Cold hardiness mechanisms and modeling: existing approaches and future avenues.

Cold hardiness models are useful tools to predict cold damage in plants, such as those produced by unseasonal temperature cycles or by increased cold exposure. Although development of these models started about five decades ago, their applications remain limited. We describe the main paradigms driving the different types of cold hardiness models (empirical to process-based), their similarities and differences. Among the existing paradigms, process-based models are built to translate physiological mechanisms into mathematical functions over a broad range of climatic conditions, thus making them more accurate for studying the effect of climate change. Different approaches have been developed in predicting cold hardiness: (1) empirical relationships between temperature and cold hardiness; (2) phenological processes controlling acclimation and deacclimation rates; (3) phenological and physiological processes predicting cold hardiness through the osmo-hydric approach; and (4) molecular regulation driving the metabolic drivers of cold hardiness. For the first three approaches, we describe the context, the experimental and field observations that defined their frameworks as well as their limitations. To increase the realism of cold hardiness models, we describe the potential of a fourth approach, based on the perception of environmental signals, how it translates into cold acclimation/deacclimation and provide recommendations to develop this framework.