Winter wheat phenology assessed at apex and canopy scales responds differently to thermal time, photoperiod and environment, with implications for yield

Winter wheat phenology is commonly modelled using externally visible crop stages, although reproductive development begins earlier at the shoot apex and may respond differently to environmental cues than canopy progression. This study tested whether phenology modelling in winter wheat depends on developmental scale, predictor choice and local field environment. Across three growing seasons, seven cultivars were evaluated at two contrasting field sites differing in soil texture and water regime. Shoot apex development was assessed using the Waddington scale, whereas canopy-visible development was scored using BBCH. Both developmental scales were related to growing degree hours (GDH) and accumulated daylength using asymptotic, beta-growth, Gompertz, logistic and Richards nonlinear models. Model performance differed among developmental scales and predictors. Overall, the beta-growth model provided the strongest support according to AICc and RMSE, although Gompertz and Richards models showed comparable support in some GDH-based relationships. GDH was a consistent predictor of developmental progression, particularly for BBCH, whereas accumulated daylength contributed more strongly to Waddington-scale development than to canopy-level progression. Site conditions modified developmental timing, with the more drought-prone sand site showing distinct phenological responses in several cultivar–scale combinations. Phenological timing also had agronomic relevance: canopy-level timing around stem elongation, together with year, site and photoperiod background, explained grain yield variation more strongly than early shoot apex timing, although WS 2.0 timing was also significant. These results show that winter wheat phenology is scale dependent, environmentally contingent and not governed by a single universal developmental model.