Propagation of Magma-Filled Cracks

Publication Year


Journal Article

The mechanism of magma transport at depth influences direction magma moves, the distance it travels before freezing, the degree to which it communicates chemically with the host rock, the form of surficial volcanism, and ultimately the growth of oceanic and continental crust. Commonly envisioned transport processes include porous flow in partially molten and deformable source rock, flow through fractures in elasticlbrittle rock, and diapiric ascent (typically of granites) through viscous rock. Of these, transport in fractures, or dikes, is the most efficient means of moving magma through cold lithosphere. Porous flow is an option only if there has been sufficient advection of heat to raise the rock temperature above the solidus (a possibility beneath Hawaii, for example). Although the sheet-like form of dikes is less advantageous thermally than the equidimensional form of diapirs, except for possibly the most viscous rhyolites, this is more than offset by the fact that transport rates depend upon the magma viscosity, rather than the host rock viscosity. 

Dike intrusion is also the transport mechanism that permits the most direct comparison between theory and observation. In part, this is because many dikes carry so little heat that magmatic and host rock structures produced during intrusion are preserved until exposed by erosion. The short timescale of intrusion also allows near-real-time seismic and geodetic monitoring of dikes in active volcanoes. In addition, the close kinship between dikes and artificial hydrofractures, used in the oil and geothermal industries, provides an economic incentive for monitoring field-scale experiments that are relevant to dike propagation. Finally, the nominally planar geometry of dikes permits idealizations that make a detailed theoretical treatment practical. While analogous equations describing porous flow in the source region have also been proposed, observations available to constrain these theories are fewer and less direct. The situation is worse for granitic diapirism; because of the complexity of the processes envisioned, and the tendency for later deformations to obliterate earlier ones, even its existence is debated.

I begin this review by highlighting the observations of ancient and modern dikes that lay the groundwork for current modeling. The emphasis then shifts to the dominant physical processes involved host rock fracture and deformation (elastic and inelastic), magma flow, and heat transfer. Previous review articles on some of these topics include those by Pollard ( 1987) for the solid mechanics, Lister & Kerr ( 1991) for coupling the fluid flow and elastic deformation, and Delaney ( 1987) for the heat flow. Where possible this review attempts to provide a framework for thinking about important but poorly understood processes such as dike initiation, the role of dike propagation in the ascent of granitic magmas, and earthquakes accompanying magma transport. 

Annual Review of Earth and Planetary Sciences
287 - 336
Date Published
Short Title
Annu. Rev. Earth Planet. Sci.