volcanoes are situated in the

volcanic arcs that border trenches. Volcanoes also 1ie along mid-

ocean ridges, but ocean water hides most of them. The voicanoes

ofvolcanic arcs and mid-ocean ridges are plate-boundaryvolca-

noes, in that they formed as a consequence of move-

ment along the boundary. Not all volcanoes on Earth

are plate-boundary volcanoes, however. For example,

the big island of Hawaii, a large volcano, lies in the

middle of the Pacific Plate, and Yellowstone National

Park, site of recent volcanic activity, lies in the inrerior

ofthe North American Plate. Worldwide, geoscientistshave identified about one hundred volcanoes that existas isolated points and are not a consequence ofmove-ment at a plate boundary. These are called hot-spot

volcanoes, or simply hot spots (Fig.2.2q. Most hot

spots are located in the interiors of plates, away from

the boundaries, but a few grew on mid-ocean ridges.

What causes hot-spot volcanoes? In the early

1960s, J. Tuzo Wilson noted that active hot-spot vo1-

canoes (examples that are erui/ting or may erupt in the

future) build an island at the end of a chain of dead

volcanic islands and seamounts (formerly active volca-

(a) A , idge-- idge ' idge Lr ip e l rnct iot 'occ.rs i r - l l^e Indian Ocear.

(b) A Lranstorr- tTencr-rrarsforn t ' ip lejunct on occurs at the north end of theSan Andreas Fau t .

FIGURf; J ;14 The dots represent the locat ions of selected hot-spot volcanoes. The red l ines represent hot-spot t racks. The mostrecentvolcano (dot) rs at one end of th is t rack Some of these are ext inct , indicat ing that the plume no longer exists. Some hot snotqarefair ly recent and do not have tracks. Dashed tracks indicate places where a t rack was broken by sea-f loor spreadjnq.

noes thatwill never erupt again). This configuration is difTerentfrom that ofvolcanic arcs along convergent plate boundaries-atvolcanic arcs, al1 of the volcanoes are active. With this imas.ein mind, Wilson suggested that the position of the heat ,o.rri.causing a hot-spot volcano is fixed, relative to the moving plate.In Wilson's model, the active volcano represents the presJnt_daylocation of the heat source, whereas the chain of dead volcanicislands represents locations on the plate that were once over theheat source but progressively moveJ off.

A few years later, Jason Morgan suggested that the heatsource for hot spots is a mantle plume, a column of very hotrock rising up through the mantle to the base of the lithosphere(Fig.2.25a-d). In Morgan's model, plumes originate deep inthe mantle. Rock in the plume, though solid, is ,oft .rro.,gh toflow, and rises buoyantly because it is less dense than ,,r..olrrr.d_ing cooler rock. When the hot rock of the plume reaches thebase of the lithosphere, it partially melts (foireasons discussedin Qfinpter 4) and produces magma that seeps up through thelithosphere to the Earth's surface. The chain of extinct volcu-noes, or hot-spot track, forms when the overlying plate movesover a fixed plume. This movement slowly carries the volcano offthe top of the plume, so that it becomes extinct. A new, youngervolcano grows over the plume. While the deep-mantle plulemodel is easy to visualize, it is not the only explanation fo. hocspot tracks, and it remains controversid. So-. geologists arguethat plumes originate at shallow depths in the mantle, or thathot spots are not due to plumes at all, but to other phenomena.

The Hawaiian chain provides a clear """-pl"

of the vol_canism associated with a hot-spot track. Volcanic eruptions

occur today only on the big island of Hawaii. Other islandsto the northwest are remnants of dead volcanoes, the oldest ofwhich is Kauai. To the norrhwest of Kauai, still older volca_nic remnants are found. To the northwest of Midwav Island.the track bends in a more northerly direction, and the volcanicremnants no longer poke above sea level; we refer to this seg_ment as the Emperor seamount chain. The bend is due to achange in the direction of the Pacific plate about 40 Ma.

