Basin migration caused by slow lithospheric … migration caused by slow lithospheric extension J.W. van Wijk , S.A.P.L. Cloetingh Faculty of Earth and Life Sciences, Vrije Universiteit,

Basin migration caused by slow lithospheric extension

J.W. van Wijk �, S.A.P.L. CloetinghFaculty of Earth and Life Sciences, Vrije Universiteit, De Boelelaan 1085, 1081 HV Amsterdam, The Netherlands

Received 19 July 2001; received in revised form 15 February 2002; accepted 25 February 2002

Abstract

Sedimentary basin migration caused by low lithospheric extension rates is investigated using a two-dimensionaldynamic numerical model of the lithosphere. We find that continental breakup will eventually occur when largerextension velocities are used. The duration of rifting prior to continental breakup is dependent on the extensionvelocity. Stretching the lithosphere with lower velocities does not lead to breakup. Instead, the locus of maximumextension migrates. Deformation localizes outside the first formed basin that is in turn uplifted. This basin thenbecomes a ‘cold spot’ in the area. In this case, syn-rift cooling predominates; the lithosphere regains its strengthduring stretching instead of becoming weaker, and the lithosphere necking zone becomes stronger than surroundingregions. The transition velocity is, for the cases studied, about 8 mm/yr, while the locus of maximum thinningmigrates after about 50^60 Myr. A comparison with observations of the mid-Norwegian, Galicia and ancient SouthAlpine margins shows a close resemblance of important features. ß 2002 Elsevier Science B.V. All rights reserved.

Keywords: basins; migration; basin inversion; Voring Plateau; continental drift

1. Introduction

Numerous examples have been described in theliterature of extensional regions in which the mainlocus of extension shifted in time (e.g. [1^4]). Inthese areas, typically several sedimentary basinsare found that are closely located, with more orless the same orientation of the basin axes, and atime of formation often several tens of millions ofyears apart. The centers of lithosphere thinning inthese areas seem to have migrated in time.

A well documented example of such an area isthe mid-Norwegian passive continental margin.

Prior to the ¢nal Late CretaceousÊarly Tertiaryrifting event that led to continental breakup, themid-Norwegian VÖring margin was a¡ected byseveral rifting events ([1,3,5]). These resulted inthe formation of several sedimentary basins posi-tioned on the passive margin between the Norwe-gian mainland and the continentôcean boundary(Fig. 1). On a basinwide scale the margin may bedivided in the Permo-Triassic TrÖndelag Plat-form, situated in the east, the Jurassic^CretaceousVÖring Basin, in the center of the margin shelf,and a western extended zone related to the ¢nalstretching event that resulted in continental break-up at the PaleoceneÊocene transition [4,7,8], ad-jacent to the ocean^continent boundary. TheNordland Ridge and VÖring Marginal High areuplifted structures. Although there still exist thor-ough uncertainties on the deeper structures of the

0012-821X / 02 / $ ^ see front matter ß 2002 Elsevier Science B.V. All rights reserved.PII: S 0 0 1 2 - 8 2 1 X ( 0 2 ) 0 0 5 6 0 - 5

* Corresponding author. Fax: +31-20-6462457.E-mail address: [email protected] (J.W. van Wijk).

EPSL 6187 26-4-02 Cyaan Magenta Geel Zwart

Earth and Planetary Science Letters 198 (2002) 275^288

www.elsevier.com/locate/epsl

mid-Norwegian margin, and the a¡ected width ofthe pre-Jurassic extension phase remains unclear[9], the locus of extension is often suggested tohave shifted westward in time, from the TrÖnde-lag platform to the VÖring margin to the marginedge (e.g. [3,4,7]).

Other areas where basin depocenter shift hasbeen observed are the nonvolcanic Galicia marginand the passive rifted ancient South Alpine mar-gin (e.g. [10,11]). On the Galicia margin the loca-tion of rifting shifted from the location of theInterior Basin towards the site of future breakup,on the South Alpine margin from the Lombar-dian Basin towards the future breakup location.

1.1. Hypotheses of rift migration

Several hypotheses exist on causes of rift migra-tion, based on the principle that rifting occurs atlocations where the lithosphere is weakest. Amechanism for limiting extension at one locationhas been studied by, for example, England [12],Houseman and England [13], and Sonder andEngland [14]. They found that cooling of the con-tinental lithosphere during stretching might in-crease its strength, so that the locus of deforma-tion may shift to a region of low strain [14]. With

this mechanism, other in£uences like changes inplate boundary forces are not needed to explainobservations of basin migration.

Migration of the locus of extension is fre-quently suggested to be the consequence of multi-ple stretching phases, with intermediate periods inwhich the lithosphere is not under tensile stress.Hereby, the weakened lithosphere resulting fromthe ¢rst stretching phase needs time to cool su⁄-ciently and regain enough strength before the on-set of the next stretching event. This explanationrequires, therefore, a long period of tectonic qui-escence between successive rifting events. It wasshown by Bertotti et al. [11], in a model of thethermomechanical evolution of the South Alpinerifted margin, that the thinned parts of the margincould indeed be stronger than the rest of the mar-gin. An explanation for this is that, after litho-spheric thinning, the proportion of stronger man-tle material vs. weaker crustal material is larger incomparison to non-thinned lithosphere [11]. Thishypothesis goes together with a sudden time-de-pendent change in the magnitude of the intra-plate stress ¢eld, causing the di¡erent stretchingand non-stretching phases.

