Segmentation and cycles of crustal accretion at mid-ocean ridges: a study of the Reykjanes Ridge

Abstract

Early studies of mid-ocean ridges suggest a fundamental difference between crustal accretionary processes at slow- and fast-spreading ridges. Accretion, and the supply of melt to the crust itself, is thought to be highly episodic at slow-spreading ridges but steady-state at fast-spreading ridges. However, recent studies are beginning to question this model, with evidence for the temporal variation in crustal accretionary processes at all spreading rates emerging. This study provides evidence from bathymetry, TOBI sidescan, gravity and magnetic data, collected during different cruises to the Reykjanes Ridge, for the temporal nature of crustal accretion and its relationship to segmentation. Interpretation of TOBI images indicates that individual adjacent axial volcanic ridges (AVRs) vary in relative age, suggesting that they are at various stages of an evolutionary lifecycle, with episodic cycles of magmatic and tectonic activity. However, prior to investigating the possible effects of tectonomagmatic cycles on the crustal structure of AVRs, the effect of the Iceland hotspot on the ridge is examined. The along-axis free-air gravity anomaly is forward modelled in 2-D, revealing an along-axis increase in crustal thickness towards Iceland from 7.5 km to 10.5 km and a decrease in mantle densities from 3.30 to 3.23 g cm"^ between 57 30'N and 62 N. Calculation of the residual mantle Bouguer Anomaly (RMBA) and inversion of magnetic anomaly data, reveal intermediate-wavelength fluctuations in RMBA amplitude and magnetization intensity respectively that are attributed to hotspot pulses, with 59 N marking the southern most extent of the most recent pulse. Removal of the hotspot effect on the gravity data reveals short-wavelength RMBA lows, associated with individual AVRs, superimposed on a broad ridge-trending low. Along-AVR-axis gravity modelling shows that a number of these RMBA lows can be explained by a 200-800 m thickening of the crust and/or by the presence of 5-20% partial melt in the mid-crust. A correlation between relative AVR age and crustal structure is established, with longer, more mature AVRs having a thicker crust and shorter, younger AVRs having more partial melt in the mid-crust. Short-wavelength magnetization intensity highs, associated with younger AVRs, corroborate the TOBI age interpretations. Local spreading rate calculations reveal that total spreading rates for younger AVRs are up to 20% faster than for older AVRs over the last 1.42 Ma. On the basis of these results a model for the cyclicity of crustal accretion is presented, whereby far-field tectonic stresses result in spreading-orthogonal brittle deformation of the crust in the neovolcanic zone, and 3-D mantle upwelling, with a wavelength of -70 km, follows the ridge trend and results in second order segments that comprise ~5 AVRs. It is proposed that along-axis migration of melt within such a segment results in the observed variations in AVR age, length, RMBA amplitude, magnetization intensity and local spreading rate. The proposed model has implications for the temporal variability of crustal accretion at all spreading rates.