Some hot spots lie within continents. I.or example, severalhave been active in the interior of Africa, ancl one ,-ro* rrr-rd.r_lies Yellowstone National Park. The famous geysers (naturalsteam and hot-water fountains) of yellowstone exist becausehot magma, formed above the Yellowstone hot spot, lies notfar below the surface of the park. A few hot ,pot, li. on mid_ocean ridges. Where this happens, a volcanic island may pro_trude above sea level, because the hot spot produce, f"i

-o..magma than does a normal mid-ocean ridge. Iceland, forexample, formed where a hot spot underlies the Mid_AtlanticRidge. The extra volcanism of the hot spot built up the islandof Iceland so that it rises almost 3 km above other places on theMid-Atlantic Ridge.

I

f

5a

l ' : l# i . { f r f f E.J5 The plume hyppothesis for the format ion of

(a) A bathymetr ic map showing thehot-spot tracks of the pacif ic Ocean

,lLithosphere

CHApTEB z The Way the Earth Works: plate Tectonrcs

hot-spot tracks.

A,volcano forms on a movinqplate above a mant le plume." ' "!l!rtf 1::a31TiT'fj!ii3i3.$fr,iliti,l.l,lrjlTi,!$,;r,i_*]i..j.*i

Activevolcano #l

(c) According to the plume modei, the Hawaj ianis lan_d chain is a hot-spot t rack that formed as thePaci f jc Piate moved northwest relat ive to u ptrr l

Plate movement carrieseach successive volcanooff the hot spot.

oevelopment of a hot_spot t rackaccordtng to the plume model.

Fxfincf, erodedvolcanrc is/and

The first volcano movesoff the plume and dies.

Subnarinefan

core

Seanounfor guyof,

rl4ore s/unping Acfivevolcantcis/and

, LUhat a Qeologist lnajnes(dl As a volcano moves off the hot spot, i t gradually sinks below sea levei, due to sinking of the plate, erosion, and slumpjno

What Dr ves Plate Motion?

12 H()W DO PLATE BOUNDARIESF()RM AND DIE?

configuration ofplates and plate boundaries visible on our

today has not existed for a1l of geologic history, and will

exist indefinitely into the future. Because of plate motion,

ic plates form and are later consumed, while continents

and later split apart. How does a new divergent boundary

into existence, and how does a convergent boundary cease

exist? Most new divergent boundaries form when a continent

and separates into two continents. We call this process

A convergent boundary ceases to exist when a piece of

t lithosphere, such as a continent or an island arc, moves

into the subduction zone. We call this process collision.

af; ffiE$*frmg

A continental rift is a linear belt in which continental litho-

sphere undergoes rifting, or pu1ls apart (Fig. 2.26a). The litho-

sphere stretches hotizontally, so it thins vertically, much like a

piece of taffyyou pull between your fingers. Near the surface of

the continent, where the crust is cold and brittle, stretching causes

rock to break and faults to develop. As a consequence of fault-

ing, blocks of rock slide down, leading to the formation of a 1ow

areathat gradually becomes buried by sediment. Deeper in the

crust, and in the undedying lithospheric mantle, rock is warmer

and softer, so stretching takes place in a plastic manner without

breaking the rock. The whole region that stretches is the rift.

As continental lithosphere thins, hot asthenosphere rises

beneath the rift and starts to melt. Molten rock erupts at volca-

noes along the rift. If rifting continues for a long enough time,

the continent breaks in two, a new mid-ocean ridge forms, and

sea-floor spreading begins. The relict of the rift evolves into a

passive margin (see Fig. 2.14b).In some cases, however, rifting

stops before the continent splits in two. Then, the rift remains

as a permanent scar in the crust, defined by a belt of faults, vol-

canic rocks. and a thick layer of sediment.

Perhaps the most spectacular example of an active rift is

located in eastern Africa; geoscientists aptly refer to it as the

EastAfrican Rift (Fig.2.26b, c).To astronauts in orbit, the

rift looks like a giant gash in the crust. On the ground, it con-

sists of a deep trough bordered on both sides by high cliffs

formed by faulting. Along the length of the rift, several major

volcanoes smoke and fume; these include the snow-crested

Mt. Kilimanjaro, towering over 6 km above the savannah. At

its north end, the rift joins the Red Sea Ridge and the Gulf

ofAden Ridge at a triple junction. Another major rift, known

as the Basin and Range Province, breaks up the landscape of

the western United States between Salt Lake City, Utah, and

Reno, Nevada (Fig. 2.26d). Here, movement on numerous

faults tilted blocks of crust to form narrow mountain ranges,

59

while sediment that eroded from the blocks filled the adjacentbasins (the low areas between the ranges).

