Marine Geosciences

Geophysical and geochemical contraints on crustal accretion at the ultra-slow spreading Mohns Ridge. 

Klingelhoefer, F., and Geli, L.,  and White, R. S., Geophysical Research Letters, (submitted June 1999).

The composition of upper mantle and lower crustal material at very-slow spreading centers cannot be reliably determined by seismic studies alone. Since the range of P-wave velocities for serpentinized peridotites and
gabbros overlap, additional information provided by the major and rare earth element (REE) content of the basalts is useful to constrain interpretations of seismic data. Refraction seismic data from the very-slow spreading (16 mm/a, full rate) Mohns Ridge in the Norwegian-Greenland Sea yields a highly variable thin crust of 4.0 ± 0.5 km thickness. Analysis of S-waves suggests that Layer 3 is composed primarily of gabbro containing at most a small percentage (< 20 %) of mantle material. The Na₈ content of Mohns Ridge basalts suggests a magmatic crustal thickness of 4-5 km. 
Inversion of the REE concentrations yields a melt thickness of ¥ 5 km. This agreement between seismic and geochemical data suggests that neither large quantities of mantle material are found in the lower crust nor is a large volume of basaltic magma frozen in the upper mantle.

Crustal Structure of a Super-slow Spreading Center: a Seismic Refraction Study of Mohns Ridge,  2° N

Klingelhoefer, F., Geli, L., and Matias, L., (accepted pending revisions): . Geophysical Journal International (1999)

A series of 8 high resolution seismic refraction profiles from the ultra-slow spreading (16 mm/y full spreading rate) Mohns Ridge in the Norwegian-Greenland Sea have been treated with modern inversion methods. The profiles were shot parallel to the ridge at an off axis distance of 0-135 km corresponding to crustal ages of 0-22 Ma. The resulting models are constrained by synthetic seismograms and gravity modelling.

The crustal thickness in all profiles is well below the global average for typical oceanic crust, and shows a high variability with a mean thickness of  4.0 ± 0.5  km. This is mainly due to a very thin and variable lower crustal layer (Layer 3). Generally, the crust is thicker beneath basement highs and thinner beneath basins, implying local isostatic compensation. The top of the basement (Layer 2a) consists of a zone with low P-wave velocities (2.5 - 3.0 kms^-1). The mean thickness of this layer decreases with distance from the ridge. Beneath it lies a layer with slightly higher velocities (Layer 2b). Its thickness shows less variability along a given profile and an overall increase with age. The combined average thickness of the upper two layers remains nearly constant, indicating, that the boundary between Layer 2a and 2b might represent an alteration front.

Upper mantle velocities are generally slow, around 7.5 kms^-1. For the profile directly within the rift valley, a model without a third layer, incorporating a constant gradient up to upper mantle velocities and a model with a Moho depth inferred from neighbouring profiles and upper mantle velocity as slow as 7.2 km/s, fit the seismic and gravity data equally well. The crustal structure is not mature below the ridge. These observations support previous models suggesting the presence of low densities and velocities at about 2 km below the rift axis. Poisson's ratios determined from converted S-wave modelling are incompatible with a Layer 3 consisting of purely serpentinized peridotite. However, a volume fraction of 10-40 % serpentinite cannot be ruled out.