More than abyssal sediments, pillow basalt, differentiated gabbro and depleted peridotite sheeted dyke complexes have long been a primary identifier for oceanic lithosphere preserved in ophiolites. That assumption has recently been questioned (Robinson, P.T. et al. 2008. The significance of sheeted dyke complexes in ophiolites. GSA Today, v. 18 (November 2008), p. 4-10). Ian Gass first discovered units made up solely of dykes that intrude one another with no intervening screens of other host rocks in the Troodos ophiolite of Cyprus in 1968. Sheeted dyke complexes became widely regarded as characteristic of extensional, sea-floor spreading environments connected to basaltic magma chambers, each increment of extension being filled with magma. They have also been imaged in eroded walls of ocean fracture systems and cut through by ocean drill cores, supporting this notion. In fact, many ophiolites are devoid of sheeted complexes, despite having all the other components of mafic-ultramafic lithosphere. Robinson et al. argue that sheeted dykes only form where spreading rates and magma supply are balanced, as expected at true constructive plate margins but far less likely at other extensional zones associated with plate tectonics, such as those in back-arc basins above subduction zones. Even at true spreading centres that generate new ocean floor magma supply may not balance extension, for instance where spreading rates are slow. Moreover, a great many ophiolites show geochemical affinities that are more akin to supra-subduction magmatic processes than those that produce mid-ocean ridge basalt.

Plate tectonics in time and space

Seismic tomography becomes increasingly revealing as the capacity of supercomputers grows. On top of that, more sophisticated software allows present-day mantle cross sections to be reverse modelled with surface plate motions to reconstruct an idea of mantle dynamics back to Mesozoic times. Geophysicists at the California Institute of Technology give a taste of the possibilities from the subduction history of North America (Liu, L. et al. 2008. Reconstructing Farallon plate subduction beneath North America back to the Late Cretaceous. Science, v. 322, p. 934-938). Investigating 3-D evolution is the key to connecting rigid plate tectonics and fluid convection that has long been postulated but remains obscure. However, while reasonable reconstructions of global plate motions are possible using sea-floor magnetic stripes that go back to the Cretaceous, seismic tomography only images the mantle’s present structure. So it might seem that generating a 3-D ‘geomovie’ is more of an expensive illusion than a model of past realities.

The logic behind the modelling is that today’s mantle temperature structure – that is what tomograms show – stems from past plate activity. For instance, a deep cold, slab-like anomaly dipping eastward beneath eastern North America can reasonably be inferred to be a relic of the Farallon Plate, which formerly constituted floor of the eastern Pacific. That plate was subducted beneath the west edge of the continent until around 40 Ma, when the East Pacific Rise that had driven it was subducted. The present thermal structure shown by the tomogram has, in a sense, ‘faded’ as a result of thermal relaxation of the original anomalies by heat diffusion. Choosing geologically reasonable starting conditions for long-term evolution of a mantle segment enables iterative forward modelling to try and achieve the present set-up. While there is an element of circularity in this logic, such a dynamic model has a predictive aspect; i.e. as cold, dense material in the mantle sinks it tends to pull the surface downwards, allowing marine flooding of continental interiors. During the Late Cretaceous this did happen spectacularly in North America, and Liu et al’s model shows this. Yet sea level also rose globally at the time, thereby amplifying the inundation. Although geeomodellers will be excited by Liu et al’s developments, it is modelling and even the simplest of models is acutely sensitive to the chosen starting conditions, as meteorologists with vastly more real data at hand have discovered again and again.

See also: Steinburger, B. 2008. Reconstructing Earth history in three dimensions. Science, v. 322, p. 866-868