Steckler and Ten Brink [15] concluded, fromconsidering lithospheric strength variations inthe northern Red Sea region, that the strengthof the lithosphere controls the locations of riftingand possibly new plate boundaries. For the north-ern Red Sea region this includes a shift in thelocation of maximum deformation during time.Strength variations in this area are in£uenced byseveral factors, like thickness and composition ofthe crust, sediment thickness, and geotherm. Saw-yer and Harry [16] modeled rift migration in theUSA Baltimore Canyon trough by including theasymmetry of the pre-rift geometry. Laterally o¡-set pre-existing weaknesses in crust and uppermantle resulted, when the lithosphere was ex-tended, in a shift of the rift location.

Manatschal and Bernoulli [10] found that rift-ing migrated during Galicia and ancient Adriamargin formation in the direction of future break-up. They proposed that cooling and strengtheningof the lithosphere during rifting forced a shift inthe rift location to previously not or baselythinned areas. The process of syn-rift cooling

Fig. 1. Sketch of three rift zones of the mid-Norwegian mar-gin (after [6]).


J.W. van Wijk, S.A.P.L. Cloetingh / Earth and Planetary Science Letters 198 (2002) 275^288276

has also been proposed to explain the formationof some of the narrow^wide margin pairs of theSouth Atlantic [17] and basin migration on theancient South Alpine passive rifted margin [11].

1.2. Migration by continuous slow lithosphericextension

In this study basin depocenter migration bycontinuous slow lithospheric extension is investi-gated. When the lithosphere is extending fast, theupwelling of warm mantle material in the areaundergoing lithosphere necking is almost adia-batic. When the lithosphere is stretched with low-er rates, syn-rift (lateral) cooling of the neckingarea will play a more important role. Intuitively,it can be reasoned that a set of conditions mayexist where syn-rift cooling prevails. This wouldbe the case when the lithosphere is stretched withprogressively smaller rates. At some point, thelithosphere would regain its strength upon stretch-ing rather than further losing it, and the neckingzone could become stronger than its surroundingregions [12^14]. Hence, during continued stretch-ing, migration of the necking area could be ex-pected [14].

In this study, the lithosphere possessing an ini-tially symmetric upper mantle weakness is ex-tended in a visco-elastic plastic ¢nite element mod-el. When large extension velocities are used,focusing of deformation takes place, causing neck-ing and eventually continental breakup. This willhereafter be referred to as ‘standard’ rifting. Adi¡erent evolution of localization takes placewhen the lithosphere is extended with smaller ve-locities. In these cases, the necking area may startmigrating, and hence, prevent continental break-up. The results of the modeling will be comparedto observations of basin migration at the mid-Nor-wegian, Galicia and ancient South Alpine margins.

2. Modeling approach

To study lithosphere extension, a two-dimen-sional ¢nite element model is used [18,19]. Theprogram is based on a Lagrangian formulation,which makes it possible to track (material) bound-

aries, like the Moho, in time and space. A draw-back of the Lagrangian method is that it is notsuitable for solving very large grid deformationproblems. This is a problem in analysis of exten-sion of the lithosphere, which is often accompa-nied by large deformations. The elements mightbecome too deformed to yield accurate or stablesolutions. In order to overcome this problem the¢nite element grid is periodically remeshed.

2.1. Lithosphere deformation

In the numerical model the base of the litho-sphere is de¢ned by the 1300‡C isotherm. Undersuch conditions, approximately the upper half ofthe thermal lithosphere behaves elastically on geo-logical time scales, while in the lower half stressesare relaxed by viscous deformation. This visco-elastic behavior is well described by a Maxwellbody [20], resulting in the following constitutiveequation for a Maxwell visco-elastic material :

_OO ¼ 12W

c þ 1E

dcdt

ð1Þ

in which _OO is strain rate, W is dynamic viscosity, cis stress and E is Young’s modulus [20]. For aNewtonian £uid the dynamic viscosity W is con-stant. In the lithosphere, however, non-linearcreep processes prevail [21,22], and the relationbetween stress and strain rate can be describedby:

_OO ¼ Acn exp

3QRT

ð2Þ

where A, n and Q are experimentally derived ma-terial constants [21,22], n is the power law expo-nent, Q is activation energy, R is the gas constantand T is temperature.

The state of stress is constrained by the forcebalance:

9 Wc þ bg ¼ 0 ð3Þ

where g is gravity and b is density.In this model it is assumed that mass is con-

served and the material is incompressible. Thecontinuity equation following from the principle


J.W. van Wijk, S.A.P.L. Cloetingh / Earth and Planetary Science Letters 198 (2002) 275^288 277

of mass conservation for an incompressible me-dium is:

9 v!¼ 0 ð4Þ

In the model, the density is dependent on thetemperature following a linear equation of state:

b ¼ b 0ð13KTÞ ð5Þ

where b0 is the density at the surface, K is thethermal expansion coe⁄cient and T is tempera-ture.

Besides visco-elastic behavior, processes of frac-ture and plastic £ow play an important role indeformation of the lithosphere. This deformationmechanism is active when deviatoric stresses reacha critical stress level. Here the Mohr^Coulombcriterion is used as a yield criterion to de¢ne thecritical stress level. The Mohr^Coulomb strengthcriterion is de¢ned as:

Md nM9c3c n tan B ð6Þ

where dn is the shear stress component, cn is thenormal stress component, c is the cohesion of thematerial and B is the angle of internal friction[23]. Stresses are adjusted every time step whenthe criterion is reached. Frictional sliding andfault movement are not explicitly described bythis criterion, analogous to [24].

