A Discriminant Model for Detecting Resistivity Variation in Sea Bed Logging Survey Data

Sea-Bed Logging (SBL) is the application of marine Controlled-Source Electro-Magnetic (CSEM) method to detect the presence of a resistive layer beneath the sea bed such as hydrocarbon (HC). Basically the method depends on the large resistivity variation between the hydrocarbon reservoir and the surrounding layers with different resistivity values. Therefore, the ability to detect the presence of resistivity variation is very important in processing SBL data. In this paper, discriminant analysis technique is applied to the simulated phase versus offset SBL data. Two SBL models with (1000Ω) and without (1Ω) hydrocarbon were simulated. The objective is to find a discriminant model that can classify the simulated data into two groups based on the assigned resistivity values to the two SBL models. Discriminant function analysis is a statistical technique used to predict membership in two or more mutually exclusive groups. Data from the two SBL models are used to develop the discriminant model. Wilks' lambda is used to test the significance of the discriminant function as a whole, while measure of F-value for a variable is used to indicate its statistical significance in the discrimination between groups and extent to which that variable makes a unique contribution to the prediction of group membership. The results obtained suggest that discriminant analysis technique has a potential to classify the SBL data according to the resistivity variation.