Vis/NIR spectroscopy is a promising tool to predict fruit set and chemical thinner response
Apple chemical thinning occurs during a critical 3- to 4-week period where growers make multiple applications of plant bioregulators to induce fruitlet abscission and optimize crop load.
However, differences between persisting and abscising fruitlets are not visible for up to two weeks following thinner application, which complicates management decisions.
The fruit growth rate model (FGM) is currently the only tool available to growers to predict fruit set following a thinner application, utilizing repeated measurements of fruitlet diameter to distinguish between persisting and abscising fruitlets.
Accurate predictions with the FGM are typically seen 7-10 days following thinner application; however, this process is time sensitive and labor intensive.
The goal of this project was to determine if a portable, hand-held Vis/NIR spectrometer (Felix F-750; Felix Instruments; Camas, WA) can predict abscission rates earlier than the FGM. An experiment was conducted in a commercial Honeycrisp/M. 26 orchard in Dana, NC. Fruitlet diameter and spectrometer data were collected on each fruitlet in 20 tagged clusters on 5 trees, measurements occurred -4, -1, 3, 5, 7, 9, and 12 days after a chemical thinner application (600 ppm Carbaryl + 5 ppm NAA). Classification models (partial least squares regression, random forests, and XGBoost) were built from captured absorbance spectra, each wavelength a predictor variable, for each measurement date.
The FGM overestimated persisting fruitlets until 9 days after thinner was applied.
Random forests models had greater than 88% accuracy and 94% area under the receiver operating characteristic curve for each measurement date.
These results indicate that Vis/NIR spectroscopy has the potential to rapidly predict fruit set before and after a thinner application, providing growers with more flexibility to optimize thinner applications.
Use of this technology and associated models to predict apple fruitlet abscission will be validated in future research.
Larson, J.E. and Kon, T.M. (2023). Vis/NIR spectroscopy is a promising tool to predict fruit set and chemical thinner response. Acta Hortic. 1360, 253-258
DOI: 10.17660/ActaHortic.2023.1360.32
https://doi.org/10.17660/ActaHortic.2023.1360.32
DOI: 10.17660/ActaHortic.2023.1360.32
https://doi.org/10.17660/ActaHortic.2023.1360.32
abscission, agricultural modeling, crop load management, fruit growth, machine learning, Malus × domestica
English
1360_32
253-258
- Division Physiology and Plant-Environment Interactions of Horticultural Crops in Field Systems
- Division Precision Horticulture and Engineering
- Division Horticulture for Development
- Division Plant Genetic Resources and Biotechnology
- Division Temperate Tree Fruits
- Division Temperate Tree Nuts
- Division Tropical and Subtropical Fruit and Nuts
- Division Vegetables, Roots and Tubers
- Division Vine and Berry Fruits