Exploring the Spatiotemporal Evolution of the Eastern China Plateau in the Mesozoic Through Machine Learning

The possible uplift of a plateau in eastern China during the Mesozoic was closely associated with variations in crustal thickness. High-Sr and low-Y granites, regarded as products of thickened lower crust melting, have been used to infer the evolution of the plateau, but their formation conditions and implications remain debated. We trained a crustal thickness estimation model by employing machine learning methods and integrating multiple geochemical features. The model achieved an R2 of 0.87 on the test set and was consistent with previous crustal thickness constraints in external validations in the Gangdese orogen and Andes arc. Based on the model and igneous rock data from the North China Craton and its surrounding orogenic belts, we reveal that crustal thickening during the Middle Jurassic facilitated the uplift of a proto-plateau centered on the Jiaodong–Liaoning region. By the Late Jurassic, the plateau matured, with a crustal thickness of 55 ± 3 km, a paleo-elevation of ~4.2 km, and likely coverage spanning the entirety of the studied areas in eastern China. During the Early Cretaceous, extensional tectonics led to the plateau’s collapse, ultimately evolving into plains. This study provides new evidence for the existence and evolution of the Mesozoic plateau in eastern China.