Articles
MODELLING THE DATES OF F1 FLOWERING STAGE IN APPLE TREES, AS A TOOL TO UNDERSTANDING THE EFFECTS OF RECENT WARMING ON COMPLETION OF THE CHILLING AND HEAT REQUIREMENTS
Article number
817_14
Pages
153 – 160
Language
English
Abstract
Since the end of the 1980s, trends toward advances in apple flowering time have been noted in long-term records of flowering dates (by 7-8 days in France). Using models, an objective was to understand the impact of global warming on flowering time.
Dates of flowering for cv. ‘Golden Delicious’ were collected in four areas from the Walloon region in Belgium to the South-East of France.
The flowering model was composed of a chilling sub-model and a heating sub-model.
The input data consisted of observed dates of F1 stage.
A software package automatically optimised the parameters in different combinations of seven chilling and three heat temperature functions.
This was achieved by maximizing the R2 value between the observed and simulated dates.
Thereby three models explaining 82-86% of F1 date variability were selected.
These three models were validated from flowering data not used for parameter optimisation.
Also, they were used to simulate recent yearly dates of completion of the chilling effect and the yearly duration of the heat requirement.
Results showed trends towards a delay for the completion of chilling both in Belgium and France by 2-6 days.
Moreover, they illustrated similar trends towards shorter duration of the heat requirement (by 11-13 days). Hence, these results support the suggestion that global warming exerted two opposing effects simultaneously in Europe since the end of the 1980s: a slower/later mean completion of the chilling requirement and a shorter mean duration of the heat requirement.
A more marked effect on completion of the heat requirement, which might explain flowering advances, may have resulted from more pronounced warming from January to April, corresponding to the heat phase, than from October to January corresponding to the chilling phase.
Dates of flowering for cv. ‘Golden Delicious’ were collected in four areas from the Walloon region in Belgium to the South-East of France.
The flowering model was composed of a chilling sub-model and a heating sub-model.
The input data consisted of observed dates of F1 stage.
A software package automatically optimised the parameters in different combinations of seven chilling and three heat temperature functions.
This was achieved by maximizing the R2 value between the observed and simulated dates.
Thereby three models explaining 82-86% of F1 date variability were selected.
These three models were validated from flowering data not used for parameter optimisation.
Also, they were used to simulate recent yearly dates of completion of the chilling effect and the yearly duration of the heat requirement.
Results showed trends towards a delay for the completion of chilling both in Belgium and France by 2-6 days.
Moreover, they illustrated similar trends towards shorter duration of the heat requirement (by 11-13 days). Hence, these results support the suggestion that global warming exerted two opposing effects simultaneously in Europe since the end of the 1980s: a slower/later mean completion of the chilling requirement and a shorter mean duration of the heat requirement.
A more marked effect on completion of the heat requirement, which might explain flowering advances, may have resulted from more pronounced warming from January to April, corresponding to the heat phase, than from October to January corresponding to the chilling phase.
Authors
J.M. Legave, I. Farrera, M. Calleja, R. Oger
Keywords
dormancy, phenology, models, climate change, database
Online Articles (43)