Integrative Multivariate Analysis of Pisum sativum L. Genotypes Using Agro-Morphological Traits under Contrasting Semi-Arid Highland Agro-Ecosystems

Understanding yield formation and environmental adaptation in pea ( Pisum sativum L.) is essential for improving productivity under semi-arid highland conditions. This study evaluated eleven pea genotypes using morphological characterization and multivariate analyses across contrasting Northern and Central Highland environments of Yemen. Qualitative and quantitative descriptors revealed clear phenotypic variation among genotypes and successfully distinguished improved cultivars from local landraces. Analysis of variance showed significant differences (P ≤ 0.05) among genotypes for all measured traits, indicating substantial genetic variability. Grain yield ranged from 0.40 to 1.20 t ha⁻¹ in the Northern Highlands and from 0.50 to 1.80 t ha⁻¹ in the Central Highlands, with Yahsb and Amran consistently exhibiting superior performance across environments. Principal component analysis indicated that yield variation under favorable conditions was mainly associated with seed weight (SW) and number of pods per plant (NPP), whereas phenological traits contributed more strongly under Northern Highland conditions. Correlation analysis revealed strong positive associations of grain yield with seed weight, pod length, pod width, and number of pods per plant, although the magnitude of these relationships varied between environments. Path coefficient analysis showed that seed weight exerted the greatest direct effect on grain yield in the Central Highlands (0.703), while pod length had the strongest direct effect in the Northern Highlands (0.568). Hierarchical clustering and heatmap analyses consistently grouped Yahsb, Amran, YG50075, and YG50083 into a distinct cluster across both environments, reflecting their close phenotypic similarity and superior agronomic performance. These findings demonstrate that yield determination in pea is environment-dependent and highlight Yahsb and Amran as promising genotypes for breeding and cultivation in semi-arid highland agroecosystems.