The displacement ¢eld is obtained by solvingEqs. 1^6. We used 2560 straight-sided seven-node triangular elements, with a 13-point Gaus-sian integration scheme. As the time discretizationschemes used are fully implicit, the system is un-conditionally stable. However, the exactness ofthe solution remains dependent on the time stepsize, which is why the Courant criterion was im-plemented.

Processes like sedimentation and erosion arenot incorporated in the modeling. They do a¡ectthe evolution of a rift basin and rift shoulders, andcan change the strength of the lithosphere [25].

2.2. Thermal evolution

The temperature ¢eld in the lithosphere is cal-

culated every time step using the heat £ow equa-tion:

bcpdTdt¼ D jkD jT þH ð7Þ

where the density b is de¢ned by Eq. 5, cp isspeci¢c heat, k is conductivity and H is crustalheat production. Temperatures are calculated onthe same grid as the velocity ¢eld, and advectionof heat is accounted for by the nodal displace-ments.

2.3. Con¢guration and initial and boundaryconditions

The model domain is divided into an uppercrust, a lower crust and a mantle lithospherepart (Fig. 2). The layers have been assigned di¡er-ent rheological parameters; a granite upper crust,diabase lower crust and olivine mantle were chos-en [22]. The rheological parameters are listed inTable 1. In order to facilitate localization of de-formation, the crust is thickened by 2 km in thecenter of the domain. The imposed Moho surfaceis assumed to be a linear feature parallel to thefuture basin axis. This dip in the Moho causes alocal weakness in the upper mantle and locallyreduces the lithospheric strength. Henk [24] de-scribes rifting of a lithosphere with this con¢gu-ration as postconvergent extension that occursafter thermal equilibration of the thickened crust.In the mid-Norwegian margin the Caledonian Or-

Fig. 2. Initial model con¢guration and horizontal deviatoricstress ¢eld. vext is extension velocity. Upon stretching of thelithosphere, the width of the model domain increases whileits thickness decreases, preserving volume.



ogeny probably resulted in a crust that was thick-ened before the onset of rifting [5].

The model is not pre-stressed; deformation isdriven by velocity boundary conditions. The rightand left sides of the model domain are pulled witha constant velocity v (Fig. 2). This constant veloc-ity boundary condition implies that the strain ratedecreases with time, which is representative forbasin-forming processes. The range of the con-stant extension rates tested is between V3 andV30 mm/yr. This falls within the range ofpresent-day plate velocities as obtained by usingthe Global Positioning System [26]. The surface isunconstrained and on the base of the model avertical velocity component is prescribed calcu-lated from the principle of volume conservation.

Temperatures are calculated using the heat £owEq. 7. The initial geotherm is in steady state. Thetemperature at the surface is 0‡C, and at the baseof the model (125 km depth) it is 1333‡C. Theheat £ow through the right and left sides of thedomain is zero. The crustal heat production isconstant (Table 1).

3. Results

Several tests were performed in which the litho-sphere was extended with di¡erent constant veloc-ity boundary conditions ranging from rather largevelocities of V30 mm/yr to extension velocities ofV3 mm/yr. In these tests the model setup (seeFig. 2) is similar. The results suggest that twoprincipally di¡erent patterns of development ex-ist : when the lithosphere is extended with a totalvelocity less than V8 mm/yr, localization of de-formation evolves distinctly di¡erently from whenthe lithosphere is stretched with higher rates.

To check the possibility that the horizontal sizeof the model domain might control any of theprocesses, some tests were performed with muchlarger horizontal model sizes as well. The horizon-tal size of the model domain proved to be of noin£uence on the results.

3.1. Rifting resulting in continental breakup

3.1.1. Evolution of lithosphere deformationAs the rift structures resulting from the larger

boundary velocity models (‘standard neckingcases’) do not di¡er signi¢cantly, one representa-tive test in which the lithosphere was stretchedwith a total velocity of about 16 mm/yr is dis-cussed here. When the lithosphere is extended,the deformation localizes in the center of the do-main where the initial mantle weakness was intro-duced. Thinning of the crust and mantle litho-sphere concentrates here, and mantle materialstarts to well up (Fig. 3). Thinning of the crustand mantle lithosphere continues, eventually re-sulting in continental breakup, after V27 Myrof stretching. Continental breakup is here de¢nedas occurring when the crust is thinned by a factor20, but another factor or another de¢nition forcontinental breakup could also have been chosen.This de¢nition corresponds to about 40^50% ofextension of the model domain at continentalbreakup. The thinning factors of the crust (L)and mantle (N) parts of the lithosphere are shownin Fig. 4. The base of the mantle lithosphere is the1300‡C isotherm, and L and N are de¢ned to bethe ratio between the initial and the present thick-

Table 1Material parameter values

Parameter Value

Density 2700 (u.c.), 2800 (l.c.), 3300(m.l.) [kg/m3]

Thermal expansion 1U1035 [K31]Crustal heat production 1U1036

[W/m3]Speci¢c heat 1050 [J/kg/K]Conductivity 2.6 (crust), 3.1 (mantle)

[W/m/K]Bulk modulus 3.3U1010 (crust), 12.5U1010

(mantle) [Pa]Shear modulus 2U1010 (crust), 6.3U1010

(mantle) [Pa]Power law exponent n 3.3 (u.c.), 3.05 (l.c.), 3.0 (m.l.)Activation energy Q 186.5 (u.c.), 276 (l.c.), 510 (m.l.)

[kJ/mol]Material constant A 3.16U10326 (u.c.), 3.2U10320

(l.c.), 7.0U10314 (m.l.) [Pa3n/s]Friction angle 30‡Dilatation angle 0‡Cohesion factor 20U106 [Pa](Initial) thickness of modeldomain H

125 [km]



ness of the crust or mantle lithosphere respec-tively.

