GLACTAL CRUSTAL REBOUND, SEA LEVELS AND SHORELTNES 1157

crust may introduce some variability, older passivemargins typically have heat flows similar to oldoceanic crust, whereas young margins, such as inthe Red Sea, have high heat flow.

Gas Hydrates

Gas hydrates, solid composires of biogenically de-rived gasses (mainly merhane) combined with waterice, may be present in marine sediments, especiallyat continental margins, under the correct pressureand temperature conditions. Gas hydrates are detec-ted by direct sampling, and inferred from seismicreflection data when a strong bottom-simulatingreflector (BSR) is produced by the seismic velocitycontrast between the gas hydrate and the sedimentbelow. Given a depth (and thus pressure) of abottom-simulating reflector, its temperature can bepredicted from known relationships and a heat flowcalculated. Alternatively, heat flow data can be usedto estimate the bottom-simulating reflector temper-ature. The volume of hydrocarbons contained inmarine gas hydrates is large. Changes in eustatic sealevel (hence pressure) or bottom seawater temper-atures could result in their release and thus increasegreenhouse gasses and affect global climate.

Further ReadingHumphris S, Mullineaux L, Ziercnberg R and Thomson

R (eds) (1995) Seafloor hydrothermal systems, phys-ical, chemical, biological, and geologicai inreractions.Geophysical Monograph s 91: 425 -44 5.

Hyndman RD, Langseth MG and Von Herzen RP (1987)Deep Sea Drilling Project geothermal measurements:a review. Reuiew of Geophysics 25: 1.563-1.582.

Langseth MG, Jr and Von Herzen RP (1,971) Heat flowthrough the floor of the world oceans. In: Maxwell AE(ed.) The Sea, vol. fV, parr 1., pp.299-352. New York:\7iley-Interscience.

Lowell RP, Rona PA and Von Herzen RP (1995) Seafloorhydrothermal systems. Journal of Geopbysical Re-search '100:

327-352.Parsons B and Sclater JG (1977) An analysis of the

variation of ocean floor bathymetry and heat flow withage. Journal of Geophysical Research 82:803-827.

Pollack HN, Hurter SJ et al. (1993) Heat flow from theearth's interior: analysis of the global data set. Reuiewof Geophysics 3I: 267-280.

Stein CA and Stein S (1992) A model for the globalvariation in oceanic depth and heat flow with litho-spheric age. Nature 359 1.23-729.

Von Herzen RP (1987) Measurement of oceanic heatflow. In: Sammis C and Henyey T (eds) Methods ofExperimental Physics - Geophysics, vol. 24, Parr B,pp. 227-263. London: Academic Press.

Wright JA and Louden KE (eds) (1989) CRC Handbooþof Seafloor Heat Flou. Boca Raron: CRC Press.

land subsidence or of an increase in ocean volume.A major scientific goal of sea-level studies is toseparate out these two effects.

A number of factors contribute ro rhe instabilityof the land surfaces, including rhe tecronic up-heavals of the crust emanating from the Earth'sinterior and the planet's inability to support largesurface loads of ice or sediments without under-going deformation. Factors contributing to theocean volume changes include the removal or addi-tion of water to the oceans as ice sheets wax andwane, as well as addition of water into the oceansfrom the Earth's interior through volcanic activity.These various processes operate over a range oftimescales and sea level fluctuations can be expectedto fluctuate over geological time and are recorded asdoing so.

The study of such fluctuations is more than ascientific curiosiry because its outcome impacts ona number of areas of research. Modern sea level

K. Lambeck, Australian National University,Canberra, ACT, Australia

Copyright O eOOt Academic press

doi:1 0. I 006/rwos.2001 .041 I

IntroductionGeological, geomorphological, and instrumentalrecords point to a complex and changing relationbefween land and sea surfaces. Elevated coral reefsor wave-cut rock platforms indicate that in somelocalities sea levels have been higher in the past,while observations elsewhere of submerged forestsor flooded sites of human occupation atrest to levelshaving been lower. Such observations are indicatorsof the relative change in the land and sea levels:raised shorelines are indicative of land having beenuplifted or of the ocean volume having decreased,while submerged shorelines are a consequence of

In: Encyclopaedia of Ocean Sciences, (J.H. Steele, K.K. Turekian, S.A. Thorpe, Eds.). Academic Press, San Diego, 2, 1157-1167.

1'r58 GLACTAL CRUSTAL REBOUND, SEA LEVELS AND SHORELTNES

change, for example, must be seen against thisbackground of geologically-climatologically drivenchange before contributions arising from the actionsof man can be securely evaluated. In geophysics, oneoutcome of the sea level analyses is an estimate ofthe viscosity of the Earth, a physical property that isessential in any quantification of internal convectionand thermal processes. Glaciological modeling ofthe behavior of Iarge ice sheets during the last coldperiod is critically dependent on independent con-straints on ice volume, and this can be extractedfrom the sea level information. Finally, as sea levelrises and falls, so the shorelines advance and retreat.As major sea level changes have occurred duringcritical periods of human development, reconstruc-tions of coastal regions are an important part inassessing the impact of changing sea levels onhuman movements and settlement.

