Genotypic Adaptation of Durum Wheat to Contrasting Water Regimes: Linking Phenology, Chlorophyll Status, Yield Components, and Grain Yield

Water scarcity is a recurrent constraint to durum wheat production in semi-arid Mediterranean-type environments, yet genotype performance can change sharply with the timing and severity of seasonal drought. This study quantified phenological, chlorophyll-related, yield-component, and grain-yield responses of 45 durum wheat genotypes grown during 2021–2022 and 2022–2023 under rainfed and supplementary-irrigated field conditions in Diyarbakır, southeastern Türkiye. The experiment comprised equal-sized plots with four replications under each water regime. Supplementary irrigation was delivered by drip irrigation in five applications beginning at stem elongation. Precipitation from sowing to heading was 187.0 mm in 2021–2022 and 285.5 mm in 2022–2023, and April was warmer in the first season (17.9 °C versus 14.7 °C). Grain yield was significantly affected by growing season, water regime, genotype, and all two- and three-way interactions (p < 0.01), although diagnostic tests indicated heterogeneous residual variance among environments and the combined fixed-effect ANOVA is therefore interpreted cautiously. Mean grain yield under rainfed conditions was 466.5 and 3254.3 kg ha−1 in the first and second seasons, respectively, compared with 4208.1 and 5560.9 kg ha−1 under supplementary irrigation. The exceptionally low first-season rainfed yield represents a near-crop-failure environment; stand-count data were not available to separate establishment losses from subsequent drought effects. Pooled drought indices descriptively ranked Fırat-93 highest for STI (0.638) and GMP (3901.4 kg ha−1), whereas Şahinbey had the highest pooled YSI (0.518); these index differences were not subjected to inferential significance testing and should not be interpreted as statistically distinct tolerance classes. Grain yield was associated with SPAD and yield components in pooled genotype means, but these associations were environment dependent. The results characterize genotype responses across four site-year × water-regime combinations and provide hypotheses for validation in broader multi-location and multi-year trials.