Evaluating Grain Yield of Barley Genotypes under Water-Limited Conditions Using Stress Indices and Multivariate Analysis

Abstract
Introduction: Drought stress is one of the most critical factors limiting barley production in arid and semi-arid regions worldwide, leading to significant reductions in both growth and yield. Given the increasing intensity and frequency of drought in many agricultural areas, identifying drought-tolerant genotypes is essential for achieving sustainable productivity. However, due to the complex and polygenic nature of grain yield, breeding for drought stress faces a significant challenge regarding the trade-off between yield potential under optimal conditions and survival under stress. In this context, utilizing drought tolerance indices and multivariate statistical methods, such as Principal Component Analysis (PCA), serves as an effective approach for interpreting phenotypic response patterns, reducing data dimensionality, and precisely identifying superior genotypes across both stress and non-stress environments.
Materials and Methods: In this study, 15 barley cultivars and promising lines were evaluated under full irrigation and drought stress equivalent to 50% of crop water requirement during the 2024–2025 growing season at a research farm in southern Iran. Grain yield was measured under both conditions. Ten drought tolerance indices were calculated to assess genotype sensitivity and yield stability. After confirming data normality using the Kolmogorov–Smirnov test, analysis of variance was performed using SAS software version 9.4. Correlation analysis, factor analysis, and genotype scatter plots were performed using R software. Biplot analysis was also conducted based on the results of principal component analysis (PCA).
Results: Combined analysis of variance revealed that the effects of environment, genotype, and their interaction on grain yield were significant at the 1% probability level, indicating differential responses of genotypes to water availability. Under non-stress conditions, line WB9816 (4.43 kg/plot) and cultivars Golchin, Nobahar, and Zahak exhibited superior grain yield. Under drought stress, cultivar Zahak and line WB9819 produced the highest yields (3.01 and 2.94 kg/plot, respectively), whereas Behdan exhibited the lowest yield (1.14 kg/plot). The MP, GMP, HM, STI, and YI indices showed strong positive correlations with yield in both environments, identifying them as the most efficient criteria for screening high-yielding genotypes. Conversely, although TOL and SSI indices classified Behdan as “tolerant” due to its minimal yield reduction, they were not effective indicators for agronomic selection on their own, given the genotype’s low absolute yield. Principal Component Analysis (PCA) showed that the first two components explained 61.57% and 37.61% of the total variance, respectively, accounting for a cumulative 99.18%. PC1 represented productivity and high-yield potential, while PC2 distinguished genotypes based on stress sensitivity. Based on the biplot and the Fernandez 3D plot analysis, the genotypes Zahak, WB9819, WB9810, and Sahra clustered in region A (high yield in both conditions), while cultivars such as Rashin and Mahoor were located in region D (low yield in both conditions).
Conclusion: This study demonstrated that multivariate evaluation based on combined indices significantly enhances the precision of genetic screening compared to traditional single-factor methods. Consequently, cultivar Zahak (as the superior commercial variety for yield potential under harsh conditions) and line WB9819 (as a promising line with minimal yield reduction under stress and high stability) were identified as the most drought-tolerant genotypes. Both genotypes, particularly line WB9819, are recommended as promising candidates for future multi-environment and multi-year evaluations.