ffiq*H*frsfr*m

India was once a small, separate continent that lay far to thesouth of Asia. But subduction consumed the ocean betweenIndia and Asia, and India moved northward, fina1ly slamminginto the southern margin of Asia 40 to 50 million years ago.Continental crust, unlike oceanic crust, is too buoyant to sub-duct. So when India collided with Asia, the attached oceanic

plate broke off and sank down into the deep mantle, whileIndia pushed hard into Asia, squeezing the rocks and sedimentthat once lay between the two continents into the 8-km-highwelt that we now know as the Himalayan Mountains. Duringthis process, not only did the surface of the Earth rise, but the

crust became thicker. The crust beneath a collisional moun-

tain range can be up to 60 to 70 km thick, about twice thethickness of normal continental crust.

Geoscientists refer to the process during which two buoyant

pieces oflithosphere converge and squeeze together as collision(Fig. 2.27). Some collisions involve two continents, whereassome involve continents and an island arc. When a collisionis complete, the convergent plate boundary that once existed

between the two colliding pieces ceases to exist. Collisions yieldsome of the most spectacular mountains on the planet, such as

the Himalayas and the Alps. They also yielded major moun-

tain ranges in the past, which subsequently eroded away so thattoday we see only their relicts. For example, the AppalachianMountains in the eastern United States formed as a conse-

quence of three collisions. After the last one, a collision betweenAfrica and North America around 280 million years ago, NorthAmerica became part of the Pangaea supercontinent.

213 WHAT DRIVES PLATE MOTION?WeVe now discussed the many facets of plate tectonics theory(see Geology at a Glance, pp. 62-63). But what drives platemotion? When geoscientists first proposed plate tectonics, they

thought the process occurred simply because convective flow in

the asthenosphere actively dragged plates along, as ifthe plateswere rafts on a flowing river. Thus, early images depicting platemotion showed simple convection ce1ls-e11iptica1 flow paths

rilr1jI

Ii ' i#{ i i4f i : i " t# Dur ing the process of r i f t lng, l i thosphere stretches.

Time 1

Basin

(b) The East Afr jcan Rift is growing today. The Red Seastarted as a r i f t .

^ Exposedl.rl Prdcambrran % rransforn

- // ,l4d-o..on '

;j.,qift\ .1dge -Sartara J

Desert \ftledr'ferranedi-.,

5ea'

L' /l/t

(d) The Basin and Range province is a r i f t . Faul t ing bounds the narrownorth-south-trending mountarns, separated by basins. The arrowsindicate the direct ion of strercnrnq.

60(c) Astronauts can see how r i f t ing has openedup gul fs on ei ther s ide of the Srnai peninsula.

Time 4

(al When cont inental l i thosphere stretchesand thins, faul t ing takes place, and voicanoes erupt. Fventual ly,the cont inent spl i ts in two anc a new ocean basin forms.

6u/f of

2.13 What Drrves Plate Motion?

FlcUt iS 3"; i I Cont inental col l is ion. Ridge-push force develops because mid_ocean ridqes lieat a higher elevation than the adjacent abyssal plains ofthe ocean(Fig.2.28a). To understand ridge-push force, imagine you havea_glas_s containing aTayer ofwater over a layer of honey. By tiltingthe glass momentarily and then returning it to its uprigirt posiltion, you can create a temporary slope in the boundary betweenthese substances. While the boundary has this slope, gravitycauses the elevated honey to push against the glass udj"..,rito th!side where the honey surface lies at lower elwation. The seom_etry of a mid-ocean ridge resembles this situation, for the ,i.fr..of the sea floor is higher along a mid-ocean ridge axis than inadjacent abyssal plains. Gravity causes the elevated lithosphere atthe ridge axis to push on the lithosphere that lies farther do- th.axis, making it move away. As lithosphere moves away from theridge axis, new hot asthenosphere rises to fill the gap. Note thatthe upward movement of asthenosphere beneath a mid_oceanridge is a consequence ofsea-floor spreading, nor the cause.

