The distinctive characteristics of hydrothermalism along fast and slow-spreading ridges must be related to differences in the cycling and relative importance of volcanic, tectonic and hydrothermal processes along ridge systems with varying magma supply rates. It is therefore instructive to attempt to reconcile the observations with the Sinton and Detrick [1992] models of magma chambers, since they represent the heat source that drives the hydrothermal circulation. Along fast-spreading ridges, a melt lens, <1-2 km wide and 1-2 km below the surface, overlies a zone of crystal mush that is in turn surrounded by a transition zone of solidified crust with some isolated pockets of magma [ Sinton and Detrick, 1992]. The melt lens provides a shallow heat source to drive hydrothermal circulation cells in the overlying crust, with fluid flow focused along shallow faults and fissures. Dike injections and fissure eruptions from the melt lens have two important effects. First, the newly injected material is cooled rapidly by circulation of large volumes of seawater resulting in widespread, diffuse flow, which decreases dramatically within a short period of time. Second, the permeability structure of the upper crust will be modified with every volcanic event, resulting in frequent reorganization of the fluid flow pathways, and thereby limiting the size of hydrothermal structures that can be constructed. In this scenario, cycling between volcanic, tectonic and hydrothermal processes is rapid, and the hydrothermal activity is dominated more by volcanic, than by tectonic, processes.
In contrast, at slow-spreading ridges, there is no evidence for a steady-state magma body; rather, Sinton and Detrick [1992] envision a deeper, dike-like mush zone and broad transition zone. Volcanic eruptions are intermittent and widely dispersed, and are coupled in time to injection events of magma from the mantle. In addition, listric faults bordering the rift valley may extend into the brittle-ductile transition within the partially molten mush. In this situation, maintenance of hydrothermal circulation is likely to be fault-controlled rather than volcanically-controlled, since construction of large hydrothermal deposits requires episodic renewal of hydrothermal activity in the same location over long periods of time. In one working model, those normal faults which remain active and hence maintain their high permeability act as the preferred conduits for fluid flow beginning at the neovolcanic zone and continuing throughout the median valley as the crust spreads. Even though these channels may become sealed due to mineral precipitation, flow is rejuvenated each time there is movement along the fault. At the border of the rift valley, activity along selected faults that form the median valley walls extends down into the molten mush zone, thereby providing access to the heat source to continue driving the circulation system.
Such models are based on extremely limited datasets and will need considerable refinement as more information becomes available, especially for slow-spreading (low magma supply) ridges. In the next few years, acoustic mapping and imaging surveys using improved high-resolution systems will provide valuable information to better define the detailed relations between hydrothermal, volcanic and tectonic processes.