Facilitating hydrocarbon exploration from earth system models

A method comprising:

• accessing, by a processor, training data including information relating to a modern feature of interest, the training data compiled from direct observations, remotely determined data, or a combination thereof, and including a global magnitude map representing the information relating to the modern feature of interest;
• compiling, by the processor, parameter data from at least one modern model simulation that impacts the modern feature of interest, the modern model simulation comprising a hydrosphere model, an atmospheric model, a heliosphere model, a geosphere model, a cryosphere model, a biosphere model, a general circulation model, an earth system model, or a combination thereof;
• providing the training data and the compiled parameter data from the at least one modern model simulation to a machine-learning model;
• generating, by the processor, a predictive model from the machine-learning model by training the machine-learning model on the training data and the compiled parameter data from the at least one modern model simulation;
• outputting, by the predictive model, a predicted feature of interest in a past time period by applying the predictive model to compiled parameter data from at least one existing historical earth system factor model simulation that impacts the feature of interest in the past time period, wherein the predicted feature of interest includes a global magnitude map;
• based on the predicted feature of interest in the past time period output by the predictive model, calculating a feature of interest score for each of a plurality of geographic areas of interest;
• determining a suitability for exploration of at least one of the plurality of geographic areas of interest based on the feature of interest scores;
• and
• indicating the suitability for exploration of the at least one of the plurality of geographic areas of interest relative to at least another one of the plurality of geographic areas of interest.

A non-transitory computer-readable medium comprising program code that is executable by a processor for causing the processor to:

• access training data including information relating to a modern feature of interest, the training data compiled from direct observations, remotely determined data, or a combination thereof, and including a global magnitude map representing the information relating to the modern feature of interest;
• compile parameter data from at least one modern model simulation that impacts the modern feature of interest, the modern model simulation comprising a hydrosphere model, an atmospheric model, a heliosphere model, a geosphere model, a cryosphere model, a biosphere model, a general circulation model, an earth system model, or a combination thereof;
• generate a predictive model from a machine-learning model by training the machine-learning model on the training data and the compiled parameter data from the at least one modern model simulation;
• output, by the predictive model, a predicted feature of interest in a past time period by applying the predictive model to compiled parameter data from at least one existing historical earth system factor model simulation that impacts the feature of interest in the past time period, wherein the predicted feature of interest includes a global magnitude map;
• based on the predicted feature of interest in the past time period output by the predictive model, calculate a feature of interest score for each of a plurality of geographic areas of interest;
• determine a suitability for exploration of at least one of the plurality of geographic areas of interest based on the feature of interest scores;
• and
• indicate the suitability for exploration of the at least one of the plurality of geographic areas of interest relative to at least another one of the plurality of geographic areas of interest.