Shallow lakes are one of the most complex aquatic systems and are known to shift between a macrophyte-dominated clear-water state and a phytoplankton-dominated turbid state. We developed a decision-tree classification based on the normalized difference vegetation index (NDVI) to classify the major vegetation cover types in shallow lakes and documented the vegetation in Gehu Lake using Landsat time series images from 1984 to 2018. The area of aquatic vegetation in Gehu Lake (a large shallow lake in China) showed a significant decreasing trend (R-2 = 0.839, p < 0.001) from 1984 to 2018, which was significantly positively correlated with water clarity (R-2 = 0.864, p < 0.001) and negatively correlated with the total nitrogen (TN) and total phosphorus (TP) concentrations (R-2 = 0.704, p < 0.001; R-2 = 0.724, p < 0.001, respectively). A complete regime shift from a clear macrophyte-dominated state to a turbid phytoplankton-dominated state was observed in 2004. Our results revealed that regime shifts in shallow lakes can be divided into two stages: changes in important variables make the lake ecosystems more susceptible and into a critical stage; in the critical stage, any shock events or actions by external drivers may induce a regime shift. Abnormal and large-area changes in aquatic vegetation may be used as early warning signals of the degradation of ecosystem resilience and impending regime shifts. Remote sensing is powerful for monitoring the dynamics of ecosystems and understanding the regime shifts in shallow lakes. Strengthening long-term ecological monitoring and developing new monitoring technologies may improve ecosystem management and conservation in shallow lakes

作者:Xu, X (Xu, Xuan); Zhang, YB (Zhang, Yibo); Chen, Q (Chen, Qiao); et al. ECOLOGICAL INDICATORS,2020,DOI: 10.1016/j.ecolind.2020.106285