Environmental and Agronomic Factors Associated with Yield Variation in Two Soybean Cultivars across Multiple Locations

Soybean ( Glycine max (L.) Merr.) yield reflects interactions among climatic conditions, soil properties, and agronomic traits that vary among cultivars. This study evaluated the medium-late-maturing cultivar ‘Seonpung’ and the medium-maturing cultivar ‘Seonyu2’ across eight locations in South Korea over two years. Representative variables were selected by hierarchical clustering based on Pearson correlation and principal component analysis, followed by multiple regression of yield and three yield-related traits: pod number, branch number, and hundred-seed weight. Significant Location × Cultivar ( F = 4.51, p < 0.01) and Location × Year × Cultivar ( F = 7.51, p < 0.001) interactions were detected for yield. Among the 12 yield models, only the Seonpung agronomic model was significant (adjusted R 2 = 0.51), in which days after sowing to R6 was negatively associated with yield ( B = -18.97). In contrast, seven of the 12 pod number models were significant, mainly in Seonpung and the combined dataset (adjusted R ² values up to 0.45), indicating pod number was more consistently associated with environmental and agronomic factors than final yield. Theoretical yield based on seed number, hundred-seed weight, and plant density was not significantly correlated with observed plot yield ( r = 0.20; ratio range = 0.54–6.42), suggesting individual-plant measurements may not fully capture plot-scale yield variation. Overall, environmental and agronomic factors were more clearly associated with yield-related traits, particularly pod number, than with final yield, which reflected more complex cultivar-specific responses across locations and years.