Identifying spatial and temporal dynamics and driving factors of cultivated land fragmentation in Shaanxi province

The cultivated landscapes in China have shown a tendency toward fragmentation, leading to a substantial decrease in the productivity of cultivated land and hindering agricultural modernization. Shaanxi province, a prominent agricultural region in China, has become severely affected by cultivated land fragmentation (CLF). To ensure national food security and promote sustainable ecological protection strategies, it is crucial to understand the spatial and temporal dynamics as well as the underlying drivers of CLF. This study focused on Shaanxi province, Northwest China, and identifyied a suite of landscape indices suitable for analyzing the evolution and driving factors of CLF. The objectives of this research are to identify the optimal spatial scale and driving factors of CLF. Using a suite of landscape pattern indices (class area, patch density, mean perimeter-area ratio, aggregation index, largest patch index, and landscape shape index), the semivariogram model (SVM) in geostatistics, the moving window method in Fragstats v.4.2–64, and Geodetector, we investigated the spatio-temporal dynamics and influencing factors of the CLF in Shaanxi province. The optimal granularity scale and extent scale were 90 m and 3000 m, respectively. Shaanxi province exhibited a high level of CLF, with the composite cultivated land fragmentation index (CLFI) increasing from 0.598 in 2000 to 0.622 in 2020, corresponding to an annual average increase of 0.11%. This increase poses a considerable threat to the sustainable and healthy development of agricultural resources in the region. Population density, landform type, and elevation were identified as the three primary factors of CLF in Shaanxi province, with influence powers of 0.572, 0.414, and 0.414, respectively. CLF has resulted from the combined influence of multiple factors, including natural, economic, and social factors. This study established a CLF identification model, classified CLF types, quantitatively reflected CLF processes at the pixel level, and revealed CLF types, areas, and spatial distribution characteristics. This provides not only the foundation for the quantitative evaluation of CLF processes in spatial and temporal terms but also a theoretical reference for the improvement of cultivated land resources and reform the of rural land systems.