Evaluation of new mathematical models for estimation of single olive leaves area

Abstract

Abstract. The study was conducted on “Khoderi” olive cultivar planted in Jableh Region-Latakia province, Syria in 2017 in order to evaluate some mathematical models adapted for olive single leaf area estimation. Leaf samples were taken from the middle of one-year branches. Actual areas of the leaves were measured using Adobe Photoshop CS5. Leaf dimensions (length and width) were measured accurately. Coefficients of determination were estimated for the relation between leaf dimensions and the actual area. The best coefficient of determination was between the natural logarithm of the product (leaf length × leaf width) and the natural logarithm of leaf area (R2= 0.962). Linear regression equation of the mentioned relation was fitted and evaluated. The accuracy of the new model (A=e0.9509ln LW – 0.2867) was compared to other models commonly used for olive single leaf area estimation. The comparison showed no significant differences between leaf area obtained by the new model and the actual leaf area values (p=0.01), whereas significant differences were found for the other models. The new model showed the lowest Root Mean Square Error (RMSE) and high efficiency in estimating olive leaf area of “Khoderi” cultivar in two different environments; the same results were obtained for olive cultivar “Picholine” the French. We recommend the new model for olive single leaf area estimation.