A Dynamic Model of Rifting Between Galicia Bank and Flemish Cap During Opening of the North Atlantic Ocean.

Transcription

1 A Dynamic Model of Rifting Between Galicia Bank and Flemish Cap During Opening of the North Atlantic Ocean Dennis L. Harry 1 and Shelly Grandell 2 1 Department of Geosciences, Colorado State University, Fort Collins, CO, (dharry@cnr.colostate.edu) 2 Department of Geology, Adams State College, Alamosa, CO

2 Abstract A finite element model is used to simulate Late Jurassic through Early Cretaceous rifting between the Flemish Cap and Galicia Bank continental margins. The model results shows that variations in the thickness of the continental crust on these margins at wavelengths greater than about 75 km can be explained as a consequence of the interaction of two preexisting weaknesses in the lithosphere. A weakness in the crust, attributed to structural fabrics in the Variscan front, controls the location of crustal extension during the early stages of rifting in the model. This results in formation of a broad rift basin similar to the Galicia Interior Basin. A deep seated weakness located 110 km further west, attributed to the thick crust beneath the central Variscan orogen, controls the location of mantle necking. Extension in this region is initially diffuse, but accelerates and becomes more focused with time. 13 m.y. after rifting begins, the locus of crustal extension shifts from the region of pre-weakened crust into the region of pre-weakened mantle. This marks the end of subsidence in the Galicia Interior Basin and the onset of subsidence in the Flemish Cap and Galicia Bank marginal basins. Extension in these areas continues for another 12 m.y. before continental breakup. The asthenosphere does not ascend to depths shallow enough for decompression melting to begin until < 5 m.y. before the onset of seafloor spreading. The model predicts that all late stage syn-rift magmatism during this period is limited to within 45 km of the rift axis, and production of melt thicknesses greater than 2 km is restricted to within 35 km of the rift axis. Mantle potential temperatures of 1250!C to 1275!C, ca. 30!C cooler than normal, result in km thick oceanic crust at the time of breakup, in general agreement with the 2-4 km thick crust observed adjacent to these margins. 1

3 Introduction This paper describes the results of a finite element numerical simulation of Late Mesozoic and Early Cenozoic rifting on the Galicia Bank and Flemish Cap segments of the Iberia and Newfoundland conjugate continental margins. The modeling goal is to match the long-wavelength (> 75 km) crustal thickness variations on the margins, the timing and duration of rifting, the magmatic history of the margins, timing and location of regional shifts in extensional tectonism, and the regional subsidence history. This paper focuses on a conjugate transect between seismic profiles ISE-1 on the Iberian margin and SCREECH-1 on the Newfoundland margin (Fig. 1) (Funck et al., 2003; Henning et al., 2004; Hopper et al., 2004)]. The modeling technique used in this paper is based on the STRCH95 two-dimensional finite element modeling program, which is a descendant of the STRCH program developed by Dunbar (1988). STRCH95 is used to model extensional processes on a lithosphere scale. The program allows for complex initial conditions to describe the pre-extension structure of the crust and rheology and temperature of the lithosphere. The simulations discussed in this paper focus on how such pre-existing features might have controlled the pattern of extension and shifts in the loci of extension with time on the Galicia Bank/Flemish Cap margins, and how such extension is accommodated without producing significant amounts of syn-rift magmatism. In the following sections we describe the geological and geophysical observations that constrain the numerical models discussed in this paper and the modeling targets. We then describe the modeling algorithm. Finally, we describe the results and discuss their implications within the context of Galicia Bank/Flemish Cap rifting in particular and rifting on non-volcanic margins in general. 2

4 Model Constraints The history and structure of the Galicia Bank and Flemish Cap continental margins is described in detail elsewhere in this volume. Here, we briefly summarize those elements of rift history and margin structure that contribute directly to constraining the numerical simulations presented in this paper. Timing - Extension on the Iberian Peninsula and Newfoundland occurred in several distinct phases. The first phase occurred in the Newfoundland and Galicia Interior Basins during the Late Triassic-Early Jurassic periods (Fig. 1) (Ziegler, 1989; Murillas et al., 1990; Tucholke et al., 1989; Foster and Robinson, 1993; Tankard and Welsink, 1987). This was followed by a period of tectonic quiescence lasting until Late Jurassic time. The second phase of extension began in the Late Jurassic Epoch and lasted until late Valanginian time, when the locus of extension shifted seaward of the Flemish Cap and Galicia Bank (Murillas et al., 1990; Tucholke et al., 1989). This led to the third phase of extension, a period of lithospheric stretching and crustal thinning west of Galicia Bank and east of Flemish Cap in Hauterivian through Aptian time. This last phase of extension continued until continental breakup and formation of new oceanic crust between the time of seafloor spreading anomalies M3 and M0 (Srivastava et al., 2000; Hopper et al., 2004; Funck et al., 2003; Minshull et al., 2001). In this paper, we focus on the Late Jurassic to Early Cretaceous phases of extension. These phases involved an initial stage of extension in the Galicia Interior Basin, a later stage of extension seaward of the unextended continental crust on Flemish Cap and the slightly extended crust on Galicia Bank, and, finally, the onset of seafloor spreading. The total duration of rifting (excluding the Triassic through Early Jurassic phase in the interior basins) lasted approximately 25 m.y., beginning in Late Tithonian time (ca. 143 Ma) and ending in the early Aptian age (ca. 118 m.y.) (Hopper et al., 2004; Minshull et al., 2001). The first ca. 12 m.y. of extension 3