Slab-pull force, the force that subducting, downgoingplates apply to oceanic lithosphere at a convergent margin, arisessimply because lithosphere that was formed more than i0

-i11io,.years ago is denser than asthenosphere, so it can sink into the

ir ;# i l811 l : . i ! - : i Forces dr iv ing plate mot ions. Both r idge push andslab pul l make plates move.

of convecting asthenosphere-beneath mid-ocean ridees. Atfirst glance, this hypothesis looked pretty good, but o,, .lor..examination, it failed. Among other reasons, it is impossible todraw a global arrangement of convection cells that can exolainthe complex geometry of plate boundaries on Earth. Gradually,geoscientists came to the conclusion that convective flow withinthe asthenosphere occurs, but does not directly drive motion. Inotherwords' hot asthenosphere does rise in some places and sinkin others, probably because of ternperature contrasts, and platemotion overall is a manifestation of this convection, bui theIocal directions of this convective flow do not necessarilv definethe local directions of plate motion. Today, geoscientists favorthe hypothesis that two forces-ridge-push force and slab-pullforce-strongly influence the motion of individual plares.

{a} Subduction consumes an oceanic plateunti l two continents coi l ide.

Collisionalmountain belt

(bl After the col l l is ion, the oceanrc pratedetaches and sinks into the mantle. Rock caught in thecotlrsron zone gets broken, bent, and squashed, and forms a mounta jn ranqe

(a) Ridge push develops because the region of a r i f t is elevated. Like a wedgeof honey with a sloping surface, the mass of the r idge pushes srdeways.

(bl Slab pul l develops because r i thosphere is denser than the underrv inoasthenosphere, and sinks l ike a stone in water (though much more slowlv)l

:

**

"* ,-l:

* 1.*---ea*rr . -

Ho1-rpotvotcano

s

t plateboundary

Subducting oceaniclithosphere

Continental lithosphere

r l

er port ion of the Earth is a relat ively r ig id layer cal led the l i thosphere. l t consists of the crustic or cont inental) and the uppermost mant le. The mant le below the l i thosphere is relat lvelv

plast ic ( i t can f low) and is cal led the asthenosphere. The di f ference in behavior (r ig id vs. p last ic)between l i thospher ic mant le and asthenospher ic mant le is a consequence of temperature-theformer is coolerthan the lat ter . Cont inental l i thosphere istypical ly about 150 km thick, whi le oceanicl i thosphere is about 100 km thick. (Note: they are not drawn to scale in th is image.)

According to the theory of p late tectonics, the l i thosphere is broken into about twenty platesthat move relat ive to each other. Most of the mot ion is accommodated by s l id ing along plate bound-ar ies ( the edges of p lates); p late inter iors stay relat ively unaffected by th is mot ion. There are threekinds of p late boundar ies.

1. Divergent boundar ies: Here, two plates move apart by a process cal led sea-f loor spreading. A mid-ocean r idge del ineates a divergent boundary. Asthenospher ic mant le r ises beneath a mid-ocean

Oceali tho

l ,l lI

l

:i ii

;{;

, ' . , l l "i'-+*.,t

h, l

Divergent plate boundary

Inactive (extinct)not-spot volcano

v",

hi,,, , ,ridge and partially melts, forming magma. Themagma rises to create new o.u*i" ciust. TheI it h os p h e ri c m a ntl e t h i c ke n s p ro g r" r, i.r*ui |'u *"yfrom the ridge axis as the plate cools.

2. Convergent boundaries: Here, two plates movetogether, and one plate subdu"t, t"n"utnanother ( i t s inks down into the mant le) . Ontyoceanic l ithosphere can subduct. At the Eartn,ssurface, the boundarv hretrnreon ,r.^ -,",^-^l-ma rked ov u 0"" p_ ol: j""lffiil jl: :ffi llXH::

t ion, melting above the Oo*ngoing Jfu't lpro-duces magma that rises to forrn-a uo"t"unia "r".3. Transform boundar ies: Here. one plate s l idessideways past another, wi thout the'creat ion ota new plate or the subduct ion of an otJ on".The boundary is marked by a targe t"r f r ,

" i_"_tul" o.? which s l id ing occurs. Transform tounO_aries l ink segments of mid_ocean , iOgu".JnuV

may also cut through cont inental t i tnJsphere.A point at which three plate boundar ies meetis called a triple junction. f i, is t ig"re sf,l in),

" ,r,ofujunct ion where three mid_ocu"n n-Jg", . ; " i . * r . lur"two cont inents col l ide, a col l is ional mountain bel tforms. This happens because cont inental crust istoo buoyant to be subducted. At a cont inental r i f t , acont inent stretches and may Or"at in iwo. 'Ri ta, ur"marked by the existence of many tautts. t t a cont i_nent breaks apart , a new mid_ocean r iOg; Jeuelops.

