TECTONICS, VOL. 17, NO. 6,PAGES 938-954, DECEMBER 1998 A new multilayered model for intraplate stress-induced differential subsidence of faulted lithosphere, applied to rifted basins R. T. v a n Balen, Y. Y. ~ o d l a d c h i k o v , 'a n d S . A. P. L. C l o e t i n g h Tectonics/StructuralGeology Group, Faculteit der Aardwetenschappen,Vrije Universiteit, Amsterdam Abstract. In-plane horizontal stresses acting on predefomed lithosphere induce differential flexural vertical motions. A high-precision rccord of these motions can be found in the sedimentary record of rifted basins. Originally, it was proposed that rifted basins experience flank uplift and basin center subsidence in response to a compressive change of inplane stress, which agrees well with observed differential motions. Subsequently published models predicted that the vertical motions may be opposite because of the flexural state of the lithosphere induced by necking during extension. However, the total, flexural and permanent, geometry of the lithospherc underlying the rifted basin is the controlling parameter for the in-plane stress-caused vertical motions. The largest part of this preexisting geometry is caused by faulting in the uppermost brittle part of the crust and ductile deformation in the underlying pans of the lithosphere. Wc present a new multilayered model for stress-induced differential subsidence, taking into account the tectonically induced preexisting geometry of the lithosphere, including faults in the upper crust. As continental lithosphere may exhibit flexural decoupling due to a weak lower crustal layer, the new multilayer in-plane stress model discriminates the geometrics of the separate competent layers. At a basin-wide scale, the new model predicts that a compressive change of in-plane force results in basin center subsidence and flank uplift, confirming the original hypothesis. Compared to all previous models, the new model requires a lower horizontal stress level change to explain observed differential vertical motions. 1. Introduction Thc global stress field, inferred from, for example, earthquake focal mechanism solutions, insitu stress measurements, and borehole breakout analyses, has persistent trends, which are uniform over large areas and can be correlated to plate motions [Cloetingh and Wortel, 1985; Zoback et al., 1989; Zoback, 1992; Miiller et al., 1992; Richardson, 1992; Coblenn and Richardson, 1995; Lindholm et al., 19951. The intraplate stress field originates largely from 'Now at Department of Earth Sciences, Eidgenossische Technischc Hochschule, Ziirich, Switzerland. Copyright 1998 by the American Geophysical Union Paper number 1998TC900003. 0278-7407/98/1998TC900003$12.00 the integrated result of lithospheric density heterogeneities and processes operating at plate boundaries, that is, slab pull and ridge push lWortel et al., 1991; Richardson. 1992; Coblentz and Richardson, 19951. Paleostress analyses (from stylolites, fault striations, etc.) demonstrate that intraplate stress orientations and magnitudes can change on a timescale of a few million years [Philip, 1987; Letouzey, 1986; Bergerat, 1987; De Ruig et aL, 1991; Csontos et al., 19911. Futthermore, folding of oceanic [Cloetingh and Wortel, 1985; Stein et al., 1989; Beekman et al., 19961 and continental lithosphere [Stephenson and Cloetingh, 1991; Martinod and Davy, 1992; W o r t and Agarwal, 19961 points to very large magnitudes of intraplate stresses of up to several hundred megapascals [Stephenson and Cloetingh, 1991: Cloetingh and Burov, 19961. In rifted sedimentary basins, a changing horizontal intraplate stress level causes flexural subsidence and uplift, as long as the stress level is lower than the limit defined by strength of the lithosphere and its weakness zones (see section 6). The induced relative sea level changes are recorded as onlap and offlap events in the basin stratigraphy LCloetingh et al., 19851. Therefore flexural uplift arid subsidence caused by intraplate stress level changes provide a tectonic mechanism for sea level changes which interferes with truly eustatic sea level changes. As a consequence, flexural deformations due to continuously changing intraplate stress magnitudes and directions can provide a plausible explanation for relative sea level changes which apparently occurred during the Mesozoic [Cloetingh et al., 19891, as inferred from (seismo) stratigraphic data [see Haq et al., 19871. Thus far, no other known mechanism (e.g., glaciwustasy) can sdlisPd~lorily explain the inferred bigb-frequency Mesozoic sea level changes [Pitman and Golovchenko, 19831, although there is some evidence for the existence of polar ice [Ziegler, 19901. Numerical forward modeling of thc Notth Sea basin [Kooi and Cloetingh, 19891, the Barents Sea basin [Reemst et al., 19941, the Beaufort-Mackenzie basin [Tang and Lerche, 19921 and the Pannonian basin [Van Balen er al., 19981 has demonstrated that a changing level of intraplate stress can explain the observed differential subsidence or relative sea levt.1 changes. These rifted basins all experienced enhanced subsidence in the basin's center in response to compression (or a relative sea level rise) during their postrifl phase, confnuing the predictions oTCluetingh et al. [19851. In this paper a new model for stress-induced subsidence and uplift in rifted basins during their postrift evolution is yrcsented, based on recent insights inlo the rheology and flexural behavior of the lithosphere [e.g., Burov and Diament, VAN BALEN ET AL.: A NEW MODEI. FOR STRESS-INDUCED SUBSIDENCE 19951 and the influence of existing faults on its flexural state [Kusznir et al., 1991; Spadini a d Podlodchikov, 19961. In contrast to previous work on this subject, our elastic thinplate model takes into account the permanent deformation of the lithosphere caused by extension, that is, normal faulting and ductile deformation. As the presence of weak layers in the continental lower crust and a more detailed shape of the deformed lithosphere, including faults, are taken into account, the proposed new thin-plate model also differs from the finite element model utilized by Cloetingh et al. [1985]. Our model is based on the elastic thin-plate approach and does not incorporate viscoelasticity because, in general, flexural subsidence profiles for rifted lithosphere predicted by elastic thin-plate models are in good agreement with observations [e.g., Wans et al., 19821. Furthermore, the values for the extra parameters in viscoelastic thin-plate models operating on a million year timescale are not well constrained [Watts et al., 19821. However, if required, the approach outlined in this paper can be applied to extend existing viscoelastic thin-plate models wlricl~include the effecl of in-plane horizontal s k s s e s [e.g., Lambeck, 19831. In the first part of the paper, we outline the major dirferences between existing models and our new model. Subsequently, we review lithospheric flexure, with special emphasis on decoupled flexure of continental lithosphere. Next, our new thin-plate model for in-plane stress-induced differential motions is outlined. Particular attention is paid to the specific response of faulted lithosphere and its effect on the sedimentary record. This is followed by a comparison of the elastic thin-plate soliltion and the results of finite element modeling. In-plane stress-induced differential subsidence patterns are presented for several fault and basin configuratinns. Finally, we discuss the relationship between observed differential subsidence of the Pannonian and Noah Sea basins and the Norwegian margin and the major fault systems occurring in their basement. uplift uplift A A I v subsidence rn preexisting shape differential subsidence Figure 1. The effect of an in-plane force on a predeformed elastic plate. The predeformed geometry can be flexural, permanent, or a combination of both. Arrows indicate vettical movements caused by the compressive in-plane force. prdicled opposile differential ve~ticalmotions for the same stress change, w e n in the abseuce of a sediment load. This completely different result was attributed to the lateral variation in plate thickness by Cloetingh et al. [1985]. However, as will be shown below, the geometry of the elastic plate is the cause for this response. According to the thin-plate theory, a horizontal in-plane force applied to a predeformed plate gives rise to moments, causing additional flexural hending of the plate (Figure 1) [Timoshenko and Woinowrky-Krieger, 19591. The momen: equals the product of the in-plane force and the slope of the midplane of the elastic plate. The slope of the midplane is caused by preexisting deformation and by bending due to the in-plane force itself. The preexisting deformation can be permanent, flexural, or a combination of both. However, what is the preexisting deflection of the midplane of a thin plate representing the lithosphere of a rifted basin? This difficulty is illustrated by the differences in published elastic thin-plate models. Karner [I9861 applies the in-plane force 2. Previous Research change to the preexisting deflections of the sedimentThe effect of horizontal in-plane stress variations on the basement interface and the Mohorovicic discontinuity and subsidence of rifted sedimentary basins was first investigated adds the two resulting deflections to obtain a net deflection of by Cloetingh et al. [1985], who applied a two-dimensional the lithosphere. His justification for this procedure is that elastic finite element model. The behavior of two different these two surfaces represent density interfaces, which can model sets was studied. One set of models employed a potentially be affected by lithospheric in-plane stress. uniform elastic layer, whose thickness was increased with age Subsequently published thin-plate models [Cloetingh, 1988; according to a squarc root function. The sewud set of n~odels Kooi and Cloetingh, 1992; Kamer et aL, 