Hydrothermal venting in magma deserts : the ultraslow-spreading Gakkel and Southwest Indian Ridges

Abstract

Detailed hydrothermal surveys over ridges with spreading rates of 50–150 mm/yr have found a linear relation between spreading rate and the spatial frequency of hydrothermal venting, but the validity of this relation at slow and ultraslow ridges is unproved. Here we compare hydrothermal plume surveys along three sections of the Gakkel Ridge (Arctic Ocean) and the Southwest Indian Ridge (SWIR) to determine if hydrothermal activity is similarly distributed among these ultraslow ridge sections and if these distributions follow the hypothesized linear trend derived from surveys along fast ridges. Along the Gakkel Ridge, most apparent vent sites occur on volcanic highs, and the extraordinarily weak vertical density gradient of the deep Arctic permits plumes to rise above the axial bathymetry. Individual plumes can thus be extensively dispersed along axis, to distances >200 km, and ∼75% of the total axial length surveyed is overlain by plumes. Detailed mapping of these plumes points to only 9–10 active sites in 850 km, however, yielding a site frequency F s , sites/100 km of ridge length, of 1.1–1.2. Plumes detected along the SWIR are considerably less extensive for two reasons: an apparent paucity of active vent fields on volcanic highs and a normal deep-ocean density gradient that prevents extended plume rise. Along a western SWIR section (10°–23°E) we identify 3–8 sites, so F s = 0.3–0.8; along a previously surveyed 440 km section of the eastern SWIR (58°–66°E), 6 sites yield F s = 1.3. Plotting spreading rate (us) versus F s, the ultraslow ridges and eight other ridge sections, spanning the global range of spreading rate, establish a robust linear trend (F s = 0.98 + 0.015us), implying that the long-term heat supply is the first-order control on the global distribution of hydrothermal activity. Normalizing F s to the delivery rate of basaltic magma suggests that ultraslow ridges are several times more efficient than faster-spreading ridges in supporting active vent fields. This increased efficiency could derive from some combination of three-dimensional magma focusing at volcanic centers, deep mining of heat from gabbroic intrusions and direct cooling of the upper mantle, and nonmagmatic heat supplied by exothermic serpentinization.