5 (Tithonian-Valanginian) involved the Galicia Interior Basin and Newfoundland Basin. The last ca. 13 m.y. of extension was primarily focused seaward of the Galicia Bank and Flemish Cap. Postrift crustal structure A reconstruction of the Newfoundland and Galicia continental margins at the time of anomaly M0, near the onset of seafloor spreading, is shown in Fig. 2. Unextended portions of the continental crust on Iberia and Newfoundland are km thick (Banda, 1988; Diaz et al., 1993; Mendes Victor et al., 1980). The pre-rift crystalline crust on the Iberia margin thins westward to less than 10 km-thick beneath the 100 km-wide Galicia Interior Basin (Pérez-Gussinyé et al., 2003), increases to approximately 20 km beneath the Galicia Bank (Gonzalez et al., 1999; Pérez-Gussinyé et al., 2003), and then thins progressively westward over a 100 km-wide region between the eastern edge of the Galicia Bank and the rift axis, where the oldest oceanic crust is emplaced. The pre-rift crust reaches a minimum thickness of 2-5 km at its seaward limit (Gonzalez et al., 1999; Chian et al., 1999). On the Newfoundland margin, the crust is approximately 30 km thick beneath the relatively unextended Flemish Cap, and thins progressively eastward to ca. 1.5 km over a ca. 80 km-wide region between the eastern edge of the Flemish Cap and the rift axis (Funck et al., 2003). Mantle exposed in a ca. 10 km-wide region on the seafloor west of Galicia Bank is thought to have been exhumed by late-stage low angle faulting during the last 1-2 m.y. of extension (Fuegenschuh et al., 1998; Pérez-Gussinyé et al., 2003; Pickup et al., 1996; Reston, 1996; Sibuet, 1992). Extension rate - Extension rates on the Newfoundland and Iberia margins vary locally at the scale of a few tens of km, probably reflecting different timing between movements on individual fault systems. In a regional sense, extension rates are best constrained on the Iberia Abyssal Plain. Here, regional extension rates are estimated to be ca. 10 mm/yr during rifting (Minshull et al., 2001) and mm/yr at the onset of seafloor spreading (Whitmarsh et al., 4

6 2001a; Russell and Whitmarsh, 2003). Assuming symmetric spreading, this corresponds to a whole-extension rate of 20 mm/yr during rifting. This is consistent with area balancing of the cross section in Fig. 2. Restoration of the cross section to a uniform crust thickness of 32 km requires an extension rate of 20 mm/yr to produce the observed amount of stretching in 25 m.y. Pre-rift lithospheric structure, rheology, and thermal constraints Breakup between the Galicia Bank and Newfoundland occurred within a region roughly coincident with the trend of the Late Paleozoic Variscan orogen in this region (Capdevila and Mougenot, 1988; Ziegler, 1989). Focusing of extensional deformation in this region was most likely a result of orogenic weakening of the lithosphere caused by the increased thickness of the crust beneath the orogen (Fig. 3). The thickness of the crust in the central Variscan orogen near the rift axis is estimated to have been ca km thick prior to extension (Pérez-Gussinyé et al., 2003), 2-5 km thicker than the crust in the unextended parts of the Iberia Peninsula and Newfoundland. Assuming the thermal parameters estimated by Tejero and Ruiz (2002) for the crust in the Duero Basin in the Iberian Peninsula and a simple two-layer rheological model following empirical ductile flow laws for quartz-diorite in the crust and wet dunite in the mantle (Table 1), this difference in crustal thickness would weaken the lithosphere in the central Variscan orogen by up to 15% in comparison to the surrounding regions (Fig. 3). The strength of the nominal lithosphere predicted by this rheological model is 7.0x10 12 N m -1 (Fig. 3a), which is moderately weaker than the 8x10 12 N m -1 lithospheric strength estimated by Tejero and Ruiz (2002) for the central Iberian Peninsula and is consistent with the heat flow and thermal structure of the Iberian lithosphere estimated by Fernandez et al. (1998). Magmatic history Both the Flemish Cap and Galicia Bank margins are flanked by unusually thin oceanic crust, ranging from km thick (Hopper et al., 2004; Whitmarsh et al., 5

7 1996), suggesting a low volume of melt production immediately following continental breakup. On the Galicia margin the oceanic crust increases to normal thickness (approximately 7 km) within 15 km seaward of the oldest oceanic crust (Whitmarsh et al., 1996). At the estimated spreading half-rate of mm/yr (Whitmarsh et al., 2001a), this suggests that the thermal regime of the mantle rapidly evolved to that of a typical mid-ocean ridge system within ca m.y. after the onset of seafloor spreading. The melt production history after the onset of seafloor spreading on the Newfoundland margin is more complicated. Here, the oldest oceanic crust is ca. 3-4 km thick and thins seaward to < 1.3 km over a ca. 50 km wide region (Hopper et al., 2004), indicating a magma starved environment during the first ca m.y. of seafloor spreading. Near-normal oceanic crust (> 6 km thick) located 3 km seaward of the thinnest oceanic crust on the Flemish Cap margin (Funck et al., 2003; Hopper et al., 2004) indicates that the transition from magma starved seafloor spreading to normal seafloor spreading occurred in about m.y. Recovery of late syn-rift volcanic and plutonic rocks seaward of the oldest oceanic crust on both the Newfoundland and Iberian margins suggests that the last stages of rifting were accompanied by minor amounts for magmatism, just prior to the onset of seafloor spreading [Leg 210 Shipboard Scientific Party, 2003; Whitmarsh et al., 2001b; Manatschal and Bernoulli, 1999). Regardless of the complexities of late rift-stage and early spreading phase melt production, there is broad consensus that for most of the rift history the margins were amagmatic, or at least involved very little melt production. Volcanism that occurred prior to the onset of seafloor spreading appears to have been limited in volume and space, being confined to within 20 km or so of the location where oceanic crust is first produced. 6

