Extensional Tectonic Structures of the upper crust
Small extensional faults

Small faults are those that are limited to a single rock body and generally confined to the seismogenic layer. A seismogenic layer is a layer that produces earthquakes. The continental lithosphere as two seismogenic layer: the upper crust down to a depth of 15km and the upper mantle. Deformation surrounding the fault does not extend very far and is limited to a three dimensional ellipse around the fault plane. Deformation here is asymmetric, as the hanging wall displays antiformal rollover and the footwall has a synformal uplift. Another form of a small fault is a blind fault (figure1), and that is one which is enclosed within a rock volume, without penetrating any free surfaces.

figure 1. 

Things change when the fault is allowed to intersect a free surface (figure 2), as this effects the deformation of the hanging wall. Subsidence is caused at the top of the hanging wall that is exposed, as there is no holding support from the rock that would have normally been above it.

figure 2. 

Metamorphic Core Complexes

Metamorphic core complexes (figure 3)  are a result of small or large faults with an initially high angle that have been extremely rotated to form a domino-style series of fault blocks as found by Lister & Davis (1989). The steep seismically active large faults enter the aseismic layer (a layer beyond which seismic data has been recorded) where the border between a brittle/ductile crust exists. The fault detaches from the seismogenic layer at a shear zone enduring greater pressures and temperatures in this aseismic layer.At higher crustal levels, stretching occurs and isostatic uplift begins above the shear zone lifting the fault up from depth. As the footwall rises, it progressively tilts and domes at the surface. An initial fault angle of 60º should have rotated 10-15 º (Buck (1988)). What would be visible at the surface is a footwall containing the mylonitic detachment fault (mylonite=extremely deformed) overlying metamorphic rocks - the metamorphic core complex.
An increase in metamorphic grade should be seen moving from the top to the bottom of the gently sloping footwall. The mylonites of the footwalls should juxtapose the adjacent hangingwall in which non-mylonites should be visible.

figure 3. 

Fault linkage

Over time small faults grow and gradually intersect one another forming a network of faults that can act as one large fault (regardless if some of the faults are inactive), (Watterson (1986)). Physical linkages between the individual faults are known as accommodation or transfer zones where strain from one fault is laterally transferred from to another by a transform fault. Soft-linkage is another accepted mechanism where the joining of faults occurs by ductile strain (Larsen 1988, Morley et al. 1990, Roberts & Jackson 1991, Roberts et al. 1991, Stewart & Hancock 1991, Walsh & Watterson 1991, Yielding & Roberts 1992). The faults are physically linked by a fault bridge or relay ramp (figure 4) striking perpendicular to the dip of the whole basin.
These fault bridges would serve as a natural ramp for sediment dispersal, but over time these bridges will eventually break apart forming transform faults.

figure 4. 

Large extensional faults

Large faults are those that can penetrate through the seismogenic layer. Large faults are predominantly planar faults, such as normal faults, though listric faults are not uncommon in accompanying extension. Jackson & White (1989) and Roberts & Jackson (1991) recorded from their tests that strike lengths of active normal faults have not been found to extend past 25 km, and that these large faults could be made up from a number of fault segments.

Such models such as the flexural cantilever model by Kusznir et al (1987) are used to explain the flexural response of the crust in a seismic event causing a large normal fault. This will be explained in the tectonic stretching models section, but a simplified version of this model applies to this section pertaining to structures. When a basin is stretched normal faults may occur, and over time with ongoing stretching these fault planes may tilt (figure 5). A series of these faults are known as rotational normal faults and make up what is known as the domino model of faulting (the geometrically simplified flexural cantilever model) as shown by figure 6. In this model all fault planes have the same angle, moving towards gentler dips in a passive rotation. However, this model comes to a problem when looking at the basin margins, as these are giant footwalls that have no possible way of rotating or moving freely. This model also comes across another unexplainable problem when the fault blocks of the basin are of different sizes, which is why it is an extremely simplified version of the flexural cantilever model.

figure 5. 

                                                                                                                              figure 6. 

Gravity-driven extensional faults

These structures are less related to lithospheric lengthening, but moreso the translation of material in crustal stretching. Generally, this leads to a series of listric faults (faults that initially dip steeply, but gradually lessen until the dip becomes sub horizontal).
Rocks that are relatively unconsolidated or are weak will not have the same elasticity as consolidated rocks and will tend to slump, and fault under gravity.