Aarhus Institute of Advanced Studies, Aarhus University, Høegh-Guldbergs Gade 6B, DK-8000, Aarhus C, Denmark.
Rocky Mountain Biological Laboratory, P.O. Box 519, Crested Butte, Colorado, 81224, USA.
Ecology. 2017 Mar;98(3):647-655. doi: 10.1002/ecy.1690.
Time series have played a critical role in documenting how phenology responds to climate change. However, regressing phenological responses against climatic predictors involves the risk of finding potentially spurious climate-phenology relationships simply because both variables also change across years. Detrending by year is a way to address this issue. Additionally, detrending isolates interannual variation in phenology and climate, so that detrended climate-phenology relationships can represent statistical evidence of phenotypic plasticity. Using two flowering phenology time series from Colorado, USA and Greenland, we detrend flowering date and two climate predictors known to be important in these ecosystems: temperature and snowmelt date. In Colorado, all climate-phenology relationships persist after detrending. In Greenland, 75% of the temperature-phenology relationships disappear after detrending (three of four species). At both sites, the relationships that persist after detrending suggest that plasticity is a major component of sensitivity of flowering phenology to climate. Finally, simulations that created different strengths of correlations among year, climate, and phenology provide broader support for our two empirical case studies. This study highlights the utility of detrending to determine whether phenology is related to a climate variable in observational data sets. Applying this as a best practice will increase our understanding of phenological responses to climatic variation and change.
时间序列在记录物候如何响应气候变化方面发挥了关键作用。然而，将物候响应回归到气候预测因子中，存在着发现潜在虚假气候-物候关系的风险，因为这两个变量也随着年份的变化而变化。通过年份进行去趋势化是解决这个问题的一种方法。此外，去趋势化可以分离物候和气候的年际变化，因此去趋势化的气候-物候关系可以代表表型可塑性的统计证据。我们使用来自美国科罗拉多州和格陵兰岛的两个开花物候时间序列，对开花日期和两个已知在这些生态系统中很重要的气候预测因子（温度和融雪日期）进行去趋势化。在科罗拉多州，所有的气候-物候关系在去趋势化后仍然存在。在格陵兰岛，75%的温度-物候关系在去趋势化后消失（四个物种中的三个）。在这两个地点，去趋势化后仍然存在的关系表明，可塑性是开花物候对气候敏感性的一个主要组成部分。最后，创建了不同的年、气候和物候之间相关性强度的模拟，为我们的两个实证案例研究提供了更广泛的支持。这项研究强调了去趋势化在确定观测数据集中物候是否与气候变量相关的实用性。应用这一最佳实践将提高我们对物候对气候变化的响应的理解。
