Field Evaluation of Data-based Pruning Severity Levels (PSL) on Mechanical Harvesting of Apples

摘要

The increasing cost and decreasing availability of skilled orchard labor are growing concerns for orchardists, and have led to interests in potential mechanical solutions for pruning and harvesting. A critical first step in developing mechanical solutions is the establishment of precision orchard systems based on specific pruning rules. In this study, we evaluated pruning severity levels (PSL) on apple fruit removal efficiency (FRE) and harvested fruit quality using a shake-and-catch vibratory harvesting system. Approximately 300 branches from 50 randomly selected apple trees were manually subjected to five data-based PSL in field conditions. The experiment was conducted in a commercial 'Jazz' orchard with trees trained in a vertical fruiting-wall architecture. Five PSL were studied, including variable length (PV) for which shoots were pruned to a length determined by their basal diameters in a linear-ratio (LR-PSL). In P-10, P-15, and P-23 treatments, shoots were pruned to be no longer than 10, 15, and 23 cm, respectively, in fixed-lengths (FL-PSL). In addition, a control set (PC) with regular pruning was used. The highest FRE was observed in the PSL of P-10 (90%) and P-15 (91%), followed by PV (88%). No significant difference was observed regarding harvested fruit quality. Canopy intercepted photosynthetically active radiation (PAR) ratio and crop load of P-23 and PV were 0.67, 1.90 kg and 0.51, 2.46 kg, respectively. They were also improved compared with PC (0.51 of PAR and 1.64 kg). In summary, field results verified that LR-PSL of PV would efficiently facilitate trellis-trained 'Jazz' apple trees to achieve generally higher FRE with such mechanical harvesting methods, and intercepted PAR and fruit yield are not sacrificed. This optimized PSL would also serve as a proof-of-concept for the further development of a mechanical or autonomous pruning machine, while maintaining a profitable crop level in such orchards.