A dataset of remote sensing images and interpretation data for typical floods in Central and Western Asia and Northwest China along the Belt and Road

As an important region along the Belt and Road, Central and Western Asia and Northwest China are jointly facing severe challenges to regional sustainable development caused by flood disasters. Given the limitations of existing flood disaster information in the region, including its limited diversity and insufficient analytical extraction, this study constructs a dataset of flood remote sensing image and interpretation based on radar images of typical floods that occurred between 2017 to 2022 in Central and Western Asia and Northwest China along the Belt and Road. An object-oriented method was applied to extract water body information before and after flood events, and change detection techniques were used to determine the flood inundation extent. The results were further validated using optical imagery. The resulting dataset aims to address the limitations of existing flood disaster information, which is incomplete and lacks sufficient detail. This dataset includes two components: a historical flood disaster dataset and a flood interpretation dataset: (1) The historical flood disaster dataset includes typical historical flood disaster records from Central and Western Asia and Northwest China along the Belt and Road, corresponding Excel data, and DEM data. (2) The flood interpretation dataset includes SAR image data covering the extent of flood inundation, optical image data, and interpreted annotation data. The overall spatial coverage of flood events ranges from 22°42′25″N to 47°22′52″N and 39°21′54″E to 101°17′24″E, including 10 typical flood disaster samples. By incorporating Landsat optical imagery for auxiliary validation and refining the results through manual interpretation feedback, the accuracy of flood inundation extent extraction has been significantly improved. The proportion of correctly classified samples consistently exceeds 85%, effectively reducing information omissions and misclassifications, and demonstrating the reliability and completeness of the dataset. The dataset is derived from clearly defined data sources and undergoes strict quality control procedures. It provides abundant sample data for hydrological research, disaster loss assessment, and disaster risk evaluation in Central and Western Asia and Northwest China, serving as a valuable data resource for flood monitoring and disaster management in these regions.