Abstract:

To clarify the impact of stalk-cutting harvest on the quality of upper tobacco leaves after curing, this study utilized ‘Yunyan 87’ as the test material. Various multivariate data analysis methods, including principal component analysis (PCA), partial least squares discriminant analysis (PLS-DA), variable importance in projection (VIP), and Pearson correlation analysis, were employed to investigate the differences in appearance quality, physical properties, chemical composition, and sensory quality between stem-leaf and leaf-by-leaf cured tobacco from six towns in Hubei Province. The results indicated that after excluding outlier indicators, the PLS-DA model showed R²_X=0.578, R²_Y=0.997, Q²_PLS-DA=0.96, demonstrating the model’s effectiveness. A validation through 200 iterations of response sorting yielded R²_intercept and Q²_intercept values of 0.903 and -0.0125, respectively, indicating a good model fit. A total of 27 variables with significant differences were selected based on VIP values greater than 1.0, including reducing sugars, total sugars, sensory quality, and various amino acid contents. Among these, six variables—total sugars, reducing sugars, total sensory weight score, aroma quality, off-flavors, and irritation—exhibited positive effects, while the remaining 21 variables showed negative effects. In summary, the stalk-cutting and curing is beneficial for increasing the content of total sugars, reducing sugars, and certain sensory factors in the upper leaves, but it is detrimental to the increase of total nitrogen and amino acid contents.

Key words: stalk-cutting harvest, upper tobacco leaves, cured tobacco leaves, multivariate data analysis, quality evaluation, physical characteristics, chemical components, sensory quality