Slab gap versus Slab window
One of the interesting discussions surrounding the initiation of the Pacific - North American plate coupling is what happens when an active spreading center (the East Pacific Rise in this case ) reaches a subduction zone. Fig 1 shows the three views through time and this page will discuss these various ideas.
Initially it was thought that spreading centers would continue to spread as they were subducted beneath continents. However, the whole concept of spreading centers is that hot asthenosphere upwells to the surface, cools ( usually by seawater ) and passively spreads sideways to accommodate new asthenosphere. If the spreading center were beneath a continent there would be little removal of heat due to the thermal blanket effect of the crustal mass above. Since accretion at the spreading center has stopped, then the spreading itself is effectively shut down.
A view brought forward by Dickinson and Snyder ( 1979 ) suggests that as the spreading center approaches North America and stalls, the slab breaks at some point to form an area of no slab beneath the continent or a slab window. This slab window continued to grow as the Mendocino and Rivera triple junctions propagated north and south respectively. The exact aerial extent of the window through time is a function of the plate vectors of the three plates involved ( NA-PAC-Farallon ) but it is usually simplified to be triangular ( as in fig 1 )
This process neatly explains the apparent magmatic arc shut down as the San Andreas developed. At the time of initiation ( ~28Ma ) the arc was active again after its previous dormancy during the Laramide orogeny but began to switch off in good agreement with the proposed slab window widening. A synthesis of this is shown in Fig 2.
Dickinson and Snyder ( 1979 ) also linked the slab break away to upwelling asthenosphere which may have heated the base of the lithosphere ( Dickinson 1997 ) and caused ( or at least aided ) increased extension in the Basin and Range. What this hypothesis fails to address is the thermal state of the subducting plate as it reaches the trench.
This was looked at in Severinghaus and Atwater ( 1990 ) where they concentrated on the thermal properties of the subducting plate and its variation along the North American plate boundary. The central argument is that as the small microplates, Monterray and Arguello approached and slowed near the subduction zone, their downgoing slabs were getting progressively younger and warmer. Obviously if a slab is very young when it is subducted it takes a proportionally short amount of time to heat back up to asthenospheric values as is observed at the present day Juan de Fuca subduction slab. It is argued that instead of a slab window developing west to east ( while also propagating N-S ) there is a slab gap that develops from the east which then also prorogates north and south with the Mendocino and Rivera triple junctions. This is because at a certain time after subduction the slab would become indistinguishable from the asthenosphere surrounding it.
Severinghaus and Atwater ( 1990 ) used the parameter S to get some handle on the variation in thermal geometry through time ( 50 Ma to present ). Figure 3 shows this geometry with the shaded area being S values of 2 and below.
The authors suggest that an S value of 10 is an approximate assimilation point for short/medium length slabs but this does not change the overall geometry of the slab gap. It can be seen from figure 3 that the northern edge of the gap is well defined due to the large thermal contrast across the Mendocino fracture zone. This is not the case at the southern edge where small thermal contrasts, stalled slab pieces and less pronounced clear edges are further complicated by incomplete plate reconstructions due to possible overriding of the North American plate ( Page and Brocher, 1993 ). The exact boundary is therefore ill defined.
An indirect implication of the this work shows that the Laramide flat slab needed in many hypotheses to drive continental deformation could work from a purely thermal standpoint.
Overall, the slab gap is only a modified version of the slab window and both can be used to explain the cessation of arc volcanism in a N-S axis but it would be interesting to know if there were cessation from east to west ( gap ) or visa versa ( window ). Unfortunately as the arc was effectively a line ( not even a particularly coherent line ) this would be difficult to establish. The theory that mantle upwelling follows the establishment of a slab window/gap and this in turn had some effect on Western US extension is uncertain without further work ( on slab dips, spatial geometrys, rheology, and density structure ) as detailed in Sonder and Jones ( 1999 ).