Overestimation of the effect of climatic warming on spring phenology due to misrepresentation of chilling

Huanjiong Wang, Chaoyang Wu, Philippe Ciais, Josep Peñuelas, Junhu Dai, Yongshuo H. Fu, Quansheng GeNature Communications 2020 · DOI

TL;DR

This paper reveals that over half of current models used to predict plant spring phenology are flawed because they misrepresent how winter chilling affects plant dormancy. These inaccurate models lead to significant overestimations of how much earlier spring events will occur due to climate warming, potentially impacting our understanding of ecosystem responses and carbon cycles.

What Problem Does This Paper Address?

Spring phenology, the timing of biological events like leaf unfolding and flowering, is a critical indicator of ecosystem health and is highly sensitive to climate change. While warming temperatures generally advance spring events, winter chilling also plays a crucial role. Plants require a certain amount of winter chilling to break dormancy and prepare for spring growth. This relationship is generally understood as inverse: more chilling means less heat is needed to trigger spring events (lower Heat Requirement, HR). However, the accuracy of models predicting this relationship, and consequently future phenology, is questionable. This paper addresses the critical gap in understanding whether current chilling models correctly capture this negative Chilling Accumulation (CA)-HR relationship, and what the implications are for future phenological predictions under climate change.

Methods

The researchers evaluated 12 different chilling models (C1-C12) using long-term phenological data from 30 perennial species across Europe. They calculated chilling accumulation (CA) and heat requirement (HR) for each species, station, and year using these models. The core of their methodology involved assessing whether the CA-HR relationship predicted by each chilling model aligned with the established physiological principle of a negative correlation (i.e., more chilling leads to lower HR). They classified models as "valid" if they consistently showed this negative relationship and "invalid" otherwise. To further validate their findings, they compared the performance of these models in simulating past phenological trends and predicting future changes under different climate scenarios (RCP 4.5 and RCP 8.5). They also explored the impact of different forcing models and data sources on the CA-HR relationship.

Datasets

  • Phenological Data: Pan European Phenology Project (PEP725) database, containing long-term in situ observations of leaf unfolding and flowering for 30 perennial species at 15,533 phenological stations across central Europe (1951-2018).
  • Climatic Data (Past): E-OBS v19.0eHOM daily gridded temperature data (0.1 x 0.1° resolution) for 1950-2018.
  • Climatic Data (Future): HADGEM2-ES model daily temperature data (0.5 x 0.5° resolution) for 2019-2099 under RCP 4.5 and RCP 8.5 scenarios, bias-corrected using the ISIMIP method.
  • Station-based Climatic Data (for scale effect assessment): Global Historical Climatology Network (GHCN) daily temperature data.

Key Insights and Ideas

  • The "Chilling Paradox": Winter warming, while seemingly beneficial, can paradoxically delay spring phenology by reducing the necessary chilling accumulation. This is because plants need a period of cold to break dormancy, and insufficient chilling increases the heat required to trigger spring events.
  • Model Validity is Crucial: The study highlights that not all chilling models are created equal. A significant portion (7 out of 12) of commonly used models fail to capture the fundamental negative CA-HR relationship.
  • Freezing Temperatures Matter: The primary reason for the invalidity of many models appears to be their failure to account for the effectiveness of freezing temperatures (<0°C) in breaking plant dormancy. Some models only consider temperatures above freezing, leading to an underestimation of chilling in colder winters.
  • Overestimation of Warming Impacts: The use of invalid chilling models leads to a substantial overestimation of how much earlier spring phenology will occur due to climate change. This has implications for predicting ecosystem productivity and carbon sequestration.

Key Results

  • Model Performance: More than half (7 out of 12) of the evaluated chilling models were found to be invalid because they showed a positive or non-significant CA-HR relationship, contradicting established physiological principles.
  • Future Phenology Overestimation: Invalid chilling models predict an advance in spring phenology by the end of the century that is significantly larger than that predicted by valid models. This overestimation can be as much as 7.6 days under RCP 4.5 and 20.0 days under RCP 8.5.
  • Spatial Trends in Chilling: While many models showed a decrease in chilling accumulation across Europe due to warming, some models (C7-C9) even predicted increasing trends, likely due to their specific temperature thresholds and how they handle freezing temperatures.
  • Model Accuracy: Valid chilling models, when used in a process-based model, generally produced more accurate simulations of past phenological trends and spatial gradients compared to invalid models.

Limitations and Future Work

  • Model Simplification: The study acknowledges that the CA-HR relationship might be more complex than a simple linear fit, and that plant acclimation to different climatic conditions could influence this relationship. However, their analysis did not find a significant difference in CA-HR relationships between low and high latitudes, suggesting their approach was robust.
  • Experimental Validation: The authors suggest that further experimental work is needed to confirm their findings, particularly regarding the effectiveness of temperatures below 0°C on dormancy break. This is challenging because plants need to acclimate to survive freezing temperatures.
  • Model Development: The paper strongly advocates for the development and adoption of more accurate chilling models that correctly incorporate the role of freezing temperatures and the negative CA-HR relationship. This is crucial for improving the reliability of future climate change impact assessments on vegetation.

Summary updated

Abstract

Spring warming substantially advances leaf unfolding and flowering time for perennials. Winter warming, however, decreases chilling accumulation (CA), which increases the heat requirement (HR) and acts to delay spring phenology. Whether or not this negative CA-HR relationship is correctly interpreted in ecosystem models remains unknown. Using leaf unfolding and flowering data for 30 perennials in Europe, here we show that more than half (7 of 12) of current chilling models are invalid since they show a positive CA-HR relationship. The possible reason is that they overlook the effect of freezing temperature on dormancy release. Overestimation of the advance in spring phenology by the end of this century by these invalid chilling models could be as large as 7.6 and 20.0 days under RCPs 4.5 and 8.5, respectively. Our results highlight the need for a better representation of chilling for the correct understanding of spring phenological responses to future climate change.
BibTeX
@inproceedings{wang2020overestimation,
  author = {Huanjiong Wang and Chaoyang Wu and Philippe Ciais and Josep Peñuelas and Junhu Dai and Yongshuo H. Fu and Quansheng Ge},
  title = {{Overestimation of the effect of climatic warming on spring phenology due to misrepresentation of chilling}},
  booktitle = {Nature Communications},
  year = {2020},
  doi = {10.1038/s41467-020-18743-8},
}

Loading related content...

Figure