Hot-spot volcanoes form above of i , rn"" of hotmant le rock that r ise f rom near th; ; ; r l_rnun,,"boundary. As a plate dr i f ts over a t rot ,pot , l t leavesa chain of ext inct volcanoes.

. l' : l

nlc

lc

le

$F

Asthenosphere

Mantle/plume

t- II

CHAPTER 2 The Way the Earth Works: Plate Tectontcs I

II

of plate A with respect to plate B, then we are talking about Irelative plate velocity. But if we describe the movement of

Iboth plates relative to a fixed point in the mantle, then we are

Ispeaking of absolute plate velocity (Fig.2.29).

ITo determine relative plate motions, geoscientists measure I

the distance of a known magnetic anomaly from the axis of a Imid-ocean ridge, and then calculate the velocity of a plate reia- Itive to the ridge axis by applying this equation: plate velocity =

How fast do plates move? It depends on yourframe of reference.

To illustrate this concept, imagine two cars speeding in the

same direction down the highway. From the viewpoint of a

tree along the side of the road, car A zips by at L00 km an

hour, while car B moves at B0 km an hour. But relative to car

B, car A moves at only 20 km an hour. Likewise' geologists

use two different frames of reference for describing plate veloc-

ity (velocity = distance/time)' If we describe the movement

64

asthenosphere (Fig. 2.28b). Thus, once an oceanic plate starts to

sink, it gradually pu11s the rest of the plate along behind it' like

an anchor pulling down the anchor line. This "pu11" is the slab-

pull force.

distance from the anomaly to the ridge axis divided by the age

of the anomaly. The velocity of the plate on one side of the

ridge relative to the plate on the other is twice this value.

To estimate absolute plate motions' we can assume that the

position of a mantle plume does not change much for a long

214 THE VELOCITY OF PLATE MOTIONS time. If this is so, then the track of hot-spot volcanoes on the

plate moving over the plume provides a record of the plate's abso-

lute velocity and indicates the direction of movement. (In real-

ity, piumes are not completely fixed; geologists must use other'

more complex methods to calculate absolute plate motions.)

Working from the calculations described above' geologists

have determined that relative plate motions on Earth today occur

at rates of about 1 to 15 cm per year. But these rates, though

small, can yield large displacements given the immensity of geo-

logic time. At a rate of 10 cm pet yeat, a plate can move 100 km

Fi{ j r* t i } t r " l l . i : * Relat ive plate veloci t ies: the blue arrows show the rate and direct ion at which the plate on one side of the boundary is

moving with respect to the plate on the other s ide. Outward-polnt ing arrows lndicate spreading (divergent boundar ies), inward-point ing

arrows indicate subduct ion (convergent boundar ies), and paral le l arrows show transform mot ion. The length of an arrow represents the

veloci tv. Absolute plate veloci t ies: the red arrows show the veloci ty of the plates wi th respect to a f ixed point in the mant le.

+ Absolute plate motions + Re:ative plate motions(5.5 cm per Year)

.|JJ- Convergent boundary-

Qidoe - Transform

Chapter Summary

FIGURH 2.3{} Due to plate tectonics, the map of Earth 's surface slowly changes

breakup, of Pangaea dur ing the past 400 mi l l ion years.

4ffiMa

l5O Ma

h a million years. Can we detect such slow rates? Until the last

decade, the answer was no. Now the answer is yes, because of

satellites orbiting the Earth with global positioning system

(GPS) technology. Automobile drivers use GPS receivers to

find their destinations, and geologists use them to monitor plate

motions. If we calculate carefully enough, we can detect dis-

placements of millimeters per year. In other words, we can now

see the plates movel

Taking into account many data sources that define the

motion of plates, geologists have greatly refined the image of

continental drift that Wegener tried so hard to prove nearly a

Chaptmr Smntx'x"xm*"P

r Alfred Wegener proposed that continents had once been

joined together to form a single huge supercontinent (Pan-

gaea) and had subsequently drifted aPart. This idea is called

continental drift.

r Wegener drew from several different sources ofdata to sup-

porthis hypothesis: (1) the correlation ofcoastlines; (2) the

distribution of late Paleozoic glaciers; (3) the distribution

of\atePaleozoic equatorial climatic belts; (4) the distribu-

tion of fossil species; (5) the match-up of distinctive rock

assemblages that are now on opposite sides of the ocean

were adjacent on Pangaea.