Tectonics and Sea Level Ghange

Major causes of land movements are the tectonicprocesses that have shaped the planet's surface overgeological time. Convection within the high-temper-ature viscous interior of the Earth results in stressesbeing generated in the upper, cold, and relativelyrigid zone known as the lithosphere, a layer some50-100 km thick that includes the crust. Thisconvection drives plate tectonics - the movementof large parts of the lithosphere over the Earth'ssurface - mountain building, volcanism, and earth-quakes, all with concomitant vertical displacementsof the crust and hence relative sea level changes.The geological record indicates that these processeshave been occurring throughout much of theplanet's history. In the Andes, for example, CharlesDarwin identified fossil seashells and petrified pinetrees trapped in marine sediments at 4000melevation. In Papua New Guinea, 1.20)))-year-oldcoral reefs occur at elevations of up to 400m abovepresent sea level (Figure 1).

One of the consequences of the global tectonicevents is that the ocean basins are being continuallyreshaped as mid-ocean ridges form or as ocean floorcollides with continents. The associated sea levelchanges are global but their timescale is long, of theorder 107 to 108 years, and the rates are small, lessthan 0.01mm per year. Figure 2 illustrates the glo-bal sea level curve inferred for the past 600 millionyears from sediment records on continental margins.The long-term trends of rising and falling sea levelson timescales of 50-100 million years are attributedto these major changes in the ocean basin configura-tions. Superimposed on this are smaller-amplitudeand shorter-oeriod oscillations that reflect more re-

Figure I Raised coral reefs from the Huon Peninsula, PapuaNew Guinea. In this section the highest reef indicated (point 1)is about 340m above sea level and is dated at about 125000years old. Elsewhere this reef attains more than 400m elev-ation. The top of the present sea cliffs (point 2) is about 7000years old and lies at about 20 m above sea level. The intermedi-ate reef toos formed at times when the rate of tectonic uolift wasabout equal to the rate of sea level rise, so that prolongedperiods of reef growth were possible. Phoiograph by Y. Ota.

gional processes such as large-scale volcanism in an

ocean environment or the collision of continents.

More locally, land is pushed up or down epi-sodically in response to the deeper processes. Theassociated vertical crustal displacements are rapid,resulting in sea level rises or falls that may attaina few meters in amplitude, but which are followedby much longer periods of inactivity or even a rcIax-ation of the original displacements. The raised reefsillustrated in Figure 1, for example, are the result ofa Iarge number of episodic uplift events each oftypically a meter amplitude. Such displacements aremostly local phenomena, the Papua New Guineaexample extending only for some 100-150km ofcoastline.

The episodic but local tectonic causes of thechanging position between land and sea can usuallybe identified as such because of the associated seis-mic activity and other tell-tale geological signatures.The development of geophysical models to describethese local vertical movements is still in a state ofinfancy and, in any discussion of global changes insea level, information from such localities is best setaside in favor of observations from more stableenvironments.

Glacial Gycles and Sea Level Ghange

More important for understanding sea level changeon human timescales than the tectonic contributions- important in terms of rates of change and in termsof their globality - is the change in ocean volumedriven by cyclic global changes in climate fromglacial to interglacial conditions. In Quaternarytime, about the last two million years, glacial andinterglacial conditions have followed each other on

l:t' .ì

GLACTAL CRUSTAL REBOUND, SEA LEVELS AND SHORELTNES 1159

oco

É õ

- Ø600 E

o400 ò

o

200 9c)(Ú

^ ( I )

500 360 290 250 210Time (million years Bp)

-200

Figure 2 Global sea level variations through the last 600 mill¡on years estimated from seismic stratigraphic studies of sedimentsdeposited at continental margins. Redrawn with permission from Hallam A (1984) Pre-Quaternary sea-level changes. AnnualReview of Earth and Planetary Science 12: 205-243.

