Cold Hardiness

Cold tolerance in woody perennials changes dynamically across dormancy progression; however, whether relative genotype hardiness hierarchy remains stable across the transition from endodormancy to bud break has not been systematically tested. We evaluated whether cold hardiness rankings are conserved or reorganized across developmental states in sweet cherry (Prunus avium L.). Midwinter lethal temperatures (LT), reflecting dormancy-phase deep supercooling, were quantified using differential thermal analysis during standardized sampling windows across three seasons (2019–2022). Spring freeze resistance was determined at tight cluster (BBCH 54), when deep supercooling is no longer detectable, using controlled-freezing LT50 assays. A 29-genotype dataset evaluated over two years and an independently analyzed 12-genotype subset evaluated across three years were classified into phenotypic groups using discriminant analysis of principal components (DAPC) based on midwinter LT and tight-cluster LT50. Winter and spring field damage (FD) were incorporated as independent in situ validation metrics. Cold hardiness hierarchy reorganized between dormancy and tight cluster in both datasets. In the 29-genotype analysis (K = 4 groups), the most winter-susceptible group exhibited the lowest LT50 at tight cluster, and group rankings differed significantly between seasons. In the 12-genotype, three-year subset (K = 3 groups), this directional inversion was reproduced across all seasons. Season × Group interactions were highly significant, winter–spring rank correlations were negative, and substantial rank shifts and extreme tercile reversals were observed. Winter FD consistently mirrored midwinter LT structure, and spring FD mirrored LT50 structure across years, confirming that seasonal reorganization was reproducible in both controlled freezing assays and independent in situ injury measurements. These results demonstrate that genotype-level cold hardiness rankings are not conserved across dormancy progression. Midwinter LT reflects dormancy-phase supercooling performance, whereas tight-cluster LT50 reflects freeze resistance expressed after developmental reactivation and loss of deep supercooling. Cold hardiness in deep-supercooling woody perennial floral buds is developmentally state-structured, with genotype rankings reorganized across dormancy transitions, requiring stage-specific evaluation of freezing tolerance.