Methods and systems for subsurface modeling employing ensemble machine learning prediction trained with data derived from at least one external model

A computer-implemented method for creating a model of a subsurface geological formation, comprising:

• training a machine learning ensemble comprising a plurality of machine learning models using training data as well as additional data derived from an external model, the additional data including an estimate of a value of a target variable derived from a cross validation of the value of the target variable at a location used to construct a respective machine learning model of the plurality of machine learning models, wherein such additional data provides training information that supplements the training data, and wherein the additional data derived from the external model comprises the estimate of the value of the target variable;
• constructing an extended vector of the training data, the extended vector comprising:
• a first secondary variable including a first observation at the location;
• a second secondary variable including a second observation at the location, the second secondary variable having a non-linear relationship with the first secondary variable;
• and
• the estimated value of the target variable at the location;
• using the trained machine learning ensemble to create the model based on the extended vector, the model relating to fluid flow in the subsurface geological formation;
• and
• modifying a drilling operation location in the subsurface geological formation based on the model.