The intent of this lithospheric modeling experiment was to test the response of initial models with different lithospheric weaknesses to extension through different multiphase extension "paths." In this paper, we present only certain models that illustrate our conclusions. First, however, a description of the full "generic" model suite follows.
Each model was initially 800 km long and underwent 500 km of extension over the duration of the model run. These values are estimates of extension across the Newfoundland and Iberian Margins (Tett, 1993). (These values reflect our supposition that the crust underlying the Newfoundland Basin and the Iberia Abyssal Plain is of continental affinity.) Each model was run for 120 m.y., which was approximately the length of time from the beginning of the first rifting phase (about 230 Ma) to the initiation of seafloor spreading (about 110 Ma) on the Newfoundland-Iberia conjugate margins. All of the generic models were symmetrical.
It should be noted at this point that the 110 Ma date for continental breakup is based on an unconformity found at Deep Sea Drilling Program (DSDP) Site 398 at the Aptian/Albian boundary (Groupe Galice, 1979)—about 110 Ma. This unconformity may reflect the breakup on the nearby Galicia Margin to the north, however. This date was intended only to approximate the date of initial seafloor spreading on these margins and, as will be seen below, was not used in designing the more margin-specific model.
Although all models had the same total amount of extension over the same total duration of time, they had different instantaneous extension rates during that interval. Over its 120-Ma span, each model had one rifting rate for the first 25 Ma, a different rate for the next 45 Ma, and yet another rate for the final 50 Ma. These durations were chosen because they approximate the durations of the first (late Triassic) rifting phase, the resting phase, and the second (Late Jurassic to Early Cretaceous) rifting phase on the Newfoundland-Iberian Margins. Each sequence of extension rates for the three phases was called a rifting "path."
Five different rifting paths were used (Table 3, Fig. 2). The extension rate (and hence the amount of extension) for the middle phase was zero in paths 1, 2, 3, and 5; this represented the period in which there is thought to have been no divergent motion between North America and Iberia (about 215 to 160 Ma). The paths were intended to simulate different distributions of the total extension between the first and second rifting phases. Path 4 is a constant extension rate for the entire duration of rifting.
Three different initial models were tested, each of which contains a different type of lithospheric weakness. The weakness types were taken from Dunbar and Sawyer (1989).
The first initial model, "MW," contains a mantle weakness, formed by a section of crust 5 km thicker than the flanking "normal thickness" crust (Fig. 3). The crust in this model was represented entirely by a quartz diorite rheology. In the models considered here, the mantle and lower crust lie mainly in the ductile field, where yield strength varies inversely with temperature (Fig. 4). At identical temperatures, the quartz diorite rheology is much weaker than the dunite rheology. Hence, where a crustal welt replaced mantle dunite with crustal quartz diorite in a model, the vertically integrated strength of a column of lithosphere was lowered.
The second initial model, "CW," contains a crustal weakness (Fig. 3). Here, a section of quartz diorite crust was replaced with the wet granite rheology. The thickness of this weakness ranged from 10 to 20 km. The CW initial model is weakened by the same amount as model MW (Fig. 4).
The third initial model, "BS" (both symmetric), contains both a mantle and a crustal weakness, which are symmetrical about the same point (Fig. 3). The sizes of the weaknesses were the same as in the individual MW and CW models, although in the BS model the crustal weakness made up 25% to 50% of the crust, independently of the crustal thickness. The BS model is quite a bit weaker than the previous two initial models (Fig. 4).
Stretching the three different starting models using five different extension paths produced 15 models in total.