8 Modeling Targets On the basis of the above discussion, our numerical simulations of rifting between Galicia Bank and Flemish Cap are constrained by 1) the pre-rift structure of the crust, including a presumably weak and heterogeneous crust in the Variscan orogen embedded between older strong Precambrian crust of Canada and the Iberian peninsula, 2) the pre-rift crustal thickness, taken to be 32 km by comparison to the thickness of the modern unextended crust beneath Iberia and Canada, 3) radiogenic heat production of the crust, taken to be similar to that of basement rocks exposed in Iberia (Table 1). The models seek to reproduce 1) the post-rift crustal thickness variation across the margins at length scales greater than ca. 75 km, particularly moderately extended crust beneath the Galicia Interior Basin, unextended crust beneath the Flemish Cap and slightly extended crust beneath Galicia Bank, and highly extended crust beneath the marginal rift basins, 2) the lack of magmatism during rifting, 3) the duration of rifting, 4) the amount of magmatism required to produce thin oceanic crust at the time of breakup, and 5) the subsidence history of the margin, including Late Jurassic through late Valanginian subsidence in the Galicia Interior Basin and late Valanginian through early Aptian subsidence seaward of Flemish Cap and Galicia Bank. The finite element model formulation used in these simulations does not explicitly account for formation of detachment faults, and so does not reproduce exhumation of the peridotite ridge west of Galicia Bank. Modeling Method The finite element modeling program used in this research, STRCH95, is the most recent version of the FORTRAN 90 computer program STRCH developed by Dunbar (1988). The STRCH95 algorithm successively solves the two-dimensional heat conduction/advection/generation equation 7

9 Equation 1: & T ' c % # $ K# T " A & t and the Navier-Stokes equation for flow in a visco-plastic lithosphere Equation 2: ' du % ' F ( # P ")# 2 u, & t where u is the particle velocity, T is temperature, ' is the density of the rock, c is the specific heat, K is thermal conductivity, A is the volumetric heat production rate, F is body force per unit mass, P is pressure, and ) is viscosity. The physical properties ', c, A, and K for each element in the finite element mesh are specified by the user. Viscosity is defined by the rheology of each element, and is pressure, temperature, and strain rate dependant. Viscosity is determined by empirical relations derived from experimental rock deformation studies. Initially, a uniform strain rate,*!, is assumed throughout the model. An effective viscosity, ) eff, for each element is determined from the estimated stress + and strain rate: Equation 3: ) + %! * eff. The stress in the above equation is taken to be the weaker of either a ductile (power law creep) or plastic (a linear pressure-dependant yield criterion) rheology: Equation 4 (ductile): 1!. + % * / 0 A, - 1/ n e Q / nrt c Equation 5 (plastic): + % S 0 " Bz where z is depth and R is the Universal Gas Constant. Once the effective viscosity is determined, the Navier-Stokes equation is solved. This results in an updated estimate of the strain rates in the model, which are used to revise the estimated effective viscosity of each element. The process is repeated until the strain rate estimates on successive iterations agree to 8

10 within a convergence criterion prescribed by the user. The strain rates are used to determine velocities at each node in the mesh, and the geometry of the model mesh is updated by stepping forward the locations of each node in the mesh a finite amount of time (time steps of 10,000 years duration are typical). The heat equation is then solved to update the thermal structure, using the new mesh geometry as a boundary condition and the thermal state of the model at the previous time step as an initial condition. The new model temperatures and the strain rates from the past time step are used to calculate a revised estimate of the element viscosities to begin the next time step. Accuracy of the model solution is assessed empirically by comparing model results obtained with coarse and fine meshes and coarse and fine time step sizes. For a given family of models, a coarse mesh and time step size is initially selected. The mesh size and time step size are then halved and the model re-run. Refinement of mesh size and time step size continues until successive models produce the same result. If the model becomes highly deformed (individual elements develop large aspect ratios), a singularity develops in the linear system of equations that define the finite element problem being solved. When this occurs, the model is remeshed. During remeshing, domain boundaries that define different materials used in the model (e.g., the model edges and the contacts between different rock types) are preserved. Within each domain, new elements are defined that have an aspect ratio as close to unity as possible. Temperatures, pressures, and strain rate information within the model are preserved during the remeshing process. Model Description A basic assumption behind the modeling strategy used here is that pre-existing strength heterogeneities in the lithosphere determine the location and distribution of loci of extension 9

