1. Biodiversity is structured in space and time, yet our understanding about the temporal variation of biological communities is still limited. Recent work suggests that temporal β diversity should be lowest in unpredictable systems with low seasonality, as expected for tropical ecosystems. However, this hypothesis remains largely unexplored.
2. Here, we analysed the temporal dynamics of Atlantic Forest coastal stream caddisfly assemblages sampled simultaneously along four years, resulting in 99 unique site-time combinations. Given that the region we studied is expected to have low climatic seasonality and low environmental harshness, we expected to find: (1) unseasonal variation in community composition; (2) equal contribution of nestedness and turnover to β diversity; (3) weak correlation between variation in community composition and in environmental variables; and (4) temporal β diversity similar to that expected by neutral dynamics.
3. Using wavelet analysis we found that, in spite of general temperature stability, precipitation was highly seasonal but unpredictable at the monthly scale. We found high temporal β diversity, with a larger contribution of nestedness due to an overall decrease in abundance and diversity during months characterised by heavy rain events. We also found that community composition was moderately structured in space, suggesting that both local in-stream characteristics and dispersal limitation drive community reassembly after heavy rain events.
4. Null model analysis indicated that observed short-term β diversity was not distinct from that expected by chance (i.e. produced by null simulations). This suggests an important role of stochastic colonisation and dispersal limitation in the temporal organisation of Atlantic Forest stream assemblages.
5. Since heavy rain events strongly disassemble communities that are then randomly recolonised, and considering the current scenario of climate change, increased extreme events of precipitation could lead to unpredictable changes in biodiversity in tropical regions.

中文翻译：

---

随机殖民动态可能是大西洋森林沿海溪流群落时间β多样性的主要驱动力

1. 生物多样性是在空间和时间上构建的，但我们对生物群落的时间变化的理解仍然有限。最近的工作表明，正如热带生态系统所预期的那样，在季节性低的不可预测的系统中，时间 β 多样性应该是最低的。然而，这一假设在很大程度上仍未得到探索。
2. 在这里，我们分析了四年内同时采样的大西洋森林沿海溪流石蛾组合的时间动态，产生了 99 个独特的站点时间组合。鉴于我们研究的地区预计气候季节性低，环境恶劣程度低，我们预计会发现：（1）群落组成的非季节性变化；(2) 嵌套和周转对β多样性的贡献相等；(3) 群落组成变化与环境变量之间的相关性弱；(4) 时间 β 多样性类似于中性动力学所预期的多样性。
3. 使用小波分析，我们发现，尽管总体温度稳定，但降水具有很强的季节性，但在月尺度上不可预测。我们发现高时间 β 多样性，由于在以大雨事件为特征的月份中丰度和多样性总体下降，嵌套性的贡献更大。我们还发现群落组成在空间中结构适中，这表明局部河道特征和扩散限制都推动了暴雨事件后群落重组。
4. 零模型分析表明，观察到的短期 β 多样性与偶然预期的（即由零模拟产生的）没有区别。这表明随机殖民化和扩散限制在大西洋森林河流组合的时间组织中起着重要作用。
5. 由于暴雨事件强烈地瓦解了然后随机重新定居的社区，并考虑到当前的气候变化情况，极端降水事件的增加可能导致热带地区生物多样性发生不可预测的变化。