Extensional Tectonic Models 
There are largely two main models when considering extensional tectonics, both offering quite different proposals. The McKenzie model incorporates pure shear, while the Wernicke  model is based upon a simple shear regime.

The McKenzie model

The McKenzie model McKenzie (1978) though old and has had some slight modifications over time still holds and remains very relevant. McKenzie basically proposes that within a given vertical column the continental lithosphere will be evenly stretched, so that there is an instantaneous initial stretch associated with subsidence.

 figure 7.  ( The McKenzie model)

The stretching factor 'B' (in McKenzie's initial subsidence equation) is the length of deformation that occurs determining the amount of lithosphere that can be vertically thinned during an extensional episode.

As an effect of the lithospheric stretching, isostatic compensation occurs causing a mantle upwelling of the asthenosphere.
Once the thermal anomaly beneath the stretched lithosphere begins cooling, it is this increase in density weighing down the lithosphere from below causing another episode of subsidence, called thermal subsidence. This second subsidence occurs exponentially and much more slowly than the rapid initial subsidence, as it is trying to maintain the isostatic equilibrium while the upwelled asthenosphere cools. Lithospheric stretching should have ceased as thermal subsidence occurs.
McKenzie's model quite simply explains the extension of the lithosphere as stretched basins that would sag in the middle forming the rift with a small amount of faulting. Figure 8 shows the process of McKenzie's model.

 figure 8.

McKenzie's model relies on time-dependent stretching on which the assumptions are based, and is basically depth-independent. The time dependent assumption is that stretching and subsidence occurs instantaneously, before cooling of the asthenosphere and thermal subsidence occurs.

Slow or fast stretching is affected by the amount of strain built up in the lithosphere, and how much heat can be lost in the lithosphere by conduction. The slower the stretching rates, the longer it takes for initial subsidence to occur, thus extending into the period of time where the upwelled asthenosphere begins to cool and thermal subsidence occurs. Because of this large time frame, the period of initial subsidence is increased and thus causes a decrease in the amount of time that thermal subsidence occurs.

McKenzie's model suggests that a weakly stretched basin fits an instantaneous time frame if the stretching factor (B) is less than 1.5. If the stretching factor is greater than 2.5, this means that there is a large amount of stretching taking place and hence may not be instantaneous. In general, for stretching to be instantaneous, the model fits a period of stretching of around 10Ma, but anywhere of up to 30Ma is still acceptable.

McKenzie's model is one of pure shear in that the homogeneous lithosphere is stretched uniformly to form a symmetrical basin with faulting in the brittle crust to accommodate stretching. However there is another pure shear model by Rowley and Sahagian (1986) that suggests that in an episode of stretching, the stretching factor may vary between the crust and lithosphere causing a physical decoupling of the two layers.

If the crust stretches more than the lithosphere, this would extend the amount of initial subsidence and decrease the amount of thermal subsidence that would take place. If the lithospheric mantle stretching exceeded the amount of crustal stretching (figure 9), initial subsidence would at worst be hindered by a thermal upwelling and subsidence.

figure 9. 

The Wernicke Model
The Wernicke (1985) model is based upon a simple shear regime which means the basin is stretched asymmetrically by a large scale detachment fault extending from the upper crust to the lower lithosphere and even asthenosphere, causing extension.

In this model (figure 10) crustal lengthening is far from the area of thermal upwelling and so the area underneath it does not accommodate much thermal subsidence. 

figure 10. 

The 'proximal' area of the upper crust will undergo an initial subsidence but no lithospheric extension occurs. Lithospheric behaviour at the 'distal' end of the shear zone may undergo a slight uplift. Buck et al (1988) has found that the footwall of the shear zone experiences minor uplift from lateral heat of the asthenosphere, that eventually sinks back to normal height after thermal cooling. Total subsidence between the simple and pure shear models are relatively the same, depending on the amount of crustal thinning.

Wernicke (1985) has only described the simple shear model based on observations of a basin and range area in western USA in qualitative terms only. Though the simple shear model is another accepted possibility of causing rift basins, quantitative tests are yet to yield results applicable to other areas.

Both Wernicke and McKenzie models are quite valid solutions to extensional tectonics with McKenzie's pure shear actually being applicable to basins worldwide. It is likely that there are cases where both simple shear and pure shear act together to aid in continental rifting where a large amount of extension occurs.

This is known as the flexural cantilever model (Kusznir et al (1987)) shown by figure 11,  as  mentioned in the tectonic structures section. Firstly flexure occurs as a  response to an overloading weight in the crust such as a mountain chain  or  sedimentary basin, causing symmetric flexure. The lithosphere being a viscous fluid layer under the brittle/ductile  crust  kind of bends like an elastic beam in the middle to accommodate this load,  and to maintain elastic equilibrium. In the case of tectonic extension,  flexure would be asymmetric due to a downward pressure from the hanging wall  of a fault.

The flexural cantilever model takes into account that with  tectonic extension there would be several large faults all producing flexure, and that there would be areas of overlapping deformation. Thus  taking all this into account, the model uses simple shear accounting for  deformation in the upper crust, with the lower crust and mantle  lithosphere  undergo pure shear deformation. This model accounts for planar faults in the  upper crust, but can include listric faults and that for both faults,  the  footwall is uplifted as the hanging wall subsides.

Zones of deformation coincide when faults are closely spaced, as the footwall of one fault  becomes the hanging wall of another. As lengthening occurs, the fault blocks  rotate to attain gentler dips- as mentioned previously this is known as  the  domino model of faulting. Over time erosion  takes place, shortening  the footwall crests and eventually the thermal margin between the brittle  and  ductile crusts relaxes, bringing boundary back to a pre-faulted state.

figure 11.