140

timescales of the order of 104-105 years. Duringinterglacials, climate conditions were similar tothose of today and sea levels were within a fewmeters of their present day position. During themajor glacials, such as 20 000 years ago, large icesheets formed in the norrhern hemisphere and theAntarctic ice sheet expanded, taking enough waterout of the oceans to lower sea levels by berween 100and 150m. Figure 3 illustrates the changes in globalsea level over the last 130000 years, from the lastinterglacial, the last time that condirions weresimilar to those of today, through the Last GlacialMaximum and to the present. At the end of the lastinterglacial, at L20 000 years ago, climate began tofluctuate; increasingly colder conditions were

-60

-80

-1 00

-120

0 20 40 60 80 100 120Time (thousand years Bp)

Figure 3 Global sea level variations (relative to present) sincethe time of the last interglacial 120 000-130 000 years ago whenclimate and environmental conditions were last similar to thoseof the last few thousand years. Redrawn with permission fromChapell J et al. (1996) Reconciliation of Late euaternary sealevel changes derived from coral terraces at Huon peninsulawith deep sea oxygen isotope records. Earth and planetarvScience Letters 1 41b: 227 -236.

reached, ice sheets over North America and Europebecame more or less permanent features of the land-scape, sea levels reached progressively lower values,and large parts of today's coastal shelves were ex-posed. Soon after the culmination of maximum gla-ciation, the ice sheets again disappeared, withina period of about 10 000 years, and climate returnedto interglacial conditions.

The global changes illustrated in Figure 3 are onlyone part of the sea level signal because the actualice-water mass exchange does not give rise to a spa-tially uniform response. Under a growing ice sheet,the Earth is stressed; the load stresses are transmir-ted through the lithosphere to the viscous underly-ing mantle, which begins to flow away from thestressed arca and the crust subsides beneath the iceload. ìØhen the ice melts, the crusr rebounds. Also,the meltwater added to the oceans loads the seafloorand the additional stresses are transmitted to themantle, where they tend to dissipate by driving flowto unstressed regions below the continents. Hencethe seafloor subsides while the interiors of the conti-nents rise, causing a tilting of the continental mar-gins. The combined adjustments to the changing iceand water loads are called the glacio- and hydro-isostatic effects and together they result in a com-plex pattern of spatial sea level change each time icesheets wax and wane.

Observations of Sea Level GhangeSince the Last Glacial MaximumEvidence for the positions of past shorelines occursin many forms. Submerged freshwater peats and treestumps, tidal-dwelling mollusks, and archaeologicalsites would all point ro a rise in relative sea level

-¿uEõ0)

.ÚoØo

E(Ú(¡)Í.

440 410

R Ij o l

G{

fr ,ä i"j, ii \ ? . i b i l

I oÞ r , f to l i {o 9 i / l i ¡I o o - o . ¡ q l d , Ej. ,/ dç j"l;'m-" ìJ 4';q r _ \ ¡ ô -

ì p.o

Eáo

1-I60 GLACIAL CRUSTAL REBOUND, SEA LEVELS AND SHORELINES

(A)r ì L' 10

ì 200Èã

! rsoc)õf rooõE

3U

40

Ec ^ ^o ltu

FCÚ

õ ¿ v0)

. cp 1 0cg)

0

c

c.F6 - t u

oõc -¿t)(¡)

cØ -30

0

t r ¡n

cF

I -zooõc).E -goc)

c

Angermanlandnorthern Sweden

,I+

Ã

Andoyar _ northern Noruvayï + +

+

+

'¡+ - +'''"------ ---"--Þ-------

+ Y

ú -' r +

Firth of ForthScotland

t++++{++++t+*++*

140

120

100

80

60

20

5050

1 0 L20(B)

' -+'++rF'rr

, + -t -

+

Dungeness, Kentsouthern England

-20

_ J U 't ¿(c)

-oo'.1 0

-cu -1 2 1 0

l - - ¡--¡--¡-- l - -rJ 01 ". ' r " ' . I

4 2 0 6 5 4ïime (1000 radiocarbon years BP)

Figure 4 Characteristic sea level curves observed in different localities (note the different scales used). (A) From Angermanland,Gulf of Bothnia, Sweden, and from Hudson Bay, Canada. Both sites l¡e close to centers of former ice sheet, over northern Europeand Nor-th America respectively. (B) From the Atlantic coast of Norway and the west coast of Scotland. These sites are close toformer ice-sheet margins at the time of the last glaciation. (C) From southern England and Virginia, USA. These sites lie outside theareas of former glaciation and at a distance from the margins where the crust is subsiding in response to the melting of the nearbyicesheet- (D) Two sites in northern Australia, one for the Coral Sea coast of Northern Queensland and the second from the Gulf ofCarpentaria some 200km away but on the other side of the Cape York Peninsula. At both localities sea level rose until about 6000years ago before peaking above present level.