19931 basically all asume that the preexisting deformation of the midplane is adopted a laterally variable elastic thickness, representing the basinward decrease of the effective elastic thickness inferred caused by the flexural bending of the lithosphere due to thermal, sediment and necking loads (Figures 2a and 2h). for passive margins [e.g., Caldwell nnd Turcotte, 1979; Watfs el aL, 19821. The results for the uniform elastic thickness Pemanel~ldeformation of the lithosphere due to nor~nal model showed that an increase of compressional in-plane faulting in the upper crust and extension-drivcn ductilc flow in the lower crust and mantle, which also deform the force (stress times elastic plate thickness) causes uplift of the midplane, is not included. Therefore, as these models do not basin flank, contemporanous with subsidence of the basin center, thus inducing a relative sea level change with different incorporate all the preexistiug deformation of the lithosphere, sign and magnitude across the basin system. In this model the predictions of these models cannot bc correct. For basin modeling purposes, the models of Kooi and Cloetingh [I9921 result, the magnitude of differential subsidence is controlled by the thickness of the elastic plate and the distribution and and Kamer et al. 119931 improve the older kinematic basin magnitude of the applied sediment load. The finite element evolution models by taking into account that, because of a models employing a laterally variable elastic plate thickness deep level of necking during extension, the lithosphere can be 940 VAN BALEN ET AL.: A NEW MODEL FOR S'IRESS-INDUCED SUBSIDENCE rifted basin strenath Figure 3. The schematic location of the lithospheric competent layers underlying a rifted basin. On the right, a schematic continental yield strength envelope is depicted. Figure 2. Predefonned shapes of the elastic thin plate representing the lithosphere of two models proposed in the literature and our model. (a) Predeformed shape of the elastic plate in the first generation of elastic thin-plate models. The predeformed shape is caused by flexure in response to thermal and sediment loads. Consequently, the predeformed shape is downward bend. (b) Predeformed shape of the elastic plate of basin evolution models taking into account necking of the lithosphere during extension. In this model, depending on the depth of the level of necking, additional buyoancy forces rel;dted to the necking can cause an upward bend pd~efornwd shatx. (cj Our nreferred mcwlcl The nredeformd %haw(ifthe elastic plate results from dcf&ation due to -tectonic processes and the flexural response to thermal, sediment, and necking-related loads. in an upward state of flexure [Weissel and Karner, 1989; Braun and Beaumont, 1989a; Chkiy et a/., 19921. This leads to different predictions for lithospheric thinning and heat flow upon forward modeling of rifted basins. However, the flexural state of the lithosphere is irrelevant in thin-plate models for determining the response to changing in-plane forces. The only relevant parameter is the geometry of the midplane at the time of in-plane force changes (Figure Zc), and this geometry is caused by many processes, including the flexural response to necking of the lithosphere during extension. As will be shown in section 4, inclusion of a preexisting permanent deformation of the lithosphere, induced by extension processes, controls the predicted deflections to a large extent. Because of its rheological structure, the continental lithosphere cannot be approximated by a single elastic plate. Depending on crustal thickness and temperature conditions, mechanical weak layers residing in the lower crust may cause flexural decoupling between the upper crustal and subcrustal competent layers [McNutt et al., 1988; Ranalli, 1994; Burov and Diament, 1995; Cloeringh and Burov, 19%; Ter Voorde et al., 19981. As these two competent layers deform differently during extension (Figure 3), our new model fm stress-induced deflection includes flexural decoupling. Before presenting our model, we first briefly review the rheology and flexural behavior of oceanic and continental lithosphere. 3. Effective Elastic Thickness and Rheology of Oceanic and Continental Lithosphere For an elastic thin-plate, fiber stresses resulting from bending are zero at the plate center and increase linearly in magnitude outward according to a gradient defined by the Young's modulus, Poisson's ratio, and the curvature of the plate (Figure 4a) [Timoshenko and WoinowskyKrieger, 19591. The bending moment in the plate is obtained from the integral with depth of the fiber stresses with respect to the midplane of the plate. The flexural rigidity D of an elastic plate is defined as the ratio of the bending moment and curvature of the bent plate. For an elastic thin plate, the flexural rigidity is a function of the plate's thickness T and its elastic properties [Timoshenko and WoinowskyKrieger, 19591: E T ~ D= 12(1-v2) where E is Young's modulus and v is Poisson's ratio. However, fibcr stresses of a lithospheric plate are lir~iited by its yield strength [McNutr and Menard, 1982; McNutr er al., 1988; Ranalli, 1994; Burov and Diament, 19951, which varies with depth according to Byerlee's law for brittle failure and ductile creep laws [Kirby, 1983; Carter and Tsenn, 1987; fiber stress fiber stress 5 4 l?igure 4. (a) pibcr in a bent single elastic thin plate. (h) Fiber stress profile in a multilayer elastic thin platc according to the decoupling model of McNun et al. [1988]. VAN BALEN ET AL.: A NEW MODEL FOR STRESS-INDUCED SUBSIDENCE Ranalli, 1987; Zoback et al., 1993; Cloetingh and Burov, 19961. Therefore the bending moment in a lithospheric plate is much smaller compared to that of an equally thick, purely elastic plate bent to the same curvature. The observed bending of a lithospheric plate can, however, still be used to compute its effective elastic thickness 7,. Furthermore, T, can be computed from the vertical stress distribution for a specific part of the lithosphere if its rheological state and curvature are known [Burov and Diament, 19951. Although u~uallynot considered when T, is computed, scresses from other sources than bending contribute to yielding of thc lithosphere (i.e., thermal stress, in-plane stress, etc.) thus limiting the magnitude of the fiber stresses and thereforc rcducing the value of the T, [Kooi and Cloetingh, 1992; Burov and Diament, 1995, 1996; Cloetingh and Burov, 19961. On the basis of observed lithospheric deflections, values for T,have bccn determined from glacial rebound data, foreland basin shapes, seamount-induced ocean floor deflections, bending of subducted slabs, and the shapes of rifted sedimentary basins [McNun and Menard, 1982; Watts et al., 1982; McNun et al., 1988; Cloetingh and Burov, 19961; see Burov and Diament [I9951 for an extensive overview. For oceanic lithosphere, which is characterized by only one mechanically strong layer, computation and interpretation of T, is relatively straightfonvard [McNutt and Menard, 1982; Burov and Diamnt, 19961. In contrast, for continental lithosphere, depending on composition, crustal thickness, temperature, and stress conditions, interpretation and computation of T, is complicated by the possible presence of one or more weak layers, splitting it up into several mechanical strong parts [McNutt et al., 1988; Lobkovsky and Kerchman, 1991; Ranalli, 1994; Burov and Diament, 1995, 19961. Such weak layers are mainly located in the lower crust [Kirby, 1983; Carter and Tsenn, 1987; Ranalli, 1987; McNun rt ul., 1988; Burov and Diament, 1995, 1996; Cloetingh and Burov, 19961. Therefore the competent layers are residing in the upper crust and upper p a t of the subcrustal lithosphere (Figure 3). This system of two competent layers separated by a weak layer can lead to a decoupled flexural behavior. In the decoupled elastic thin-plate model for continental lithosphere, originally proposed by McNutt et al. [I9881 and extended by Burov and Diament's [I9951 leaf spring model, the lower crustal weak layers are incapable of sustaining fiber stresses created by bending (Figure 4b), assuming that such stresses result in instantanuous flow of lower crust. In this approximation the shear and horizontal stresses applied to one of the competent layers will not be transferred to the other competent layer [Burov and Diament, 19961, and lower crustal shear allows for flexural slip between the upper crustal and subcrustal competent layers. Owing to limited lower crustal flow, the vertical motions of the competent layers are assumed to be the same, that is, the competent layers behave as a single flexural system. The total rigidity of the decoupled system is given by the sum of the separate rigidities of the competent layers. Therefore the total T, of a decoupled system is given by LMcNun et al., 1988; Burov andDiament, 19951: (1) where h,, h , ..., h, is elastic thicknesses of the n separate competent layers. 94 1 The rigidity of a decoupled system is much smaller than the rigidity of a welded system of the same thickness bent to the same curvature. Therefore the decoupling model can explain that the T, of the lithosphere is generally larger than the theoretical T,of the crust alone but is smaller than the sum of crust and subcrustal T, [Burov and Diament, 1995,19961. The rheological state of the lithosphere controls whether or not the crust and the subcmstal lithopshetic mantle are ~nechanicallydecoupled, it depends mainly on the thickness of the crust and the thcrmal stnte of the lithosphere. For old (cold) lithosphere, the critical thickness is about 35 km, which coincides with the average crustal thickness. ' h i s suggests that most of the continents are in a decoupled mode [Burov nnll Dinment, 19961. About 75% of observed continental T, can bc explained by flexural decoupling [Burov and Diament, 1995, 1996: Cloetingh and Burov, 19961. 