Little is known about whether changes in lake ecosystem structure over the past 150 years are unprecedented when considering longer timescales. Similarly, research linking environmental stressors to lake ecological resilience has traditionally focused on a few sentinel sites, hindering the study of spatially synchronous changes across large areas. Here, we studied signatures of paleolimnological resilience by tracking change in diatom community composition over the last 2000 years in four Ecuadorian Andean lakes with contrasting ecoregions. We focused on climate and anthropogenic change, and the type of biological responses that these changes induced: gradual, elastic, or threshold. We combined multivariate ordination techniques with nonlinear time-series methods (hierarchical generalized additive models) to characterize trajectories of community responses in each lake, and coherence in such trajectories across lakes. We hypothesized that remote, high-elevation lakes would exhibit synchronous trends due to their shared climatic constraints, whereas lower elevation lakes would show less synchronous trends as a consequence of human density and land-cover alteration. We found that gradual and elastic responses dominated. Threshold-type responses, or regime shifts, were only detected in the less remote lake, after a long period of gradual and elastic changes. Unexpected synchrony was observed in diatom assemblages from geographically distant and human-impacted lakes, whereas lakes under similar broad-scale environmental factors (climate and ecoregion) showed asynchronous community trajectories over time. Our results reveal a complex ecological history and indicate that Andean lakes in Ecuador can gradually adapt and recover from a myriad of disturbances, exhibiting resilience over century to millennial timescales.

中文翻译：

---

热带安第斯湖泊的生态恢复力：古湖沼学观点

在考虑更长的时间尺度时，关于过去 150 年湖泊生态系统结构的变化是否是前所未有的，我们知之甚少。同样，将环境压力源与湖泊生态恢复力联系起来的研究传统上集中在几个哨点，阻碍了对大面积空间同步变化的研究。在这里，我们通过跟踪过去 2000 年厄瓜多尔安第斯山脉四个具有对比生态区的湖泊中硅藻群落组成的变化，研究了古湖沼恢复力的特征。我们专注于气候和人为变化，以及这些变化引起的生物反应类型：渐进、弹性或阈值。我们将多元排序技术与非线性时间序列方法（分层广义加法模型）相结合，以表征每个湖泊中群落响应的轨迹，以及跨湖泊的此类轨迹的连贯性。我们假设偏远的高海拔湖泊由于共同的气候限制而表现出同步趋势，而低海拔湖泊由于人类密度和土地覆盖变化而表现出较少的同步趋势。我们发现渐进和弹性反应占主导地位。经过长时间的渐进和弹性变化后，仅在不太偏远的湖泊中检测到阈值型反应或政权转变。在来自地理遥远和受人类影响的湖泊的硅藻组合中观察到意外的同步，而在类似的大尺度环境因素（气候和生态区域）下的湖泊随着时间的推移显示出异步的群落轨迹。我们的研究结果揭示了一个复杂的生态历史，并表明厄瓜多尔的安第斯湖泊可以逐渐适应并从无数干扰中恢复，在一个世纪到千年的时间尺度上表现出弹性。