The 59,000 km long global mid-ocean ridge system is the site of formation of 20 km3 of oceanic crust yearly. Two-thirds of all heat loss from the interior of our planet is through the ocean floors, 40% of this amount is focused through the ridge. Activity involves complex interactions among a number of processes occurring over wide ranges of depths and lateral distances, including melting of the earth's mantle, delivery of the molten rock to a crustal magma chamber, cooling of the magma intrusion by hydrothermal circulation and volcanic eruption, chemical exchange between hot rock surrounding the magma chamber and the overlying seawater, and even the establishment of exotic biological communities near hydrothermal vents at the ridge axis. These features justify the expanding scientific interest in the study of the ridge.
Transient controlled-source electromagnetics (CSEM) is a geophysical exploration technique capable of determining the electrical conductivity beneath fast-spreading segments of the mid-ocean ridge. Geological structure beneath the mid-ocean ridge that is readily accessible to transient CSEM exploration is located at crustal levels and includes the axial magma chamber and its associated zones of partial melt and hydrothermal activity. Seismic images of the top several kilometers beneath the fast spreading East Pacific Rise (EPR) between 9-13°N have already been obtained. Multi-channel reflection profiles place strong constraints on the geometry of the top of the axial magma chamber but refraction data provide only coarse estimates of the sub-surface temperature, distribution of partial melt and porosity, parameters required to distinguish between proposed petrological models of the ridge. Electrical conductivity is a strong indicator of all these critical parameters and therefore CSEM methods are well-suited to improve the estimates and help characterize the ridge environment.
In this thesis, a pair of forward modeling computer programs have been developed to design ridge-going experiments and assist interpretation of mid-ocean ridge transient CSEM data sets, as they become available. The programs may also be used to evaluate the transient CSEM technique as it might be applied to investigate other tectonically active regions of the seafloor. One program rapidly computes the theoretical response, as a function of time, of an arbitrary, two dimensional earth to a sudden switch-on of electric current in a line source of electromagnetic energy. The other program is more advanced, requires more computer time, and is referred to as a 2.5-D program because it can handle excitation of the earth by a more realistic, finite source.
The programs solve the forward problem as follows. Electromagnetic boundary value problems based on the governing Maxwell's equations are solved by the finite element method in the Laplace frequency s-domain. The calculated electromagnetic field components are then transformed into the time domain by means of the Gaver-Stehfest algorithm. In the 2.5-D program, Maxwell's equations are additionally Fourier transformed in the direction parallel to the strike of the 2-D conductivity structure, and field components are computed in the along-strike wavenumber q-domain. Following the calculation, inverse transforms are performed to obtain the along-strike spatial variations of the field components. The codes have been validated through comparisons with known analytic solutions in which the earth is modeled as a uniformly conducting half-space. Convergence of the finite element approximation is found to be O(h), where h measures the size of the triangles comprising the finite element mesh. An extrapolation formula is described by which numerical solutions on progressively finer meshes are combined. The formula permits great accuracy to be attained in the computed field components, using relatively coarse meshes.
A numerical study of the performance of an idealized transient CSEM system at the East Pacific Rise has been carried out using the 2-D code. The system consists of an infinite source located 5 km west of the ridge axis, and seafloor magnetic field sensors placed at various distances across the ridge crest. The source is oriented with respect to the strike of the ridge so as to produce only the H-polarization mode of electric current flow. The results indicate that this system can detect the axial magma chamber and the associated zones of hydrothermal activity and partial melt by monitoring two electromagnetic response parameters, the diffusion time T and the response amplitude B max , as a function of transmitter/receiver separation. These response parameters are easily extracted from measured data and are diagnostic of the sub-surface electrical conductivity. The presence of a highly conductive magma chamber slows and attenuates signals diffusing beneath the ridge, increasing T and decreasing Bmax. Hydrothermal circulation in the highly fractured, extrusive basalt layer has the same effect on the data for receivers placed within 3 km of the ridge axis, but very little effect elsewhere. Inferences made from the numerical results suggest that a horizontal electric dipole (HED) of moment 10 4 A•m and receivers sampling the seafloor magnetic field at 10-25 Hz with a sensitivity of 1 pT/s over a time window extending to 10 s are sufficient to detect these crustal targets.
Interpretation of transient CSEM data requires forward modeling using a more realistic, finite source. The 2.5-D code is capable of achieving this. Sample field patterns produced in the vicinity of the ridge by a sudden switch-on of electric current in a horizontal electric dipole (HED) are computed. The patterns illustrate diffusion, in three spatial dimensions and time, of various along-strike electromagnetic field components through typical mid-ocean ridge structures. The results demonstrate the utility of the 2.5-D code, i.e. its potential for interpreting data from a transient CSEM ridge-going experiment.