780 Thousand Years of Upper‐Crustal Construction at a Melt‐Rich Segment of the Ultraslow Spreading Southwest Indian Ridge 50°28′E
Jie Chen, Mathilde Cannat, Chunhui Tao, Daniel Sauter, Marc Munschy
Abstract
Abstract Melt supply at the ultraslow‐spreading Southwest Indian Ridge (SWIR) has been shown to vary from nearly amagmatic, leading to ultramafic seafloor, to magmatically robust, producing fully volcanic seafloor. The center of the SWIR 50°28′E segment represents a magmatically robust endmember. High‐resolution bathymetry and backscatter, near‐bottom magnetic data, and visual observations were acquired to infer spatiotemporal variations in upper‐crustal construction over the past 780 kyr. Tectonic strain inferred from observed faults represents <8% of the total plate divergence during this period. Mapped seafloor and lava morphologies show a prevalence of smooth seafloor and smooth hummocky seafloor (64% mapped area), corresponding to a dominance of high‐effusion‐rate, sheet‐lobate lavas, over hummocky seafloor (29%) that mostly comprises low‐effusion‐rate pillow lavas. This prevalence is most pronounced within a ∼5 km‐long domain at the segment center, while hummocky seafloor prevails to its east and west, indicating a substantial along‐axis decrease in average eruption rate. Across‐axis changes of seafloor morphologies and faulting pattern indicate two successive cycles of upper‐crustal construction. These cycles last ∼300 kyr, interpreted as waxing and waning magmatic phases. During waxing phases, shallow axial melt lenses (AMLs) form beneath the segment center, feeding narrow diking systems that build high‐eruption‐rate smooth domal volcanos. During waning phases, lower frequencies of melt replenishment may cause the AML to become deeper and ultimately disappear. This configuration triggers diking and hummocky‐dominated eruptions over a wider axial domain, with more faults. We propose that modes of upper‐crustal construction at slow and ultraslow mid‐ocean ridges are influenced by melt supply more than spreading rate.