11 during rifting (e.g., Dunbar and Sawyer, 1989). Weaknesses in the upper and middle crust (crustal weaknesses) tend to produce short-lived extensional provinces during the early stages of rifting, whereas deep seated weaknesses (mantle weaknesses) control the location where lithospheric necking develops and, hence, the location of eventual continental breakup. If these weaknesses are offset from one another, interior rift basins (controlled by the crustal weakness) often form landward of the deep offshore marginal basin (controlled by the mantle weakness) (e.g., Harry and Sawyer, 1992). The models presented here include both of these types of weakness in the lithosphere (Figs. 3 and 4). Mantle weaknesses are simulated as regions of thickened crust, which is attributed to the root beneath the Variscan orogen prior to rifting. This is represented with a simple triangular shaped region of increased crustal thickness. Key variables describing the mantle weakness are the width of the region of thickened crust and the amount that the crust is thickened. A crustal weakness is simulated by a region in which both the plastic and ductile yield strengths are decreased relative to the surrounding crust. In the plastic regime, this involves a reduction of S 0 (equation 5). For ductile behavior, both the temperature dependant terms (Q c /n) and temperature independent terms (A -1/n ) are reduced by similar amounts (equation 4). The crustal weakness may be attributed to lithologic variations in the upper and middle crust or to pre-existing faulting near the Variscan front. Key variables describing this weakness in the model are the amount by which the yield strength is decreased, the width and thickness of the weakened crust, and the position of the crustal weakness relative to the mantle weakness. All of these variables were iteratively adjusted to produce a model that best fits the geometry and rift history of the Flemish Cap and Galicia Bank margins. Other modeling parameters, including heat production, specific heat, thermal conductivity, extension rate, and nominal thickness of the 10

12 crust (outside the Variscan orogen) were held fixed (Table 1). The criterion used to determine when rifting ends in the models is the requirement that the crust thin to ca. 2 km at the rift axis, which is similar to the thinnest continental crust on the Galicia Bank and Flemish Cap margins (Chian et al., 1999; Funck et al., 2003; Gonzalez et al., 1999). Results Results of a typical model are shown in Figs The model simulates in general the present structure of the Galicia Bank/Flemish Cap conjugate margins. In particular, the model predicts extreme crustal thinning beneath conjugate offshore rift basins that form during the late stages of rifting, formation of an interior basin on the eastern shelf during the early stages of rifting, and a region of less extended crust between the offshore and interior rift basins (Fig. 6). Most of the models examined in this study produced these basic attributes. Exceptions were models in which either the crustal or mantle weakness was very small (typically involving a reduction of yield strength in the crustal weakness of less than 5% or a change in crustal thickness in the mantle weakness of less than 2 km). In those models (not shown here), the dominant weakness controls rifting from the outset, resulting in formation of a nearly symmetric rift basin centered on the dominant weakness. All other models that were examined, involving changes in yield strength in the crustal weakness ranging from 5-30% and changes in crustal thickness in the mantle weakness ranging from 2-5 km, produced results that were generally similar to that shown in Figs However, changes in the relative locations and magnitudes of the two weaknesses produces a wide variation in the widths and depths of the interior rift basin and deep offshore basins, the duration of extension in the interior basin, the time elapsed before breakup, the amount and pattern of crustal thinning, and the presence or absence of a region of relatively unextended crust between the deep offshore and interior rift basins. As discussed 11

13 below, changes in these parameters also have a significant affect on the magmatic history of the model. The model shown in Figs. 5-7 is based on the variables that provide the best fit to observations on the Galicia Bank/Flemish Cap conjugate margins (Table 2). During the first 10 m.y., extension results in rapid thinning of the crust in the region encompassing the crustal weakness (Figs. 5 and 6a). Crustal thinning also occurs above the mantle weakness, but at a much slower rate. Tectonic subsidence in the region of weakened crust begins almost immediately after the onset of extension, marking the first stages of formation of the interior rift basin (Fig. 6b). The region above the mantle weakness, where the deep offshore basin ultimately forms, does not subside below sea level until approximately 5 m.y. later. The two basins are separated by a structural high that undergoes moderate crustal thinning (from 32 km to ca. 20 km) during the first 5 m.y. of extension, and only a minor amount of thinning (ca. 3 km) afterward. This pattern of crustal thinning and subsidence can be understood in terms of the distribution of extension, illustrated by the model strain rates (Fig. 7). Extension in the crust is broadly distributed across the region encompassing the crust and mantle weaknesses during the first 10 m.y., but is concentrated most strongly in the area of the crustal weakness (Fig. 7). Strain in the mantle is more evenly distributed across the pre-weakened region, and is coupled to deformation in the middle and upper crust by narrow regions of relatively high strain. Offsets in the initially vertical element boundaries show that the high strain region in the lower crust behaves as a subhorizontal detachment during this time (Fig. 5). Because extension in the mantle is initially distributed over a wide region, asthenospheric upwelling is broad and the rate of lithospheric thinning is relatively slow. As extension progresses, strain in the mantle begins to focus strongly beneath the mantle weakness. After 10 m.y., this deformation begins to propagate 12