4. New Stress-Induced Differential Subsidence Model The preexisting deformation of an elastic thin plate affects the response of the plate to a change in the magnitude of the in-plane force [Timoshenko and Woinowsky-Krieger, 19591. For a single elastic plate an increase in the amount of the farfield in-plane force amplifies the existing bend w,, whereas a decrease of force will flatten it. The flexural response is given by [Timoshenko and Woinowsky-Krieger, 19591 where w is force-induced deflection, F is in-plane force change, D is rigidity, p, is density of asthenosphere, pa, is density of material filling in the depression (sediment and water) and g is gravitational acceleration. In the derivation of (2), w, repxsents the preexisting shape of the midplane of the thin platc bcforc the forcc is applied [Timoshenko and Woinowsky-Krieger, 19591. The preexisting defamation can be permanent, elastic, or a combination of both. The total deflection of the midplane is given by ( w + w, ). In the decoupling model, however, the lithosphere consists of two or more superimposed elastic thin plates. This can he accounted for in the flexure equation hy noting that each elastic plate carries p a t of the total in-plane force applied to the lithosphere [Cloetingh and Burov, 19961. Each of these separate forces acts on a deformed midplane, with the shape of the preexisting deformations differing between the individual plates forming the composite beam. For example. the stress-induced deflection for a continental lithosphere consisting of two elastic plates, one residing in the upper crust and one residing in the upper subcrustal mantle, is described by the following equation: + w(Pm - Ptill)8 = 0 (3) where D is rigidity of decoupled elastic plate system, Fc is inplane force change in the crustal layer, Fm is in-plane forcc change in the lithospheric mantle, w, is preexisting deflection of crustal midplane, w,,, is preexisting deflection of subcrustal mantle midplane and w is deflection of the lithosphere caused 942 VAN BALEN ET AL.: A NEW MODEL FOR STRESS-INDUCED SUBSIDENCE by the in-plane force. Assuming an elastic response and no stress relaxation in the fibers, Fc and Fm can he related to the total force F acting on the lithosphere and the thicknesses of the upper crustal and subcrustal competent layers: F m h= F A Fc=- h h,+ h, hc+hm where h, is uppercrustal competent layer thickness and h,, is subcrustal fiber thickness. An approximation for the shape of the preexisting deformation of the crustal elastic plate is given by the basement topography. However, the basement topography must be corrected for erosion which is likely to take place during the synrift and early postrift phase of the basin, particularly in the area of the rift shoulders. As the midplane is located about halfway between the upper and lower boundary of the upper crustal competent layer, it is displaced downward by half the magnitude of erosion (i.e., the bottom of the layer is not modified by erosion): w, = hwement topography + erosion amount / 2 The topography of the Moho discontinuity is probably the best approximation for the shape of the preexisting deformation of the subcrustal flexural midplane, as it is located close to the strongest part of the mantle lithosphere predicted by rheological modeling using yield-strength envelopes [e.g., Cloetingh and Burov, 19961. For basin nrodding purposes, both the Moho and the basement topography have the advantage that they can be observed and mapped on deep-seismic profiles. In the twolayer mudel, the flexural effect of a change in the magnitude of an in-plane force depends on the curvature of the midplanes of the lithospheric competent layers. As outlined above, those parts of the continental lithosphere contributing to its flexural strength are mainly located in thc upper crust and the uppermost part of the subcrustal mantle. During extension processes, these parts are permanently deformed by extensional brittle faulting and ductile flow. As demonstrated by the results of deep seismic profiling [Holliger and Klemperer, 1990: Bois, 1993; Posgay et al., 19961, dynamic models for lithosphere extension [Zuber and Parmentier, 1986; Braun and Beawnont, 1989b; Bassi et al., 19931 and tectonostratigraphic forward modeling [Cloetingh et al., 19951, the upper crust is generally displaced downward, and the lithospheric mantle is displaced upward because of lithospheric extension. Therefore the preexisting shapes of the upper crustal and subcrustal fibers are opposite, as is the contribution of these fibers to the total in-plane force-induced deflection. However, on deep seismic profiles of rifted sedimentary basins, the Moho is relatively smooth compared to the basement topography [Holliger and Klemperer, 1990; Buis, 19931, although subcrustal faults can also occur [Flack et al., 1990; Posgay et al., 19961. During the posuift phase of basin subsidence the Moho relief is reducing because of thermal subsidence. Moreover, lower crustal flow, occuning on a large timescale, may contribute to flattening of the Moho during the postrift stage [Kaufman and Royden, 1994; Ter Voorde et al., 19981. If the Moho topography has a negligible curvature, then the preexisting shape of the midplane of the submstal competent layer contrib~~tes little to the differential venical motions of the lithosphere upon the buildup of in- plane force. In contrast, the shape of the sediment-basement interface is generally highly irregular. This shape is an approximation for the geometry of the upper crustal competent layer. Therefore total basinal in-plane forceinduced deflections should be largely determined by the effects of the geometry of the midplane of the upper crustal competent layer. As a result, an increase in compressive inplane force will in most rifted basins cause a downward dcflcction of the center of the rifted sedimentary basin, because the preexisting upper nustal midplane geometry is bent downward. As can be observed on seismic profilcs and in outcrops, the upper crustal part of the lithosphere is permanently deformed by discrete faults, causing a rugged hwement topography. In fact, in rifted basins most of the deformation of the upper crustal competent layer results h m faulting. Therefnre, at first sight, the curvature of the upper crustal flexural midplane locally attains extremely high values, particularly along the rift margins which are marked by large border faults with throws of a few kilometers. However, owing to the presence of faults, the upper crustal midplane is discontinuous, and therefore the curvature of the midplane is not defined at the position of faults. A new theory is required in order to resolve this problem. 4.1. Faulted Lithosphere and Stress-Induced Deflection A fault in the upper crust causes an offset in the upper crustal competent layer. For vertical loading, the iaulted elastic plate behaves like an equally thick continuous plate [e.g., Kusznir et al., 19911, because the loads are acting perpendicular to the plate. As faults and fractures are widespread in the lithosphere, this is a basic assur~rplionfor all models dealing with lithosphcric flcxure in response to vcrtical loading. Howcvcr, for horizontal loading the venical offset causes an additional moment at the fault, because the in-plane stress is tmnsmitted through the overlap zone of the two pans of the plate which are separated by the fault. The overlap zone, located at the fault, is narrower than the plate, leading to a slight concentration of in-plane stress. The horizontal in-plane stress profile in the overlap zone is not in equilibrium with respect to the midplanes of thc two snbplates, that is, at the overlap zone the integral of in-plane stress with respect to the midplanes of the subplates is nonzero. This causes a moment at the fault (Figure 5). Seismotectonic and deep seismic reflection data suggest that fundamental basement faults in rifted sedimentary basins are planar and are restricted to the brittle, cold, topmost pan of the lithosphere, corresponding to the upper crust [King et al., 1988; Jackron and White, 1989; Kusznir et aL, 1991; Roberts and Yielding, 1991; Van Wees et aL, 19961. Therefore, in the model we constructed a venical fault that is assumed to cut the entire upper crustal competent layer. The results of finite element modeling, presented in section 4.2, demonstrate that the assumption of a vertical fault plane does not significantly influence the prediction as long as the fault dip is higher than 63'. Fundamental planar basement faults must have the largest effect on stress-induced deflections, as they cause most of the deformation of the upper crustal competent layer in rifted basins. However, elastic solutions for Fdulls no1 culling the entire upper crust also exist [Savage VAN BALEN ET AL.: A NEW MODEL FOR STRESS-INDUCED SUBSIDENCE force -- midplane elasticplate Figure 5. The in-plane force-induced moment at a fault from the of the faulted thin plate, The moment results from disauilibrium of the horizontal in-olane stresses with respect to the'midplanes of the subplates. A and Guohua, 19851 and possibly can be extend* to include variations in in-plane force levels. by a faulted thin plate is The hrfield in-p'ane force constant. Using this balance of horizontal in-plane force and that the the geometry of the faulted Plate, it can be total in-plane force-induced moment at the position of a fault equals minus the magnitude of in-plane force times fault offset, -F u [Van Balen and Podladchikov, 19981. Using gmeral solutions for the thin-plate equation [e.g., Hetdnyi, 19461, ht:flexural profile caused by the in-plane forceinduced moment can bc found unalytically for several scenarios, including variable rigidity [Van Balen and Podladchikov, 19981. For a constant rigidity the elastic flexural profile due to the moment is described by w=- Fu 4Dap sin(xp) e-''la B=- (4) where w is deflection, D is flexural rigidity, F is in-plane force (positive for compression), u is fault displacement, a and B are flexural parameters, Ap is density difference between infilling and compensating material, x is 0 corresponds to the fault position, and g is gravitational acceleration. The value of u is positive if the downthrown block is on the left side of the fault; otherwise, it is negative. Typical profiles predicted by this equation are presented in section 4.2. 