| Abstract |
As the second-largest sedimentary basin in China, the Ordos Basin holds considerable importance for the optimization and sustainable development of the regional energy structure through its geothermal resources. However, the accurate assessment of these resources has been hampered by the sparse and uneven distribution of terrestrial heat flow measurements, which restricts further exploration and development. To address this limitation, this study employs an XGBoost machine learning model to predict the heat flow across the basin. We integrated global and Chinese heat flow datasets with key geological and geophysical parameters—including gravity anomalies, magnetic anomalies, and Moho depth—to construct a model training dataset. The model was validated using reliable Class A heat flow data from the North China Craton. The resulting regional-scale heat flow prediction provides critical support for the future exploration, development, and efficient utilization of geothermal resources in the region. |