Dynamic Topography: Elephant or Menagerie?
Often we will say of workers exploring different parts of a problem that they are feeling different parts of an elephant, a reference to the tale of the blind men describing an elephant by feeling different parts of it, each describing the elephant differently. The notion is that the totality of the elephant requires integrating the observations. But what if the blind men are all feeling different animals? You’d get a pretty confused looking animal if you combined all those descriptions. This might be the case with dynamic topography.
Of late, dynamic topography has become something of a touchstone that might explain a number of different phenomena. In a recent workshop GG got to sit in on, it was invoked for things from the unexpectedly shallow depth of really old seafloor, the elevation of southern Africa, renewed erosion in parts of the Appalachians, and erosion and deposition for short segments of a continental margin. (Unmentioned or at least unemphasized was the role of dynamic topography in explaining the drownings and emergences of continents that have produced the sedimentary record across cratons). If these are all facets of the same phenomenon, then it would be one of the most significant facets of earth science. But is this the case?
(We will leave aside for the moment just what is encompassed by varying definitions of dynamic topography)
Consider a few of the situations where dynamic topography is invoked. How about variations in bathymetry of seafloor? This is arguably the most robust place to look because seafloor is mostly monogenetic, being created at spreading centers once before being destroyed. The main exception are volcanic areas within plates like Hawaii that can reasonably be excluded. Most of the topography of seafloor is produced by cooling of the oceanic crust and upper mantle with time, but the simplest model for such cooling is that of a cooling infinite half space. It turns out that model works well out to ages of about 80 million years but beyond that, the sea floor to shallower than expected. Is this a signal from the deep mantle? Probably not.
Seismological study of the whole of the Pacific ocean floor shows that the plate ceases to cool like a half space somewhere out around 70-80 million years of age (see above). While there are arguments about what process this represents, it seems clear that the cause of the deviation from a half space is in preventing the plate from cooling further (some ideas have to do with layered convection, plate models instead of half spaces, and development of convective instabilities that “drip off” the colder lower part of the plate).
How about Southern Africa? It certainly seems anomalous, standing well over a kilometer above sea level with no major tectonic deformation seen in the past few hundred million years. It sits over a seismically slow part of the lower mantle, seemingly an excellent candidate for being supported by hot material rising off the core-mantle boundary. But the geologic history of the region contains some differing suggestions, in particular the indication that there might have been warming of the continental lithosphere associated with or following eruption of a major flood basalt. Major erosion of the region started in the late Mesozoic, indicating that this topography has been in place for some time. Given the evidence for warming of the lithosphere at about the same time that significant erosion starts, it might be this “dynamic topography” also represents some alteration of lithosphere.
Much shorter wavelength patterns trouble geodynamic modellers. Imagine a big block of gelatin and you push on it from one side with a finger. The gelatin on the other side of the block doesn’t have a finger-shaped piece poke out, it just bows out broadly. So if you want to have relatively narrow upwellings and downwellings at the earth’s surface, you have to have the forcings be relatively shallow. Its not that such things couldn’t exist, it is that the models used to generate predictions of dynamic topography tend to produce much broader features. So perhaps these ups and downs, as, for instance, Roberts and White (2010) argued for Africa, are reflecting some different phenomenon.
Many of the signals being interpreted for recent changes in dynamic topography lie within the geomorphology and stratigraphy of tectonically quiescent areas such as the Appalachians. Interpreting these signals is also challenging because the changes in the climate system over the past 3 million years mean that changes in erosion rates are quite likely, both from changes in precipitation patterns and fluctuations in sea level as glacial ice waxed and waned.
All of this suggests that there are a number of phenomena we aren’t really understanding at present; whether dynamic topography is an important part of the description of many of them remains to be seen. The flip side is to ask, where would we expect dynamic topography to be strongest? Generally this appears to be on top of subduction zones, where the weight of a descending slab should greatly affect surface topography. And yet the topography of back arc areas seems to be little affected by such a weight (e.g., Billen and Gurnis, GJI, 2003; Wheeler and White, Tectonics, 2002).
So while it seems likely that density variations within the deeper mantle have some influence on surface topography, the scale and magnitude of such processes remains in question. We are always tempted by Occam’s Razor to take a single explanation for multiple phenomena rather than multiple explanations, but this is no more than a guideline, especially if our explanation has enough degrees of freedom to cover many observations (but perhaps not all of them at once).