(D)

+,t ,

T

Ottawa lslandHudson Bay, Canada

++

+

J

Cape CharlesVirginia, USA

T{tT T ' f T

r r F ï {

i lT '+

r t t.T

T

Lrft

, ï 'f-

+ Karumba

+1ï

GLAGIAL CRUSTAL REBOUND, SEA LEVELS AND SHORELINES 1161

since the time of growth, deposition, or construc-tion. Raised coral reefs, such as in Figure 1, whalebones cemented in beach deposits, wave-cut rockplatforms and notches, or peats formed from saline-loving plants would all be indicative of a falling sealevel since the time of formation. To obtain usefulsea level measurements from these data requiresseveral steps: an understanding of the relationshipbefween the feature's elevation and mean sea levelat the time of growth or deposition, a measurementof the height or depth with respect to preserit sealevel, and a measurement of the age. All aspects ofthe observation present their own peculiar problemsbut over recent years a substantial body of observa-tional evidence has been built up for sea levelchange since the last glaciation that ended at about20 000 years ago. Some of this evidence is illustratedin Figure 4, which also indicates the very significantspatial variability that may occur even when, as isthe case here, the evidence is from sites that arebelieved to be tectonically stable.

In areas of former major glaciation, raised shore-lines occur with ages that are progressively greaterwith increasing elevation and with the oldest shore-lines corresponding to the time when the region fustbecame ice-free. Two examples, from northernSweden and Hudson Bay in Canada, respectively,are illustrated in Figure 4A. In both cases sea levelhas been falling from the time the area becameice-free and open to the sea. This occurred at about9000 years ago in the former case and about 7000years later in the second case. Sea level curves fromlocalities just within the margins of the former icesheet are illustrated in Figure 48, from westernNorway and Scotland, respectively. Here, immedi-ately after the area became ice-free, the sea level fellbut a time was reached when it again rose. A localmaximum was reached at about 6000 years ago,after which the fall continued up until the present.Farther away from the former ice margins the ob-served sea level pattern changes dramatically, as isillustrated in Figure 4C. Here the sea level initiallyrose rapidly but the rate decreased for the last 7000years up to the present. The two examples illus-trated, from southern England and the Atlanticcoast of the United States, are representative oftectonically stable localities that lie within a fewthousand kilometers from the former centers of gla-ciation. Much farther away again from the formerice sheets the sea level signal undergoes a furthersmall but significant change in that the present levelwas first reached at about 6000 years ago and thenexceeded by a small amount before returning to itspresent value. The two examples illustrated in Fig-ure 4D are from nearby localities in northern Aus-

tralia. At both sites present sea level was reachedabout 6000 years ago after a prolonged period ofrapid rise with resulting highstands that are small inamplitude but geographically variable.

The examples illustrated in Figure 4 indicate therich spectrum of variation in sea level that occurredwhen the last large ice sheets melted. Earlier glacialcycles resulted in a similar spatial variability, butmuch of the record before the Last Glacial Max-imum has been overwritten by the effects of the lastdeglaciation.

Glacio-Hydro-lsostatic ModelsIf, during the decay of the ice sheets, the meltwatervolume was distributed uniformly over the oceans,then the sea level change at time ¿ would be

L(,(t) : (change in ice volume/ocean surface area)

x p¡lp* [1a]

where pi and pw are the densities of ice and water,respectively. (See end of article - symbols used.)This term is the ice-volume equivalent sea level andit provides a measure of the change in ice volumethrough time. Because the ocean area changes withtime a more precise definition is

tlbl

where A6 is the area of the ocean surface, excludingareas covered by grounded ice. The sea level curveillustrated in Fþre 3 is essentially this function.However, it represents only a zero-order approxi-mation of the actual sea level change because of thechanging gravitational field and deformation of theEarth.

In the absence of winds or ocean currents, theocean surface is of constant gravitational potential.A planet of a defined mass distribution has a familyof such surfaces outside it, one of which - the geoid- corresponds to mean sea level. If the mass distri-bution on the surface (the ice and water) or in theinterior (the load-forced mass redistribution in themantle) changes, so will the gravity field, the geoid,and the sea level change. The ice sheet, for example,represents a large mass that exerts a gravitationalpull on the ocean and, in the absence of otherfactors, sea level will rise in the vicinþ of the icesheet and fall farther away. At the same time, theEarth deforms under the changing load, with twoconsequences: the land surface with respect to

1162 GLACTAL CRUSTAL REBOUND, SEA LEVELS AND SHORELTNES

which the level of the sea is measured is time-dependent, as is the gravitational attraction of thesolid Earth and hence the geoid.

The calculation of the change of sea level result-ing from the growth or decay of ice sheets thereforeinvolves three steps: the calculation of the amountof water entering into the ocean and the distributionof this meltwater over the globe; the calculation ofthe deformation of the Earth's surface; and the cal-culation of the change in the shape of the gravi-tational equipotential surfaces. In the absence ofvertical tectonic motions of the Earth's surface, therelative sea level change L((E,t) at a site E and timet can be written schematically as

a((E, t ) : L(" ( t ) - t L(1@,t ) tz l

where L(r(E,l) is the combined perturbation fromthe uniform sea level rise term [1]. This is referredto as the isostatic contribution to relative sea levelchange.

