DISCUSSION 

Composition and Grain-Size Variations

The compositional variations associated with grain size in the Leg 149 sand samples (Fig. 7, Fig. 8) could be attributed to a number of factors. First, this relationship could be a function of different provenance for finer vs. coarser grained turbidites: the turbidites fed from more proximal canyons may be coarser, whereas more distant canyons would supply finer turbidites to the same site, because of the longer transport distance. Second, the source region for the coarser and finer turbidites may be the same geographic region of the shelf, with grain-size variations a function of sediment derivation from different shelf depositional systems; for example, the medium, well-rounded quartzose sand might be derived from beach to shelf environments, whereas the finer, more angular, feldspathic sand may better tie to fluvial-deltaic environments. Compositional and textural differences among samples would therefore be linked to transport history. Third, given a uniform source of both coarse and fine turbidites, an inherent difference in size fractions could be associated with source-rock textural variations (e.g., relative crystal sizes of feldspar and quartz in source rocks) or with relative abrasion rates during transport (e.g., preferential breakdown of feldspar into a finer fraction because of the presence of cleavage). A combination of transport history and depositional environment controls on sand composition seems most plausible for the Iberia Margin turbidites. Results from a study of Pennsylvanian sandstones in Colorado by Kairo et al. (1993) support this interpretation, in that they link higher sand maturity in the foreshore area, as compared with offshore and alluvial regions, to the abrasion and mechanical breakdown of feldspar. Furthermore, studies of modern sand indicate that the framework mineralogy of beach sand is more mature than associated fluvial sand (Sedimentation Seminar, 1988).

Sand Provenance

The Leg 149 sites are located in an embayment of the Iberia Abyssal Plain, between the Galicia Bank and the Estremadura Spur (Fig. 1). General bathymetry around the sites suggests that terrigenous components within the cored successions were most likely derived from the Iberia Margin, or perhaps the Galicia Bank. Bathymetry of the Iberian continental slope to the east of the Leg 149 transect shows a series of submarine canyons and valleys that potentially funnel(ed) sandy sediment out to the Iberia Abyssal Plain, at least back through the Pliocene, and perhaps throughout the Tertiary.

For the most part, the monomineralic and lithic components of the Leg 149 sand and sandstone are consistent with derivation from felsic plutonic and metamorphic terrains, similar to the Hercynian basement rocks (Variscan Belt) of the northwestern Iberian Peninsula (Galicia-Tras-Os-Montes Zone; Azor et al., 1994; Capdevila and Mougenot, 1988). Precambrian to Paleozoic felsic plutonic and metamorphic rocks crop out over much of the drainage basins of the major rivers (i.e., Rio Duero, Rio Minho, and Rio Mondego) that presently empty onto the Iberian shelf north of latitude 40°N (Fig. 12). As discussed below, similar basement rocks also comprise the Galicia Bank, and so in terms of composition both the Galicia Bank and Iberian mainland (massif) are plausible sources of sand. However, there is no evidence to suggest that the Galicia Bank was emergent during the early Cenozoic (Winterer et al., 1988), and therefore it was not likely a major source of Cenozoic sand at Leg 149 sites. Thus, the Iberian mainland most likely supplied the terrigenous component within the Cenozoic sequences cored on Leg 149.

However, the Mesozoic picture of sand provenance is less clear. The shelf and onshore regions to the east-southeast of the Leg 149 transect were dissected into a series of horsts and grabens during rift-related extension in the late Triassic/Early Jurassic, forming, in part, the Lusitanian Basin (Wilson and Exton, 1979; Montenat et al., 1988); faults within this region were rejuvenated during the Late Jurassic and Early Cretaceous (Wilson and Exton, 1979; Montenat et al., 1988). This extension and dissection may have served to trap mainland-derived sand, limiting its offshore distribution. During this same time frame, sand was shed to the west off the emergent basement highlands of the Galicia Bank and deposited by turbidity currents at Leg 103 sites (Fig. 1; Winterer et al., 1988; Johnson, 1988). This sand is quartzofeldspathic with a minor lithic component, and its composition suggests a basement source composed of felsic plutonic and metamorphic rocks (Winterer et al., 1988; Johnson, 1988) which is not only similar to dredged samples from the Galicia Bank (Capdevila and Mougenot, 1988), but also to the rocks of the Iberian mainland (Winterer et al., 1988).

