Physiological plasticity and structural adaptation explain yield-stability trade-offs in bread wheat under contrasting moisture regimes

Drought stress is a major constraint to wheat production in dryland environments, and improving yield stability requires the identification of genotypes and traits associated with adaptation to water-limited conditions. This study aimed to investigate the relationships among morphological, anatomical, physiological, and yield-related traits associated with drought adaptation and to identify wheat genotypes exhibiting superior yield stability under contrasting moisture regimes. Twelve bread wheat genotypes were evaluated under supplementary irrigation and dryland conditions over four consecutive growing seasons (2019–2023) using a split-plot arrangement based on a randomized complete block design with three replications. some agronomic, physiological, anatomical, and yield-related traits were assessed, including grain yield (GY), yield components, canopy temperature depression (CTD), stomatal conductance (gs), photosynthetic rate (Pn), stem diameter (STD), and number of vascular bundle strands (NVVS). Best linear unbiased predictions (BLUPs), drought tolerance indices (DTIs), correlation analysis, principal component analysis, LASSO regression, and hierarchical clustering were used to characterize genotype responses and identify traits associated with drought adaptation. dryland conditions reduced grain yield by 26.5% relative to supplementary irrigation. The greatest reductions occurred in CTD (46.7%, P < 0.001), stomatal conductance (36.7%, P < 0.001), photosynthetic rate (21.9%, P < 0.01), and NVVS (15.6%, P < 0.05), whereas grain length, grain width, spikelet number, and stem diameter showed comparatively greater stability. The corresponding DTI values were 0.46 for CTD, 0.62 for gs, 0.74 for GY, and 0.77 for Pn. LASSO regression identified grain number per square meter (coefficient = 0.962), thousand-grain weight (0.838), and spikelets per spike (0.386) as the strongest positive predictors of grain-yield drought tolerance. Stability analysis identified Hashtrood and Waran as the most promising genotypes, combining high yield performance with superior stability under contrasting moisture regimes. Hashtrood showed broad adaptation across environments, whereas Waran maintained consistent performance under dryland and supplementary irrigation conditions. These findings demonstrate that integrating physiological and agronomic traits can enhance the identification of drought-adapted wheat genotypes and provide valuable selection criteria for breeding programs targeting dryland environments.