Irregular Bearing and Climate in Olive

Longterm yield records are important reference databases useful for calibration and validation of models for climate analysis. The evaluation of olive tree yields for Jaen province, the main production area in Spain, during a 21 years period of system stasis (constancy of variety and agro-techniques), revealed (a) a remarkable interaction between rainfall and previous year production with a significant non-additive term and (b) the fact that the previous year yield does not show any descriptive power of the current-year production. These evidences raise several questions about the rightness of alternate bearing as driving variable in olive orchard models, being the time series distributed in irregular fashion among conditions of Medium (M), Low (L) and High (H) fruit load. In the Jaen data set, the interaction between an aleatory factor (annual rainfall) and an equilibrium determined by two biological variables (flower fertilization and previous year shoot elongation) is statistically proved. These biological variables depend on the rainfall level. As a consequence, yield time series are not auto-correlated and yield modelling should measure productive efficiency at least in 3 classes of fruit load in order to avoid the risk of finding stochastic yield frequency. So the models of olive orchard production should record previous year fruit load, current year fruit load and water availability in order to describe up to 70% of the variance within IPCC fingerprint high confidence limits (i.e. above a probability level of 0.67).