新疆自然保护区景观格局时空演变与生态风险定量评价

Quantifying Spatiotemporal Landscape Patterns and Ecological Risks in Nature Reserves of Xinjiang

摘要: 目的：基于土地利用变化背景，定量识别新疆自然保护地景观格局与生态风险的时空演变特征及其空间关联性，为分区管控与生态安全格局优化提供依据。方法：基于1992、2000、2010和2020年土地利用/覆被数据，利用Fragstats 4.2提取景观格局指数构建景观生态风险模型，结合空间自相关方法分析生态风险的时空分布特征和空间集聚特征。结果：① 1992~2020年土地利用结构以裸地(57%)和草地(30%)为主，转化集中于裸地–草地–灌丛之间，灌丛比例由4.6%增至7.2%。② 低生态风险区面积占比始终超过24%，总体呈上升趋势。中等及以上等级面积分别净增723.87 km²、499.68 km²和884.52 km²，高质量生境占比1.34%。③ 综合生态风险指数长期维持在0.08以下，1992~2010年高风险区面积显著下降，2010~2020年局部区域出现回升。④ Moran’s I指数稳定在0.5以上，呈显著正空间自相关，高–高与低–低风险区块状集聚特征明显。结论：研究期内新疆自然保护地景观结构整体稳定，生态风险处于低值区间，风险演变与裸地–草地–灌丛结构调整密切相关。2010年后局部风险回升表明景观格局对气候波动与局地扰动具有阶段性响应特征，需加强重点区域动态监测与分级管控。

Abstract: Objective: This study aims to quantitatively characterize the spatiotemporal evolution and spatial correlation of landscape patterns and ecological risks in Xinjiang’s protected areas against a backdrop of land use change, thereby providing a scientific basis for zonal management and the optimization of ecological security patterns. Methods: Landscape ecological risk indices were constructed using landscape pattern metrics derived from land use/cover data for 1992, 2000, 2010, and 2020 via Fragstats 4.2. The spatiotemporal distribution and spatial clustering characteristics of ecological risk were subsequently analyzed using spatial autocorrelation methods. Results: The results showed that: (1) From 1992 to 2020, the land use structure was dominated by barren land (57%) and grassland (30%), with primary conversions occurring among barren land, grassland, and shrubland. The proportion of shrubland increased from 4.6% to 7.2%. (2) Low ecological risk zones consistently accounted for more than 24% of the study area and showed a general upward trend over time. The extents of moderate, high, and very high risk classes increased by 723.87 km², 499.68 km², and 884.52 km², respectively. High-quality habitats comprised only 1.34% of the total area. (3) The comprehensive ecological risk index remained consistently below 0.08. A significant decline in high-risk areas was observed from 1992 to 2010, followed by a localized rebound between 2010 and 2020. (4) Moran’s I index remained above 0.5, indicating significant positive spatial autocorrelation, with distinct agglomerations of high-high and low-low risk patches. Conclusion: Over the study period, the landscape structure of Xinjiang's protected areas remained relatively stable, with ecological risks confined to a low range. Risk evolution was closely linked to the dynamic adjustments among barren land, grassland, and shrubland. The localized increase in risk post-2010 suggests a periodic response of the landscape pattern to climatic fluctuations and local disturbances, underscoring the need for enhanced dynamic monitoring and differentiated management in key areas.