On quantifying the apparent temperature sensitivity of plant phenology

• Published in: The New phytologist Vol. 225; no. 2; pp. 1033 - 1040
• Main Authors: Keenan, Trevor F., Richardson, Andrew D., Hufkens, Koen
• Format: Journal Article
• Language: English
• Published: England Wiley 01.01.2020
• Subjects: autumn; bud‐burst; climate; ecological function; ecosystem; Environmental Sciences; growing degree days; land surface; Life Sciences; Methods; phenology; seasonal variation; spring; temperature; uncertainty; warming; spring; ecosystem; bud-burst; growing degree days; warming; land surface; autumn
• ISSN: 0028-646X; 1469-8137; 1469-8137
• DOI: 10.1111/nph.16114

• Many plant phenological events are sensitive to temperature, leading to changes in the seasonal cycle of ecosystem function as the climate warms. To evaluate the current and future implications of temperature changes for plant phenology, researchers commonly use a metric of temperature sensitivity, which quantifies the change in phenology per degree change in temperature. • Here, we examine the temperature sensitivity of phenology, and highlight conditions under which the widely used days-per-degree sensitivity approach is subject to methodological issues that can generate misleading results. We identify several factors, in particular the length of the period over which temperature is integrated, and changes in the statistical characteristics of the integrated temperature, that can affect the estimated apparent sensitivity to temperature. • We show how the resulting artifacts can lead to spurious differences in apparent temperature sensitivity and artificial spatial gradients. Such issues are rarely considered in analyses of the temperature sensitivity of phenology. • Given the issues identified, we advocate for process-oriented modelling approaches, informed by observations and with fully characterised uncertainties, as a more robust alternative to the simple days-per-degree temperature sensitivity metric. We also suggest approaches to minimise and assess spurious influences in the days-per-degree metric.