Articles
Modeling Japan’s Yoshino cherry blooming with the PhenoFlex framework
Article number
1434_12
Pages
83 – 90
Language
English
Abstract
The annual blooming of Japan’s Yoshino cherries is a unique natural spectacle that plays an important part in Japan’s culture.
Process-based modeling can be applied to decipher how trees respond to temperature, by accurately representing the trees’ winter dormancy phases.
This period of suspended bud development consists of two temperature-dependent phases, endodormancy and ecodormancy, whose successful completion eventually leads to flowering.
We applied the recently developed PhenoFlex modeling framework to model Yoshino cherry bloom phenology.
PhenoFlex is a process-based model that combines the Dynamic Model for chill accumulation and the Growing Degree Hours Model for heat accumulation.
We calibrated individual PhenoFlex models with temperature and phenology data from 47 locations across Japan, which were collected from 1954 to 2022 by the Japan Meteorological Agency.
The resulting 47 models had, on average, an RMSE of 2.1 days in predicting bloom dates of Yoshino cherry at their corresponding locations, suggesting that PhenoFlex can reliably predict the spring phenology of Yoshino cherries.
In a further step, we tested the flexibility of PhenoFlex by calibrating models using a leave-one-out approach and testing prediction accuracy at left-out location whose data had not been used for model calibration.
The models showed a comparatively good performance for prediction of bloom date at most “untrained” locations, except for Kagoshima, as the warmest location.
This study indicates that the framework is also suitable for predicting bloom dates at new locations, suggesting possible applicability of PhenoFlex for future spring phenology forecasting of Yoshino cherry.
Process-based modeling can be applied to decipher how trees respond to temperature, by accurately representing the trees’ winter dormancy phases.
This period of suspended bud development consists of two temperature-dependent phases, endodormancy and ecodormancy, whose successful completion eventually leads to flowering.
We applied the recently developed PhenoFlex modeling framework to model Yoshino cherry bloom phenology.
PhenoFlex is a process-based model that combines the Dynamic Model for chill accumulation and the Growing Degree Hours Model for heat accumulation.
We calibrated individual PhenoFlex models with temperature and phenology data from 47 locations across Japan, which were collected from 1954 to 2022 by the Japan Meteorological Agency.
The resulting 47 models had, on average, an RMSE of 2.1 days in predicting bloom dates of Yoshino cherry at their corresponding locations, suggesting that PhenoFlex can reliably predict the spring phenology of Yoshino cherries.
In a further step, we tested the flexibility of PhenoFlex by calibrating models using a leave-one-out approach and testing prediction accuracy at left-out location whose data had not been used for model calibration.
The models showed a comparatively good performance for prediction of bloom date at most “untrained” locations, except for Kagoshima, as the warmest location.
This study indicates that the framework is also suitable for predicting bloom dates at new locations, suggesting possible applicability of PhenoFlex for future spring phenology forecasting of Yoshino cherry.
Authors
J.N. Bauer, L. Caspersen, E. Luedeling, H. Yamane
Keywords
Prunus × yedoensis, sakura, phenology, dormancy, process-based
Online Articles (52)