A conceptual framework for simultaneous optimization of integrated climate scenario data and phenology models

Area of Science:

• Agricultural Science
• Climate Science
• Data Science

Background:

• Global food security relies on climate-resilient crops.
• Accurate crop phenology prediction under climate change is crucial.
• Existing models often neglect uncertainties from climate data and model interactions.

Purpose of the Study:

• To develop a novel dynamic crop phenology model framework.
• To simultaneously integrate uncertainties from phenology models, climate data, and their interaction.
• To improve predictions of crop phenology under future climate scenarios.

Main Methods:

• Developed a dynamic crop phenology model framework using a large dataset (2500+ locations, 80+ years) of winter wheat phenology and environmental data.
• Integrated up to seven environmental covariates with defined response curves.
• Employed an effective model selection process using Swiss climate scenario projections.

Main Results:

• The model framework successfully integrates multiple sources of uncertainty.
• Predictions indicate a 22-day earlier heading for winter wheat by 2070-2099 under future climate scenarios.
• Ignoring climate data uncertainties leads to contradictory phenological predictions.

Conclusions:

• Considering all three uncertainty sources (model, climate data, interaction) is vital for accurate phenology prediction.
• The developed framework enhances the reliability of crop response assessments to climate change.
• This work provides a robust tool for predicting future crop phenology and informing agricultural strategies.