In a first approximation, the Earth's response toa global force is that of an elastic outer sphericalIayer (the lithosphere) overlying a viscous orviscoelastic mantle that itself contains a fluid core.\When subjected to an external force (e.g., gravity)or a surface load (e.g., an ice cap), the planet experi-ences an instantaneous elastic deformation followedby a time-dependent or viscous response witha characteristic timescale(s) that is a function ofthe viscosity. Such behavior of the Earth is welldocumented by other geophysical observations: rhegravitatíonal attraction of the Sun and Moon raisestides in the solid Earth; ocean rides load the seafloorwith a time-dependent water load to which theEarth's surface responds by further deformation;atmospheric pressure fluctuations over the conti-nents induce deformations in the solid Earth. Thedisplacements, measured with precision scientificinstruments, have both an elastic and a viscous com-ponent, with the latter becoming increasingly im-portant as the duration of the load or forceincreases. These loads are much smaller than the iceand water loads associated with the major deglaci-ation, the half-dally ocean ride amplitudes beingonly lTo of the glacial sea level change, and theyindicate that the Earth will respond ro even smallchanges in the ice sheets and to small additions ofmeltwater into the oceans.

The theory underpinning the formulation ofplanetary deformation by external forces or surfaceloads is well developed and has been tested againsta range of different geophysical and geological ob-servations. Essentially, the theory is one of formula-ting the response of the planet to a point load and

then integrating this point-load solution over rheIoad through time, calculating at each epoch thesurface deformation and the shape of the equipoten-tial surfaces, making sure that the meltwater is ap-propriately distributed into the oceans and that thetotal ice-water mass is preserved. Physical inputsinto the formulation are the time-space historv ofthe ice sheets, a description of the ocËan basins fromwhich the warer is extracted or into which itis added, and a description of the rheology, orresponse parameters, of the Earth. For the lastrequirement, the elastic properties of the Earth,as well as the density distribution with depth,have been determined from seismological andgeodetic studies. Less well determined are theviscous properties of the mantle, and the usualprocedure is to adopt a simple parametrization ofthe viscosity structure and to estimate the relevantparameters from analyses of the sea level changeitself.

The formulation is conveniently separated intotwo parts for schematic reasons: the glacio-isostaticeffect representing the crustal and geoid displace-ments due to the ice load, and the hydro-isostaticeffect due to the water load. Thus

AUE,I) : A("( t) -r L(¡(ç, t) + A(*(E,t) t3l

where L(,(E,r) is the glacio-isostatic and L(*(E,t)the hydro-isostatic contribution to sea level change.(In reality the two are coupled and this is includedin the formulation.) If L(7@,/) is evaluated every-where, the past water depths and land elevationsH(E,t) measured with respect to coeval sea level aregiven by

H(E,t ) : Ho@) - L((ç , t ) l4 l

where Ho(rp) is the present water depth or landelevation at Iocatíon E.

A frequently encountered concept is eustatic sealevel, which is the globally averaged sea level atany time f. Because of the deformation of the sea-floor during and after the deglaciation, the isostaticterm Á( (E,t) is not zeÍo when averaged over theocean at any time /, so that the eustatic sea levelchange is

^(" , , ( r ) : L(" ( t ) + (A( t {E, t } )o

where the second term on the right-hand side de-notes the spatially averaged isostatic term. Note thatL("(t) relates directly to the ice volume, and nota(""" (¿).

GLACTAL CRUSTAL REBOUND, SEA LEVELS AND SHORELTNES 1163

The Anatomy of the Sea LevelFunction

The relative importance of the two isostatic terms ineqn. [3] determines the spatial variability in the sealevel signal. Consider an ice sheet of radius that ismuch larger than the lithospheric thickness: the lim-iting crustal deflection beneath the center of the loadis lpil p^ where l is the maximum ice thickness andp- is the upper mantle density. This is the localisostatic approximation and it provides a reasonableapproximation of the crustal deflection if the load-ing time is long compared with the relaxation timeof the mantle. Thus for a 3 km thick ice sheet themaximum deflection of the crust can reach 1km,compared with a typical ice volume for a large icesheet that raises sea level globally by 50-100m.Near the centers of the formerly glaciated regions itis the crustal rebound that dominates and sea levelfalls with respect to the land.

