Average compositions and analyses range for clinopyroxene, olivines, orthopyroxene, and spinel from the MARK area at DSDP Site 45-395A, and ODP Site 153-920

Mineral compositions of residual peridotites collected at various locations in the Mid-Atlantic Ridge south of the Kane transform (MARK area) are consistent with generally smaller degrees of melting in the mantle near the large offset Kane transform than near the other, small offset, axial discontinuities in the area. We propose that this transform fault effect is due to along-axis variations in the final depth of melting in the subaxial mantle, reflecting the colder thermal regime of the ridge near the Kane transform. Calculations made with a passive mantle flow regime suggest that these along-axis variations in the final depth of melting would not produce the full range of crustal thickness variations observed in the MARK area seismic record. It is therefore likely that the transform fault effect in the MARK area is combined with other mechanisms capable of producing crustal thickness variations, such as along-axis melt migration, the trapping of part of the magma in a cold mantle root beneath the ridge, or active mantle upwelling.

(Table 2) 230Th, 230Th xs and 232Th concentrations in dissolved, particulate and total (dissolved+small particles) for all KEOPS stations, Marion Dufresne cruise MD145

In the context of the KErguelen Ocean and Plateau compared Study (KEOPS, 19 January-13 February 2005), particle dynamics were investigated using thorium isotope measurements over and off the Kerguelen plateau. Dissolved and particulate 230Th and 232Th samples were collected at nine stations. Dissolved excess 230Th concentrations (230Thxs) vary from 0.5 to 20.8 fg/kg and particulate 230Thxs concentrations from 0.1 to 10.0 fg/kg. Dissolved and particulate 232Th concentration ranges are 16.8-450.2 pg/kg and 3.8-502.8 pg/kg, respectively. The 230Thxs concentrations increase linearly with depth down to the bottom at most of the plateau stations and down to 1000 m at the off-plateau stations. This linear trend is observed down to the bottom (1550 m) at Kerfix, the open-ocean "upstream" station located west of the Kerguelen plateau. A simple reversible scavenging model applied to these data allowed the estimation of adsorption rate constant (k1~=0.2-0.8 per year), desorption rate constant (k-1~=1-8 per year) and partition coefficients (average K=0.16±0.07). Calculated particle settling velocities S deduced from this simple model are ca. 500 m/year at most of the plateau stations and 800 m/year at all the off-plateau stations. The plateau settling velocities are relatively low for such a productive site, compared to the surrounding HNLC areas. The difference might reflect the fact that lateral advection is neglected in this model. Taking this advection into account allows the reconstruction of the observed 230Thxs linear distributions, but only if faster settling velocities are considered. This implies that the 1D model strongly underestimates the settling velocity of the particles. In the deep layers, the occurrence of intense boundary scavenging along the escarpment due to bottom sediment re-suspension and interaction with a nepheloid layer, yielding a removal of ?50% of the Th stock along the northwestward transect, is suggested.