Dynamic modeling of cold hardiness in tea buds by imitating past temperature memory.

Background and aims Most perennial plants memorize cold stress for a certain period and retrieve the memories for cold acclimation and deacclimation, which leads to seasonal changes in cold hardiness. Therefore, a model for evaluating cold stress memories is required for predicting cold hardiness and for future frost risk assessments under warming climates. In this study, we develop a new dynamic model of cold hardiness by introducing a function imitating past temperature memory in the processes of cold acclimation and deacclimation. Methods We formulated the past temperature memory for plants using thermal time weighted by a forgetting function, and thereby proposed the dynamic model of cold hardiness. We used the buds of tea plants (Camellia sinensis (L.) O. Kuntze) from two cultivars, Yabukita and Yutakamidori, to calibrate and validate this model based on 10 years of observed cold hardiness data. Key results The model captured more than 90% of the observed variations in cold hardiness and predicted accurate values for both cultivars, with root mean square errors of ~1.0 °C. The optimized forgetting function indicated that the tea buds memorized both short-term (recent days) and long-term (previous months) temperatures. The memories can drive short-term processes such as increasing/decreasing the content of carbohydrates, proteins, and antioxidants in the buds, as well as long-term processes such as determining the bud phenological stage, both of which vary with cold hardiness. Conclusions The use of a forgetting function is an effective means of understanding temperature memories in plants and will aid in developing reliable predictions of cold hardiness for various plant species under global climate warming.