The total (integrated) strength of the litho-sphere during rifting is shown in Fig. 5. The cen-ter of the model domain where the initial mantleweakness was imposed is the weakest part ; thiscontinues to be so until breakup. The values ofthe integrated strength of continental lithospherevary between 1012 and 1013 N/m, the higher values

characterize Precambrian shields [21,27]. Thestrength of the lithosphere in the model falls with-in this range. The strength decreases with time,caused by thinning and heating of the lithosphereas a consequence of stretching, the non-linearrheology, and decreasing strain rates.

3.1.2. Relation between rift duration and extensionvelocity

The evolution of the localization of deforma-tion in other cases with higher constant boundaryvelocities is comparable to that discussed above.One zone develops where deformation concen-trates, and thinning continues, eventually result-ing in continental breakup in all cases. The dura-tion of rifting until the breakup depends on theextension velocity. When the lithosphere isstretched with larger velocities, it takes less timeto reach continental breakup (Fig. 6). When thetotal extension velocity is less than V8 mm/yr,stretching of the lithosphere does not lead to con-tinental breakup. This is discussed in more detail

Fig. 4. Evolution of thinning factors of crust (L) and mantlelithosphere (N) for case vextW16 mm/yr. Breakup after V27Myr. Width of the model domain (horizontal axis) vs. time(vertical axis). The width of the model domain increases asthe lithosphere is extended.

Fig. 3. Thermal evolution of lithosphere for case vextW16 mm/yr, 2, 10 and 20 Myr after stretching started. Note the changinghorizontal and vertical scales in the panels indicating the changing sizes of the model domain upon stretching.

Fig. 5. Evolution of lithosphere strength (vH0 c(z) dz), in N/m, for case vextW16 mm/yr.



in Section 3.2. The dependence of rift duration onpotential mantle temperature is discussed in VanWijk et al. [28]. The shape of the rift shows noclear dependence on the velocity boundary condi-tion tested here (see also Bassi [29]).

The tendency for the lithosphere to neck (or tofocus strain) is weaker with decreasing extensionvelocity. When the stresses exerted on the litho-sphere are smaller, the rate of localization is alsoslowed down. The mantle upwelling is slower andsyn-rift lateral conductive cooling plays a moreimportant role. When the lithosphere is stretchedat high rates, the upwelling of mantle material isfast (almost adiabatic) with little or no horizontalheat conduction. The fast rise of hot material fur-ther reduces the strength of the lithosphere in thecentral region, with the consequence that defor-mation and thinning are even further accelerated.The result is a short rifting period and rapid con-tinental breakup.

3.2. Rift migration

3.2.1. Thermal evolutionWhen extension is characterized by lower veloc-

ities, the lithosphere reacts di¡erently. The resultsof one representative test (total extension rate ofV6 mm/yr) have been selected to illustrate this.The lithosphere is stretched relatively slowly inthis case for a period of more than 100 Myr.The thermal evolution of the lithosphere is shownin Fig. 7. During the ¢rst 30 Myr after onset ofstretching, deformation localizes in the center ofthe domain where the lithosphere was initially

weakened (Fig. 2). Mantle material wells up anda sedimentary basin is formed. Then, as litho-spheric stretching proceeds, temperatures beginto decrease (see panels 45 Myr and 50 Myr;Fig. 7), in contrast to what happened in the stan-dard necking case shown in Fig. 3. Developmentof the upwelling zone ceases and the lithospherecools in the center of the domain. Cooling of thecentral zone continues, while after 70 Myr an in-crease in temperature is visible on both sides ofthe central zone. These new upwelling zones fur-ther develop (Fig. 7, 110 Myr panel) and two newbasins are formed adjacent to the ¢rst basin. Tem-peratures in the lithosphere below the ¢rst-stagebasin are now lower than temperatures in the sur-rounding lithosphere; a ‘cold spot’ is present inan area that previously underwent extension (seealso Fig. 8). Surface heat £ow values re£ect thisthermal structure; the surface heat £ow values arelower in the ¢rst-stage basin (66 mW/m2) than inthe surrounding areas (75 mW/m2) at 110 Myr.

3.2.2. Lithosphere thinningThe thinning factors of the crust and mantle

lithosphere are shown in Fig. 8. Thinning of thecrust starts in the central weakness zone of thedomain, as expected. One basin is formed, witha maximum thinning factor of V1.85 for thecrust. Crustal thinning in the central basin contin-ues until about 65 Myr after the onset of stretch-ing, at which time the locus of thinning shiftstowards both sides of the ¢rst basin. The loci ofmaximum thinning of the new ‘rifts’ are at a dis-tance of about 500 km from the center of the ¢rstrift. The ¢rst necking zone is not further thinned,although stretching of the lithosphere continues.Instead, two new basins have developed, belowwhich further thinning of the lithosphere is ac-commodated. The thinning factor of the mantlelithosphere re£ects this behavior. During the V45Myr after stretching started the upwelling mantlematerial causes N to be larger in the central zoneof the domain than in its surroundings. After thistime, however, temperatures decrease rapidly inthis area (see also Fig. 7). New upwelling zonesdevelop on both sides and N decreases in the re-gion of the ¢rst basin and increases where the newbasins develop.

Fig. 6. Rift duration to continental breakup as a function oftotal extension velocity (stars). Rifting eventually results inbreakup at larger extension rates, while at lower rates syn-tectonic cooling prevails and the rift migrates before breakupis reached (see text).



