The measurements of radiogenic heat production suggest that heat flow over the thinned continental crust of the Galicia Bank Margin should be higher than it is on the thinned continental (or oceanic) crust of the Iberia Basin, given that their lithospheric ages and crustal thicknesses are the same. To test this prediction, heat-flow measurements from Leg 149 (Sawyer, Whitmarsh, Klaus, et al., 1995) are compared with previously published values for Galicia Bank (Louden et al., 1991). The heat flow at Sites 897, 898, and 900 is recalculated from the original data using linear least-squares fits to values of temperature vs. Bullard depth (Fig. 4, Table 2). Bullard depths (RS) are calculated using the linear least-squares fit to thermal conductivity (ks) vs. subseafloor depth (z) for all data from Leg 149 (Fig. 5), using the relationship (Louden and Wright, 1989):
As is typical for most such determinations of heat flow, there is a large uncertainty produced by the large variation in ks and few measurements of temperature.
The resulting heat-flow values for Leg 149 (Table 2) are shown in Figure 6, as a function of distance along the Leg 149 transect. Before comparing these values with measurements across Galicia Bank, we assess the possibility that focusing and dispersal of heat flow through contrasts between the low-conductivity postrift sediment and higher conductivity basement may have disturbed the conductive heat flow measured at the seafloor. This process is modeled in Figure 6 using a finite element method (C. Jaupart, pers. comm., 1989). We assume a constant basal heat flux of 50 mW/m2 and constant basement conductivity, kb = 2.5 W/m·K. The conductivity of the sediment is assumed to following the relationship, ks = 2.25- e-0.43z, which gives values of 1.25 W/m·K at the seafloor, 1.48 W/m·K at z = 600 m (consistent with the observed conductivities in Fig. 5), and 1.83 W/m·K at maximum sediment thicknesses of ~2 km (Fig. 6). Sub-bottom topography for the two layers is constrained by the three multichannel profiles shown in Figure 2, using assumed mean sediment (vs) and water (vw) velocities of 2.0 and 1.507 km/s, respectively. Mean values of vs are consistent with core measurements (Sawyer, Whitmarsh, Klaus, et al., 1995), as well as with the measured depths to basement at Sites 897, 898, and 900. The result of this modeling shows that the topographic correction can be large but only very near Site 901, where the basement rises nearly to the seafloor. Seafloor measurements of heat flow at the other sites, therefore, should be representative of local crustal values.
A comparison between observed heat flow across Goban Spur and the Leg 149 sites in the Iberia Basin is given in Figure 7B. This comparison shows that similar values for both margins are observed in the oceanic/serpentinite ridge domain, but landward values on transitional crust at Site 900 are 16-23 mW/m2 higher than values landward of the serpentinite ridge on Galicia Bank. These variations in heat flow are interpreted in terms of variations in heat production using the pure-shear, depth-independent model of continental extension (McKenzie, 1978). In this model, heat flow and subsidence are parameterized as a function of
, which defines the fractional amount of initial vertical thinning of both lithosphere and crust. Following Voorhoeve and Houseman (1988), the effects of radiogenic heat production are included within the model, by addition of the quantity, H h [1 - h/2L]/, which linearly reduces as a function of
-1, where H = rate of radiogenic heating (µW/m3), h = thickness of radiogenic crust (km), and L = lithospheric thickness (km).
Theoretical values of heat flow vs. ln , which result from this model (Louden et al., 1991), are shown in Figure 7A for a variety of ages. These show the competition between the increase in heat flow caused by the thinning of the oceanic lithosphere vs. the reduction in heat flow caused by the thinning of the more radiogenic continental crust. By 130 Ma, an age representative of the Iberia Margin (Shipboard Scientific Party, 1993), the expected variation in heat flow across the margin is controlled primarily by the contribution from crustal heat production. The residual mantle effect of <5 mW/m2 is not very sensitive to reasonable uncertainties in its rifting age.
In Figure 7B, we show the effect of changes in radiogenic heating in the models at an age of 130 Ma. Values of
= 5 are assumed for the Leg 149 sites, which is consistent with observed crustal thicknesses of ~6 km (Whitmarsh et al, 1990). It is clear that the higher heat flow at Site 900 can be explained only by high values of radiogenic heat production (3-6
µW/m3) and certainly not the low value of 0.21 µW/m3
as measured, even if one increases the assumed thickness of the radiogenic layer (h). Larger topographic effects produced by off-profile variations in the basement are not consistent with site survey reflection profiles. Although the Leg 149 heat-flow measurements are not conclusive because of their large uncertainties, surface measurements of heat flow along the profile (Sibuet et al., 1994) are consistent with a region of elevated heat flow at this site. Other possibilities are that the measured heat production at Site 900 is not representative of the complete crustal section or that some other process has disturbed the surface flux. Both these possibilities require further study.
