Nature reserves (NRs) play a pivotal role in minimizing habitat loss and protecting wild animals and plants, which are critical for human ecological security. However, focusing only on the construction of ecological security patterns of NRs without understanding their ecological security early-warning situations and their driving factors may fail to achieve protection goals. This study constructed an ecological security early-warning framework and index system based on the Driving force-Pressure-State-Impact-Response (DPSIR) framework model. The gray model (GM) was used to predict the ecological security early-warning situation, and the Geo-detector model was applied to explore the driving factors of the ecological security early-warning system in the Baishuijiang National Nature Reserve (BNNR). The results showed that the average ecological security index (ESI) value increased from 0.2796 in 2005 to 0.3171 in 2017, with an average increase of 11.82%. The ecological security early-warning index (ESEWI) value increased from 0.3171 in 2018 to 0.3622 in 2030, which was an average increase of 12.46%. These results indicated that the ecological security situation continually improved from 2005 to 2030. By 2030, the number of towns with a “no warning” grade increased to four, the number of towns with an “extreme warning” grade was zero, and the proportion of areas with early-warnings decreased from 100% to 33%. The q values of per capita forest land areas and per capita grassland areas were both 0.9334, which indicated that environmental characteristic factors were the primary driving factors in ecological security early-warning. Our results demonstrated that the ecological security early-warning index system based on the DPSIR model and grey model can well prediction ecological security situation and provide scientific support for the ecological protection and management of NRs.

自然保护区（NRs）在减少栖息地丧失和保护野生动植物方面发挥着举足轻重的作用，这对人类生态安全至关重要。然而，只关注自然保护区生态安全格局的构建，而不了解自然保护区的生态安全预警态势及其驱动因素，可能无法实现保护目标。本研究基于驱动力-压力-状态-影响-反应（DPSIR）框架模型构建了生态安全预警框架和指标体系。利用灰色模型（GM）预测白水江国家级自然保护区生态安全预警态势，应用Geo-detector模型探索白水江国家级自然保护区生态安全预警系统的驱动因素。结果显示，平均生态安全指数（ESI）值从2005年的0.2796上升到2017年的0.3171，平均增幅为11.82%。生态安全预警指数（ESEWI）值从2018年的0.3171上升到2030年的0.3622，平均增幅为12.46%。这些结果表明，从2005年到2030年，生态安全形势持续好转。到2030年，“无预警”等级的乡镇数量增加到4个，“极端预警”等级的乡镇数量为0，所占比例有预警的地区从 100% 下降到 33%。这 平均增幅为 12.46%。这些结果表明，从2005年到2030年，生态安全形势持续好转。到2030年，“无预警”等级的乡镇数量增加到4个，“极端预警”等级的乡镇数量为0，所占比例有预警的地区从 100% 下降到 33%。这 平均增幅为 12.46%。这些结果表明，从2005年到2030年，生态安全形势持续好转。到2030年，“无预警”等级的乡镇数量增加到4个，“极端预警”等级的乡镇数量为0，所占比例有预警的地区从 100% 下降到 33%。这人均林地面积和人均草地面积q值为0.9334，说明环境特征因子是生态安全预警的主要驱动因素。研究结果表明，基于DPSIR模型和灰色模型的生态安全预警指标体系能够很好地预测生态安全态势，为自然保护区的生态保护和管理提供科学支撑。