Rocks retain a record of the Earth's magnetic field as it existed

at the time the rocks formed. This record is called paleomag-

netism. By measuring paleomagnetism in successively older

rocks, geologists discovered apparent polar-wander paths.

Apparent polar-wander paths are different for different con-

tinents. This observation can be expiained by movement of

Here we see the assembly, and later the

century ago. We can now see how the map of our planet's sur-

face has evolved during the past 400 million years (Fig. 2.30),

and even before.

continents with respect to each other, while the Earth's mag-

netic poles remain roughly fixed.

Around 1960,Harry Hess proposed the hypothesis of sea-

floor spreading. According to this hypothesis, new sea floor

forms at mid-ocean ridges, above a band of upwelling man-

tle, then spreads symmetrically away from the ridge axis'

Eventually, the ocean floor sinks back into the mantle at

deep-ocean trenches.

Geologists documented that the Earth's magnetic field

reverses polarity every now and then' The record of rever-

sals is called the magnetic-reversal chronology.

The proof of sea-floor spreading came from the interpretation

of marine magnetic anomalies and drilling of the sea floor'

The lithosphere, the rigid outer layer of the Earth, is bro-

ken into discrete plates that move relative to each other.

Plates consist of the crust and the uppermost (cooler)

mantle. Lithosphere plates effectively float on the under-

lying soft asthenosphere. Continental drift and sea-floor

spreading are manifestations of plate movement.

25O Ma

e F*ApTHR ? The Way the Earth Works: Plate Tectonlcs

Most earthquakes and volcanoes occur along plate boundar-

ies; the interiors of plates remain relatively rigid and intact.

There are three types of plate boundaries-divergent, con-

vergent, and transform-distinguished from each other by

the movement the piate on one side of the boundary makes

relative to the plate on the other side.

Divergent boundaries are marked by mid-ocean ridges. At

divergent boundaries, sea-floor spreading takes place, a pro-

cess that forms new oceanic lithosphere.

Convergent boundaries are marked by deep-ocean trenches

and volcanic arcs. At convergent boundaries, oceanic lith-

osphere of the downgoing plate is subducted beneath an

overriding plate.

tansform boundaries are marked by large fauits at which

one plate slides sideways past another. No new plate forms

and no old plate is consumed at a transform boundary.

tiple junctions are points where three plate boundaries

intersect.

Hot spots are places where volcanism occurs at an isolated

volcano. As a plate moves over the hot spot, the volcano

moves off and dies, and a new volcano forms over the hot

spot. Hot spots may be caused by mantle plumes.

A large continent can split into two smaller ones by the

process of rifting. During rifting, continental lithosphere

stretches and thins. If it finally breaks apartt ^

new mid-

ocean ridge forms and sea-floor spreading begins.

Convergent plate boundaries cease to exist when a buoyant

piece of crust (a continent or an island arc) moves into the

subduction zone. When that happens, collision occurs.

Ridge-push force and slab-pu11 force contribute to driving

plate motions. Plates move at rates of about 1 to 15 cm peryear.Modern satellite measurements can detect these motions.

ffiffikry_

f:T**'*,; *:yri j L**1 t:i1q1r*

1. What was Wegener's continental drift hypothesis? What

was his evidence?

2. How do apparent polar-wander paths show that the

continents, rather than the poles, have moved?

3. Describe the basic characteristics of mid-ocean ridges,

deep-ocean trenches, and seamount chains.

4. Describe the hypothesis of sea-floor spreading.

5. How did the observations of heat flow and seismicity

support the hypothesis ofsea-floor spreading?

