60-year (1964 to 2024) retrogressive thaw slump inventory in the Beiluhe region of the central Qinghai–Tibet Plateau

The dataset is retrogressive thaw slump inventories with annual intervals in the Beiluhe region of the central Qinghai–Tibet Plateau from 1964 to 2024. It contains the boundaries of thaw slumps delineated yearly based on the KH-4 and -9 images from 1965 to 1984, Landsat imagery archives from 1986 to 2022, Worldview images from 2007 to 2014, SPOT5 images from 2009 to 2011, RapidEye images in 2015, and PlanetScope images from 2016 to 2024. The inventories were compiled based on manual delineation and deep learning. We merged all RTSs from the annual inventory and grouped those with spatial overlaps. Each group was assigned a unique identifier (‘uid’). Then, we assigned a secondary identifier (‘sid’) to distinguish separate RTSs within the same group. The group identifier together with the secondary identifier formed the final ‘id’, which is in the form of ‘uid_sid’. We retained the same ‘id’ to maintain tracking continuity for reactivated thaw slumps (those recurring at the same location). To fill the temporal gaps when the satellite image archives were incomplete, we reserved the RTS boundaries in the previous year and labeled the RTS as ‘low’ confidence. The name of the file is “RTS_YYYY”, with the ‘YYYY’ representing the year of the boundaries. Relevant attribute tables include area (units: m2), ground identifier, secondary identifier, and the unique identifier, with the corresponding names of the table fields 'area', 'uid', 'sid', ‘id’, and 'confidence’.