Evaluation of the long-term effect of ecosystem recovery projects is critical for future ecological management and sustainable development. The Three-North Shelterbelt (TNS) is a large-scale afforestation project in a crucial region of China. Numerous researchers have evaluated the vegetation ecological quality (VEQ) of the TNS using a single vegetation indicator. However, vegetation ecosystems are complex and need to be evaluated through various indicators. We constructed the vegetation ecological quality index (VEQI) by downscaling net primary productivity, leaf area index, fractional vegetation cover, land surface temperature, vegetation moisture, and water use efficiency of vegetation. The spatiotemporal characteristics and main contributing factors of VEQ in the TNS from 2000 to 2020 were investigated using SEN+Mann−Kendall, Hurst exponent, geographical detector, and residual trend analysis testing. The results suggest that VEQ in the TNS showed an improving trend over the 21-year study period. The areas with significant improvements were concentrated in the central and eastern parts of the TNS. Significant deterioration occurred only sporadically in various urban areas. Characteristics of future unsustainable VEQ trends could be detected across the TNS. Precipitation, vegetation type, soil type, elevation, and solar radiation exhibited the greatest impact on VEQ throughout the TNS. Human activities (e.g., afforestation and government investments) were the dominant factors and had a relative contribution of 65.24% to vegetation area change. Our results provide clues for assessing environmental recovery and sustainable development in other regions.

Small lakes (areas between 0.01 km² and 1 km²) on the Qinghai–Tibet Plateau (QTP) are prone to fluctuations in number and area, with serious implications for the surface water storage and water and carbon cycles of this fragile environment. However, there are no detailed long-term datasets of the small lakes of the QTP. Therefore, the intra-annual changes of small lakes in the Qilian Mountains region (QMR) in the northeastern part of the QTP were investigated. The small lake water bodies (SLWB) in the QMR were extracted by improving existing commonly used waterbody extraction algorithms. Using the Google Earth Engine platform and 13,297 Landsat TM/ETM + /OLI images, the SLWB of the QMR were extracted from 1987 to 2020 applying the improved algorithm, cross-validation and manual corrections. The reliability, uncertainty and limitations of the improved algorithm were discussed. An intra-annual small lake dataset for QMR (QMR-SLD) from 1987 to 2020 was released, containing eight attributes: code, perimeter (km), area (km²), latitude and longitude, elevation (m), area error, relative error (%), and subregion.