Fig. 9. (A) Evolution of relative surface topography for mi-grating rift case, vextW6 mm/yr. Gray lines indicate the posi-tions of the corresponding synthetic subsidence curves de-rived from this panel, shown in B. (B) Synthetic subsidencecurves for three locations indicated in A: in the ¢rst-stagebasin (right panel), outside this basin but in the new basin(left panel) and in the transition zone (middle panel).

6

Fig. 8. Evolution of thinning factors for crust (L) and mantlelithosphere (N) for case vextW6 mm/yr. No continental break-up. The asymmetry (upper panel) is caused by the asymme-try of the ¢nite element grid that was used. We used triangu-lar-shaped elements that resulted in a not perfectlysymmetric mesh and initial Moho topography. The wigglesin the lines (lower panel, upper right) are due to interpola-tion inexactnesses while calculating the mantle thinning.

Fig. 7. Thermal evolution of lithosphere for migrating riftcase, vextW6 mm/yr, times in Myr after stretching started.

Fig. 10. Lithosphere strength evolution for migrating riftcase, vextW6 mm/yr.



3.2.3. Tectonic subsidenceThe relative topography of the surface and tec-

tonic subsidence curves for three di¡erent loca-tions are shown in Fig. 9. The central basinreaches a maximum depth of about 760 m. Thistectonic subsidence would be further ampli¢ed bysedimentation. The basin is at its maximum depthV35 Myr after stretching started. Then the basinsubsidence commences and at both sides of thebasin surface subsides. Ziegler [30] de¢ned basininversion as the reversal of the subsidence pat-terns of a sedimentary basin in response to com-pressional or transpressional stresses. Followingthis de¢nition, no basin inversion takes placehere as the lithosphere remains in a tensional re-gime. The synthetic tectonic subsidence curve forthis basin (right panel) shows a clear reversal ofthe subsidence pattern. Here, basin uplift thustakes place in a stretching regime. The relativeuplift is considerable; with values, depending onthe location in the basin, of approximately 800^1700 m. The left panel in Fig. 9B shows the sub-sidence curve at the location of the migrated rift.After about 50 Myr subsidence starts. The centralpanel shows the subsidence curve for the ‘transi-tion zone’ between the original basin and the ba-sin formed after migration. This area displayscontinuous uplift.

3.2.4. Lithosphere strengthThe total strength of the lithosphere, obtained

by integrating the stress ¢eld over the thickness ofthe lithosphere [21], is shown in Fig. 10. The cen-tral part of the model domain is weaker than else-where until about 55 Myr after stretching started.Its minimum strength value is already reached by30 Myr. Thereafter the strength increases, but re-mains less than for the rest of the domain. From55 Myr, the smallest values of lithosphericstrength are found on both sides of the centralbasin. By comparing the strength with the thermalstructure of the lithosphere (Fig. 7), the strongdependence of the strength on the temperature iseasily demonstrated.

3.2.5. Lithosphere deformationThe horizontal and vertical components of the

velocity ¢eld are shown in Fig. 11, for several time

intervals. These were corrected for the kinematicvelocity ¢eld. In the ¢rst panels, from 25 Myrafter stretching started, one strong central upwell-ing zone is present (see the vertical z componentof the velocity ¢eld) with two weaker downwellingzones on either side. Material for the upwellingzone is drawn into the zone from deep in thelithosphere; in the horizontal or x component ve-locity panel it is visible that mantle materialmoves towards the left on the right side of thecenter and towards the right on the left side ofthe center, so the movement of the material istowards the center, deep in the lithosphere. Thethickness of this horizontal convergence zone isabout 20 km.

By 50 Myr after stretching started, the upwell-ing zone becomes split into two zones. While thelithosphere is still extending with the same veloc-ity, the maximum velocities in the horizontal andvertical direction become reduced by almost oneorder of magnitude. Mantle material still wells up,but the upwelling is considerably weaker. After 60Myr two separate, new upwelling zones are active,and the upwelling region has been reversed toform a downwelling zone. In the horizontal veloc-ity component panel a similar split is visible ; twoconverging zones have developed. By 110 Myr thetwo upwelling zones are fully developed, and thehorizontal and vertical velocities are once againincreasing. The x component of the velocity ¢eldshows that the ¢rst-stage basin that was formed iseventually uplifted with respect to its surround-ings; as part of the two upwelling systems, crustaland mantle material converges in the area of for-mer basin formation. The response of the litho-sphere to other low extension velocities is similarto the case described above. The single upwellingsystem eventually splits into two new upwellingzones. The ¢rst-stage basin is uplifted and twoseparate, new rifts are formed.

3.2.6. Sequence of eventsComparing Figs. 7, 10 and 11, we observe that

the strength in the center of the domain alreadystarts to increase (at 35 Myr), while the upwellingzone does not split into two zones until V50 Myrafter stretching began. Temperatures in the up-welling area start to drop from 35 Myr, which



initiates the re-strengthening of the thinned litho-sphere. Considerable syn-rift conductive coolingof the warm upwelling mantle material seems tocause this reduction in temperature. The upwell-

ing is so slow that syn-rift cooling can play amajor role. The strength of the lithosphere, how-ever, attains its minimum value in the center partof the domain until the upwelling zone has begun

Fig. 11. Corrected velocity ¢eld (see text) in the lithosphere at 25, 50, 60 and 110 Myr after stretching began, for migrating riftcase, vextW6 mm/yr. (A) Horizontal component of the velocity ¢eld. Velocities are positive to the right, negative to the left.(B) Vertical component of the velocity ¢eld. Velocities are positive upwards, negative downwards. Velocities in m/s. 1e-10 means1U10310. Note that a larger part of the model domain is shown here than in the other ¢gures for a better illustration.



to split up, and crustal thinning in this regioncontinues until V60 Myr. Renewed crustal thin-ning on both sides of the ¢rst basin is delayed forV20 Myr. It starts when the two new upwellingsystems have fully developed.