14 upward, creating two regions of focused extension in the upper crust (the interior rift basin and marginal basin) that are separated by a region of relatively slow strain forming the structural high between the two basins. At this time, extension and subsidence is waning in the interior rift basin and accelerating in the marginal basin. By 15 m.y., extension throughout the lithosphere begins to focus exclusively in the offshore region (Fig. 7). Crustal thinning and subsidence ceases in the interior rift basin and on the structural high (Fig. 6). Because extension is more focused, lithospheric necking accelerates (Figs. 5 and 7). The lithosphere beneath the rift axis thins to approximately half its initial thickness 20 m.y. after the onset of extension, and proceeds rapidly to breakup in the next 5 m.y. Discussion The thickness of the crust in the model generally agrees with the long wavelength (> 75 km) crustal thickness variations on the Flemish Cap and Galicia Bank margins along the SCREECH-1 and ISE-1 seismic profiles (Fig. 6a). The duration of the main extensional episode in the interior rift basin is approximately 13 m.y., in agreement with the late Tithonian through Valanginian phase of rifting in the Galicia Interior Basin (Murillas et al., 1990; Tucholke et al., 1989). The model predicts that deep water depths (> ca. 1 km) did not become established in the offshore rift basin seaward of Galicia Bank until ca. 12 m.y. after the onset of extension. Breakup is predicted in early Aptian time (ca. 118 Ma), 25 m.y. after the onset of extension, in agreement with the estimated onset of seafloor spreading at around anomaly M0 time (Srivastava et al., 2000; Hopper et al., 2004; Funck et al., 2003; Minshull et al., 2001). The magmatic consequences of rifting in the model are assessed in terms of the timing, volume, and spatial distribution of melt produced by adiabatic decompression melting in the upwelling asthenosphere (McKenzie and Bickle, 1988). We neglect conductive heat loss from 13

15 the rising mantle, which may be significant in slow spreading systems such as the Newfoundland-Iberia rift (Bown and White, 1995; Pederson and Ro, 1992; Whitmarsh et al., 2001a; Whitmarsh et al., 2001b), so our melt calculations should be considered to be upper bounds. Key parameters in the melt calculation are the shape of the ascending asthenospheric diaper, which is dictated by the pattern of lithospheric thinning in the model, and the mantle potential temperature, which is a variable in the model. A series of models were developed using mantle potential temperatures ranging from 1250!C to 1300!C as a basal boundary condition (Fig. 4). Other parameters in these models were identical to those in Tables 1 and 2. Each of these models produced patterns of extension and subsidence very similar to that shown in Figs The only significant differences were the predicted melt production history. Maximum melt thicknesses range from 3.1 to 6.5 km at the rift axis at the time of breakup (Fig. 8). The models using a 1250!C and 1275!C mantle potential temperature produce a melt thickness of 3.1 km and 4.7 km at the time of breakup, respectively. This is in good agreement with the 3-4 km thick oceanic crust adjacent to the Newfoundland and Galicia Bank margins (Hopper et al., 2004; Whitmarsh et al., 1996), suggesting a mantle potential temperature during rifting somewhere between these two values. Since syn-rift conductive cooling of the asthenosphere was neglected, this should be considered a minimum estimate of the asthenospheric potential temperature prior to and during rifting. It seems unlikely, then, that the mantle potential temperature was more than ca. 30!C cooler than the 1280!C global average calculated by McKenzie and Bickle (1988). In any case, the presence of normal 7 km-thick oceanic crust within about 15 km of the oldest oceanic crust on the Galicia Bank margin (Whitmarsh et al., 1996) requires that mantle potential temperatures reached ca. 1280!C within about m.y. after breakup, assuming a mm/yr seafloor spreading half rate (Whitmarsh et al., 2001a; 14

16 Russell and Whitmarsh, 2003). We deem it most likely that the asthenosphere potential temperature was close to the global average prior to rifting. Moderate conductive cooling (less than ca. 30!C) of the upwelling asthenospheric diaper would be sufficient to account for the unusually thin oceanic crust at the time of breakup, and the near-normal mantle potential temperature would explain the rapid transition to normal midocean ridge melt production soon after seafloor spreading began. This scenario is generally consistent with the spatial and temporal distribution of magma production observed on the margins. In the 1250!C and 1275!C potential temperature models, melt production does not begin until 2.5 m.y. or 5 m.y. before the onset of seafloor spreading, respectively. No melt is predicted further than 45 km landward of the rift axis in either model, and melt thicknesses in excess of 2 km is restricted to within 35 km of the rift axis. Minor late-stage syn-rift magmatic episodes have been documented further south on the Iberian margin (Whitmarsh et al., 2001b; Manatschal and Bernoulli, 1999) and Newfoundland margin (Tucholke et al., 2004), and similar evidence in the form of small MORBlike volcanic edifices and gabbroic intrusions emplaced on and within exhumed subcontinental mantle has been reported on the Adrian fossil nonvolcanic rifted margin exposed in the Swiss Alps (Manatschal and Bernoulli, 1999; Muntener et al., 2000; Muntener and Piccardo, 2003). We surmise that a minor amount of late stage magmatism, within the range of volumes predicted by the models here, is likely to be typical of nonvolcanic rifted margins such as Newfoundland- Iberia. Summary Finite element models simulating rifting between Galicia Bank and Flemish Cap invoke pre-existing weaknesses in the lithosphere to account for the shift in the locus of extension from 15