4.2. Finite Element Modeling In this section, the flexural profile predicted by (4) is compared to elastic finite element solutions. Two sets of models are analyzed: the first set employs a 10 km thick elastic plate, and the second set has a plate thickness of 5 km. These values for the elastic thickness are warranted by T, estimates from the shapes of deflections around normal faults in rift settings [King et al., 1988; Kusznir et al., 1991; Roberts and Yielding, 19911. Rifting processes occur at the same timescale as in-plane stress level changes, that is, a few My 943 (in fact, rifting is caused by an in-plane stress magnitude which exceeds the tensile strength of upper crustal lithosphere). Furthermore, flexural bending due to normal faulting occurs at the same spatial scale as fault-moment bending. Therefore we assume that both normal faulting during rifting and flexural bending due to fault moments are controlled by the same T, (see also section 4.4). All models apply 100 MPa far-field compression and a fault throw of + I km, which is within the limits defined by the yield strength of the lithosphere [Kusznir and Park, 1984; Savage and Guohua, 19851 and values deduced from studies on folding of oceanic [Cloetingh and Wortel, 1985; Beekman et al., 19961 and cu~lline~ilal lithosphere [Lambeck, 1983; Cloetingh and Burov, 19961. Apart from thc intraplatc strcss, applied boundary conditions are zero vertical displacement at the lateral boundaries of the model (Figure 6). An array of springs at the bottom of the plate simulates the buoyancy conditions. The upper side of the model is a free surface. In order to reduce the effect of the lateral boundary conditions, the models have a total length of 1500 km. For some models, the plate is not allowed to thicken under the application of compressive stress thus mimicking the thin-plate assumption made in the derivation of the general elastic flexure equation [Timoshenko and Woinow,vb-Krieger, 19591.. The adopted values for Young's modnlus, Poissonk ratio and the density contrast between compensating are the same for the analytical and finite element models, that is, 7x101Upa,0.25, and 90(1 kg respectively. me applied density contrast is based on the density of fresh, surlicial, hardly compacted sediments with a density of about 1800 kg m-3 and a lower crustal density of 2700 kg m ~ 3 , For the km thick model the analytical solution is to finite element results for models employing 1 km by 1 km and 2 km by 2 km quadratic finite elements, respectively (Figure 7a). Both nu&cal models predict the same result, indicating that the element size is not influencing the prediction. The difference between the analytical result and numerical solutions is up to 10% at the maxima of the flexural nrofile. The finite element model in which the date is not allowed to thicken predicts a flexural profile which nearly coincides with the analytical prediction. For the 5 km thick plate Lhe analytical solution is compared to finite element models with 1 km by 1 km and 2 km by 1 km elements, respectively (Figure 7b). Both these models predict the same deflection, indicating again that the element size docs not influence the predictions. The maximum difference between numerical and analytical solutions is now up to 14%. The finite element model without plate thickening deviates ,,-; 6. Outline of geometry and boundary conditions used in the finite element nefault displacement is represented by a step in the upper and lower boundaries of the plate. Vertical springs at the bottom of the plate simulate buoyancy conditions. VAN BALEN ET AL.: A NEW MODEL FOR STRESS-INDUCED SUBSIDENCE 944 analvtical solution o N ........ 2x2kmFE fixed dy FE firmre 7. Comparison of analvtical predictions for deflection caused bv a fault moment to results of finite cl&ncnt modelini employing q;adrat~i elements. The models cmploy a iaull throw of I km and an in-plane stress ot IW MPa i l ktlobar). (3) K e s u l ~of~ 10 km thtck elastic fin~teelement model, ctn~lovineI X I km and 2x2 km sized elements and.lx1' km elements without vertical thickening compared to inaiytical prediction employing the same horizontal in-plane stress and elastic parameters. (h) Analytical results compared to \&ou; tinite element snlutlons for i 5 km thick plate. The numerical models inludf 1x1 km and 2x1 km si7eJ elemenl.~, I X I km elements w~thnutvenisal thickcntng, and a fault with a dip of 63'. Scc text tior discussion. slightly (7%) from the analytical solution. We also tested the influence of fault inclination. As shown by the result, a numerical model employing a 63" fault angle does no1 differ significantly from the numerical models employing a vertical fault plane. As predicted by the analytical solution (equation (4)). the stress-induced deflection scales linearly with fault throw u. T%is is confirmed hy numerical modeling: plates with fault throws of 2 km and 3 km subjected to 100 MPa stress predict deflections with magnitudes which are 2 and 3 times larger, respectively (in Figure 7b the resulting flexural displacements are divided by 2 and 3). On the basis of the numerical modeling results, we conclude that the analytical elastic thin-plate equation (4) is accurate enough for basin modeling purposes and that fault dip does not significantly influence the predicted flexural profile as long as it is larger than 63O. 4.3. Predicted Flexural Profiles for Multiple Fault Moments and Varinble Rigidity In-plane force-induced flexural profiles for fault systems can be obtained by applying the method of elastic superposition [HetPnyi, 19461, that is, by adding up the flexural profiles for the single fault solution (Figure 8a). The flexural profile for a moment occuning at a single normal a half graben, was obtained by applyitig a fault, represe~~ting 5 km.T,, a Pdult throw of 2 km and an in-plane compression of 100 MPa carried by the whole plate (resulting in an in-plane force of 5x10" N m-'). As shown by (4), the magnitude of the differential subsidet~cecaused by the fault moment scales linearly with applied in-planc forcc and fault displacement. Therefore the flexural profiles presented below can also result h m a combination of larger fault throws with less in-plane stress and vice versa. The elastic parameters have the same values as in the previous models. The predicted flexural profile for the half graben is characterized hy a maxirnurn uplift of about 75 m at the footwall and an equal amount of subsidcncc at thc hangingwall (Figurc 8a). I h e two maxima are separated by 50 km. Therefore the slope of the differential subsidence profile is at its steepest position equal to 0.17'. The differential subsidence profile predicted for a system consisting of two master normal faults dipping in opposite directions and spaced 50 km apart, delineating a simple rift, shows almost 150 m subsidence in the graben and 75 rn uplift at the flanks. The larger amount of subsidence in the maben is caused bv positive interference of the two flexGal profiles induced by h e fault moments. The maximum slope in this profile is 0.23". A model involving a set of five master normal faults spaced 25 km apart and having throws in the same direction predicts a flexural profile with maxima of 140 m Figure 8c). Finally, a model combining two opposing sets of five master faults predicts a maximum subsidence in the graben of about 210 m and a maximum uplift at the flanks of 170 m (Figure 8d). The maximum slope in this profile is 0.17". The resulting profiles show that closely spaced multiple-fault systems generate positive interference, causing enhanced compression-induced subsidence at a scale largcr than the individual tilted fault block scale, that is, basin-wide. 945 VAN BALEN ET AL.: A NEW MODEL FOR STRESS-INDUCED SUBSIDENCE I -100 -50 0 50 , l , I , -100 -50 100 I , l 0 , l 50 I I 100 distance (km) distance (km) 0 0 7 - u - 1 1 1 -100 -50 1 1 1 1 0 50 distance (kin) 1 1 1 1 100 u - I , I , / -100 -50 , l I I 0 50 distance (km) I l ~ l - 100 Figure 8. In-plane stress-induced differential subsidence profiles (continuous line), or relative sea level change, for different fault combinations. Faults have tbmws of 2 km. The elastic plate has a thickness of 5 km and is subjected to a 100 MPa compressive stress. Dashed lines indicate the faulted basement profile (vertical scale bas to be multiplied by 10). (a) Resulting profile for a single fault. The compression induces 75 m uplift of the footwall and an equal amount of subsidence of the hanging wall. This profile is used to construct the predictious for the other scenarios by applying etdslic superposition. @) Simple rift with two master faults spaced 50 km apart. Predicted footwall uplift (flank) is about 75 m;stress-induced subsidence in ihe graben center equals 150 m. ( c ) Five faults with displacement in the same direction, spaced 25 km apan. m e maximum uplift and subsidence are both about 140 m. The resulting differential subsidence pattern shows an almost linear trend across the faults, with a slope of 0.16". (d) Results for two opposing sets of five faults. Within a set, faults are spaced 25 km apart and have the same tbmw direction. The two sets are spaced 50 km apart. The maximum of the predicted subsidence exceeds 210 m, whereas maximum uplift is about 170 m. m e trends of differential subsidence across the fault sets are almost linear, with slopes of 0.17'. 