Remote sensing based assessment of land degradation neutrality across China: Multi-scale patterns and drivers

Land Degradation (LD) compromises essential ecosystem services and threatens human wellbeing, posing a significant challenge to global sustainable development. Although existing research has extensively evaluated Land Degradation Neutrality (LDN), assessments are often limited to single scales or local contexts, overlooking spatial heterogeneity, scale effects, and the specific drivers of degradation. To address this research gap, our study conducts a multi-scale evaluation of LD across China, employing remote sensing data and the indicators outlined in the Sustainable Development Goal (SDG) 15.3.1. This assessment evaluates LD across geospatial, administrative, and climatic scales by comparing the reporting period of 2011–2020 with the 2000–2010 baseline period. Subsequently, the Geodetector model was applied to identify the scale-dependent driving factors. Under the SDG 15.3.1 evaluation framework, the area of newly improved land (1.85 × 10 6 km 2 ) exceeded that of newly degraded land (1.05 × 10 6 km 2 ), and China is identified as meeting its LDN target during the study period. However, our multi-scale analysis reveals that, at their respective scales, 68 cities, six provinces, and one climate zone failed to achieve LDN. Furthermore, the results suggest that the status of LD is strongly scale-dependent, with a spatially concentrated distribution in arid and semi-arid regions, a pattern primarily driven by rainfall variability. Overall, the results reveal the scale-dependent characteristics of LD, thereby informing the development of more tailored land management policies and ecosystem rehabilitation efforts. • Multi-scale land degradation neutrality (LDN) assessment of China was conducted • National LDN target achieved, but sub-national disparities remain • Spatial patterns of degradation and LDN achievement are mismatched • The main driver of land degradation show similarities with variations across scales • The current LDN framework is practical yet retains uncertainties and limitations