Sandstone detrital modes calculated by Johnson (1988) for 31 very fine sandstones of Cretaceous age from Leg 103 are similar to those of Cenozoic samples from Leg 149; Johnson's data cluster about a mean value of 52:36:12 %QFL, which is almost the same as the Leg 149 Pliocene mean in Figure 9 (52:38:10 %QFL). In general, Mesozoic samples from the Leg 149 sites (Fig. 11) contain significantly less feldspar than the Leg 103 Mesozoic samples. This difference is not a function of variation in techniques between the two studies, in that Johnson (1988) employed petrographic and staining methods identical to those of our study. The similarity of the Leg 103 Cretaceous samples and Leg 149 Pliocene samples could be a function of grain size, as discussed previously, or may be related to the similarity of sand source regions. Perhaps the ratio of metamorphic/plutonic exposures on the Galicia Banks "paleoland" was similar to that of northwestern Iberia during the Pliocene. Of course, other provenance factors such as climate and associated weathering intensity may have also played a role. Although Cretaceous sands derived from the Galicia Bank are dissimilar to the few Leg 149 Cretaceous sandstones analyzed for this study, given the likely spatial variations in basement lithology across the Galicia Bank, a northern source for the Leg 149 Mesozoic sand cannot be ruled out.

Temporal Changes in Composition and Passive Margin Evolution

A number of studies of modern and ancient sand have examined passive margin provenance, and of these, most focus on ancient sequences and emphasize the high quartz content of these sediments (e.g., Schwab, 1975; Dickinson and Valloni, 1980; Ingersoll and Suczek, 1979). Only one study (i.e., Marsaglia, 1991) explores the relationship of sand provenance to the rift-to-drift evolution of passive margins. The Leg 149 data presented in Table 2 are therefore noteworthy because they represent a semicontinuous Cenozoic to Mesozoic sedimentary sequence that accumulated during a rift-to-drift transition along the Iberian passive margin.

The series of temporal trends outlined in Figure 9, Figure 10, and Figure 11 and discussed above cannot be rigorously modeled, but can be linked to the tectonic evolution of this margin. The QFL trend depicted by the arrow in Figure 11 is a crude evolutionary path that more accurately consists of a major jump from Jurassic lithoclastic sand to more quartzofeldspathic post-Jurassic sand. Similar evolutionary pathways have been proposed for sand detrital modes from other tectonic settings, some of which are thought to represent unroofing trends (e.g., Dickinson, 1985; Marsaglia et al., 1992). Progressive variation within the lithic populations of the Leg 149 fine sand also suggests an unroofing trend for this margin. Mesozoic sandstones are sedimenticlastic, whereas the Cenozoic sandstones largely contain metamorphic clasts and exhibit progressively smaller lithic proportions, to the point where they are quartzofeldspathic (see data in Table 2). This pattern could grossly reflect the progressive removal of first sedimentary, and then metamorphic, cover of plutonic basement rocks. In terms of the provenance fields of Dickinson and Suczek (1979), this trend would correspond to passing from recycled orogen provenance (synrift to postrift phases?) to continental block provenance (postrift to drift phases?). Obviously, more data are needed to clarify the nature of this trend and to determine if such a trend is specific to this segment of the Iberia Margin or is more representative of passive margin sequences as a whole.

Compositional trends within the very fine sand samples may represent a regional tectonic overprint unrelated to the passive margin setting. This tectonic overprint is a Tertiary compressional phase that caused inversion of basin structures and probably uplifted onshore source terrains. The shift to more lithic-rich (Fig. 9) and quartzose (Fig. 10) sand compositions in the Oligocene and Miocene could therefore be linked to uplift of metamorphic terrains and/or drainage rearrangement, and the subsequent shift toward more quartzofeldspathic sand compositions in the Pliocene and Pleistocene could be linked to the removal of metamorphic cover and exposure of underlying plutonic sequences.

The small (n = 50) data set presented here suggests that models for passive margin sequences based on sand detrital modes are plausible and that passive margin sequences are perhaps more complex than previously thought. Of course, in our brief discussion we simplify the story in that we emphasize the large-scale tectonic controls on the Leg 149 sand detrital modes and downplay other provenance factors. For example, we do not address the possible effects of the major climatic changes that also occurred during the evolution of this margin. More subtle variations in feldspar composition could be linked to these climatic changes, particularly with respect to the intensity of weathering, but given the small sample number, we thought this somewhat imprudent.

NEXT