This is indeed observed, as illusrrared in Figure4A, and the sea level curve here consists of essential-ly the sum of two terms, the major glacio-isostaticterm and the minor ice-volume equivalent sea levelterm Â("(r) (Figure 5A). Of note is that the reboundcontinues long after all ice has disappeared, andthis is evidence that the mantle response includesa memory of the earlier load. It is the decay time ofthis part of the curve that determines the mantleviscosity. As the ice margin is approached, the localice thickness becomes less and the crustal rebound isreduced and at some stage is equal, but of oppositesign, to A(.(t). Hence sea level is constant for aperiod (Figure 5B) before rising again when therebound becomes the minor term. After globalmelting has ceased, the dominant signal is the latestage of the crustal rebound and levels fall up to thepresent. Thus the oscillating sea level curves ob-served in areas such as Norway and Scotland (Fig-ure 4B) are essentially the sum of two effects ofsimilar magnitude but opposite sign. The early partof the observation contains information on earthrheology as well as on the local ice thickness andthe globally averaged rate of addition of meltwarerinto the oceans. Furthermore, the secondary max-imum is indicative of the timing of the end of globalglaciation and the latter part of the record is indica-tive mainly of the mantle response.

At the sites beyond the ice margins it is themeltwater term A("(r) that is imporranr, but it is notthe sole factor. \Øhen the ice sheet builds up, mandematerial flows away from the stressed mantle regionand, because flow is confined, the crust aroundthe periphery of the ice sheet is uplifted. nØhen theice sheet melts, subsidence of the crust occurs in

a broad zone peripheral to the original ice sheet andat these locations the isostatic effect is one of anapparent subsidence of the crust. This is illustratedin Figure 5C. Thus, when the ocean volumeshave stabilized, sea level continues to rise, furtherindicating that the planet responds viscously to thechanging surface loads. The early part of the obser-vational record (e.g., Figure 4C) is mostly indicativeof the rate at which meltwater is added into theocean, whereas the latter part is more indicative ofmantle viscosity.

In all of the examples considered so far it is theglacial-rebound that dominates the total isostaticadjustment, the hydro-isostatic term being presentbut comparatively small. Consider an addition ofwater that raises sea level by an amount D. Thelocal isostatic response to this load is D p*lp^,where p* is the density of ocean water. This givesan upper limit to the amount of subsidence of thesea floor of about 30m for a 100m sea level rise.This is for the middle of large ocean basins and atthe margins the response is about half as great. Thusthe hydro-isostatic effect is significant and is thedominant perturbing term at margins far from theformer ice sheets. This occurs at the Australianmargin, for example, where the sea level signal isessentially determined by A("(¿) and the water-loadresponse (Figure aD). Up to the end of melting, sealevel is dominated by A("(t) but thereafter it is deter-mined largely by the water-load term A(*(E, r), suchthat small highstands develop at 6000 years. Theamplitudes of these highstands turn out to bestrongly dependent on the geometry of the water-Ioad distribution around the site: for narrow gulfs,for example, their amplitude increases with distancefrom the coast and from the water load at rates thatare particularly sensitive to the mantle viscosity.

!Øhile the examples in Figure 5 explain the gen-eral characteristics of the global spatial variability ofthe sea level signal, they also indicate how observa-tions of such variability are used to estimate thephysical quantities that enter into the schematicmodel (3). Thus, observations near the center of theice sheet partially constrain the mantle viscosity andcentral ice thickness. Obsèrvations from the icesheet margin partially constrain both the viscosityand local ice thickness and establish the time oftermination of global melting. Observations farfrom the ice margins determine the total volumes ofice that melted into the oceans as well as providingfurther constraints on the mantle response. By se-lecting data from different localities and time inter-vals, it is possible to estimate the various parametersthat underpin the sea level eqn. [3] and to use thesemodels to predict sea level and shoreline change for

1164 GLACTAL CRUSTAL REBOUND, SEA LEVELS AND SHORELTNES

(A)

200

Ë; . 1 0 0(g-c

õ_9(Úc)øo

E(ÚoE.

1 5 0

100

150

100

50

0

-50

o( l ) ^

c ) ã ,

Ã\ ='. > õoE,

_20õ( D ^ a U

9 o - R o- ' o ) - -

; ;( Ú õc)cc -1oo

l r ì

õ( I ) ^ v

),Y - r^

^\ ='Ë õ - 2 0O _JU

E.

-1 00

, ^ , -150 L(n, 20

-120

-1 40

0

-1 00

-1 50

20

0

-20

40

-1 00

-120

-140

1 0

0-10

-20

1 0t 5

-60

-80

(c)

KarumbaGulf of

Carpentaria-40 I-

1 0

(D) Time (1000 years Bp)

Figure 5 Schematic representation of the sea level curves in terms of the ice-volume equivalent function Â("(t) (denoted by ESL)and the glacio- and hydro-isostatic contributions L(1@,t), LÇ@,t) (denoted by ice and water, respectively). The panels on the leftindicate the predicted individual components and the panels on the right indicate the total predicted change compared with theobserved values (data points). (A) For the sites at the ice center where L(¡(E, Ð > ^("(Ð > L(*(E, t). (B) For sites near but within theice margin where lÂ(¡(rp, Ðl - l^("(Ðl > l^(,(E, r)1, but the first two terms are of opposite sign. (C) For sites beyond the ice marginwhere lÂ("(f) l>lL( '@,Ðl > l¿(*(q,Ð1. (D) For sites at continental margins far from the former ice sheets wherelA(*(rp,t) l > l^( i( f) | . Adapted with permission from Lambeck K and Johnston P (19SS). The viçcosity of the mantle: evidencefrom analyses of glacial rebound phenomena. In: Jackson ISN (ed.) The Earth's Mantle- Compos¡t¡on, Structure and Evotution,pp. 461-502. Cambridge: Cambridge University Press.

