A 30 m resolution dataset of soil and water conservation terraces across China for 2000, 2010, and 2020

Abstract. Terraces, as one of the most widely distributed and heavily invested soil and water conservation (SWC) measures in China, currently lack a comprehensive database containing spatiotemporal distribution and diverse classification types. This absence significantly hampers the accurate soil erosion assessment and SWC planning in China. To address this gap, we developed a two-stage mapping framework to classify various terrace measures and produced a new dataset named the Soil and Water Conservation Terrace Measures Dataset (SWCTMD). The dataset, spanning the years 2000 to 2020, was produced by integrating time-series Landsat imagery and digital elevation model data. The data incorporate SWC measure factors and four terrace types: level terraces, slope terraces, zig terraces, and slope-separated terraces. On average, the SWCTMD achieved OA of 91.7 % and F1 of 83.3 % for terraces, and 89.4 % OA and 78.9 % F1 for different terrace types, underscoring its high accuracy in terrace mapping. Comparative analysis demonstrated the superior robustness of the SWCTMD compared to existing products. This dataset demonstrated that terraces in China are predominantly concentrated in the Loess Plateau, Southwest and Southeast regions. From 2000 to 2020, the total terrace area increased by 41 594.1 km2, with slope terraces exhibiting the largest expansion, while decreases were primarily observed in peri-urban areas. Notably, the modeling results indicated that terraces had reduced soil erosion of cropland by approximately 1390 million tons in 2020. The SWCTMD can be employed to enhance the accuracy of soil erosion simulations and support long-term analysis of soil erosion trends. Furthermore, the dataset provides valuable applications for earth system modelling and contributes to research on land resource management, food security, biodiversity, and water cycle. The SWCTMD is freely available at https://doi.org/10.11888/Terre.tpdc.302400 (Duan, 2025).