4.4. Implementation of Fault Moment Solution Although it is the most important part, the contribution of fault-related moments to the differential subsidence in a rifted basin is only part of the total response of the rifted basin to changing far-field stresses. The shape of the m s t a l midplane without faults and in some exceptional cases the shape of the r~laritlemidplale (see below) also contribute. The lotal, basinwide differential subsidence pattern can be obtained by superposition of the components conmbuting to it. The shape of the upper crustal elastic layer without the faults is found when the fault throws are reversed. After the fault offsets are 946 VAN BALEN ET AL.: A NEW MODEL FOR STRESS-INDUCED SUBSIDENCE restored, the discontinuities in the basement profile are eliminated and therefore the resulting shape of the upper crustal elastic plate is differentiable. The flexural subsidence profile due to a changing in-plane forcc can be found by using a standard finite difference method. For oceanic, thick, and hot continental lithospheres, (2) should be used in combination with the fault solution (4). because, owing to the rheological state, their flexure can be effectively described by one elastic plate [Cloetingh and Burov, 1996). Stress-induced deflection for an average continental lithosphere should be calculated using (3) in conjunction with the fault solution (4). Flexure of a thin continentaL lithosphere underlain by normal or cold mantle is dominated by the mantle rheology. Therefore, in this exceptional case, the single elastic plate equation (2) should be used, and fault moment contributions can be neglected. The T,controlling the fault moments can be substantially less than the T, of (decoupled) lithosphere. Values of T, estimates from the shapes of deflections resulting from normal faulting are about 5-10 km [King et al., 1988: Kusznir et al., 1991; Roberts and Yielding, 19911. The timescales of normal faulting (e.g., rifting) and fault moments are the same, that is a few million years. Additionally, they have also the same spatial scale. Therefore they should be governed by the same T,. The T, of large-scale deflected, (decoupled) rifted continental lithosphere is in the range of 10 - 20 km [Watts er al., 1982: Burov and Diament, 19951. The low value of T, for faulting and fault moments can be explained by a combination of two effects: a reduction of T, around faults because of relaxation of fiber strcss [Kusznir et al., 19911, and the effect of the scale of the deflection or, equivalently, by the difference in load distribution [Ter Voorde et al., 19981: Sediment, thermal and necking loads typically have a basinwide distribution. This causes a relatively minor pressure gradient on the lower crustal low-viscosity channel, inducing negligible lower crustal flow on a 10-100 My timescale. Normal faulting and fault moments, however, produce more localized loads and therefore cause higher pressure gradients in the lower crust. Therefore local lower crustal flow may provide flexural compensation for normal faulting and fault moments. The T, for faulting and fault-moments is thus controlled by the effective elastic thickness of the upper crust alune. Additionally, the relaxation of fiber stress in the vicinity of faults may cause a reduction of T, compared to the upper crustal value [Kusmir er al., 19911. Moho uplift (Figure 9b). The models with a flat Moho represent a basin orginating from a deep level of necking during extension. The models with a Moho uplift represent a hasin which developed with an interrncdiatc depth of necking, located in the lower crust. Each model set contains three configurations for the total T, distribution. As the T, results from decoupling, it can be located mainly in the upper crust or mainly in the subcrustal mantle, or it can be about equally distributed between the mechanically strong layers, resulting in the same total T, according to (1). In the models, combinations of upper crustal and subcrustal competent layer thicknesses of 5 km and 14.8 km, 11.9 km and 11.9 km, and 14.8 and 5 km are applied. The result for the models adopting a flat Moho show 100 to 200 m flank uplift and 180 to 240 m basin center subsidence (Figure 9a). The model with the largest upper crustal competent layer thickness (14.8 km) predicts the largest amplitude of deflection.The models adopting a slight Moho uplift predict 150 to 200 m flank uplift and 150 to 190 m basin center subsidence, also depending on ihe competent layer thickness distribution (Figure 9b). For the model adopting thicknesses of 5 and 14.8 km for the upper crustal and subcrustal competent layers, rcspwtivcly, the separatc contributions from thc faults, the geometry of upper crust without the faults, and the deformed shape of the mantle to the total stress-induced deflection is shown in Figure 9c. In this model, upon compression, the subcrustal competent layer causes most of the flank uplift, whereas the faults induce the basin center subsidence. The results demonstrate that with increasingly deeper levels of lithospheric necking during rifting the stress-induced differential subsidence amplitude increases. Furthermore, the total amount of stress-induced subsidence and uplift depends on the thickness of the upper crustal competent layer, because the total in-plane force acting in this layer is the product of inplane stress and the thickness of the layer. Finally, the models representing a basin orginating from an intermediate level of necking during extension predict flank uplift which is greater than the basin center subsidence, leading to a considerable narrowing of the basin. The flank uplift is mainly caused by stress-induced deflection of the subcrustal upper mantle. 4.5. Predictions for Total Stress-Induced Deflection Applying the numerical procedure outlined in section 4.3, we have computed stress-induced differential subsidence and uplift of a decoupled continental lithosphere for different rifted basin scenarios. In all examples, the total T, of the decoupled lithosphere is 15 km, and the T, for the fault moments is 5 km. The adopted in-plane stress level change is 100 MPa compression. Elastic parameters have the same values as in previous examples. The dcnsity difference for lower crustal compensation is taken as 700 kg I"", asthenospheric compenstaion is calculated adopting a dcnsity diffcrcncc of 1300 kg m-'. Two opposing sets of five faults are located in the upper crust; each fault has a throw of 2 km. Two types of models are analyzed, one set has a flat Moho discontinuity (Figure 9a), while the other set adopts a slight 5.1. Southern North Sea Basin 5. Examples of Stress-Induced Subsidence and Uplift in Rifted Basins Related to Faults During the Tertiary, the overall development of the North Sea basin is characterized by subsidence rates in accordance with the poshift development following Triassic to Early Cretaceous rifring events [Ziegler, 1990; Kooi et al., 19911. During late Pliocene-Quaternary times unusually rapid subsidence and sedimentation occurred in the southern pan of the North Sea basin, leading to a Quaternary sediment thickness of up to 1 km (Figure 10). Since most sediments were deposited in a shallow-water environment, this part of the Noah Sea basin must have experienced an extremely high tectonic subsidence rute during the Quuternay [Kooi et al., 19911. Emhquake focal mechanism solutions and borehole breakout data indicate that the present-day stress field in the - VAN BALEN ET AL.: A NEW MODEL FOR STRESS-INDIJCED SUBSIDENCE c I uplift # +ZN m o o G=o . $ S T :subsidence -0 rE z t o 5 R -r Moho p - / -5114.8 f 0 I 100 200 1 I I 300 400 500 1 I distance (km) I I I !subsidence - 5114.8 11.9111.9 14.815 100 b) O E?N m o o F O 200 300 400 500 distance (km) - ---- --.- - - L _ - ' 7- I :uplift \ - 1 I I , - top of crust / - - / -7 S E -r -crust subcrust faults ~ o h o - 0 100 200 300 - __/---- $=? :subsidence C) - 5 : c - I C .-O - N r- uplift Zg? o - 11.9111.9: 14.815 - - a) -. . --- - 400 - : - 500 distance (km) Fimve 9. Predicted differential subsidence for rifted basin eeomebies orieinatine from deeo and shallow lcL.ls of lithospheric necking during extension. For e ~ c hg c o ~ ~ c t rthc y , towi?; of i 5 km i\ distrihutd oier the unner crustal and subcrucwl a)mnctcnt l ~ v e nas ~,t~rnhinations of 5 km and 14.8 km. 11.9 km and 11.9 km, a i d 14.8 and 5 km. The applied'in-plan~suess is 100 MPa. (a) The results for the models representing deen neckine show 100 to 200 m flank uolift and 180 to 240 basin center subsidence. The model with the Irrrge\t u p p crustal competent l q e r thicknes5 predicts the lugest amplirudc of strcss-induced deflection. (b) The rndels adtnuinu a slight Moho uol~h.reoresentinr an intermediate level of neckine. nredict I50 to 200m flank uplift and 155 to i90 m bash center subsidence, depending on the thickne; of the upper crustal compctcnt layer. (c) The sepnrate contributions of upper crustal faults, geometry of upper crust without faults, and shape of the subcrustal lithosphere to the total deflection for the model adopting a slight Moho uplift and thicknesses for the npper crustal and subcrustal competent layers of 5 and 14.8 km, respectively. The faults contibute most to the predicted subsidence in the basin center, whereas the subcrustal suess-induced deflection causes the largest put of thc basin flank uplift. See text for further discussion. - 948 VAN BALEN ET AL.: A NEW MODELFOR STRESS-INDUCED SUBSIDENCE Figure 10.