" " 1_¿+U U

1 0

E S L . ; .. . upper hri ln ol Forln

GLACIAL CRUSTAL REBOUND, SEA LEVELS AND SHORELINES 1165

unobserved areas. Analyses of sea-level change fromdifferent regions of the world lead to esrimares forthe lithospheric thickness of between 60-1,00km,average uppef mantle viscosity (from the base of thelithosphere to a depth of 670km) of about(1-5)1020Pas and an average lower mande viscosityof about (1-5)1022Pas. Some evidence exists thatthese parameters vary spatially; lithosphere thick-ness and upper mantle viscosiry being lower beneathoceans than beneath continents.

Sea Level Ghange and ShorelineMigration: Some Examples

Figure 6 illustrates the ice-volume equivalent sealevel function Â("(r) since the time of the lastmaximum glaciation. This curve is based on sealevel indicators from a number of localities, all farfrom the former ice sheets, with corrections appliedfor the isostatic effects. The right-hand axis indi-cates the corresponding change in ice volume (fromthe relation [1b]). Much of this ice came from theice sheets over North America and northern Eurooe

Time (radiocarbon, 1000 years BP)20 15 1 0 5 0

2 5 2 0 1 5 1 0 5 0Time (calibrated, 1000 years BP)

Figure 6 The ice-volume equivalent function Â("(t) and icevolumes since the time of the last glacial maximum inferred fromcorals from Barbados, from sediment facies from nodh-westernAustralia, and from other sources for the last 7000 years. Theactual sea level function lies at the upper limit defined by theseobservations (continuous line). The upper time scale corres-ponds to the radiocarbon timescale and the lower one iscalibrated to calendar years. Adapted with permission fromFleming K et al. (1998) Refining the eustatic sea-level curvesince the LGM using the far- and intermediate{ield sites. Earfhand Planetary Science Letters 163i 327-342; and Yokoyama Yet al. (2O0O) ïiming of the Last Glacial Maximum from observedsea-level minimum. Nature 406i 713-716.

but a not insubstantial part also originated fromAntarctica. Much of the melting of these ice sheetsoccurred within 10000 years, at times the rise in sealevel exceeding 30mm per year, and by 6000 yearsago most of the deglaciation was completed. IØiththis sea level function, individual ice sheet models,the formulation for the isostatic factors, anda knowledge of the topography and bathymetry ofthe world, it becomes possible to reconstruct thepaleo shorelines using the relation [4].

Scandinavia is a well-studied area for glacial re-bound and sea level change since the time the iceretreat began about 18 000 years ago. The observa-tional evidence is quite plentiful and a good recordof ice margin retreat exists. Figure 7 illustratesexamples for two epochs. The first (Figure 7A), at16000 years ago, corresponds to a time afrer onserof deglaciation. A large ice sheet existed over Scan-dinavia with a smaller one over the British Isles.Globally sea level was about 110m lower than nowand large parts of the present shallow seas wereexposed, for example, the North Sea and the Eng-lish Channel but also the coastal shelf farther southsuch as the northern Adriatic Sea. The red andorange contours indicate the sea level change be-tween this period and the present. Beneath the icethese rebound contours are positive, indicating thatif shorelines could form here they would be abovesea level today. Immediately beyond the ice margin,a broad but shallow bulge develops in the topogra-phy, which will subside as the ice sheet retreats. Atthe second epoch selected (Figure 7B) 10500 yearsago, the ice has retreated and reached a temporaryhalt as the climate briefly returned to colder condi-tions; the Younger Dryas time of Europe. Much ofthe Baltic was then ice-free and a freshwater lakedeveloped at some 25-30m above coeval sea level.The flooding of the North Sea had begun in earnesr.By 9000 years ago, most of the ice was gone andshorelines began to approach their present config-uration.

The sea level change around the Australian mar-gin has also been examined in some detail and it hasbeen possible to make detailed reconstructions ofthe shoreline evolution there. Figure 8 illusrratesthe reconstructions for northern Australia, theIndonesian islands, and the Malay-IndoChinapeninsula. At the time of the Last Glacial Maximummuch of the shallow shelves were exposed anddeeper depressions within them, such as in the Gulfof Carpentaria or the Gulf of Thailand, would havebeen isolated from the open sea. Sediments in thesedepressions will sometimes retain signatures of thesepre-marine conditions and such data provide impor-tant constraints on the models of sea level chanqe.

