Detecting and assessing hydrocarbon reservoirs without the need to drill test wells is of major importance to the petroleum industry. Seismic methods have traditionally been used in this context, but the results can be ambiguous. Another approach is to use electromagnetic sounding methods that exploit the resistivity differences between a reservoir containing highly resistive hydrocarbons and one saturated with conductive saline fluids. Modeling presented by Eidesmo et al. (2002) demonstrates that by using seabed logging (SBL), a special application of frequency domain controlled source electromagnetic (CSEM) sounding, the existence or otherwise of hydrocarbon bearing layers can be determined and their lateral extent and boundaries can be quantified. Such information provides valuable complementary constraints on reservoir geometry and characteristics obtained by seismic surveying. In November 2000, a full-scale trial survey was carried out from the research ship RRS Charles Darwin offshore Angola, in an area with proven hydrocarbon reserves. The project was a collaboration among Statoil, Scripps Institution of Oceanography, and the Southampton Oceanography Centre. The object was to demonstrate that SBL, developed by Statoil (Eidesmo et al., 2000; Ellingsrud et al., 2001), could direct detect hydrocarbon-filled layers in the subseafloor. The petroleum prospects offshore Angola are in a deep Tertiary basin consisting of a thick (10-20 km) sequence of prograding sands and shales. The area is characterized by allochthonous salt of Aptian age, and deepwater channel sands with petroleum potential. Well logs show sediment resistivities typically around 0.7 Ωm that rise to around 100 Ωm in petroleum reservoirs. The survey site was on the continental slope in water depths of about 1200 m, with a known petroleum reservoir about 1100 m below seafloor. Shallow salt occurs in the northeast corner of the area.

In 1995, a multidisciplinary geophysical experiment targetted the intermediate spreading Valu Fa Ridge (full rate 60 mm yr−1 ), which is centred on 22°20′S, 176°40′W in the Lau Basin. As part of this experiment, wide‐angle and normal‐incidence seismic profiles were collected both along‐ and across‐axis to determine the crustal structure of the Central Valu Fa Ridge (CVFR) and its overlap with the Northern Valu Fa Ridge (NVFR). Controlled‐source electromagnetic profiles and underway gravity, magnetic and bathymetry data were also collected.  In this paper we describe the results of forward modelling of the along‐ and across‐axis wide‐angle and normal‐incidence seismic data. An axial low‐velocity block and its underlying slightly broader zone of depressed seismic velocities (low‐velocity zone) have been identified, and these features are interpreted as corresponding to a melt lens and underlying magma chamber. The low‐velocity block is 1–2 km wide and has a first‐order upper boundary, from which large‐amplitude reflections are observed; amplitude analysis of these indicates an interconnected melt fraction. The nature of the lower boundary is more poorly constrained, as no reflection event corresponding to the base of the low‐velocity block is observed. Modelling indicates that velocities similar to those observed at the base of layer 2 within the axial region (~5.5 km s−1) are achieved by 250 m below the upper boundary, possibly suggesting a gradational lower boundary with high velocity gradient. The low‐velocity zone (LVZ) is interpreted as an ~4 km wide magma chamber delineated by a seismic velocity anomaly of −0.2 km s−1, extending down through layer 3 to within 1.5–2 km of the Moho. The velocity anomaly and dimensions of the LVZ are generally smaller than those observed at the East Pacific Rise (EPR) and Mid‐Atlantic Ridge (MAR).  The observed along‐axis continuity of the low‐velocity block is remarkable, extending from the southern tip of the CVFR to the overlapping spreading centre (OSC) with the NVFR. A low‐velocity block is modelled beneath the inside flanks (i.e. the slopes that dip into the overlap basin) of both ridges at the OSC, although the existence of a single low‐velocity block beneath the overlap basin itself cannot be ruled out. The identification of a single LVZ centred on the overlap region, rather than two merged LVZs beneath each segment, implies that the material in each low‐velocity block originates from the same crustal magma source. A reflection event from the Moho is observed from directly beneath the axis on both across‐axis profiles, which indicates that a distinct crust–mantle boundary may be formed within the axial region. Many of the observations at the Valu Fa Ridge are consistent with those at the EPR and the Reykjanes Ridge (MAR), which implies that, regardless of spreading rate, crustal accretionary processes at mid‐ocean ridges with similar magmatic budgets are also broadly similar.

Quantifying the melt distribution and crustal structure across ridge-axis discontinuities is essential for understanding the relationship between magmatic, tectonic and petrologic segmentation of mid-ocean-ridge spreading centres. The geometry and continuity of magma bodies beneath features such as overlapping spreading centres can strongly influence the composition of erupted lavas and may give insight into the underlying pattern of mantle flow. Here we present three-dimensional images of seismic reflectivity beneath a mid-ocean ridge to investigate the nature of melt distribution across a ridge-axis discontinuity. Reflectivity slices through the 9 degrees 03' N overlapping spreading centre on East Pacific Rise suggest that it has a robust magma supply, with melt bodies underlying both limbs and ponding of melt beneath large areas of the overlap basin. The geometry of melt distribution beneath this offset is inconsistent with large-scale, crustal redistribution of melt away from centres of upwelling. The complex distribution of melt seems instead to be caused by a combination of vertical melt transport from the underlying mantle and subsequent focusing of melt beneath a magma freezing boundary in the mid-crust.

The RAMESSES study (Reykjanes Axial Melt Experiment: Structural Synthesis from Electromagnetics and Seismics) targeted an apparently magmatically active axial volcanic ridge (AVR), centred on 57°45′N at the Reykjanes Ridge, with the aim of investigating the processes of crustal accretion at a slow spreading mid‐ocean ridge. As part of this multicomponent experiment, airgun and explosive wide‐angle seismic data were recorded by 10 digital ocean‐bottom seismometers (OBSs) along profiles oriented both across‐ and along‐axis. Coincident normal‐incidence seismic, bathymetry and underway gravity and magnetic data were also collected.  Forward modelling of the seismic and gravity data has revealed layer thicknesses, velocities and densities similar to those observed elsewhere within the oceanic crust near mid‐ocean ridges. At 57°45′N, the Reykjanes Ridge has a crustal thickness of approximately 7.5 km on‐axis. However, the crust is modelled to decrease in thickness slightly off‐axis (i.e. with age), which implies that full crustal thickness is achieved on‐axis and that it is subsequently thinned, most likely, by off‐axis extension. Modelling also indicates that the AVR is underlain by a thin (∼100 m), narrow (∼4 km) melt lens some 2.5 km beneath the seafloor, which overlies a broader zone of partial melt approximately 8 km in width. Thus the results of this study provide the first clear evidence for a crustal magma chamber beneath any slow spreading ridge. The size and depth of this magma chamber (the melt lens and underlying zone of partial melt) are similar to those observed beneath fast and intermediate spreading ridges, which implies that the processes of crustal accretion are similar at all spreading rates. Hence the lack of previous observations of magma chambers beneath slow spreading ridges is probably temporally related to the periods of magmatic activity being considerably shorter and more widely spaced in time than at fast and intermediate spreading ridges.