Introduction to XJTU-LoessSlide/HWAL/392 database

Loess landslides are widespread geohazards on the Chinese Loess Plateau and can generate long runout, broad impact zones, and substantial risk to settlements and infrastructure. Open and consistently structured landslide inventories are needed for regional impact assessment, empirical runout evaluation, probabilistic modeling, and data-centric geohazard studies. This short communication introduces the XJTU-LoessSlide/HWAL/392 database, a curated inventory of 392 loess landslides in Baoji City, Shaanxi Province, China. The database contains landslide height (H), maximum width (W), planimetric area (A), length (L), together with location and lithology metadata. The landslide data were compiled from field investigation, remote sensing interpretation, digital elevation information, and GIS-based inventory analysis. The database covers wide ranges of landslide geometry, with H from 10 to 350 m, W from 7 to 2100 m, A from 0.024 to 709.36 × 10³ m², and L from 5 to 1400 m. Exploratory analysis shows strong positive dependence among the primary geometric variables, particularly between A and L, W and L, A and W, and H and W. The database is intended to support loess landslide impact-zone assessment, regional risk management, and future machine-learning and probabilistic modeling of landslide runout behavior.