3.3. Multiple rifts

In the model, rifting in the original basin istaken over by rifting in two new zones. This isthe case in an almost perfectly symmetric litho-sphere con¢guration. In the Earth however, suchperfectly symmetric situations are very unlikely(e.g. because of pre-existing structural inhomoge-neities, deep crustal structures, etc.). Therefore,renewed rifting is likely to take place at one pre-ferred side of the original basin.

When the rate at which the lithosphere is ex-tended is larger than a certain ‘critical’ value, ex-tension may eventually end in continental break-up. Smaller extension rates will not result inbreakup, instead, when extension continues, thelocus of maximum extension shifts, the old basinis abandoned and rifting concentrates in otherareas. This may lead to a region of lithosphereextension with several rifts. The basin that isformed ¢rst is (partly) uplifted while the next ba-sin is formed. This cycle can be repeated one orseveral times, resulting in, for example, a thirdbasin. Between the cessation of rifting in the¢rst-stage basin and the onset of rifting in thesecond-stage basin a period of some tens of mil-lions of years is predicted with the setup andrheological parameters used. The rifts may almostborder on one another; a small zone without sig-ni¢cant subsidence or uplift and an inverted oruplifted region separate the basins. The criticalextension rate which separates the two localiza-tion modes (resulting in continental breakup andresulting in rift migration) is about 8 mm/yr. Thedependence of these parameters on initial andrheological conditions, however, needs further in-vestigation.

While in the modeling the lithosphere under-goes constant extension, separate periods of local-ized deformation or necking can be distinguished.From the ¢nal situation alone it is di⁄cult todetermine whether the area of lithosphere exten-

sion with multiple rift zones is the result of onestretching episode or multiple stretching phases.In order to generate rift migration, suddenchanges in the intra-plate stress ¢eld are not re-quired, provided that the rate at which the litho-sphere is extended is low.

The results of this study are qualitatively inagreement with previous studies on limited exten-sion (e.g. [12^14]). Sonder and England [14] foundthat just after the onset of extension the strengthof the lithosphere decreases slightly, while aftersome time cooling increases the strength. Breakupis expected at high strain rates, the locus of de-formation shifts at lower strain rates. The transi-tion is predicted between strain rates of 10314 and10316 s31. England [12] found that the duration ofrifting until breakup is probably less than 10^20Myr, our modeling study predicts that longer riftdurations are possible. Houseman and England[13] predicted that the maximum extension factorbefore extension ceases is probably 6 1.5; we ¢ndslightly larger values before migration takes place(LW1.85). Shifting of the location of extension isalso found in modeling studies by Bassi [29,31],Bassi et al. [32] and Hopper and Buck [33], onsmaller scales and time spans.

4. Comparison with observations of migrating rifts

In the mid-Norwegian margin, three rift zonesare present (see Fig. 1); these formed during aPermo-Triassic episode (290^235 Ma), a Late Ju-rassicÊarly Cretaceous episode (170^95 Ma) anda Late CretaceousÊarly Tertiary episode (75^57Ma) [3,5,7], respectively, all under E^W to SE^NW directed extension [9]. The time gap betweenthe successive rifting events was V20^60 Myr.Maximum crustal thinning factors range from1.6 for the Permo-Triassic TrÖndelag Platform,to about 2.6 for the Late JurassicÊarly Creta-ceous VÖring Basin [6]. In comparison with themodeling results, the time gap between the succes-sive rifting events is of the same order, and thecrustal thinning factors also correspond favorably.

The model experiments predict an uplift of(part of) the ¢rst-stage basin, and re-thickeningof the thinned crust. On the mid-Norwegian mar-



gin both the Nordland Ridge and the VÖringMarginal High are uplifted structures, as well asthe western part of the VÖring Basin (Fenris Gra-ben, Hel Graben and Fulla Ridge) and the west-ern part of the TrÖndelag Platform [5]. The Nord-land Ridge and western part of the TrÖndelagPlatform experienced uplift probably in the EarlyCretaceous [5,8]. This uplift is suggested to be rift-£ank uplift from the VÖring Basin formation [5].The VÖring Marginal High and the western partof the VÖring Basin experienced regional upliftprobably in the Early Tertiary [5]. This uplift isfrequently suggested to be related to the thermal(or magmatic underplating) in£uence of the Ice-land mantle plume during continental breakup(e.g. [34,35]). The observed inversions and upliftsare considerable; maximum estimations rangefrom 1500 (Nordland Ridge) to 2400 m (VÖringMarginal High). The modeling predicts uplift,starting at the end of the formation of the ¢rst-stage basin and continuing during the ¢rst part ofthe formation of the second-stage basin. These pre-breakup uplifted structures of the mid-Norwegianmargin may have experienced basin shift-relateduplift. Other pre-breakup contractional structuresat the Norwegian continental shelf have been sug-gested to be related to the Alpine Orogeny [36].

