ALBERT

Description: Deep-towed geophysical surveys require precise knowledge of navigational parameters such as instrument position and orientation because navigational uncertainties reflect in the data and therefore in the inferred geophysical properties of the subseafloor. We address this issue for the case of electrical conductivity inferred from controlled source electromagnetic data. We show that the data error is laterally variable due to irregular motion during deep towing, but also due to lateral variations in conductivity, including those resulting from topography. To address this variability and quantify the data error prior to inversion, we propose a 2-D perturbation study. Our workflow enables stable and geologically reliable results for multicomponent and multifrequency inversions. An error estimation workflow is presented, which comprises the assessment of navigational uncertainties, perturbation of navigational parameters, and forward modelling of electric field amplitudes for a homogeneous and then a heterogeneous subseafloor conductivity model. Some navigational uncertainties are estimated from variations of direct measurements. Other navigational parameters required for inversion are derived from the measured quantities and their error is calculated by means of error propagation. Some navigational parameters show direct correlation with the measured electric fields. For example, the antenna dip correlates with the vertical electric field and the depth correlates with the horizontal electric field. For the perturbation study each standard deviation is added to the navigational parameters. Forward models are run for each perturbation. Amplitude deviations are summed in quadrature with the stacking error for a total, laterally varying, data error. The error estimation is repeated for a heterogeneous subseafloor model due to the large conductivity range (several orders of magnitude), which affects the forward model. The approach enables us to utilize data from several components (multiple electric fields, frequencies and receivers) in the inversion to constrain the final model and reduce ambiguity. The final model is geologically reasonable, in this case enabling the identification of conductive metal sulphide deposits on the seafloor.