Retooling Supervised Machine Learning for Data-Driven Hydrothermal Resource Assessments
US Geological Survey
Previous moderate- and high-temperature geothermal resource assessments of the western United States utilized data-driven methods and expert decisions to estimate resource favorability. Although expert decisions can add confidence to the modeling process by ensuring reasonable models are employed, expert decisions also introduce human and, thereby, model bias. This bias can present a source of error that reduces the predictive performance of the models and confidence in the resulting resource estimates. By using the same input data with several machine learning approaches, we create new resource favorability without the bias of the expert decisions. In so doing, we find that machine learning can produce resource favorability maps at least as well as, if not better than, the methods dependent upon expert decisions.