Various embodiments described herein provide methods of hydrocarbon management and associated systems and/or computer readable media including executable instructions. Such methods (and by extension their associated systems and/or computer readable media for implementing such methods) may include obtaining geophysical data (e.g., seismic or other geophysical data) from a prospective subsurface formation (that is, a potential formation or other subsurface region of interest for any of various reasons, but in particular due to potential for production of hydrocarbons) and using a trained machine learning (ML) system for direct hydrocarbon indicators (DHI) analysis of the obtained geophysical data. Hydrocarbon management decisions may be guided by the DHI analysis.

Methods, systems, and computer-readable media for processing seismic data. The method includes receiving a plurality of seismic traces representing a subterranean domain, and receiving an implicit stratigraphic model of at least a portion of the subterranean domain. The method also includes selecting an iso-value in the implicit stratigraphic model, and defining, using a processor, a geologically-consistent interval in the implicit stratigraphic model based at least partially on a position of the iso-value in the implicit stratigraphic model. The method further includes calculating one or more attributes of the plurality of seismic traces in the interval.

Systems, computer-readable media, and methods for generating machine learning training data by obtaining reservoir data, determining subsections of the reservoir data, labeling the subsections of the reservoir data to generate labeled reservoir data, and feeding the labeled reservoir data into an artificial neural network. The reservoir data can be labeled using analysis data or based on interpretive input from an interpreter.

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, to generate a custom seismic surface and volume attribute. In one aspect, a method includes receiving a seismic cube and a seismic surface, and the seismic cube includes traces recorded at receivers deployed to collect seismic data. The seismic surface is picked on the seismic cube. Seismic wavelets are extracted with a selected length from the seismic cube along an intersection with the seismic surface for each spatial coordinate associated with the seismic surface. A reference wavelet is determined. A surface attribute map is generated based on comparing each of the seismic wavelets to the reference wavelet. A productivity of the seismic surface is evaluated using the surface attribute map.

A method of obtaining a relative amplitude preserved seismic volume acquired in a time-domain for a subterranean region of interest and transforming it into a low-frequency monospectral amplitude volume. The method further determines a seismic attenuation volume from the relative amplitude preserved seismic volume acquired in the time-domain. Furthermore, the method generates a low-frequency monospectral amplitude map for a surface of interest by averaging the low-frequency monospectral amplitude volume over a depth-window around the surface of interest, and generates a seismic attenuation map for a surface of interest by averaging the seismic attenuation volume over a depth-window around the surface of interest. The method further determines an attribute map based on the seismic attenuation map and the low-frequency monospectral amplitude map for the surface of interest, and determines a presence of gas in the subterranean region of interest based on the attribute map.

Methods and systems, including computer programs encoded on a computer storage medium can be used to preserve edges while performing Gaussian grid smoothing of noise components in subsurface grids to generate geological maps. A subsurface grid is generated from data indicating properties of subsurface formations. A weighting grid is generated by: i) receiving seismic data representing the subsurface formations; ii) generating seismic attributes associated with discontinuities in the subsurface formations; and iii) assigning a particular weight value to weighting grid points that the seismic attributes associated with discontinuities in the subsurface formations indicate the presence of a discontinuity. The subsurface grid is processed by iteratively computing local averages of grid points in the subsurface grid using a compact Gaussian filter weighted by values in the weighting grid. A geological map of subsurface formations is generated based on the filtered subsurface grid.

Systems and methods are provided for a horizon patch extraction process and in particular, to receiving seismic trace data of a plurality of seismic events of a subterranean volume, selecting a first seismic trace based on the seismic trace data of the plurality of seismic events, the first seismic trace including a plurality of seismic onsets, determining a depth, an amplitude, and a first thickness of a first seismic onset of the first seismic trace, determining a second thickness between the first seismic onset and a second seismic onset, determining a third thickness between the first seismic onset and a third seismic onset, and generating a horizon patch based on the depth, the amplitude, and the first thickness of the first seismic onset, the second thickness between the first seismic onset and the second seismic onset, and the third thickness between the first seismic onset and the third seismic onset.

A method for training a backpropagation-enabled segmentation process is used for identifying an occurrence of a sub-surface feature. A multi-dimensional seismic data set with an input dimension of at least two is inputted into a backpropagation-enabled process. A prediction of the occurrence of the subsurface feature has a prediction dimension of at least 1 and is at least 1 dimension less than the input dimension.

A method for training a backpropagation-enabled regression process is used for predicting values of an attribute of subsurface data. A multi-dimensional seismic data set with an input dimension of at least two is inputted into a backpropagation-enabled process. A predicted value of the attribute has a prediction dimension of at least 1 and is at least 1 dimension less than the input dimension.

A method that includes transforming a relative amplitude preserved 3D seismic volume acquired in the time-domain into a plurality of isofrequency volumes, extracting from the plurality of isofrequency volumes a first isofrequency spectral amplitude volume and a second isofrequency spectral amplitude volume. The method further includes determining an attribute volume from the two isofrequency spectral amplitude volumes, and determining a presence of gas in a subterranean region of interest based on the attribute volume.