GEOTHERMAL PLAY DEVELOPMENT USING MACHINE LEARNING AND MAGNETOTELLURICS DATA

This talk will demonstrate a coupling strategy between an unsupervised machine learning technique, non-negative matrix factorization with k-means clustering (NMFk), and inverted magnetotellurics (MT) data to find geothermal drilling locations and depth. We used geothermal data of the Tularosa Basin of south-central New Mexico. NMFk automatically finds four groups/signatures representing latent structure in the dataset. Post-processing of results helps find potential highly prospective locations for geothermal resources near White Sands Missile Range and McGregor Range at Fort Bliss. MT is used to detect the potential depth of geothermal prospects at McGregor Range based on apparent resistivity structures/layers in the subsurface. The McGregor Range consists of three resistivity layers and two resistivity structures. MT also helps identify that the western portion of the McGregor Range has thick, low resistivity earth materials. Moreover, the same portion has a high thermal gradient, and the reservoir at 2km depth may contain high-temperature geothermal resources. Thermal gradient data possesses high uncertainty. Even if the thermal gradient is lower than the reported value, the western McGregor Range may still have high-temperature geothermal resources at a greater depth than 2km.