Leaf area estimation of gherkin plants from linear leaf dimensions

Resumo

The gherkin is a widely consumed species in the Northeast Brazil, where its production is mainly derived from naturally occurring or wild plants. The leaf area of plants plays a crucial role in their development and productivity, as it is directly related to photosynthetic capacity. Through their leaves, plants absorb sunlight and convert light energy into chemical energy, which is essential for growth and biomass production. Thus, the objective of this study was to propose regression models to estimate the leaf area of gherkin using linear dimensions of the leaves. Two experiments were conducted, one between January and April (summer-autumn) with two gherkin cultivars (‘Caipira do Norte’ and ‘Liso Calcutá’ for test and validation – dependent data) and other between May and August (autumn-winter) 2021 only with the cultivar ‘Caipira do Norte’ (for validation – independent data). In the summer-autumn experiment, the relationships between individual leaf area (LA), as the dependent variable, and leaf length (L), width (W), or the L×W product, as independent variables, were analyzed using both linear and power regression models. These models were developed individually for each cultivar, as well as for the two cultivars jointly (grouped data). Statistical indicators, including the Pearson’s linear correlation coefficient (r), coefficient of determination (R²), Willmott’s agreement index (d), and root mean square error (RMSE), were used as criteria for selecting the best models. In the validation between observed and estimated values, the best estimates of individual LA of gherkin were obtained using the L×W product as an independent variable. The grouping of two gherkin cultivars (‘Caipira do Norte’ and ‘Liso Calcutá’) into a single model was possible. Based on higher accuracy and lower errors, the linear (LA = 0.7296×L×W; r = 0.9769, R² = 0.9543, d = 0.9882, and RMSE = 8.94) and power (LA = 1.0024×(L×W)^0.9440; r = 0.9772, R² = 0.9549, d = 0.9883, and RMSE = 8.88) models, using grouped data, are indicated for individual LA estimation of the gherkin plants.