Seismic Facies classification and TOC prediction using seismic attributes and Hjorth parameters, in the Groningen field in northeastern Netherlands

The use of seismic attributes for characterization and exploration is crucial. This significance increases with the growing number of attributes being used, developed, or newly derived, regardless of their relation to previous ones. Facies classification heavily depends on these attributes in conjunction with other predictive processes. A key focus of these processes is predicting organic carbon content using seismic data, particularly through seismic attributes. The study of organic carbon content has gained attention from researchers because it provides critical information for identifying source rocks and kerogen-bearing formations, which are vital for future drilling operations. In this paper, we employed well-known seismic attributes alongside Hjorth parameters—originally developed for analyzing time-series data in medical applications like EEG signal processing—as innovative seismic attributes for prediction and classification tasks. The most significant of these Hjorth parameters include activity, mobility, and complexity. We applied these parameters through machine learning models, primarily using random forests. The results from integrating these parameters with machine learning models were impressive, achieving a prediction and classification accuracy of up to 93%. Furthermore, this approach provided more accurate information than analyses performed without these parameters.