生态系统状态跃变的早期预警信号及检测

生态系统在气候变化和土地利用及人类活动等的影响下其状态会由某一稳态转变到另一稳态。由于环境压力的复杂性、非线性、随机性等特征,往往导致状态转变表现为非线性、突变、跃变等特点。准确界定系统状态跃变的拐点或阈值点存在很大的挑战,而捕捉接近临界拐点前的生态系统结构和属性上的变化特征作为早期预警信号是切实可行的。早期预警信号理论经历理论框架构建、方法确立、机理认知等近半个多世纪的探索,已经由最初的通过仅依赖检测临界点恢复力的速率减慢、方差增加、系统自相关增强等统计学信号过度到更加多样化的检测方法,如检测系统组分属性的变化特征,诊断系统组分各属性之间的关系变化,系统组分的性状变化、系统组分网络结构变化等等,并且试图整合多信号提高预警的精确性。利用来自自然生态系统的长时间高密度数据集和空间代替时间的数据集,基于多度及性状信号的早期预警,结合稳定性、临界恢复力的减速、以及统计参数的指示作用对系统跃变进行早期诊断和预警是预测生态学的主旨。早期预警信号的深入研究不仅能够完善已有理论的不足,同时还能够为生态系统的保护和管理提供切实有效的理论指导。

Testing on the early warning signals in approach to ecosystem state transition

Shifts in ecosystem states are common phenomena as ecosystems in general experience climate fluctuations and land use changes, and anthropogenic perturbations. Such environmental pressures are characterized by complexity, non-linear, and stochasticity which usually result in non-linear transitions and catastrophic shifts in ecosystem states. Studies on early warning signals (EWSs) focus on capturing the information from ecosystem structure and attributes as the system is approaching to the critical transition under the control of environmental stresses. For instance, large scale spatial regular configuration of vegetation patches could be as early warning signals to indicate the imminent transition of grassland to desertification. Moreover, labyrinth of bushy vegetation, striped pattern of bushy vegetation, labyrinth of perennial grasses, spots of patches, and regular maze patterns of shrubs and trees have been recorded in natural ecosystems under drought and resource stresses. Meanwhile, some statistical parameters in the models can also present some generic characteristics as the system approaches to the state transition, which were treated as signals indicating ecosystem catastrophic transitions. All such efforts contribute to constructing the framework of the classic EWSs theory. The signals, such as critical slowing down, increase of variance in fold bifurcations, and high autocorrelations, were used to test catastrophic shifts in ecosystem states in real nature. Furthermore, as considering the mechanisms relevant to ecosystem state transition, self-organized patchiness underlying positive feedback between resources and consumers was considered as an important mechanism linking spatial patch configurations and catastrophic shifts. After the rapid development in the past half century, the theory of early warning signals is making progress towards the direction with diverse testing methods and in trying to integrate multiple signals. All such efforts aim to improve the accuracy of predictions for ecosystem state shifts. Ecosystem state shifts can cause severe losses of ecological and economic resources, and restoring a desired state may require high investment. Therefore, neglect of the possibility of shifts to alternative stable states in ecosystem may have heavy costs to society. The key purpose of the predictive ecology is to identify early warning signals indicating the upcoming of ecosystem crisis by integrating abundance-based and trait-based signals into stability, critical slowing down, and statistical parameters. Long-time series data sets and space-for-time data sets are absolutely necessary. Linkage between self-organized patchiness and catastrophic shifts cannot only bridge gaps in EWSs theory but also provide help to biodiversity conservation and ecosystem management.