The TreeChill model: A new framework for predicting the impact of erratic winter weather on trees

Abstract

Perennial plants rely on exposure to low winter temperatures to break dormancy, aligning reproductive timing with local weather conditions. Current chill accumulation models quantify this exposure using only air temperature as input, leading to reduced accuracy as climate becomes more variable and unpredictable. This study introduces a novel framework that replaces air temperature with tree temperature as a continuous functional parameter to improve the reliability of chill accumulation calculations. Over three winters, we measured tree temperature and collected weather data in sweet cherry orchards in distinct climate areas. Tree temperature exceeded air temperature, particularly on sunny days, leading to a 15 % reduction in chill accumulation. We developed the TreeChill model, which predicts tree temperature using publicly available weather data with high precision. The model has a coefficient of determination of 0.930 and a standard deviation of residuals of 2.02 °C. The difference in chill accumulation calculated using predicted versus measured tree temperature was only 0.4 chill portions. Tree temperature provided more accurate insights into dormancy-related processes than air temperature. Variations in bloom progression within sweet cherry tree canopies corresponded to differences in tree temperature, and bloom timing was precipitated by prolonged winter chill exposure at the branch level, followed by localized short-term heat exposure. This study integrates macro- and micro-climatic data and plant physiological information and can be used in different climate-related prediction models to enhance their precision. The model will make tree temperature readily available to researchers and stakeholders, facilitating the development of climate change mitigation strategies in agricultural, natural and urban systems.