Multi-scale analysis of cultivated land use system resilience and its obstacle factors in Hunan Province of China

Cultivated land use system resilience can hold strategic significance for the national food security in sustainable agriculture. This study aims to conduct research on the resilience of cultivated land use systems and the spatiotemporal heterogeneity of the obstacle factors on the ecological risks in the agricultural economy. Hunan Province was selected as the study area. A resilience evaluation index system was established under the "Pressure-State-Response" (PSR) framework. Quantum Genetic Algorithm-Projection Pursuit (QGA-PP) model, kernel density estimation, and obstacle factor diagnosis were employed to consider the multi-scale "province-agricultural area-county" system. A systematic investigation was made into the resilience evolution patterns and constraint mechanisms in the cultivated land use system. Results indicated that: (1) Temporal Dynamics of Resilience: The resilience level of the cultivated land use system exhibited a phased fluctuation pattern in the regions. The composite resilience index demonstrated an initial increase from 1.86 in 2008 to a peak of 2.18 in 2015, followed by a gradual decline to 2.06 by 2022; South-Central Hunan also demonstrated the resilience enhancement, while the absolute gaps widened between South-Central and Western Hunan. The Changsha-Zhuzhou-Xiangtan metropolitan exhibited aggravated polarization, compared with the convergence trends in the Dongting Lake basin. (2) The Xiangxi hilly areas were featured by a highly resilient farmland, due to weak human disturbances and efficient disaster response. Meanwhile, the high resilience values were shifted into the Central-Southern Hunan water-saving areas, thus benefiting from the superior cultivated land resources and greater investment. In contrast, the Changsha-Zhuzhou-Xiangtan metropolitan demonstrated a weaker system resilience, due to the protection of arable land to resist disturbances and conflicts. (3) The obstruction degrees among the provinces were ranked in descending order of: pressure layer > response layer > state layer at the criteria level. Notably, the decreasing trend was observed in the obstruction degree of the response layer, while that of the state layer also displayed an increasing tendency. The key obstruction factors included the precipitation anomaly value, self-calibrated Palmer Drought Severity Index, per capita arable land area, agricultural fiscal expenditure per unit of cultivated land area, and agricultural disaster governance effectiveness. (4) Spatial heterogeneity of key constraint factors was observed in the diverse agricultural zones. A three-level differentiated resilience strategy was proposed to improve the cultivated land use systems in the province-agricultural area-county. The finding can provide a scientific reference for the three-level synergistic cultivated land protection (province–agricultural zone–county) in Hunan Province.