Geophysical joint inversion requires the setting of a few parameters for optimum performance of the process. However, there are yet no known detailed procedures for selecting the various parameters for performing the joint inversion. Previous works on the joint inversion of electromagnetic (EM) and seismic data have reported parameter applications for data sets acquired from the same dimensional geometry (either in two dimensions or three dimensions) and few on variant geometry. But none has discussed the parameter selections for the joint inversion of methods from variant geometry (for example, a 2D seismic travel and pseudo-2D frequency-domain EM data). With the advantage of affordable computational cost and the sufficient approximation of a 1D EM model in a horizontally layered sedimentary environment, we are able to set optimum joint inversion parameters to perform structurally constrained joint 2D seismic traveltime and pseudo-2D EM data for hydrocarbon exploration. From the synthetic experiments, even in the presence of noise, we are able to prescribe the rules for optimum parameter setting for the joint inversion, including the choice of initial model and the cross-gradient weighting. We apply these rules on field data to reconstruct a more reliable subsurface velocity model than the one obtained by the traveltime inversions alone. We expect that this approach will be useful for performing joint inversion of the seismic traveltime and frequency-domain EM data for the production of hydrocarbon.

Vertical electrical sounding (VES) data acquired with the Schlumberger configuration is popularly used to image the electrical resistivity variation with depth at a single azimuth. Apart from the random subjective choice of the single azimuthal direction by the field geophysicists, important hydrological information such as fracture orientation and anisotropic coefficients needed for understanding resultant groundwater flow direction are by design lost in the process. Panoramic (0°–360°) azimuthal VES data were acquired at two data points at the Federal University of Technology, Akure (FUTA) at angular step of 15°, making a total of 24 data sets per data point. Each azimuthal VES data was inverted using equal number of layers in order to confirm the presence of anisotropy, quantify the anisotropic coefficients and image the orientation of fracture at a particular depth. Little to large apparent resistivity data and model suggested the presence of anisotropy which otherwise would have been lost in a single azimuthal survey. Elliptical fit of each layer azimuthal inverted resistivity was used to quantify the fracture orientation and coefficient of anisotropy with depth. From the results, it is established that anisotropy is present only at the near-surface: and the anisotropic coefficient increases from the surface to 7m. The result also showed the presence of an isotropic unit from 8m to the fresh basement. In agreement with existing published results on the geology of the area, the majority of the fractures trend North West and North East at stations 1 and 2 respectively. We hope that the methodology will foster detailed 3D panoramic imaging of the fracture network within and outside the study location, which will help in designing better groundwater management scheme and understanding resultant groundwater flow direction for contaminant and pollutant prevention and for flood control.

We provide a MATLAB computer code for training artificial neural network (ANN) with Nþ1 layer (N-hidden layer) architecture. Currently, the ANN application to solving geophysical problems have been confined to the 2layer, i.e. 1-hidden layer, architecture because there are no open source software codes for higher numbered layer architecture. The restriction to the 2-layer architecture comes with the attendant model error due to insufficient hidden neurons to fully define the ANN machines. The N-hidden layer ANN has a general architecture whose sensitivity is the accumulation of the backpropagation of the error between the feedforward output and the target patterns. The trained ANN machine can be retrieved by the gradient optimization method namely: Levenberg-Marquardt, steepest descent or conjugate gradient methods. Our test results on the Poisson's ratio (as a function of compressional and shear wave velocities) machines with 2-, 3-and 4-layer ANN architectures reveal that the machines with higher number of layers outperform those with lower number of layers. Specifically, the 3-and 4-layer ANN machines have ! 97 % accuracy, predicting the lithology and fluid identification in the oil and gas industry by means of the Poisson's ratio, whereas the 2-layer ANN machines poorly predict the results with as large error as 20 %. These results therefore reinforce our belief that this open source code will facilitate the training of accurate N-hidden layer ANN sophisticated machines with high performance and quality delivery of geophysical solutions. Moreover, the easy portability of the functions of the code into other software will enhance a versatile application and further research to improve its performance.

Joint inversions of coincident geophysical data are usually constrained to produce more reliable subsurface models. Structural, petrophysical, model parameter correlation, empirical, and transforms are some of the published constraints. The Gramian constraint provides a broad mathematical framework for implementing the aforementioned constraints. The Gramian constraint is formed from the determinant of the inner products of the model parameters involved. Previous works have used the Gramian constraint to invert multimodal parameters of different geophysical methods. But there has not been any extension of Gramian‐constrained joint inversion to mono‐model parameter from similar geophysical methods, for example, a similar conductivity or resistivity model from time‐ and frequency‐domain airborne electromagnetic methods. I implement the Gramian‐constrained joint inversion of time‐ and frequency‐domain airborne EM (AEM) data. This implementation allows the Gramian constraint to enhance the linear correlation of the model parameter between the two methods as the number of iterations increases. Improvement of the final joint inversion results over the standalone models is noticeable for both 3% noise‐contaminated synthetic and field data experiments. The field data jointly inverted are the high moment time‐domain SkyTEM data and frequency‐domain RESOLVE helicopter EM data acquired over the salinized Bookpurnong Irrigation District in South Australia in 2006 and 2008, respectively.