Abstract The Tibetan Plateau, known as the ‘Third Pole’ of the Earth, has become a hotspot for landform classification studies due to its young tectonics, complex terrain and diverse landforms. Strong internal and external forces have shaped highly distinctive landscapes, posing significant challenges to landform classification. Previous studies mainly focused on landform morphology, whereas classifications that integrate morphology and genesis remain limited. Existing genetic classifications of the Plateau are based mainly on visual interpretation, hardly meeting the demand for large‐scale automated landform classification. To address this issue, this study selected plains within the Tibetan Plateau as the study area and developed an automated classification method for plain genetic types by integrating multisource data. The results show that: (1) plains account for 25% of the Tibetan Plateau, with a mountain‐to‐plain ratio of about 3:1. (2) Fluvial and periglacial processes are the dominant external forces shaping the plains, with fluvial and periglacial plains comprising 51.73% and 26.07% of the total plain area, respectively, followed by lacustrine plains (10.55%), arid plains (7.55%), aeolian plains (3.21%) and loess plains (0.89%). (3) Accuracy evaluation results indicate that the classification accuracy for different genetic types of plains ranges from 75% to 91.89%, with an overall classification accuracy of 85.33%. Comparison with the 1:1 000 000 Geomorphological Atlas of China confirms that the overall distribution patterns are consistent, and the results of this study provide finer detail. The proposed hierarchical classification strategy and multisource data fusion framework provide a transferable approach for landform genetic classification in complex geomorphic regions.
Wei et al. (Thu,) studied this question.