17 the Galicia Interior Basin during the first half of the rifting episode to the deep offshore rift basin during the second half of the rifting episode. Two forms of weaknesses were examined: an upper mantle weakness, created by excess crustal thickening over a 120 km wide region that is attributed to the central part of the pre-rift Variscan orogen, and a crustal weakness created by reducing the yield strength of the upper and middle crust in an 80 km wide region that is attributed to preexisting structural fabrics in the eastern Variscan front. The model results were found to be robust over excess crust thicknesses ranging from 2-5 km and upper and middle crust weakening ranging from 5 to 30%. Under these circumstances, all models produced an early interior rift basin followed by a shift in extensional deformation to a deep offshore basin where continental breakup ultimately occurs. Details of the width and depth of the basins, the time at which deformation shifts from the interior rift basin to the offshore basin, and the pattern of crustal thickness at the time of breakup are determined by the relative positions, shape, and magnitudes of the weaknesses imposed in the model. Model parameters in Tables 1 and 2 produced the general features of the last 25 m.y. of rifting on the Flemish Cap and Galicia Bank continental margins, including 13 m.y. of extension in the Galicia Interior Basin, 12 m.y. of extension in the Galicia Bank and Newfoundland Cap marginal basins, and formation of moderately extended crust on Galicia Bank. The model geometry at the time of breakup approximates structural features of greater than about 75 km wavelength on these margins. The nonvolcanic nature of rifting, ultimate production of thin (2-4 km thick) oceanic crust, and rapid transition to generation of normal thickness (7 km) oceanic crust within m.y. after breakup requires a mantle potential temperature of 1250!C to 1275!C, roughly 30!C cooler than the global average at midocean ridges. This modest amount of cooling is attributed 16

18 to conductive cooling of the ascending asthenosphere during the late stages of rifting. At these mantle temperatures, the model predicts production of km new oceanic crust at the time of breakup. Syn-rift magmatism is limited to the last < 5 m.y. of extension and to within 45 km of the locus of continental breakup. Melt thicknesses of greater than 2 km are restricted to within 35 km of the locus of continental breakup. Acknowledgements This research was supported by National Science Foundation grant OPP We thank the organizers of the InterMargins Modeling of Extensional Deformation of the Lithosphere workshop held in Pontresina, Switzerland July, 2004 for providing a venue for discussions that led to this modeling study, and the many participants of the workshop who provided data used to constrain the models. Shelly Grandell received support from the Ronald E. McNair Postbaccalaurate Achievement Program for this research. References Banda, E., Crustal parameters in the Iberian Peninsula. Physics of the Earth and Planetary Interiors 51: Bown, J.W., and White, R.S., Effect of finite extension rate on melt generation at rifted continental margins. Journal of Geophysical Research 100: 18,011-18,030. Capdevila, R., and Mougenot, D., Pre-Mesozoic basement of the western Iberian continental margin and its place in the Variscan belt. Proceedings of the Ocean Drilling Program, Initial Reports 103: Chian, D., Louden, K.E., Minshull, T.A., and Whitmarsh, R.B., Deep structure of the ocean-continent transition in the southern Iberia Abyssal Plain from seismic refraction 17

19 profiles: Ocean Drilling Program (Legs 149 and 173) transect. Journal of Geophysical Research 104: Diaz, C.J., Gallart, J., Cordoba, D., Senos, L., Matias, L., Surinach, E., Hirn, A., and Maguire, P., A deep seismic sounding investigation of lithospheric heterogeneity and anisotropy beneath the Iberian peninsula. Tectonophysics 221: Dunbar, J.A., Kinematics and Dynamics of Continental Breakup [Ph.D. thesis]. Austin, TX, University of Texas. Dunbar, J.A., and Sawyer, D.S., How preexisting weaknesses control the style of continental breakup. Journal of Geophysical Research 94: Fernandez, M., Marzan, I., Correia, A., and Ramalho, E., Heat flow, heat production, and lithospheric thermal regime in the Iberian Peninsula. Tectonophysics 291: Foster, D.G., and Robinson, A.G., Geological history of the Flemish Pass basin, offshore Newfoundland. AAPG Bull. 77: Fuegenschuh, B., Frotzheim, N., and Boillot, G., Cooling history of granulite samples from the ocean-continent transition of the Galicia margin: Implications for rifting. Terra Nova. 10: Funck, T., Hopper, J.R., Larsen, H.C., Louden, K.E., Tucholke, B.E., and Holbrook, W.S.., Crustal structure of the ocean-continent transition at Flemish Cap: Seismic refraction results. Journal of Geophysical Research. 108: doi: /2003jb Gonzalez, A., Cordoba, D., and Vales, D., Seismic crustal structure of Galicia continental margin, NW Iberian Peninsula. Geophysical Research Letters 26: Harry, D.L., and Sawyer, D.S., A dynamic model of extension in the Baltimore Canyon trough region. Tectonics 11:

20 Henning, A.T., Sawyer, D.S., and Templeton, D.C., Exhumed upper mantle within the ocean-continent transition on the northern West Iberia margin: Evidence from prestack depth migration and total tectonic subsidence analyses. Journal of Geophysical Research 109: doi: /2003jb Hopper, J.R., Funck, T., Tucholke, B.E., Larsen, H.C., Holbrook, W.S., Louden, K.E., Shillington, D., and Lau, H., Continental breakup and the onset of ultraslow seafloor spreading off Flemish Cap on the Newfoundland rifted margin. Geology 32: Manatschal, G., and Bernoulli, D., Architecture and tectonic evolution of nonvolcanic margins: Present-day Galicia and ancient Adria. Tectonics 18: McKenzie, D., and Bickle, M.J., The volume and composition of melt generated by extension of the lithosphere. Journal of Petrology 29: Mendes Victor, L.A., Hirn, A., and Veinante, J.L., A seismic section across the Tagus Valley, Portugal: possible evolution of the crust. Annales Geophysicae 36: Minshull, T.A., Dean, S.M., White, R.S., and Whitmarsh, R.B., Anomalous melt production after continental break-up in the southern Iberia Abyssal Plain. In R.C.L. Wilson, R.B. Whitmarsh, B. Taylor, B., and N. Froitzheim, eds., Non-volcanic Rifting of Continental Margins: A Comparison of Evidence From Land and Sea, pp Geological Society of London. Muntener, O., Hermann, J., and Trommsdorff, V., Cooling history and exhumation of lower-crustal granulite and upper mantle (Malenco, eastern Central Alps). Journal of Petrology 41:

