A method of detecting an earthquake, comprises, by a processor and memory circuitry: obtaining motion data based on data collected by one or more sensors, filtering the motion data to obtain filtered data FD, wherein the filtering comprises, within at least one range of frequencies representative of an earthquake, amplifying one or more frequencies of the motion data within this range, wherein the one or more frequencies are more amplified relative to other frequencies within this range, comparing, at least once, data representative of FD to at least one threshold, if this comparison meets an alerting criteria, generating an alert indicating that an earthquake has been detected.

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.

A system can determine a location for future wells using machine-learning techniques. The system can receive seismic data about a subterranean formation and may determine a set of seismic attributes from the seismic data. The system can block the set of seismic attributes into a set of blocked seismic attributes by distributing the set of seismic attributes onto a geo-cellular grid representative of the subterranean formation. The system can apply a trained machine-learning model to the set of blocked seismic attributes to generate a composite seismic parameter. The system can distribute the composite seismic parameter in the subterranean formation to characterize formation locations based on a predicted presence of hydrocarbons.

Systems and methods include a computer-implemented method for determining normalized apparent power. Drilling acoustic signals corresponding to a time domain and generated during drilling of a well. A fast Fourier transformation (FFT) is performed using the drilling acoustic signals to generate FFT data. Normalized FFT data is generated using normalization parameters and a drill string rotation rate record of a drill string used to drill the well. The drill string rotation rate is received during drilling. Normalized apparent power is determined from data points of a predetermined top percentage of the normalized FFT data within a lithological significant frequency range. The normalized apparent power is a measure of the power of the drilling acoustic signals and it is a function of the amplitude and frequency of the normalized FFT data. The lithological significant frequency range is a frequency range within which the drill sounds are more closely related with lithology.

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 of operating a reservoir simulator can include performing a time step of a reservoir simulation using a spatial reservoir model that represents a subterranean environment that includes a reservoir to generate simulation results for a first time where the simulation results include a front defined by at least in part by a gradient at a position between portions of the spatial reservoir model; predicting a position of the front for a subsequent time step for a corresponding second time using a trained machine model; discretizing the spatial reservoir model locally at the predicted position of the front to generate a locally discretized version of the spatial reservoir model; and performing a time step of the reservoir simulation using the locally discretized version of the spatial reservoir model to generate simulation results for the second time.

Systems and methods are provided for generating a model for detection of seismic events. In this regard, one or more processors may receive from one or more stations located along an underwater optical route, one or more time series of polarization states of a detected light signal during a time period. The one or more processors may transform the one or more time series of polarization states into one or more spectrums in a frequency domain. Seismic activity data for the time period may be received by the one or more processors, where the seismic activity data include one or more seismic events detected in a region at least partially overlapping the underwater optical route. The one or more processors then generate a model for detecting seismic events based on the one or more spectrums and the seismic activity data.

A method of identifying events includes obtaining an acoustic signal from a sensor, determining one or more frequency domain features from the acoustic signal, providing the one or more frequency domain features as inputs to a plurality of event detection models, and determining the presence of one or more events using the plurality of event detection models. The one or more frequency domain features are obtained across a frequency range of the acoustic signal, and at least two of the plurality of event detection models are different.

An ensemble of machine learning models is trained to evaluate seismic and risk-related data in order to evaluate, value, or otherwise rank various prospective hydrocarbon reservoir (“prospects”) of a field. A classification machine learning model is trained to classify a prospect or region of a prospect based on the exploration risk level. From the seismic data, a frequency-filtered volume (FFV) for each prospect is calculated, where the FFV is a measure of reservoir volume which takes into account seismic resolution limits. Based on the risk classification and FFV, prospects of the field are ranked based on their economic value which is a combination of the risk associated with drilling and their potential reservoir volume.

A system can receive seismic data that can correlate to a subterranean formation. The system can derive a set of seismic attributes from the seismic data. The seismic attributes can include discontinuity-along-dip. The system can determine parameterized results by analyzing the seismic data and the seismic attributes using a deep learning neural network. The deep learning neural network can include a dilation module. The system can determine one or more fault probabilities of the subterranean formation using the parameterized results. The system can output the fault probabilities for use in a hydrocarbon exploration operation.