7.7

EY

æ . 1 ?c)Ef

õ38.6 I

-r. \¿: .a."'. ' : '

GLOBAL MARINE POLLUTION 1167

Lagoons. Mangroves. Past Climate From Gorals,Rocky Shores. Salt Marshes and Mud Flats. SaltMarsh Vegetation. Sandy Beaches, Biology of. SeaLevel Change.

Fuñher ReadingLambeck K (1988) Geophysical Geodesy; The Slow

Deformations of the Earth. Oxford: Oxford UniversityPress.

Lambeck K and Johnston P (19991 The viscosity of themantle: evidence from analysis of glacial-reboundphenomena. In: Jackson ISN (ed.) The Earth'sMantle - Composition, Structure and Euolution.pp. 461-502. Cambridge: Cambridge University Press.

Lambeck K, Smither C and Johnston P (1998) Sea-levelchange, glacial rebound and mantle viscosity fornorthern Europe. Geophysical Journal International134: 102-144.

Peltier l7R (1998) Postglacial variations in the level of thesea: implications for climate dynamics and solid-earthgeophysics. Reuietas in Geophysics 36: 603-689.

Pirazzolí P^ (t991) World Atlas of Holocene Seø-LeuelChanges. Amsterdam: Elsevier.

Plassche O van de (ed.) (1986) Sea-Leuel Research: AManual for the Collection and Eualuation of Døtø.Norwich: Geo Books.

Sabadini R, Lambeck K and Boschi E (1991\ GlacialIsostasy, Sea Leuel and Mantle Rheology. Dordrecht:KIuwer.

A. D. Mclntyre, University of Aberdeen,Aberdeen, UK

Copyright O 2001 Academic Press

doi:1 0.1 006/rwos.2001 .0501

Selected topics on pollution of the oceans arecovered in detail by articles listed in the cross-references at the end. This brief overview is anintroduction and guide to rhe appended list ofFurther Reading, which is designed to provide anentry to the historical and global context of theseissues.

Until recently the size and mobility of the oceansencouraged the view that they could not be signifi-cantly affected by human activities. Fresh warerlakes and rivers had been degraded for centuries byeffluents, particularly sewage, but, although fromthe 1920s coastal oil pollution from shipping dis-charges was widespread (Pritchard, L987), ir wasfelt that in general the sea could safely dilute anddisperse anything added to it. Erosion of this viewbegan in the 1950s, when fallout from the testing ofnuclear weapons in the atmosphere resulted in en-hanced levels of artificial radionuclides throughoutthe world's oceans (Park et a1.,1.983). At about thesame time, the effluent from a factory at Minamatain Japan caused illness and deaths from consump-tion of mercury-contaminatecl fish (Kutsuna, 1,968),focusing global attention on rhe potenrial dangersof toxic metals. In the early 1.960s, build-up in themarine environment of residues from syntheticorganic pesticides poisoned top predators such asfish-eating birds (Tolba et al., 1992), and in 1967the first wreck of a supertanker, the Torrey Canyon,highlighted the threat of oil from shipping accidenrs,

as distinct from operational discharges (EighthReport of the Royal Commission on EnvironmentalPollution, 1981,).

It might therefore be said that the decades of the1950s and 1960s saw the beginnings of marinepollution as a serious concern, and one that de-manded widespread control. It attracted the effortsof national and international agencies, not leastthose of the United Nations. The fear of effects ofradioactivity focused early attention, and initiatedthe establishment of the International Commissionon Radiological Protection (ICRP) (Brackley, 1990),which produced a set of radiation protection stan-dards, applicable not just to fallout from weaponstesting but also to the increasingly more relevantissues of operational discharges from nuclear feac-tors and reprocessing plants, from disposal of low-level radioactive material from a variety of sourcesincluding research and medicine, and from accidentsin industrial installations and nuclear-poweredships. Following Minamata, other metals, in par-ticular cadmium andlead,joined mercury on the listof concerns. Since this, like radioactive wastes, wasseen as a public health problem, immediate acrionwas taken. Metals in seafoods were monitored andimport regulations were put in place. As a result,metal toxicity in seafoods is no longer a major issue,and since most marine organisms are resilient tometals, this form of pollution affects ecosystemsonly when metals are in very high concentrations,such as where mine tailings reach the sea.

Synthetic organics, either as pesticides and anti-foulants (notably tributyl tin (TBT); de Mora,1996)or as industrial chemicals, are present in sea water,biota, and sediments, and affect the whole spectrumof marine life, from primary producers to mammals