- (a) Contour map of the Quatemary thickness distribution in the southern part of the North Sea basin [Kooi et al., 1991; Van den Berg, 1994). Axial labels are in hundreds of mctm. Thc maximum thickness locally exceeds 1 km. (b) Main Mesozoic fault zones and rift systems in the southem North Sea basin [Ziegler, 1990; Kooi et al., 1991; Van den Berg, 19941. Abbreviations are as follows: MF, Morray Firth: VG, Viking Graben; CG, Central Graben; DCG, Dutch Central Graben; SP, Sole Pit basin; BF, Broad Fourteens basin; HG, Horn Graben. A good correlation exists between the N-S orientated faults and the local maxima (plus sign) and minima (minus sign) of Quaternary subsidence. Hanging walls correspond to maxima; footwalls correspond to minima. See text for discussion. southem part of the Nonh Sea basin is characterized by a maximum compressive stress orientated NW-SE [Ahorner, 1975: Illies, 1975: Miiller er al., 1992; Lindhulrn et ul., 19951. The NW-SE compression is caused by a combination of Alpine collisional forces and North Atlantic ridge push. The effea uf the stress field on Quatemary subsidence was investigated by Kooi et al. [1991]. Their results indicate that the anomalous subsidence can be partly explained by a combination of a 300 MPa far field compressive stress and a decrease of T, from 20 km at the rims of the basin to about 1 2 km below the Quaternary centers of maximum deposition (giving rise to an unrealistic local in-plane stress of about 3000 MPa). The fit was improved by adopting a decrease of paleowater depth during the Quaternary. However, their model requires a high amount of far-field stress and a paleohathymetry which is not in accordance with observations. Therefore Kooi et al. [I9911 favor a model based on hypothetical focal extension below the Quaternary centers of maximum deposition because of pull-apart tectonics related to strike-slip faulting, for which no evidence exists. A three-dimensional forward model of Quaternary stress-induced subsidence by Van Wees and Cloetingh [1996], also excluding the permanent deformation of the lithosphere, utilizes a lower far-field in-plane stress level (150 MPa) compared to the twedimensional flexural model of Kooi et al. [1991]. The 3-D flexure model can explain up to 70% of Lhe Local Quatemary thickness in the DuLch Central Graben (Figure 10) but fails to predict the remaining patterns of Quaternaty deposition. Furthermore, their model also requires a substantial decrease of T, in the Dulch CenLrdl Graben area (from 15 km at the flanks to lcss than 5 kin in the center). Therefore Van Wees and Cloetingh 119961 also suggest that other tectonic mechanisms, like pull-apart extension, have played a significant role in Quaternary subsidence of the southern North Sea basin. There is, however, no evidence for major, large-scale Quaternary faulting, and the relationship between the NW-SE directed maximum compressive stress directions and the geometry of the fault system requires that the N-S orientated centers of maximum deposition are transpressed by the proposed strike-slip motions and not pulled avart. The thickness distributinn of Quatemary sediments is characterized by a saucer shaped long-wavelength component and local elongated thickness maxima and minima (Figure 10a). The largest Quaternary centers of deposition in the southem North Sea basin, located in the Dutch Central Graben area, are on the downthrown sides of N-S striking Mesozoic normal faults (Figure lob) [Kooi and Cloetingh, 1989; Van den Berg, 19941. Local minima of Quaternary thickness are on the footwalls of these faults. The strike of the N-S normal faults bas an angle of 45' to the maximum compressive stress direction. The NW-SE orientated faults, parallel to the maximum compressive stress direction, are not associated with Qualm~ary cmlers or deposition and 949 VAN BALEN ET AL.: A NEW MODEL FOR STRESS-INDUCEO SUBSIDENCE thickness minima. Furthermore, in the Central Graben and Egersund Basin of the northern North Sea basin liltle or no increase in subsidence has occurred [Hall and White, 19941; these basins strike parallel to the maximum compressive stress direction. The N-S striking faults have a component of the maximum compressive stress acting on them. Therefore we suggest that the Quaternary centers of maximum deposition and thickness minima can be explained by compressive stress-induced suhsidence related to moments acting at the position of the Mesozoic faults. Faults with throws of several kilometers (up to 7 km) are common in this par1 o l the Nonh Sea basin [Ziegler, 1990; Holliger and Klemperer, 19901. As the differential subsidence caused by the moments scales linearly with the Caul1 displacement and in-plane force, the results in Figure 8 can be used to cstimate thecontribution of the moments to the differential subsidence (a T, of about 5 km for the faulted upper crust of the Nolth Sca basin is inferrcd by Roberts and Yielding [1991]). In the centers of maximu deposition, the thickness of Quaternary sediments is about 400 m more than the regional longwavelength trend. This can, for example, he explained by a combination of fault throws of 5 km and a far-field stress of 100 MPa. The total Quaternaty thickness distribution is then the sum of small-wavelength fault-related differential subsidence and the long-wavelength Quaternary distribution. The latter can be explained by the compressive stress response of the downflexed lithosphere caused by necking, sediment, and thermal loading 5.2. Pannonian Basin The Pannonian basin is a Neogene intramontane basin system hounded in the west by the Alps, in the north, east, and southeast by the Carpathians, and in the southwest by the Dinarides. 'The Pannonian basin orginates &om contemporanous tectonic escape in the Alps, asthenospheric upwelling, and subduction rollback along the Carpathian kont [Ratschbacher et al., 1991: Horvdth, 19931. Because of changes in platc motions, the stress field in the Pannonian basin changed to compression during late Pliocene [Csontos et al., 1991; Muller et al., 19921. The maximum horizontal stress direction in the southern part of the basin is roughly SW-NE oriented and in the northern pan coincides with the European NW-SE trend [Muller er al., 19921. During late Pliocene-Quaternary times, lhe wnlral pan of the basin system experienced an acceleration of subsidence, whereas the external parts were uplifted, with the exception of the local subhasin ccnters in the southwestern and northwestern parts of the basin system (Figure I la). Dilferenlial subsidence took place without major fault reactivation [Horvdth and Cloetingh, 19961, indicating a flexural mechanism. Uplift in excess of 300 rn occurred in the Trans-Danubian central range, whereas about 1500 m of sediment accumulated in the Mako Trough and Bikks Basin during the same time interval. Using the previously published thin-plate model [i.e., Kooi and Cloetingh, 19921, the differential subsidence pattern can be explained by the buildup of the contemporanous compressive stress field. adopting a 300 MPa in-plane stress, in combination with ongoing thermal subsidence in the center of the basin system [Horvdth and Cloetingh, 1996; Van Balen et al., 19981. A comparison of the differential suhsidence pattern with the fault system in the basin shows a clear relatiouship (Figure llb), evidencing that the faults exett a first-order control on the stress-induced differential subsidence. The subsidence and uplift in the Sava and Drava Troughs in the southwest and the Little Hungarian Plain (LHP) in the northwest are perfectly fault bounded. The faults bounding the Mako Trough and Bekes Basin are overstepped by late Pliocene-Quaternary sediments, indicating a flexural rigidity which is less for the hanging walls compared to the footwalls [Van Balen and Podiadchikov, 19981. This can be explained by the high heat flow in the Mako Trough and B6k6a Basin [Pusgay er al., 19x1, causing weakening of the lithosphere. 5.3. Norwegian Margin The morphology of Norway, in combination with the almost complete absence of post-Paleozoic sediments, evidences large-scale Tertiary uplift of western Fennoscandia. Two domal uplifts in southern and noahem Norway are inferred from the present position of preserved planation surfaces [Riis and Fjelhkaar, 1992; Riis, 19961 and results from fission-track thermochronology [e.g., Rohrman et al., 1995; Riis, 19961. These uplifts are spatially separated by a low-lying region. The onshore domes correspond to narrow ullshure shelves, whereas the low-lying saddle in between thcm is associated with the broad shelf of the mid-Norwegian mnrein and Eldholm.19961. The domes . . .. are - lStuevold . . elongated and roughlv ~arallelto the Atlantic coastline. Their maximum elevation is about 2 km. The flexural nature of the domal uplifts is evidenced by their shape and the absence of major fault activity associated with their development. The observed asymmetry of the domes, consisting of a steep westward face and a more gentle slope eastward, is explained by amplified weathering on their seaward flank [Sruevold and Eldholm, 19961. The domal uplift in southem Norway is hounded in the south and west by the North Sea basin and in the noahwest by the M0re Basin and the mid-Norwegian margin (Vering Basin) (Figure 12). These basins were formed in response to Permo-Triassic to Early Cretaceous rifting. A final Paleocene ~xtensionphase in the Mere Basin and the mid-Norwegian margin caused the opening of the North Atlantic Ocean. The structural, basinward dip of strata and deposition of large clastic wedges in the Norwegian offshore sedimenrary basins record substantial Tertiary uplift and erosion of the mainland and subsidence in the basins [Sluevold er a/., 1992; Stuevold and Eldholm.1996; Riis, 19961. The flexural uplift and subsidence of the Norwegian mainland and the surrounding basins is separated in time from the opening of the North Atlantic. Several mechanisms havc been proposed, including convectivekhermal processes, innaplate compression, rift shoulder uplift, glacial erosion, mantle plumes, and phase changes. However, uone of these can satisfactorily explain the uplift and associated subsidence (for an overview see Stuevold and Eldholm [19961). In addition, the exact timing of uplift is controversial: a late Oligocene to early Miocene time is suggested by Rohrmn et al. [I9951 and Stuevold and Eldholm [1996]. Recently, using morphological evidence and offshore seismic reflection data, Riis [I9961 proposed that the Tertiary uplift and subsidence - - .. . - VAN BALEN ET AL.: A NEW