A linear model for predicting olive yield using root characteristics

Predicting yield is an important objective in agricultural research. We developed a linear regression model to predict the olive fruit yield (FY) for four olive cultivars (Sivillano, Conservolia, Zard and Clonavis) by monitoring soil moisture, response to root growth and its characteristics including root weight density (RWD), root length (RL) and root biomass (RB). Our results show the model predicts fruit yield based on a simple linear function of root characteristics (R2 = 0.85). A principal component analysis provided a meaningful combined factor (the first principal component) that showed a clear discrimination in olive fruit yield among four cultivars. The model could be applied to rapidly evaluate olive fruit yield using the measured values of root characteristics and to support decision making for orchard management.