Incorporation of environmental covariates to nonlinear mixed models describing fruit growth
Yield prediction is still a major challenge in pear production. Forecasting fruit growth after modeled curves allows predicting both potential yield and quality. This research aimed to fit multilevel no-linear mixed models (NLMM) based on logistic curves to describe pear growth in the Upper valley o...
| Main Authors: | , , , |
|---|---|
| Format: | info:ar-repo/semantics/artículo |
| Language: | Inglés |
| Published: |
Ediciones INTA
2024
|
| Subjects: | |
| Online Access: | http://hdl.handle.net/20.500.12123/16648 https://doi.org/10.58149/14h1-sp68 |
| Summary: | Yield prediction is still a major challenge in pear production. Forecasting fruit growth after modeled curves allows predicting both potential yield and quality. This research aimed to fit multilevel no-linear mixed models (NLMM) based on logistic curves to describe pear growth in the Upper valley of Rio Negro and Neuquén, Argentina. The models incorporated several thermo-accumulative indices accounting for temperature effects on fruit-growth physiology. In this way, they captured normal fruit-growth patterns and environmental variations along the growing season. The study was conducted on “William´s” pear trees for 16 seasons. Many trees and fruits were randomly selected and identified. Equatorial diameters were weekly measured with an electronic digital caliper. Climatic data was recorded for all studied seasons from INTA Upper Valley agrochemical station and thermo accumulative indexes were calculated from daily temperature. The best models were selected according to the information criteria index. Multilevel NLMM discerned and quantified the sources of stochastic variability at different levels, allowing better index criteria in comparison to models only considering a single level of variability among random effects. The incorporation of thermo accumulative indexes also increased model performance. |
|---|