21 Muntener, O., and Piccardo, G.B., Melt migration in ophiolitic peridotites: The message from Alpine-Apennine peridotites and implications for embryonic ocean basins. In Y. Dilek and P.T. Robinson, eds., Ophiolites in Earth History, pp Geological Society of London. Murillas, J., Mougenot, D., Boillot, G., Comas, M.C., Banda, E., and Mauffret, A., Structure and evolution of the Galicia Interior Basin (Atlantic western Iberian continental margin). Tectonophysics 184: Pederson, T., and Ro, H.E., Finite duration extension and decompression melting. Earth and Planetary Science Letters 113: Pérez-Gussinyé, M., Ranero, C.R., Reston, T.J., and Sawyer, D., Mechanisms of extension at nonvolcanic margins: Evidence from the Galicia interior basin, west of Iberia. Journal of Geophysical Research 108: doi: /2001jb Pickup, S.L.B., Whitmarsh, R.B., Fowler, C.M.R., and Reston, T.J., Insight into the nature of the ocean-continent transition off West Iberia from a deep multichannel seismic reflection profile. Geology 24: Reston, T.J., The S reflector off Galicia: The seismic signature of a detachment fault. Geophysical Journal International 127: Russell, S.M., and Whitmarsh, R.B., Magmatism at the west Iberia non-volcanic rifted continental margin: evidence from analyses of magnetic anomalies. Geophysical Journal International 154:1-25. Sibuet, J.-C., New constraints on the formation of the non-volcanic continental Galicia- Flemish Cap conjugate margins. Journal of the Geological Society of London 149:

22 Srivastava, S.P., Sibuet, J.-C., Cande, S., Roest, W.R., and Reid, I.D., Magnetic evidence for slow seafloor spreading during the formation of the Newfoundland and Iberian margins. Earth and Planetary Science Letters 182: Tankard, A.J., and Welsink, H.J., Extensional tectonics and stratigraphy of Hibernia oil field, Grand Banks, Newfoundland. AAPG Bulletin 71: Tejero, R., and Ruiz, J., Thermal and mechanical structure of the central Iberian Peninsula lithosphere. Tectonophysics 350: Tucholke, B.E., Austin, J.A., and Uchupi, E., Crustal structure and rift-drift evolution of the Newfoundland Basin. In A.J. Tankard H.R. Balkwill, eds., Extensional Tectonics and Stratigraphy of the North Atlantic Margins, pp American Association of Petroleum Geologists. Tucholke, B.E., Sibuet, J.-C., Klaus, A., and Shipboard Science Party, Proceedings of the Ocean Drilling Program Initial Reports 210 [online]. Available from World Wide Web: Whitmarsh, R.B., Manatschal, G., and Minshull, T.A., 2001a. Evolution of magma-poor continental margins from rifting to seafloor spreading. Nature 413: Whitmarsh, R.B., Minshull, T.A., Russell, S.M., Dean, S.M., Louden, K.E., and Chian, D., 2001b. The role of syn-rift magmatism in the rift-to-drift evolution of the West Iberia continental margin: Geophysical Observations. In In R.C.L. Wilson, R.B. Whitmarsh, B. Taylor, B., and N. Froitzheim, eds., Non-volcanic Rifting of Continental Margins: A Comparison of Evidence From Land and Sea, pp Geological Society of London. 21

23 Whitmarsh, R.B., White, R.S., Horsefield, S.J., Sibuet, J.-C., Recq, M., and Louvel, V., The ocean-continent boundary off the western continental margin of Iberia: Crustal structure west of Galicia Bank. Journal of Geophysical Research 101:28,291-28,314. Ziegler, P.A., 1989, Evolution of the North Atlantic - An overview. In A.J. Tankard and H.R. Balkwill, eds., Extensional Tectonics and Stratigraphy of the North Atlantic Margins, AAPG Memoir 46, pp Tulsa, Oklahoma, American Association of Petroleum Geologists. 22

24 Figure Captions Figure 1. Location Map. A) Newfoundland margin. B) Iberia margin. Solid lines are locations of model transects. Dots are DSDP and ODP drillsites. Shaded regions are interior rift basins: GIB, Galicia Interior Basin; LB, Lusitania Basin; FCG, Flemish Cap Graben; FPB, Flemish Pass Basin; JDB, Jeanne d arc Basin; WB, Whale Basin; SWD, South Whale Basin; HS, Horse Shoe Basin; OB, Orphan Basin. Contour interval is 1000 m. Figure 2. Cross-section of the Newfoundland-Iberia rift at the time of anomaly M0 along the trend of seismic transects SCREECH-1 on the Newfoundland margin (Funck et al., 2003; Hopper et al., 2004) and ISE-1 on the Galicia Bank margin (Henning et al., 2004). After Funck et al. (2003). Figure 3. Yield strength envelopes for lithosphere based on rheological and thermal properties in Table 1 calculated at a strain rate of s -1. A) Lithosphere with nominal 32 kmthick crust. B) Lithosphere with 35 km-thick crust. C) Lithosphere with weakened upper crust. D) Steady state geotherm. See text for discussion. Figure 4. Finite element model construction. Numbers at top and sides indicate constant temperature and constant extension rate boundary conditions. Figure 5. Finite element mesh for model that best fits Galicia Bank/Flemish Cap rifting. Figure 6. A) Crustal thickness variation in the finite element model. Bold line indicates observed thickness in cross section of Fig. 2. B) Elevation predicted by the finite element model (does not include syn- and post-rift sediment loading).

