Leaf area estimate of Pennisetum glaucum by linear dimensions

Leaf area measurements are of the main parameters used in agronomic studies to evaluate plant growth. The current study used a non-destructive method based on linear leaf dimensions (length and width) to select the regression model to estimate millet (Pennisetum glaucum) leaf area. For two millet genotype (IPA BULK 1 BF and ADR 300) 128 randomly-chosen leaves were measured at different vegetative growth stages. Measures of length and width of each leaf were made using digital calipers. Leaf area was measured using a gravimetric method. The best-fit leaf area estimation model was selected via linear, potential and gamma regression models. Leaf area values varied from 3.02 to 209.21 cm2 . The average value was 95.31 cm2 . The potential regression model exhibited lower residual sum of squares and Akaike's information criterion and similar determination coefficient and Willmott index. Thus, potential regression was more efficient in explaining the leaf area of millet, independent of the genotype, when compared to other models evaluated in this research. Length (L) and width (W) could be used in the following potential regression model to estimate millet leaf blade.(AU)

Leaf area estimate of Pennisetum glaucum by linear dimensions

Leaf area measurements are of the main parameters used in agronomic studies to evaluate plant growth. The current study used a non-destructive method based on linear leaf dimensions (length and width) to select the regression model to estimate millet (Pennisetum glaucum) leaf area. For two millet genotype (IPA BULK 1 BF and ADR 300) 128 randomly-chosen leaves were measured at different vegetative growth stages. Measures of length and width of each leaf were made using digital calipers. Leaf area was measured using a gravimetric method. The best-fit leaf area estimation model was selected via linear, potential and gamma regression models. Leaf area values varied from 3.02 to 209.21 cm2 . The average value was 95.31 cm2 . The potential regression model exhibited lower residual sum of squares and Akaike's information criterion and similar determination coefficient and Willmott index. Thus, potential regression was more efficient in explaining the leaf area of millet, independent of the genotype, when compared to other models evaluated in this research. Length (L) and width (W) could be used in the following potential regression model to estimate millet leaf blade.