On both the Galicia and South Alpine margins,rifting migrated after about 20^30 Myr of exten-sion to the location of eventual sea£oor spread-ing. The Interior and Lombardian basins attainedcrustal thinning factors of about 1.4^1.5[10,11,37], before the basins migrated. These ex-tension factors and rift durations are smaller thanthe modeling predictions. On the Galicia margin,extension of the new (second-stage) basin wascontemporaneous with uplift continentward (theGalicia Bank) [10], and Manatschal and Bernoulli[10] suggested that this might be a breakup-re-lated isostatic edge e¡ect. The similarity with themodel-predicted uplift suggests that it is possiblybasin migration related. Below the Lombardianbasin, cooling started soon after the onset of rift-ing [38]. This is in agreement with the modelingresults. The modeling results con¢rm the explan-ations of basin depocenter shifts on these marginsby [10,11], who found that syn-rift cooling of thelithosphere led to strengthening, which eventually

forced the locus of thinning to migrate to weakerand less extended areas.

The importance of understanding the interplaybetween extensional basins and inversion tectonicsand uplifted structures has been addressed byCloetingh et al. [39]. This study shows that thesefeatures, which are often related to compressionalregimes, may also occur in extensional regimes.

5. Conclusions

This modeling study predicts that basin migra-tion can be the result of constant stretching of thelithosphere with low rates. As the lithosphere isextended at progressively high rates, continentalbreakup eventually occurs. The duration of riftingprior to continental breakup is dependent uponthe extension velocity; the larger the extensionvelocity, the sooner continental breakup will oc-cur. When the lithosphere is stretched with a ve-locity lower than a ‘critical’ velocity, continuingstretching will eventually no longer result in con-tinental breakup. Instead, syn-rift cooling startsto play a major role causing the originally devel-oped necking zone to cease and new neckingzones to develop. Deformation then localizes out-side the ¢rst formed basin that is, in addition,uplifted. This basin becomes a ‘cold spot’ in thearea. A qualitative comparison with observationsof the mid-Norwegian, Galicia and ancient SouthAlpine passive continental margins shows a closeresemblance of important features. Particularlyinteresting are the observed uplifted structures;they are predicted by the modeling as well. An-other outcome of this modeling study is that onelithospheric extension episode appears to be su⁄-cient to explain separate rift locations and epi-sodes present in a single extensional area. Slowlithospheric extension appears to be a possiblecause for basin migration.

Acknowledgements

We would like to thank W.R. Buck and R.H.Gabrielsen for very useful reviews and sugges-tions, and R. Stephenson for improving the Eng-



lish and the style of the manuscript. G. Bertotti,R. van der Meer and M. ter Voorde are thankedfor discussions and suggestions, and R. Huismansand Y. Podladchikov for their large contributionto the numerical code. This is Publication20010704 of the Netherlands Research School ofSedimentary Geology.[AC]

References

[1] P. Ziegler (Ed.), Evolution of the Arctic-North Atlanticand the western Tethys, Am. Assoc. Petrol. Geol. Mem.43, 1988.

[2] P. Ziegler, Geodynamic processes governing developmentof rifted basins, in: F. Roure, N. Ellouz, V.S. Shein, I.I.Skvortsov (Eds.), Geodynamic Evolution of SedimentaryBasins, International Symposium Moscow, 1994, pp. 19^67.

[3] C. Bukovics, P.A. Ziegler, Tectonic development of theMid-Norway continental margin, Mar. Petrol. Geol. 2(1985) 2^22.

[4] E.R. Lundin, A.G. Dore¤, A tectonic model for the Nor-wegian passive margin with implications for the NE At-lantic: Early Cretaceous to break-up, J. Geol. Soc. Lon-don 154 (1997) 545^550.

[5] J. Skogseid, T. Pedersen, V.B. Larsen, VÖring Basin: sub-sidence and tectonic evolution, in: R.M. Larsen, H.Brekke, B.T. Larsen, E. Talleraas (Eds.), Structural andTectonic Modelling and Its Application to PetroleumGeology, NPF Spec. Publ. 1, Elsevier, Amsterdam,1992, pp. 55^82.

[6] J. Skogseid, O. Eldholm, Rifted continental margin o¡mid-Norway, in: E. Banda et al. (Eds.), Rifted Ocean-Continent Boundaries, Kluwer Academic, Dordrecht,1995, pp. 147^153.

[7] P. Reemst, S. Cloetingh, Polyphase rift evolution of theVÖring margin (mid-Norway): constraints from forwardtectonostratigraphic modeling, Tectonics 19 (2000) 225^240.

[8] J. Skogseid, S. Planke, J.I. Faleide, T. Pedersen, O. Eld-holm, F. Neverdal, NE Atlantic continental rifting andvolcanic margin formation, Geol. Soc. Spec. Publ. 167(2000) 295^326.

[9] R.H. Gabrielsen, T. Odinsen, I. Grunnaleite, Structuringof the Northern Viking Graben and the MÖre Basin; thein£uence of basement structural grain, and the particularrole of the MÖre-TrÖndelag Fault Complex, Mar. Petrol.Geol. 16 (1999) 443^465.

[10] G. Manatschal, D. Bernoulli, Architecture and tectonicevolution of nonvolcanic margins: present-day Galiciaand ancient Adria, Tectonics 18 (1999) 1099^1119.

[11] G. Bertotti, M. ter Voorde, S. Cloetingh, V. Picotti, Ther-momechanical evolution of the South Alpine rifted mar-gin (North Italy): constraints on the strength of passive

continental margins, Earth Planet. Sci. Lett. 146 (1997)181^193.

[12] P. England, Constraints on extension of continental litho-sphere, J. Geophys. Res. 88 (1983) 1145^1152.

[13] G. Houseman, P. England, A dynamical model for litho-sphere extension and sedimentary basin formation,J. Geophys. Res. 91 (1986) 719^729.

[14] L.J. Sonder, P.C. England, E¡ects of a temperature-de-pendent rheology on large-scale continental extension,J. Geophys. Res. 94 (1989) 7603^7619.