6. What is a marine magnetic anomaly? How is it detected?

7. Describe the pattern of marine magnetic anomalies across

a mid-ocean ridge. How is this pattern explained?

absolute plate velocity (p 6a)

abyssal plain (p. 42)

active margin (p. 48)

apparent pol ar-wander path

asthenosphere (p. 47)

hrthrrmernr ( . 4O)"* .^.- I . . . ' . . - I \ r . ' " ,

black smoker (p. 51)

collision (p 59)

continental drift (p. 34)

continental rift (p. 59)

convergent boundary (p. a9)

divergent boundary $. a9)f reet , r re .^^. (n .4,) \

global positioning system(GPS) (p 6s)

hot spot (p. 56)

hot-spot track (p. 57)

lithosphere @.a7)lithosphere plate (p. 47)

magnetic anomaly (p. a3)

magnetic declination (p. 3B)

magnetic dipole (p. 3B)

magnetic inclination (p. 3B)

magnetic pole (p. 3B)

magnetic reversal (p. aa)

magnetic-reversal chronology

@.aa)r r l4rrrrr Prul l lL \P. J/ /

marine magnetic anomaly(p.43)

mid-ocean ridge (p. 42)

paleomagnetism (pp. 37, 39)

paleopole (p. 39)Pangaea (p 3a)

passive margin (p.aB)

plate (p. 34)

plate boundary @. a7)

plate tectonics (p. 34)

relative plate velocity (p. 6a)ridge-push force (p. 61)

rifting (p. 59)sea-floor spreading (p. 3a)

seamount (p. 42)

slab-pu1l force (p. 61)

subduction (p. 3a)transform boundary

(pp. 49,56)

trench (p.a2)

triple junction (p. 56)

volcanic arc (p. 42)

Wadati-Benroff zone (p 5a)

Many l ines of evidence indicate that the posi t ion ofcont inents indeed changes over t ime, and thus thatthe map of the Earth 's surface is not f ixed. Not only docoastl ines match, but the observed distribution of rockuni ts, fossi ls. and cl imate bel ts, and evidence of pastglaciat ions al l point to the occuirence of cont inentaldrift. Drift occurs because ocean basins grow widerby the process of sea-floor spreading, or get narrowerby the process of subduction. The documentation andinterpretat ion of marine magnet ic anomal ies provedthat sea-f loor spreading does happen.

0n Further Thought

B. Did drilling into the sea floor contribute further proof of

sea-floor spreading? If so' how?

9. What are the charactqristics of a lithosphere plate?

10, How does oceanic lithosphere differ from continental

lithosphere in thickness, composition, and density?

11. What are the basic premises of plate tectonics?

12. How do we identify a plate boundary?

13, Describe the three types of plate boundaries'

14. How does crust form along a mid-ocean ridge?

15, Why is subduction necessary on a nonexPanding Earth

with spreading ridgesr

16. Describe the major features of a convergent boundary'

17. Why are transform plate boundaries required on an Earth

with spreading and subducting plate boundaries?

18. What is a triple junction?

19. How is a hot-spot track produced, and how can hot-spot

tracks be used to track the past motions of a plate?

20. Describe the characteristics ofa contintental rift, and give

examples of where this process is occurring today'

21. Describe the process of continental collision, and give

examples ofwhere this process has occurred'

22. Discuss the major forces that move lithosphere plates'

23. trxplain the difference between relative plate velocity and

absolute pl ate veiocitY'

1. Why are the marine magnetic anomalies bordering the

East Pacific Rise in the southeastern Pacific Ocean wider

than those bordering the Mid-Atlantic Ridge in the

South Atlantic Ocean?

2. The Pacific Plate moves north relative to the North

American Plate at a tate of 6 cm per year' How long will

it take Los Angeles (a city on the Pacific Plate) to move

northwards by 480 km, the present distance between Los

Angeles and San Francisco?

3. Look at a map of the western Pacific Ocean, and examine

the position ofJapan with respect to mainland Asia' Japans

older crust contains rocks similar to those of eastern Asia'

Presently, there are many active volcanoes along the length

ofJapan. With these facts in mind, explain how theJapan

Sea (the region betweenJapan and the mainland)' formed'

THE vt[\p,r s$t#ld $fi"*k*tr This image was produced by Christoph Hormann, using computer rendering techniques'

W'v,nFtnt !g,l,ws,qt1,19,,?,1?,.."5,1,g,, P,19'1l]f;S3,?,S:?r?:r?F,,f*i'l'l:li,.;?n?jii-1,,i

It shows the Caucasus