MODEL FOR STRESS-INDUCED SUBSIDENCE a) b) Figure 11. (a) Late Pliocene-Quaternary subsidence and uplift pattern in the Pannonian Basin [Horvdtk and Cloefingk, 1996; Van Balen et al., 19981. In general, the flanks of the basin system are uplifted and the central part subsides. Noticeable exceptions are the centers of the Drava and Sava Troughs and the Little Hungarian Plane (see Figure 1lb). Dotted areas indicate the foredeep, flysch belt, volcanic, and metamorphic units of the Alpine-Carpathian-Dinaide belt. (b) Nwgene basement fault pattem and subbasins of the Pannonian Basin system [Horvath, 19931. Abbreviations are as follows: DT, Drava Trough; ST, Sava Trough; LHP, Little Hungarian Plain; MT, Mako Trough; BB, BCk6s Basin; and TDCR, Trans-Danubian central range. The major fault-boundcd subbasins show subsidence or enhanced suhsideuce, whereas their flanks are uplifted, indicating that the faults exert a first-order control on the subsidence and uplift pattern. The faults bounding the Mako Trough and Rikis Basin are overstepped by thc late Pliocene-Quaternary sediments, indicating a flexural rigidity which is less for the hanging walls compared to the footwalls. This can he explained by a higher heat flow in the Mako Trough and BCkCs Basin. took place in two phases: a Palcogene phase and a P l i e Pleistocene phase, each with about an equal amount of uplift. Tlle hinge liues o l the duma1 uplifts, separating uplifted from downwarped areas, coincide almost everywhere with major basement faults. Along the North Sea coast, the hinge liue coincides with the 0ygarden Fault Zonc (Figurc 12) [Riis, 19961, which consists of two major faults, each with a throw of more than 5 km toward the basin [Gabrielsen et al., 19901. Northward, the hinge line is located at the position of the M0reTmndelag Fault Zone. On the mid-Norwegian margin, the hinge line coincides with the Bremstein and Revfallet fault complexes [Stuevold and Eldkolm, 19961. This cornlation pattern continues farther northward [Riis, 19961. Therefore it seems that the major fault zones exerted a firstorder control on the distribution of uplift and subsidence. The stress system of the Norwegain mainland and adjacent basins is comparable to that of the North Sea area [Bungum et al., 1991; Miilkr et al., 1992; Lindholm et al., 19951. Furthemore, if mantle upwelling has contributed to the domal uplift of the Norwegian mainland, it can explain part of the compressive stress distribution 1e.g.. Poliakov et al., 19931, and the total uplift may be the result of a stress-induced amplification of the mantle-caused uplift. 6. Discussion In a recent paper, Kurner et al. [I9931 slated that the developmenl of global or regional unconfotmities by in-plane force variations is difficult to accept for the following reasons: (1) localized fault reactivation dominates the response of the lithosphere lo applied Corces, (2) thc orientation and sign of the present-day stress field changes markedly across the same plate, (3) the magnitude of the inplane force required to modify the stratigraphy of an evolving basin is extremely large, and (4) the flexural deformation is governed by factors that will give very dilferent responses of thc lithosphcrc f a thc samc in-plane force. However, not every in-plane force change leads to active faulting: as long as the prevailing in-plane force is at a level lower than the integrated strength of the lithosphere underlying the basin (including weakness zones due to preexisting faults), flexural deformation of the basin dominates. For example, the latest Cretaceous inversion of the mid-Polish Trough is preceded by a flexural acceleration of subsidence. Possible analogues are the rapid Paleocene-early Eocene phase of tectonic subsidence of the Sverdmp Basin in arctic Canada [Dadlez et al., 1995; Stephemon, 19961 and the acceleration of subsidence during the Paleocene in the central Nonh Sea [Ziegler et al., 19951. These examples show that prior to fault reactivation, flexural subsidence in response to the buildup of compressional stresses occurs in the basin center. Only if the in-plane force reaches a threshold defined by the strength of lithosphere [Kusmir and Park, 19841, will fault reactivation take place. Therefore, if the in-plane force does not reach this threshhold, the lithosphere will respond to changing in-plane forces by flexure. The in-plane force of 5xlO"N m.'adopted in the models in this paper is within the limits defined by the yield strength of the lithosphere [Kusznir and Park, 1984; Savage and Guohua, 19851, and values deduced from studies on folding of oceanic [Cloetingh and Worrel, 1985; Beekn~an et al., 19961 and continental lithosphere [Lambeck, 1983; Cloetingh and Burov, 19961 VAN RALEN ET Al,.: A NEW MODEL FOR STRESS-INDUCED SUBSIDENCE 8 4 4 8 uplifted area El ~ p i i f>t 1500 m Figure 1 2 Generalized Mesozoic-Paleocene fault map of southern Nonvay and the mid-Norwegian margin [Gabrielsen er al,, 1990; Blysrad et al,, 1995: Riss, 19961, Abbreviations are as~ follows: ~ ..~~-. ~7nne: ~ MTFZ M~~~ -. . -.--. -~ @FZ. ,- - -, fivgarden -,-.--. ~- .-.. - -l- - - t, Tmndelag Fault Zone; BFC, Bremstein Fault Complex; m, Revfallet Fault Complex; NSB, North Sea basin; MB, M0re basin; VB, V0ring basin; and VMH, V@ring marginal high. Ruled areas indicate the extent of the late 01';goce&-~lioc&e uplift [Riis, 1996; Stuevold and Eldholm, 19961. The hinge line of the uplift coincides with major fault zones. . As shown by numerous studies, the orientation and sign of the intraplate stress field strongly correlates to plate motions, and the stress field has persistent trends over large areas [Cloetirigh and Wortel, 1985; Zoback et al., 1989; Zoback, 1992; Miiller et al., 1992; Richardson,l992; Coblentz and Richardson, 1995; Lindholm er al., 19951. Indeed, the amount of in-plane force in previously published models is large (up to 300 MPa); the new model presented in this paper requires lower stress levels (100 MPa). In addition, the inferred discrepancy between estimated average stress levels arising from plate boundary forces and isostatically compensated loads (less than 100 MPa) and 951 measured and calculated stresses (up to 300 MPa) can be explained by the ductile nature of the lower part of the crust which causes a redistribution of stresses applied to the whole lithosphere and their amplification in the more brittle parts of the lithosphere [Kusznir and Park, 1984: Savage and Guohua, 19851. Furthermore, the thickness over which the in-plane force is distributed should be taken into account. For example. a ridge push force of 2 . 4 ~ 1 0N ' ~ m-l applied to 80 km thick lithosphere may produce 30 MPa, whereas the same force applied to 160 thick lithosphere gives only 15 MPa [Kusznir and Park, 19841. If the same in-plane force is applied to a 15 km thick strong p m of the lithosphere, a stress of 160 MPa results. The flexural response of rifted lithosphere to an in-plane force change is dominated by the faulted shape of the upper crustal competent layer and, in exceptional cases, also the subcrustal mantle fiber. In rifted basins, the large-scale shape of the faulted upper crustal flexural midplane is always downward bend. Therefore the response of rifted basins to an in-plane force change should always be the same: basin center subsidence and flank uplift. This confirms the original results of Cloetingh et al. [1985]. The elastic thin-plate models including necking of the lithosphere during extension [Kooi and Cloetingh, 1992; Karner et al., 19931 concluded that the deeper the level of necking is during extension the less stressinduced differential vertical motion there is during the postrift phase. Also, if the level of necking during rifting was deer, enough, the flexural response to in-plane stress changes during the postrift stage would be reversed compared to a situation with a shallow level of necking during extension. However, a deeper level of h e lithospheric necking causes an incrcascd dcformed geometry of the uppcr crustal competent layer. As a result, the postrift stress-induced vertical motions increase when the synrift lithospheric necking level is deeper. Using the argument that stress-induced lithospheric deflections cause simificant tiltine of sedimentarv strata. Karner er al. [I9931 argued that parillel seismic reflectors are indicative for eustatic sea level changes. O w results demonstrate that the angle of tilting is so extremely low (less than 1") that such tilting is hardly observable on seismic sections. Therefore parallelism of seismic reflectors is not necessarily an indication of a eustatic control on onlap and offlar, configurations. After reversing the sign of the horizontal intraplate stress, our model predicts tlexural uplift in the center of a rift basin in response to tension. The upward bending of the plate causes additional tensile fiber stresses in the center of the rifi. These additional Lensile stresses augment the far-field tensile stress and therefore will facilitate further extensional faulting in the rift. As a consequence, our model predicts tbat rifting will localize in preexisting rift grabens. For the elastic stress-induced differential subsidence model, a decrease in tensional slress has the samc cffcct as an increase in compressiorral stress. As shown by our results, an increase in compressional stress leads to uplift of a basin flank. Therefore the release of tensional stress when rifting ceases is a mechanism for flank uplift. Evidence from various passive margins demonstrates that erosion of a rift shoulder causcs uplift of the nearby sedimentary basin, inducing stratigraphic offlap during the early phases of the postrift VAN BALEN ET AL.: A NEW MODEL FOR STRESS-INDUCED SUBSIDENCE evolution of a rifted basin [Van Balen et al., 19951. This can he explained by flexural compensation for erosion on a flank uplift [Van Balen et al., 19951, which amplifies the stressinduced uplift and the resulting relative sea level change. Fmally, application of the presenled equation for fault moments is not restricted to faulted elastic plates. In fact, this cquation can be used to study the effects of any steep lateral change of the geometry of a competent layer on in-plane stress-induced deflections. Such steep morphology may, for example, exist at ocean-continent boundaries. 