[15] M.S. Steckler, U.S. TenBrink, Lithospheric strength var-iations as a control on new plate boundaries: examplesfrom the northern Red Sea region, Earth Planet. Sci. Lett.79 (1986) 120^132.

[16] D.S. Sawyer, D.L. Harry, Dynamic modeling of divergentmargin formation: application to the US Atlantic margin,in: A.W. Meyer, T.A. Davies, S.W. Wise (Eds.), Evolu-tion of Mesozoic and Cenozoic Continental Margins,Mar. Geol. 102 (1991) 29^42.

[17] I. Davison, Wide and narrow margins of the BrazilianSouth Atlantic, J. Geol. Soc. London 154 (1997) 471^476.

[18] R.S. Huismans, Y.Y. Podladchikov, S.A.P.L. Cloetingh,Transition from passive to active rifting: relative impor-tance of asthenospheric doming and passive extension ofthe lithosphere, J. Geophys. Res. 106 (2001) 11271^11291.

[19] A. Poliakov, Y.Y. Podladchikov, Diapirism and topogra-phy, Geophys. J. Int. 109 (1992) 553^564.

[20] D.L. Turcotte, G. Schubert, Geodynamics: Applicationsof Continuum Physics to Geological Problems, Wiley,New York, 1982, 450 pp.

[21] G. Ranalli, Rheology of the Earth, 2nd edn., Chapmanand Hall, London, 1995, 413 pp.

[22] N.L. Carter, M.C. Tsenn, Flow properties of continentallithosphere, Tectonophysics 136 (1987) 27^63.

[23] P.A. Vermeer, R. de Borst, Non-associated plasticity forsoils, concrete and rock, Heron 29 (1984) 3^64.

[24] A. Henk, Did the Variscides collapse or were they tornapart? a quantitative evaluation of the driving forces forpostconvergent extension in central Europe, Tectonics 18(1999) 774^792.

[25] E. Burov, S.A.P.L. Cloetingh, Erosion and rift dynamics:new thermomechanical aspects of post-rift evolution ofextensional basins, Earth Planet. Sci. Lett. 150 (1997) 7^26.

[26] D.F. Argus, M.B. He£in, Plate motion and crustal defor-mation estimated with geodetic data from the Global Po-sitioning System, Geophys. Res. Lett. 22 (1995) 1973^1976.

[27] P. England, Comment on ‘Brittle failure in the uppermantle during extension of the continental lithosphere’by D.W. Sawyer, J. Geophys. Res. 91 (1986) 10487^10490.

[28] J.W. Van Wijk, R.S. Huismans, M. ter Voorde, S.A.P.L.Cloetingh, Melt generation at volcanic continental mar-



gins: no need for a mantle plume?, Geophys. Res. Lett.28 (2001) 3995^3998.

[29] G. Bassi, Factors controlling the style of continental rift-ing: insights from numerical modelling, Earth Planet. Sci.Lett. 105 (1991) 430^452.

[30] P.A. Ziegler, Compressional intra-plate deformations inthe Alpine foreland ^ an introduction, Tectonophysics137 (1987) 1^5.

[31] G. Bassi, Relative importance of strain rate and rheologyfor the mode of continental extension, Geophys. J. Int.122 (1995) 195^210.

[32] G. Bassi, C.E. Keen, P. Potter, Contrasting styles of rift-ing: models and examples from the Eastern Canadianmargin, Tectonics 12 (1993) 639^655.

[33] J.R. Hopper, W.R. Buck, The e¡ect of lower crustal £owon continental extension and passive margin formation,J. Geophys. Res. 101 (1996) 20175^20194.

[34] O. Eldholm, J. Thiede, E. Taylor, Evolution of the VÖringvolcanic margin, in: O. Eldholm, J. Thiede, E. Taylor etal. (Eds.), Proc. ODP Sci. Results 104 (1989) 1033^1065.

[35] J. Skogseid, T. Pedersen, O. Eldholm, B.T. Larsen, Tec-

tonism and magmatism during NE Atlantic continentalbreak-up: the VÖring margin, in: B.C. Storey, T. Alabas-ter, R.J. Plankhurst (Eds.), Magmatism and the Causes ofContinental Break-up. Geol. Soc. London Spec. Publ. 68(1992) 305^320.

[36] E. Vafignes, R.H. Gabrielsen, P. Haremo, Late Cretaceous-Cenozoic intraplate contractional deformation at the Nor-wegian continental shelf: timing, magnitude and regionalimplications, Tectonophysics 300 (1998) 29^46.

[37] G. Bertotti, V. Picotti, D. Bernoulli, A. Castellarin, Fromrifting to drifting: tectonic evolution of the south-Alpineupper crust from the Triassic to the Early Cretaceous,Sediment. Geol. 86 (1993) 53^76.

[38] C.A.E. Sanders, G. Bertotti, S. Tommasini, G.R. Davies,J.R. Wijbrans, Triassic pegmatites in the Mesozoic middlecrust of the southern Alps (Italy): Fluid inclusions, radio-metric dating and tectonic implications, Ecl. Geol. Helv.89 (1996) 505^525.

[39] S.A.P.L. Cloetingh, Z. Ben-Avraham, W. Sassi, F. Hor-vath, Dynamics of basin formation and strike-slip tecton-ics, Tectonophysics 266 (1996) 1^10.



Basin migration caused by slow lithospheric … migration caused by slow lithospheric extension J.W. van Wijk , S.A.P.L. Cloetingh Faculty of Earth and Life Sciences, Vrije Universiteit,

Documents