Enhanced geothermal Resource assessment using geothermal heat flow predicted with machine learning method: Application to the sedimentary basin of Henan Province

Geothermal energy is a renewable and sustainable resource crucial to the clean energy transition. Henan Province, China, has considerable geothermal potential due to its complex sedimentary basins. Traditional assessment methods, such as the volumetric approach, provide static estimates and overlook spatial heat transfer. This study proposes the Thermal Property (TP) method, a machine learning-assisted approach integrating geothermal heat flow (GHF) and vertical thermal conductivity (k) to improve heat transfer evaluation. By integrating geological, geophysical, and drilling data, the TP method improves geothermal resource assessment accuracy. Results reveal high GHF values (>75 mW/m2) concentrated in northwest and central-northern Henan, peaking at 81.2 mW/m2 in Sanmenxia. Neogene and Paleogene reservoirs show strong geothermal potential, especially in subzone Ⅰ5, where fault-controlled heat transfer dominates. Sensitivity analysis indicates that reservoir area (A) and thickness (M) are key contributors to total heat storage, while deep heat flow plays a regional role. Strong spatial correlations between resource distribution and geological features—such as Moho depth, sediment thickness, and gravity anomalies—further validate the method. High-potential zones are linked to shallow Moho, thick sediments, and negative gravity anomalies. The TP method enhances geothermal resource evaluation, supporting more effective exploration and sustainable energy development.