7. Conclusions Our new model for intraplate stress-induced subsidence takes into account flexure of a multilayered continental lithosphere and its structural configuration prior to the change in the in-plane stress regime. The preexisting deformation influences the response of a thin plate to a change in in-plane force (stress times plate thickness) considerably [Timoshenko and Woinowsb-Krieger, 19591. Because the preexisting deformation of a rifted sedimentary basin has a concave geometry, the response of a basin to compression is subsidence in the basin center and upbft of the tlanks.The preexisting gwmehy is not properly taken into account in published thin-plate models. The largest part of the preexisting deformation in rifted sedimentary basins is caused by faults in the upper crustal competent layer. The offset of the midplane of the upper crustal competent layer due to a fault causes an additional moment at the position of the fault. This moment causes suhbidence on the downthrown sidr of the fault and uplift of the footwall. The amount of differential uplift depends on the vertical throw of the fault, the magnitude of the in-plane stress, the thickness of the upper crustal competent layer. and the degree of flexural coupling between the upper crustal and upper mantle competent layers. As shown in section 4, uplift and subsidence of several hundred meters are predicted for conservative values of these parameters in a basinal setting with multiple faults. Therefore the total stress-induced deflection, or relative sea level changes, of rifted basins can be of the order of 1 km in the case of a l u g e in-plane stress level change (up to 300 MPa), large fault throws (several kilometers), and a strong upper crust in combination with a relatively weak subcrustal layer. In-plane stress level changes caused by, for instance, lithospheric plate reorganizations [e.g., Cloetingh et al., 19891 can therefore have a substantial impact on relative sea level change in a rifted basin. As the lithospheric plates form an interlocking system, plate boundary reorganizations may cause a global change in in-plane stress level, and this may induce worldwide onlap and offlap events, interpreted as custatic sca level changes [e.g., Haq et al., 19871. Examples frnm the southern North Sea basin, Pannonian basin, and the Norwegian margin demonstrate that faults do, indeed, play a key role in the pattern of observed stressinduced differential subsidence of rifted sedimentary basins during their postrift evolution. The results of our new model confirm the original hypothesis of Cloetingh et al. [1985]: an increase in the level of compressive far-field in-plane stress causes flank uplift and basin center subsidence. Acknowledgments. T h e authors want to thank Fred Beekman, Ritske Huismans, Henk Kwi, Tore Skar, and Marlies ter V w r d e for helpful discussions. Ronald van Balen was sponsored by Norsk Hydro as. and Yuri Podladchikov was sponsored by the,"Universitair Stimulerings Fonds" of the Vrije Uniuersiteit. Sergei Medvedev, Fritz Schlunegger, and P.A. Ziegler have provided useful comments. This is Netherlands Research Schwl of Sedimentary Geology publication 980301. References Ahorner, L., Present-day stress field and seismotectonic block movements along major faulr zones in central Europe, Tecronophydcr, 29.233-249.1975. Bassi, G., C.E. Keen, and P. Potter, Conmsting styles of riftinp: Mcdels and examvles from the eastern Canadian margin, T@etonier,12, 639655,1993. Beekman, F., J.M. Bull, S. Clcetingh, and R.A. Scrutton, Cnlstal fault reactivation facilitating lirhospheric luldin#buckling in the central Indian Ocean, in Modern Developmenrr in Structural Inferpretotion. Validorion and Modelling, edited by P.G. Buchanan and D.A. Nieuwland, Geol. Soc. Spec. Publ., Y9, 251263,1996. Bereeat. F.. Stnss fields in the Eumoean vlatform at t h e time of ~frica-Euraiia Lollision, Tectonics, 6.99-132,1987. Blyrtad, P.,H. Brckc, R.B. Faerseth, B.T. h n , I. Skoeseid. . and B. T0ludbakLen. Structural elements of the Nawegian continental shelf, part 11, the Norwegian Sea region, Bull. 6, 45 pp., Now. Petrol. Dir., Stavanger, Norway, 1995. Bois. C., Initiation and evolution of thc OligoMiocene rifc basins of southwestern Europe, contribution of deep seismic reflection profiling, Tedonophysic.~,226,227-252,1993. B m , J., and C. Beaumont, A physical explanation of the relationship between flank uplifrs and the breakup unconformity at rifted continental margins, Geology, 17, 760-764, 19898. Braun J., and C. Beaumont, Contrasting stylea of lithoshperic extension: Implications for differences between the Basin and Ranee province and rifted continental margins,Tn Exrensional Tecr0nic.v and Stratigmphy of the Norrh Atlantic Marghr, edited by A.I. Tankard and H.R. Balkwill, AAPG Mem, 46. 53-80, 1989b. Bungum H., A. Alsaker, L.B. Kvamme, and R.A. Hansen, Seismicity and seismatecronics of Norway and nearby continental shelf areas J. Geophys. Res., 96,2249-2265, 1991. Bumv, E.B., and M. Diament, The effective elastic thickness (T.1 of contincntd lithosvhere: What does it really mean? J. Geophys: Res., 100, 3905-3924, 1995. Burov, E.B., and M. Diament, Isostasy, equivalent elastic thickness. and inelastic rheoloev -, of continents and oceans, Geology, 24, 419422, 1996. Caldwell, J.G., and D.L. Tureotte, Dependence of the thickness of the elastic oceanic lithosphere onage, 3. Geophys. Rer.,84,7572-7576, 1979. Caner, N.L., and M.C. Tsenn, Flow P w n i e s of continental lithosuhere, Teetonopkysicr, 136, 2763,1987. CMry, J., F. Lucazeau, M. Daignikres, and J.P. Vilotte, Large uplift at rift flanks a genetic link ~ ~ with lirho~phericrigidity?, Earlh Planet Sci. Lelt.. 112,195-211, 1992. Clur~in&~, S., Inlraplak slresacs: A ,new rlerrlrnt in basin analyses, in New Per$ppeeriver in &?sin Analy.vis,edited by K. Kleinspehn and C. Paolu, PP. 305-330, Springer-Verlag, New York, 1988. Clcetingh, S., and E. Bwov, Thermomechanical srruccure of European eontinencal lithosphere: Constraints from rheological profiles and EET estimates, Geuphvr J. Im., 124, 695~723,1996. Clcetingh, S., and R. Wortel, Regional smss field of the Indian plate, Geophyc Res. Len., 12, 7780.1985. Cloetingh, S., H. McQueen, and K. Lambeck, On a rectonic mechanism for regional sea level variations, Eorrh Planet. Sci. Len., 75, 157-166, 1985. Cloetingh, S., A.J. Tankard, H.J. Welsink, and W.A.M. Jerkin$ Vail'scoastol onlapchvlts and their correlation with tectonic events. offshore eastern Canada, in Ertenrional ~?rrinicrand Srrorigrophy of the Nonh Atlantic Margins, edited by A.J. Tankard and H.R. Balkwill, AAPGMem, 46,-283-293, 1989. Clcetlngh, S., J.D. van Wees, P.A. van der Beek, and (i. Spadini, Role of pre-rift rheology in kinematics of extensional basin f a t i o n : Constraints from thermomechanical models of Mediterranean and intracratonic basins, Mar. Per. Geol., 12, 793-808, 1995. ~ ~ ~ ~ 954 VAN BALEN ET AL.: A NEW MODEL FOR STRESS-INDUCED SUBSIDENCE Implications f a sequence stratigraphy, Earth Planet. Sci. Left, 134, 527-544, 1995. Van Bale", R.T., L. Lenkey, P. IIoru6th and S.A.P.L. Clcetingh, Twodimensional modelling of strati~aphyand fluid flow in the Pannonian Basin, in Exrenrion In Collirion, edited by B. Durand, E HorvAth and M. Sbtrane, Geol. Soc. Spec. Publ., in press, 1998. Van den Berg, M.W., Neotectonics of the Roer Valley tih ryrtem Style and rate of crustal deformation inferred from syn-tectonic sedimentauon, Geol. Mijnbouw, 73, 143-156, 1994. Van Wees, I.D.. and S. Clcetingh. 3D flexure and inuaplate compression in the N a h Sea Basin, Tecronophysio,~,266,343-359, 1996. vallwee*, 1.n. S. cllletingh, md G. ne vincenre, The mle of pre-existing weak zones in basin evol~tion:Constraints Worn ZD finite element and 3D flexure modelling in Modem Developments in slmclurai Interpretation. Validation and Modelling, edited by P.G. Buchanan and D.A. Nieuwiand. Gaol ~~~, -~~~ Soc Spec. Pub!., 99,297-320.19%. Watts, A.B., G.D. Kamer, and M. Steckler, Lithospheric flexure and the evolution of sedimentary basin* Pkilor. Tmnr. R. Soc. London, Ser. A., 305,249-281.1982, Weissel, J.K.,and G.D. Kamer, Flexural uplifl of riIl narks dm lo olcclmilicnl unloading of lhr lithosphere duing extension, I. Geophvs. REL, 94,13919-13950, 1989. Wortel, M.I.R.. M.J.N. Remkes, R. Govers. S.A.P.I.. Clo~ringh,and P.T. Meijer, Dynamics of the lithosphere and the intraplate stressfield, Phrlor Tmnc R. Soc London, Ser. A, 337, 111-126.1991. Ziegler, P.A., Geological Atlas of Western and Central Europe, pp. 239, Geol. Sce. Publ., Bath, ~ n ~ l a n 19%. d, Zegler P A . S. Cloetingh, and .ID. van Weer, Dynamics of intra-plate compressional deformadon. the A l ~ i n eForeland and other S. A. P. L. Cloetin&!h Y. Y. Podladehikov. and R. T. Van Balen, Facvlteit der Aardwetenrhappen, Vrije Univeniteit, De Bociclaan 1085. 1081 HV Amsterdam the ~~~~~, Netherlands. (e-mail: cloeting@geo.vu.nl, podl@geo.vu.nl; balr@geo.vu.nl) Zoback, M . D . , et d., Upper crustal strength infened from stress measurements to 6 km deoth KTB Nature. 633 r~~~ in the ~ borehole. ~~~~~~~~~, ~ ~.M. ~~, 635, 1993. (Received Oetober30, 1996; revired Anti1 28.1998: accepted September 2,1998.) ~ ~ ~~~ ~~ ~ ~~~~ ~ ~~~ Zoback, M.L., First- and second-arder patterns of stress in the lithosphere: The World Stress Map Project, I. Geopky,~.Res, 97, 1170311728, 1992. Zoback, M.L., and World Stress Map Team OLobd pltenls UC tectonic sVes$ Nature, 341, 291-298, 1989. Zuber, M., and E.M. Pamentier, Lithaspheric necking: A dynamic model for "fl morphology, Earth Plnn~t.Sci. 12tr. 77.371-381,1986. ~~ ~~ ~~~~