FRUIT WEIGHT PREDICTION MODELS IN QUERCUS BASED ON SOME BIOMETRIC PARAMETERS

Abstract

This study aimed at developing some mathematical models for the estimation of achene weight in oak based on fruit biometric parameters and at evaluating through principal component analysis (PCA) fruit size group affinity with biometric parameters. We studied fruits of two oak species, Quercus petraea and Quercus robur. At achene level, we determined such parameters as achene length (Al), middle achene diameter (MAD), achene basis diameter (Abd), achene tip diameter (Atd), and achene weight (Aw). At cupule level, we determined such parameters as cupule height (Ch), outer cupule diameter (Ocd), inner cupule diameter (Icd), cupule all thickness (Cwt), and cupule weight (Cw). We also calculated fruit size index (FSI). Regression analysis was used to find out equations as achene weight prediction models based on studied biometric parameters. The working process followed a logical pattern suggested a general working model. We obtained a linear equation for the estimation of achene weight based on Atd in Q. robur; all other achene weight prediction equations were of the second-degree polynomial type in both Q. robur and Q. petraea. The level of statistic safety in the estimation of fruit size ranged between R2 = 0.720 and R2 = 0.962 in Q. petraea and R2 = 0.376 to R2 = 0.928 in Q. robur, respectively. Based on biometric parameters, the fruits of the two species studied were grouped into seven size classes. The PCA analysis explained the variance based on PC1 as 92.573% and on PC2 as 6.2315% in Q. petraea and as 82.041% based on PC1 and as 16.032% based on PC2 in Q. robur. For size classes Qp1-Qp3 in Q. petraea and Qr1 in Q. robur, the least variation of any size parameter leads to achene size, while size classes Qp4 and Qr7, diameter variation significantly determines achene weight increase. In size classes Qp5-Qp7 and Qr2-Qr6, fruit length variation induces a significant fruit weight increase.