25 Figure 7. Second invariant of the strain rate in the finite element model. Strain rates range from s -1 (dark blue) to s -1 (red). Figure 8. Melt production predicted by models using mantle potential temperatures of 1300!C, 1275!C, and 1250!C.

26 Table 1. Fixed model parameters. 1 Constraint Pre-Rift References Model Value Continental Crust Thickness km Perez-Gusenye et al. (2003) Oceanic Crust Thickness 3-4 km Tucholke et al. (2004) Lithosphere Thickness Fernandez et al. (1998) Surface Heat Production, A 0 3.1!Wm -2 Fernandez et al. (1998) Heat Decay Exponent, D 12 km Fernandez et al. (1998) Thermal Conductivity, K (crust) 2.5 W m -1 K -1 Tejero and Ruiz (2002) Thermal Conductivity, K (mantle) 3.4 W m -1 K -1 Tejero and Ruiz (2002) Specific Heat, Cp (crust) 875 J kg -1 K -1 Tejero and Ruiz (2002) Specific Heat, Cp (mantle) 1250 J kg -1 K -1 Tejero and Ruiz (2002) "hermal Expansion Coefficient, # (crust) 3.1x10-5 K -1 Tejero and Ruiz (2002) Thermal Expansion Coefficient$# (mantle) 3.1x10-5 K -1 Tejero and Ruiz (2002) Extension Rate 20 mm/yr Russell and Whitmarsh (2003) Brittle Yield Strength, S 0 60 MPa Bowling and Harry (2001) Slope of Brittle Failure Strenght, %$ 5 MPa/km Bowling and Harry (2001) Ductile Creep Coefficient, A (crust) 5x10-18 Pa -n s -1 Bowling and Harry (2001) Ductile Activation Energy, Qc (crust) 219 kj mol -1 Bowling and Harry (2001) Ductile Creep Exponent, n (crust) 2.4 Bowling and Harry (2001) Ductile Creep Coefficient, A (mantle) 4x10-25 Pa -n s -1 Bowling and Harry (2001) Ductile Activation Energy, Qc (mantle) 498 kj mol -1 Bowling and Harry (2001) Ductile Creep Exponent, n (mantle) 4.5 Bowling and Harry (2001) 1 Symbols are defined in text.

27 Table 2. Variable model parameters. Parameters Best Fit Model Value Mantle Weakness Width 120 km Mantle Weakness Amplitude 3 km Crustal Weakness Width 80 km Crustal Weakness Thickness (edges) 12 km Crustal Weakness Thickness (center) 26 km Strength Reduction in Crustal Weakness 25% Distance between Center of Weaknesses 110 km Temperature at Bottom of Lithosphere 1250!C, 1275!C, 1300!C

28 Fig. 1 - Harry & Grandell a) 44 N 16 W 14 W 12 W 10 W 8 W 6 W N GIB 40 N LB N b) 54 W 52 W 50 W 48 W 46 W 44 W 42 W 50 N OB N 46 N Newfoundland JDB FPB FCG Flemish Cap WB SWB 44 N HB CB N

29 Fig. 2 - Harry and Grandell West East Flemish Cap Rift Axis Serpentinized Mantle Galicia Bank Pre-rift crust Galicia Interior Basin Mantle Iberia Syn- and Post-rift Depth (km) Distance from rift axis (km)

30 Fig Harry and Grandell 0 a) b) c) d) Stress (MPa) Stress (MPa) Stress (MPa) Temperature ( C) Depth (km) Net Strength x 10 N-m 32 km thick crust Net Strength x 10 N-m 35 km thick crust Net Strength x 10 N-m Crustal weakness

31 Fig. 4 - Harry and Grandell T = 0 C Ux crust Crustal weakness Ux mantle Mantle weakness T = T m Isostatic Bouyancy Pressure

32 Fig. 5 - Harry and Grandell 0 m.y. Novia Scotia Variscan orogen Iberia Weak upper crust Crust 0 32 km Mantle 5 m.y. 122 km 10 m.y. 15 m.y. 20 m.y. 25 m.y. V.E. 5:1 500 km

33 a) Fig. 6 - Harry and Grandell Crust Thickness (km) m.y b) Elevation (km) m.y Distance from rift axis (km)

34 Fig. 7 - Harry and Grandell Novia Scotia Variscan orogen Weak upper crust Iberia 0 m.y km 5 m.y. 122 km 10 m.y. 15 m.y. 20 m.y. 25 m.y. V.E. 5:1 500 km

35 Fig. 8 - Harry and Grandell Melt Thickness (km) C, 25 m.y C, 25 m.y C, 25 m.y C, 20 m.y Distance From Rift Axis (km)