湖泊稳态转换是当今水生生态系统研究的热点与前沿，其中干旱区浅水湖泊更易受到人类活动的侵扰，有关该类湖泊稳态转换预警识别及驱动机制的研究尚属薄弱领域。本项目选择黄河流域最大的淡水湖泊乌梁素海为研究区，基于灌区-河流与湖泊连续体系统观点，采用资料收集-野外观测-遥感解译-模型构建-机制识别-稳态调控一体化方法，结合古湖沼测年与孢粉记录等技术，开展浅水湖泊水盐平衡分析，识别灌区与流域尺度多因素协同作用下稳态转换突变年份与预警因子，继而引入干旱浅水湖泊特征因子，构建湖泊生态模型，揭示水文条件与营养盐负荷交互及协同作用下的稳态转换驱动机制，界定其生态阈值范围，明晰其非线性、迟滞性、多阈值、多稳态等特征，最终采用可控驱动因子多情景模拟优化耦合分析，提出稳态转换联合调控策略，旨在丰富完善干旱区湖泊生态系统理论，为湖泊富营养化防控与修复提供科学依据与实践参考。

Regime shift, defined as rapid and dramatic change in the structure and function of lake ecosystems, is the hot topic and frontier domains of aquatic ecosystem. The shallow lake in the arid region is prone to be disturbed by human activities, however, the research on driving mechanisms of regime shifts due to human-induced pressures, such as hydrological regulation and nutrient loadings, is rarely reported. Lake Wuliangsuhai, the largest freshwater lake in the Yellow River Basin, is chosen as study area. In considering of the inflow of waste water from irrigation region and runoff from adjacent catchments, the history data and remote sensing data are collected, and in-situ sampling tests of water quantity, water quality and nutrients release from sediments are performed. Multi-proxy paleolimnological records are used to obtain long time series ecological data and determine the inflection points of regime shifts in Lake Wuliangsuhai. First, the sediment chronology is established, and macrophyte pollen as well as nutrient conditions in each dated sediment layer are identified. Then the heuristic segmentation algorithm and Pettitt test are used to determine the most significant inflection points of regime shifts. Additionally, multiple early warning indicators including variance, autocorrelation and skewness are used to test their ability to forecast the major regime shifts. The modified ecological model of PCLake is built to simulate the annual changes of the indicator of regime shifts. The combination of paleolimnological records and ecosystem modeling can overcome the data insufficiency and unravel the drivers and underlying mechanisms in ecosystem dynamics. Using model scenarios, we can further disentangle the roles of hydrological regulation and nutrient loading, disclose the nonlinearity, hysteresis, multi-stable state and multi threshold of regime shift, and suggest the feasible management strategy in the shallow lake of arid region. Overall, our study will enrich the theory of aquatic ecosystem and provide the scientific reference for the control and restoration of lake eutrophication.