Interpretation of hybrid × sowing date interaction for oil content and oil yield in sunflower

Due to the presence of an interaction, it is often difficult to recommend which sunflower hybrids should be planted based on specific sowing dates or production regions. Through this study, the hybrid × sowing date (H × SD) interaction for sunflower oil content and oil yield was investigated in field trials by a multivariate and multiplicative statistic approach such as site regression (SREG) and partial least squares (PLS) regression model. For the environmental dissection of the H × SD interaction for oil content and oil yield, six climatic variables (minimum, maximum and mean temperature, precipitation, sunshine hours and relative air humidity) in four sunflower developmental stages (V4-6 leaves, R1-budding, R5.8-flowering, and R9-physiological maturity) were used. The three-way ANOVA for both traits indicated that all main effects and first order interactions were highly significant (P < 0.01) except H × SD interaction for oil content which was significant (P < 0.05). The second order interaction for both traits was non-significant. The oil content was predominantly influenced by the hybrid (69.6%) followed by the year (10.3%) and sowing date (6.8%), while the oil yield was predominantly influenced by the year (58.8%), followed by the sowing date (12.9%) and hybrid (10.7%). Statistical models revealed that the relative air humidity in flowering and physiological maturity stage, maximum temperature and mean temperature in the budding and flowering stage are the most important causes of the H × SD interaction for the oil content. The most important causes of the interaction for oil yield were relative air humidity in flowering and physiological maturity, maximum temperature, mean temperature, sunshine hours and precipitation in flowering stage.

► Hybrid × sowing date interaction was investigated for sunflower oil content and oil yield.
► Hybrids and sowing dates were significant sources of variation for both traits.
► Multivariate statistical models have enabled the understanding of the impact of different sowing dates on oil content and yield.
► The performance of the hybrids was strongly under the influence of several climatic variables at flowering stage.